Studies of Gastric Motility in Health and Diabetes

A thesis submitted by

Julie Eva Stevens

For the Degree of

Doctor of Philosophy

Discipline of Medicine

University of Adelaide

May 2009

Chapter 1

1

1Chapter 1NORMAL GASTRIC MOTOR FUNCTION

1.1 Introduction

The stomach is responsible for receiving, storing, mixing and ultimately delivering,

ingesta into the small intestine at an optimum rate for digestion and absorption of

nutrients (Abell et al. 2008a, Horowitz et al. 1994, Horowitz et al. 2001, Horowitz et

al. 2002b, Parkman et al. 2004). This chapter presents an overview of normal gastric

motor function in humans, including the different stomach regions and their

individual functions, characteristic patterns of gastric emptying and the factors which

determine the rate at which the stomach empties a meal.

Chapter 1

2

1.2 Anatomical and functional motor regions of the

stomach

Gastric emptying is a complex process. - Ingested food is received by the stomach,

mixed with gastric secretions, ground into small particles less than 1 – 2 mm in size

and emptied into the small intestine at a rate which optimises digestion and

absorption of nutrients (Horowitz and Fraser 1995). The stomach accomplishes these

tasks by the interaction of two distinct functional motor regions, a proximal and a

distal region (Figure 1.1). The proximal stomach comprises the fundus and the upper

one-third of the corpus, and the distal stomach consists of the remaining two-thirds of

the corpus, the antrum and the pylorus. The division between proximal and distal

regions has been established by myoelectrical and motor criteria (Kelly 1980).

Transpyloric passage of gastric contents is predominantly pulsatile, rather than

continuous (King et al. 1984, Malbert and Ruckebusch 1991). As described by

Cannon, in 1911, most liquefied chyme is propelled into the duodenum as a series of

small gushes (Cannon 1911). Both forward, interrupted (antegrade) and reverse

(retrograde) flow can occur and no single motor component is believed to exert

dominant control over gastric emptying, but rather an integration of motor activity

from the fundus, corpus, antrum, pylorus and duodenum is believed to be primarily

responsible (Hausken et al. 1992, Horowitz et al. 1994, Malbert and Mathis 1994).

Chapter 1

3

Figure 1.1: Representation of the distinct anatomical and functional motor regions of the stomach with an outline of motor events during normal gastric emptying. Adapted from Rayner and Horowitz (2005).

1.2.1 Proximal stomach

The proximal stomach is primarily responsible for the receipt and storage of food.

When food is ingested, there is an initial “receptive” relaxation of the proximal

stomach, triggered in part by the act of swallowing, which reduces gastric tone. This

is followed by a more prolonged “adaptive” relaxation, known as accommodation,

which occurs in response to gastric distension and is mediated by mechanoreceptors

in the gastric wall (Cannon and Lieb 1911) (Figure 1.1). Thus, intragastric pressure is

maintained at a low level, even during food ingestion (Azpiroz and Malagelada

1987). Accordingly, intragastric pressure rises by less than 10 mmHg when the

stomach is filled with 2 litres of fluid (Mariani et al. 2004). The myoelectrical activity

Chapter 1

4

of the proximal stomach is characterised by a slow depolarisation of membrane

potential of small amplitude, which triggers a sustained, or “tonic” contraction. It is

these sustained tonic contractions of the proximal stomach which are responsible for

the regulation of intragastric pressure and the gastroduodenal pressure gradient. Both

receptive and adaptive relaxations are affected principally by noncholinergic,

nonadrenergic, and partially dopaminergic, vagal neurons which act dominantly on

the oblique muscle layer of the proximal stomach (Christensen and Torres 1975). The

tonic intragastric pressure of the proximal stomach appears to have a major influence

on the gastric emptying of low-nutrient liquids, but has only a minor role in the

gastric emptying of solids (Collins et al. 1991).

1.2.2 Distal stomach

In contrast to the proximal stomach, the distal stomach is associated with distally

propagated, cyclical changes in membrane potential which mark the onset of

peristaltic waves; thus, the distal stomach is associated with phasic, rather than tonic,

motor activity (Kelly 1980). Peristaltic waves are circular rings of contraction that

sweep aborally through the distal gastric wall and are thought to originate in the

interstitial cells of Cajal. Their amplitude, frequency and velocity of propagation are

determined by regular cyclical changes in electrical potential called slow waves. The

gastric pacemaker, situated on the greater curvature of the stomach (Figure 1.1),

depolarises at the fastest rate, of approximately 3 cycles per minute in the human, and

controls the rate of contractions in the distal stomach (Kelly 1980). The main effects

of the peristaltic contractions are to mix gastric chyme with gastric juice before

propulsion through the pylorus. Solids are initially retained in the antrum, triturated

into small particles less than 1 mm in size and mixed with gastric juice before they

are allowed to pass through the pylorus and into the duodenum. This antral peristalsis

Chapter 1

5

occurs via a repeated sequence of propulsion, squeezing and retropulsion in the distal

stomach (Figure 1.1). The distal stomach was originally thought to bear the greatest

influence on gastric emptying of solids with minimal effect on gastric emptying of

liquids, however, it is now known that the distal stomach and antral tone also play a

role in the regulation of gastric emptying of liquids (Hveem et al. 1996).

The pattern of contractile activity differs from the fasting to the fed state. In the

fasting state, a cycle of myoelectrical activity, termed the migrating motor complex

(MMC), occurs approximately every 100 minutes. The MMC, the stomach’s

“housekeeper”, allows larger nondigestible food particles to be emptied from the

stomach (Coupe et al. 1991). The MMC consists of three phases: phase I, a period of

motor quiescence, lasting approximately 40 minutes; phase II, characterised by

irregular contractions, lasting approximately 50 minutes; and phase III, consisting of

regular, high amplitude contractions at the maximal rate of 3 per minute for a much

shorter period of 5 – 10 minutes. It is during late phase II and phase III that larger

indigestible particles are emptied from the stomach into the duodenum and this is

thought to be mediated by antropyloric relaxation (Coupe et al. 1991, Smith and

Ferris 2003).

Following ingestion of food, the MMC has a different pattern of contractile activity;

during the fed state, proximal tonic contractions increase and contractions in the

antrum become irregular and of variable amplitude, duration and frequency. The fed

pattern is induced 5 – 10 min after meal ingestion, reaches its peak after 10 – 20 min

and persists as long as food is present in the stomach (Horowitz et al. 2002b).

Chapter 1

6

1.2.3 Pylorus

The pylorus, or gastroduodenal junction, lies at the junction of the antrum and the

duodenal bulb and acts as a brake, preventing passage of large (> 2 mm) particles

from the stomach into the duodenum. Both tonic and phasic contractions occur in the

pylorus across a narrow zone (< 4 mm) (Heddle et al. 1988c) (Figure 1.1).

1.3 Patterns of gastric emptying

The patterns of gastric emptying are dependent on meal composition and physical

properties; solids, high-nutrient liquids, low and non-nutrient liquids and fats all

empty from the stomach at different rates, at least in part due to differences in their

caloric density (Smith and Ferris 2003). As described by Nelsen and Kohatsu (1971),

the rate of gastric emptying (dv/dt) is a function of the difference in pressure between

the stomach (PS) and the duodenum (PD) and the resistance to flow across the pylorus

(RP) (Nelsen and Kohatsu 1971):

The stomach empties its contents only when the pressure gradient between the

stomach and duodenum is sufficient to overcome the resistance to flow across the

pylorus. Since the resistance to transpyloric flow of low-nutrient liquids is small, the

emptying rate of liquids is dependent primarily on the pressure gradient between the

stomach and duodenum, and as proximal stomach contractions are the main

regulators of intragastric pressure, the emptying of liquids may, accordingly, be

largely dependent on the proximal stomach (Kelly 1980).

Chapter 1

7

In contrast, the resistance to flow of solids and high-nutrient liquids across the

pylorus is large. The magnitude of this resistance is controlled by antral and pyloric

contractions and as such, they play a major role in the regulation of gastric emptying

of solids (Kelly 1980).

1.3.1 Solids

Following meal ingestion, digestible solids are retained within the stomach before

they begin to empty into the duodenum. The interval before emptying commences

has been termed the “lag phase” and is characteristically 20 – 60 minutes in duration

(Collins et al. 1983). During the lag phase, digestible solids are initially retained in

the proximal stomach before being redistributed to the antrum, where they are

triturated into fine particles and converted into chyme through mixing with gastric

acid and enzymes. This is followed by an emptying phase that approximates a linear

pattern (Figure 1.2). Meyer et al. (1981) demonstrated that most digestible solid

particles are less than 1 mm in diameter before they empty into the duodenum (Meyer

et al. 1981). Accordingly, digestible solids ingested in homogenised form empty

more quickly than those ingested as chunks (Hinder and Kelly 1977). Thus, the

trituration of food into fine particles is a major rate-limiting factor in the emptying

rate of solids from the stomach (Lin et al. 1992b).

The rate of solid emptying is also influenced by the liquid with which it is consumed.

Nutrient-containing liquids prolong the emptying rate of solids when consumed

concurrently (Horowitz et al. 1989a, Houghton et al. 1988). The gastric emptying rate

of homogenised liver mixed with 1% dextrose has been compared with that of

homogenised liver mixed with 10% dextrose; in both circumstances, homogenised

Chapter 1

8

liver emptied at the same rate as the liquid in which it was mixed (Hinder and Kelly

1977). Furthermore, the volume of liquid consumed concurrently with solid affects

the rate of solid emptying. In a mixed solid-liquid meal, approximately 80 % of the

liquid component of the meal empties before the solid (Horowitz et al. 1989a),

indicating that the stomach has the ability to discriminate between solids and liquids.

Figure 1.2: Scintigraphic gastric emptying curves for solid (100 g minced beef), semisolid / high-nutrient liquid (porridge / dextrose 25 %w/v) and low-nutrient liquid (beef soup). Solid, semisolid and high-nutrient liquid gastric emptying curves are characterised by a lag phase followed by a linear emptying phase, while low/non-nutrient liquids empty in a monoexponential fashion with minimal lag phase.

Indigestible solids usually empty during late phase II and phase III of the MMC

(Coupe et al. 1991). Following concurrent consumption of digestible and indigestible

solids, emptying of the digestible solids is delayed due to induction of a fed pattern

and resultant abolition of phase III contractions. It is only when digestible solids have

been completely emptied from the stomach that a fasting pattern returns and phase III

Chapter 1

9

waves are able to clear indigestible debris from the stomach (Coupe et al. 1991, Mroz

and Kelly 1977). Studies (Coupe et al. 1991, Hinder and Kelly 1977) have

demonstrated that 7 mm indigestible, plastic spheres do not empty from the stomach

before digestible solids and liquids, but do so during phase III “housekeeping”

contractions. However, recent evidence indicates that gastric emptying of larger,

indigestible solids up to a particle size of 3 – 4 mm, and perhaps even 7 mm, can

occur unrelated to phase III activity (Stotzer and Abrahamsson 2000).

1.3.2 Liquids

In comparison to solids, liquids empty much more rapidly with a minimal lag phase.

Non-nutrient and low-nutrient liquids empty fastest from the stomach and do so in a

monoexponential fashion (Figure 1.2). Posture and intragastric volume influence the

emptying of non- and low-nutrient liquids, such that the volume emptied over time is

proportional to the volume remaining in the stomach. In contrast, high-nutrient

liquids empty initially rapidly and then in a linear pattern, delivering nutrients to the

duodenum at an overall rate of 2 – 3 kcal/minute (Brener et al. 1983). The differences

in gastric emptying of non- and low-nutrient versus high-nutrient liquids can be

accounted for by small intestinal feedback inhibition (Lin et al. 1989). Accordingly,

posture and intragastric volume play relatively little roles in the gastric emptying of

high-nutrient liquids.

The distal stomach was traditionally believed to be primarily responsible for the

gastric emptying of solids, owing largely to its role in grinding and triturating ingesta

into fine particles, with minimal effect on gastric emptying of liquids (Kelly 1980).

There is now evidence to indicate that the distal stomach and antral tone also play a

role in the regulation of gastric emptying of liquids (Hveem et al. 1996).

Chapter 1

10

1.3.3 Fats

It is generally accepted that fats empty more slowly than non- or low-nutrient liquids

(Carney et al. 1995, Edelbroek et al. 1992, Horowitz et al. 1993a). This may be

explained by the fact that the rate of gastric emptying is determined to some extent by

the caloric density of the nutrient ingested. While fat (with a caloric density of ~ 9

kcal/g) is more calorie-rich than carbohydrate (with a caloric density of ~ 4 kcal/g),

when given at isocaloric concentrations (i.e. 4 g fat and 9 g carbohydrate are

isocaloric at 36 kcal), both nutrients will slow gastric emptying to a comparable

extent (Hunt and Stubbs 1975).

Chang et al. (1968) proposed that fat, due to its low density and high viscosity, has

the capacity to form a layer on top of aqueous meal components, to explain its slower

gastric emptying rate, while Houghton et al. (1990) attributed this observation to

retrograde movement of fat from distal to proximal stomach (Chang et al. 1968,

Houghton et al. 1990). It has been established that when fat is ingested as an oil phase

it layers on top of other, more dense, meal components and, accordingly, in the sitting

or erect posture, empties after them (Carney et al. 1995, Edelbroek et al. 1992,

Horowitz et al. 1993a). Hunt & Knox (1968) proposed that the gastric emptying of

fat, much like other nutrients, is dependent on lipolysis of fats into fatty acids and the

interaction of these digestion products with receptors in the small intestinal lumen

(Hunt and Knox 1968). Like carbohydrate, the magnitude of small intestinal feedback

inhibition that induces a slowing of gastric emptying is dependent on the number

(length of intestine) and location of receptors exposed (Lin et al. 1990). Medium-

chain fatty acids with 12 – 20 carbon atoms have been reported to be the most

effective inhibitors of gastric emptying (Hunt and Knox 1968). By modulating fat

digestion, lipase plays an important role in the regulation of gastric emptying of fat

Chapter 1

11

(Borovicka et al. 2000, Degen et al. 2006, Schwizer et al. 1997b). Orlistat, a potent

and specific lipase inhibitor, accelerates gastric emptying of high fat meals in healthy

volunteers (Borovicka et al. 2000) and patients with type 2 diabetes (O'Donovan et

al. 2004a, Pilichiewicz e t a l . 2003). In patients with exocrine pancreatic

insufficiency, gastric emptying of fat is abnormally rapid as a result of impaired

lipolysis of fat and, hence, a diminished small intestinal feedback response (Carney et

al. 1995, Smith et al. 1990).

1.4 Determinants of gastric emptying

1.4.1 Small intestinal feedback inhibition

The major factor controlling gastric emptying of nutrients is feedback inhibition from

chemoreceptors located in the small intestinal lumen (Horowitz and Dent 1991). The

receptors are distributed throughout the small intestine with variations in the number

and type between the duodenum, jejunum and ileum. Neural and hormonal

mechanisms are involved in triggering the process. The magnitude of feedback

inhibition is dependent on both the number and the site of small intestinal receptors

that are exposed (Lin et al. 1989) and is affected by previous nutrient exposure, e.g.

prior intake of a high-glucose diet for 4 - 7 days accelerates the gastric emptying rate

of glucose and fructose in normal, healthy subjects (Horowitz et al. 1996a) and,

likewise, gastric emptying rate of fat is more rapid following a high-fat diet

(Cunningham et al. 1991). As discussed (Chapter 1.3.3), fat-induced retardation of

gastric emptying is dependent on lipase-induced hydrolysis of fats to fatty acids.

Chapter 1

12

1.4.2 Posture

Gastric emptying of solids and liquids is influenced by gravity. Non- and low-nutrient

liquids, which are subject to minimal small intestinal feedback inhibition, empty

faster from the stomach when patients are lying on their right side, rather than in a

sitting position or lying on their left side (Burn-Murdoch et al. 1980). Conversely,

emptying of high-nutrient liquids, which activate feedback inhibition pathways from

the small intestine, are not influenced significantly by posture. After a gastric

drainage procedure (e.g. pyloroplasty or gastric jejunostomy), the small intestinal

braking pathways to the stomach are disrupted and hence, liquid emptying rates are

more rapid in the sitting rather than supine position (McKelvey 1970). Posture also

has a major effect on the intragastric distribution of oil when ingested as a mixed

oil/aqueous meal, but has relatively little effect on gastric emptying of the oil phase

(Horowitz et al. 1993a).

1.4.3 Meal temperature

The temperature of the ingested meal influences the gastric emptying rate;

temperatures that are above or below core body temperature retard gastric emptying

(Kelly 1980). Cold (4 °C) and warm (50 °C) drinks, when compared with a 37 °C

drink, have been shown to suppress and alter the organisation of antral pressure

waves, and stimulate isolated pyloric pressure waves in healthy humans, a pattern of

motility associated with slowing of transpyloric flow (Sun et al. 1995).

1.4.4 Meal volume

The meal volume can significantly alter the gastric emptying rate, particularly for

non- and low-nutrient liquids, where larger volumes accelerate gastric emptying rates

(expressed as millilitres emptied/minute) (Lin et al. 1992a). This is not quite the case

Chapter 1

13

for high-nutrient liquids, in which the rate of emptying is determined largely by small

intestinal feedback inhibition (Lin et al. 1989). Gastric emptying of solids has been

reported to be influenced by meal volume, such that larger volumes are associated

with slightly more rapid emptying. Posture influences intragastric distribution of

solids, but does not appear to affect the rate of emptying (Doran et al. 1998).

1.4.5 Gastrointestinal peptides

A number of peptides are released from the stomach and small intestine in response

to the ingestion of nutrients and these include glucagon-like peptide-1 (GLP-1)

(Barnett 2008, Nauck et al. 1993, Nauck et al. 1997b), glucose-dependent

insulinotropic peptide (GIP) (Meier et al. 2004, Nauck et al. 1993), cholecystokinin

(CCK) (Rayner et al. 2000b), peptide tyrosine tyrosine (PYY) (Savage et al. 1987)

and ghrelin (Kojima et al. 1999). They have a variety of effects on gastrointestinal

motility, which are discussed briefly below.

1.4.5.1 GLP-1 and GLP-2

GLP-1 and GLP-2 are produced predominantly in mucosal L-cells of the distal ileum

and colon, as a result of proteolytic cleavage of proglucagon. GLP-1 is expressed in

two bioactive forms of equal potency; predominantly GLP-1(7-36) amide and, to a

lesser extent, GLP-1(7-37) amide (Barnett 2008, Orskov et al. 1993). GLP-1 is

secreted in response to lipid (Feinle et al. 2003), carbohydrate (Nauck et al. 1993)

and protein (Herrmann et al. 1995). Exogenous GLP-1 has been shown to enhance

insulin secretion (Nauck et al. 1993, Nauck et al. 1997b) and suppress postprandial

glucagon secretion in a glucose-dependent manner (Nauck et al. 1993, Ritzel et al.

1995), retard gastric emptying (Flint et al. 2001, Little et al. 2006a, Naslund et al.

1999, Schirra et al. 2000) and suppress energy intake (Brennan et al. 2005, Flint et al.

Chapter 1

14

1998), collectively influencing glycaemia, particularly postprandially. Effects of

exogenous GLP-1 specific to gastrointestinal motility include fundic relaxation

(Delgado-Aros et al. 2002, Schirra et al. 2000), inhibition of antroduodenal motility

(Schirra et al. 2000) and stimulation of tonic and phasic contractions of the pylorus

(Schirra et al. 2000). In addition to delaying gastric emptying, exogenous GLP-1 has

been shown to inhibit gastric acid secretion and pancreatic exocrine secretion in

humans (Schirra and Goke 2005). Moreover, preclinical data suggest that GLP-1 may

also increase –cell mass and slow –cell apoptosis (Drucker 2003).

Studies using the potent and specific GLP-1 receptor antagonist, exendin(9-39), have

demonstrated that endogenous GLP-1, too, has a role in postprandial glycaemic

response and antropyloroduodenal motility (Edwards et al. 1999, Schirra and Goke

2005). Exendin(9-39) has been shown to increase both fasting blood glucose

(Edwards et al. 1999, Schirra et al. 2006) and postprandial glycaemia (by ~ 35 %)

following oral (Edwards et al. 1999), and intraduodenal (Schirra et al. 2006), glucose.

Exendin(9-39) also increases fasting and fed glucagon secretions, with inconsistent

effects on insulin (D'Alessio et al. 1996, Edwards et al. 1999, Kolligs et al. 1995,

Schirra et al. 2006). Exendin(9-39), when given by intravenous administration, has

been shown to inhibit antroduodenal motility and stimulate pyloric motility in

humans (Schirra et al. 2006, Schirra et al. 2009). The effects of exendin(9-39) on

gastric emptying are less well understood. In rats, intravenous exendin(9-39) has been

demonstrated to block GLP-1-induced inhibition of gastric emptying (Chelikani et al.

2005), and in mice, an acceleration of gastric emptying has been reported during

intraperitoneal exendin(9-39) (Kumar et al. 2008). There has been only one human

study investigating the effects of exendin(9-39) on gastric emptying, which reported

no difference in gastric emptying of an oral glucose load, although a relatively

Chapter 1

15

insensitive technique was used to measure gastric emptying (Salehi et al. 2008).

Chapter 5 evaluates the effect of exendin(9-39) on gastric emptying, measured by the

‘gold standard’ scintigraphy, and glycaemia, in healthy subjects.

GLP-2 exerts trophic effects on the gut mucosa (Drucker et al. 1996), however, it

does not appear to have any effect on gastric emptying or satiety (Schmidt et al.

2003).

1.4.5.2 GIP

GIP is secreted by K cells of the proximal small intestine and stimulates insulin

secretion, regulates fat metabolism and is thought to stimulate glucagon secretion

(Barnett 2008), without effects on gastric emptying or energy intake (Meier et al.

2004).

1.4.5.3 CCK

CCK, which exists in several bioactive molecular forms ranging from 4 to 58 amino

acid residues, is released from enteroendocrine I-cells of the duodenum and jejunum,

in response to the presence of nutrients (fat and protein in particular) in the small

intestine in a load-dependent manner (Hellström and Naslund 2001, Rayner et al.

2000b). CCK slows gastric emptying (Kleibeuker et al. 1988, Liddle et al. 1986,

Yamagishi and Debas 1978), inhibits antral contractions (Fraser et al. 1993, Schwizer

et al. 1997a), stimulates phasic and tonic contractions of the pylorus (Brennan et al.

2005, Fraser et al. 1993) and suppresses energy intake (Rayner et al. 2000b, Stacher

et al . 1982). Studies with the CCK receptor antagonist, loxiglumide, have

demonstrated that endogenous CCK has a physiological role in the regulation of

gastric emptying and appetite (Rayner et al. 2000b). Gastric emptying of a mixed

Chapter 1

16

solid-liquid meal was accelerated following loxiglumide administration (Borovicka et

al. 1996, Chua et al. 1994).

1.4.5.4 PYY

PYY is secreted in two forms, PYY1-36 and, to a greater extent, PYY3-36, a 34 amino

acid peptide. PYY is produced by enteroendocrine L-cells throughout the entire, but

predominantly more distal, gastrointestinal tract. Exogenous PYY reduces secretions

from the stomach and pancreas, inhibits gallbladder contractions and slows gastric

emptying (Savage et al. 1987).

1.4.5.5 Ghrelin

Ghrelin is a 28 amino acid peptide, produced as a result of cleavage of preproghrelin

(Kojima et al. 1999). It is produced predominantly in X/A-like or “ghrelin” cells of

the stomach, in particular the parietal cells of the fundus, and to a lesser extent

throughout the length of the small intestine (Wren and Bloom 2007). In contrast to

other gastrointestinal peptides, ghrelin expression is suppressed in response to meal

ingestion (Monteleone et al. 2003). When given exogenously, ghrelin has been shown

to stimulate hunger and appetite (Levin et al. 2006) and accelerate gastric emptying

in healthy humans (Levin et al. 2006), diabetic gastroparesis (Murray et al. 2005) and

idiopathic gastroparesis (Tack et al. 2005).

Chapter 1

17

1.5 Conclusions

The stomach is a complex organ with sophisticated function. Its anatomical structure

is that of a simple sack, yet this highly structured organ with distinct anatomical and

motor regions of diverse function, is capable of tight control of delivery of nutrients

from the stomach and into the small intestine. The patterns of gastric emptying are

determined largely by the composition of the ingested material and their associated

physical properties, and the factors which determine the gastric emptying rate are

complex, interrelated and numerous. Chapter 5 reports the effects of the GLP-1

receptor antagonist, exendin(9-39), on gastric emptying and glycaemia in healthy

humans.

Chapter 2

18

2Chapter 2MEASUREMENT OF GASTRIC MOTOR FUNCTION

2.1 Introduction

The first documented measurements of gastric motility were performed in the 13th

century by the Holy Roman Emperor Frederick II (1194 – 1250) using a less than

sophisticated technique that would prove unacceptable to ethics committees today. -

He ordered that two men be fed an identical, large meal and then sent one hunting and

ordered the other to rest in order to determine under which condition digestion is

aided. Frederick II had both men executed and disembowelled; he discovered that the

stomach of the rested man was empty while that of the man who had been hunting

was full (Sanaka et al. 2000).

Chapter 2

19

There are now a number of sophisticated methods to evaluate gastric motility and

novel techniques continue to emerge (Table 2.1). These include stable isotope breath

tests, magnetic resonance imaging, radiopaque marker techniques, electrical

impedance, scintigraphy, ultrasonography, absorption kinetics of orally administered

drugs, manometry, barostat and electrogastrography, and may be broadly categorised

into: (i) measurement of gastric emptying, (ii) measurement of intragastric pressure,

and (iii) measurement of gastric electrical activity. The application of these

techniques has been fundamental to the increased understanding of the physiology,

pathogenesis, diagnosis and management of both normal and disordered gastric

motility. This chapter will discuss the techniques currently in use in the evaluation of

gastric motility and their associated advantages and disadvantages, with the major

emphasis on gastric emptying.

2.2 Measurement of gastric emptying

The emergence of novel techniques in the measurement of gastric emptying has

increased our understanding of normal and disordered gastric physiology

considerably. A number of techniques to quantify gastric emptying are currently in

use and include scintigraphy, ultrasonography, stable isotope breath tests, magnetic

resonance imaging, radiopaque marker techniques, applied potential tomography and

absorption kinetics of orally administered drugs (Table 2.1). Accurate, reproducible

and sensitive tests are available in the research setting, and application of

standardised techniques to the clinical setting is also necessary to ensure the

diagnostic worth of such tests. A number of factors, which may affect the gastric

emptying rate, must be standardised and are discussed in more detail within the

context of scintigraphy, the ‘gold standard’ measure of gastric emptying. Such factors

include, but are not limited to, composition of the test meal, meal ingestion time,

Chapter 2

20

study duration, subject position, glycaemic control, duration of fasting and

concomitant medication (Abell et al. 2008b, Abell et al. 2008a).

Table 2.1 Methods in the assessment of gastric motor function

(i) Measurement of gastric emptyingScintigraphyUltrasonographyStable isotope breath testsMagnetic resonance imagingSingle photon emission computed tomographyRadiopaque marker techniquesApplied potential tomography / Epigastric impedanceAbsorption kinetics of orally administered drugs

(ii) Measurement of intragastric pressure and contractionsManometryBarostatUltrasonographyStrain rate imaging

(iii) Measurement of gastric electrical activityElectrogastrography

(iv) Measurement of gastroduodenal flowImpedanceUltrasonography

Chapter 2

21

2.2.1 Scintigraphy

Gastric scintigraphy was introduced in 1966 by Griffith et al. (Griffith et al. 1966)

and currently remains the ‘gold standard’ for the measurement of gastric emptying. It

represents the most physiologic, quantitative, noninvasive test currently available and

remains the most extensively studied and available method for measurement of

gastric emptying rates in both research and clinical practice (Bratten and Jones 2006,

Camilleri et al. 1998, Horowitz and Dent 1991, Kim et al. 2000).

Scintigraphy involves the incorporation of radiopharmaceuticals into solid, liquid, or

mixed solid/liquid meals that are subsequently ingested. Gastric emptying of solid

and liquid meals may be measured individually or, preferably, simultaneously, as

there is a relatively poor correlation between solid and liquid gastric emptying

(Collins et al. 1983, Horowitz et al. 1991, Jones et al. 1995b). The equipment

required for scintigraphic imaging consists of an external gamma camera with a large

field of view, in front of which patients usually sit, that is linked to a dedicated

computer with appropriate software for analysis. The gamma camera, fitted with a

collimator appropriate for the energy level at which the radioisotope emits, quantifies

the abdominal distribution of radioactivity over time and, hence, the rate at which the

stomach empties the radiolabelled meal. Data may be acquired dynamically from the

posterior aspect every minute for the first hour and every three minutes thereafter, for

2 – 4 hours duration, or via static anterior and posterior images obtained either

sequentially or simultaneously (via the use of a dual head gamma camera). Using

appropriate computer software, regions-of-interest are drawn around the total

stomach, which is subsequently divided into proximal and distal regions, so that the

total gastric emptying rate, as well as intragastric distribution, may be quantified

(Collins et al. 1983) (Figure 2.1). The number of counts in each region are expressed

Chapter 2

22

as “percentage retention” and plotted over “time”, thus producing a gastric emptying

curve, which provides a convenient means of expressing data, particularly when

compared against a normal range. Several parameters are commonly reported,

including the lag phase (Tlag) which represents the time period before the stomach

starts to empty and the percentage of gastric retention at specific time points post-

ingestion. For solids, the percentage of gastric retention at 100 min (T100) is a

frequently used parameter; for liquids, the time taken for half of the liquid component

of the meal to empty (T50) is often used (Collins et al. 1983, Horowitz et al. 1991,

Jones et al. 1995b).

Figure 2.1: Scintigram of stomach (posterior view), divided into total, proximal and distal regions, following ingestion of 100 g minced beef labelled with 20 MBq 99mTc-sulphur colloid.

Patients must fast for at least 8 hours, and, ideally, cease any medications which

affect gastrointestinal motility (including prokinetics, H2 receptor antagonists,

anticholinergics, narcotic analgesics, tricyclic antidepressants) for at least 48 hours,

Chapter 2

23

prior to the gastric emptying scan. Tobacco and alcohol must also be withheld for at

least 24 hours prior to imaging. Women should be studied preferably during the first

10 days of the menstrual cycle, so as to avoid hormonal effects on gastric motility

(Mariani et al. 2004). Naturally, such requirements are applicable to all techniques in

the measurement of gastric emptying.

The test meal used to evaluate gastric emptying should be of a standard size, volume,

caloric content and temperature as it is well recognised that these factors can

substantially influence the gastric emptying rate (Doran et al. 1998, Kelly 1980, Lin

et al. 1989) (Chapters 1.3, 1.4.3, 1.4.4). Patients must also be scanned in the same

position, usually sitting, as it is known that posture can influence the gastric emptying

rate, particularly that of non- and low-nutrient liquids (Burn-Murdoch et al. 1980)

(Chapter 1.4.2). It is also well established that changes in the blood glucose

concentration can have a significant impact on the gastric emptying rate (Fraser et al.

1990, Schvarcz et al. 1993, Schvarcz et al. 1997). This is of particular significance in

patients with diabetes and, accordingly, at a minimum, their blood glucose

concentrations should be measured both pre-ingestion and during the course of the

post-ingestion imaging time (Chapters 3.3.2, 3.4.3).

The solid component of the test meal is frequently radiolabelled with 99mTechnetium-

sulphur colloid (99mTc-sulphur colloid), mostly due to its low cost and wide

availability. Meyer et al. (1976) was the first to describe a radiolabelling method

using chicken liver suitable for gastric emptying studies. 99mTc-sulphur colloid was

injected intravenously into a chicken (in vivo) via a wing vein. Following uptake by

the Kupffer cells of the liver, the chicken was sacrificed; the radiolabelled liver was

excised and the counts measured (Meyer et al. 1976). This method yields a highly

Chapter 2

24

stable label; 97 % remains bound in gastric juice at 4 hours in vitro (Christian et al.

1984). The liver can thus be cut to a size that provides 20 MBq 99mTc-sulphur colloid

chicken liver, minced, mixed with 100 g minced beef, dry-fried and administered to

the patient (Collins et al. 1983). While this technique represents the ‘gold standard’, it

is neither practical nor convenient for routine use. Alternative methods for labelling

solids have been devised including injecting 99mTc-sulphur colloid into egg yolk or

albumin and mixing this with the resultant solid test meal (Shuter and Ng 2005),

however, the stability of the radiolabelled solid with such alternative measures is

reduced; the percentage bound to gastric juice at 3 hours in vitro is 82 % for 99mTc-

sulphur colloid whole egg and 95 % for 99mTc-sulphur colloid albumin (Knight 1996,

Mariani et al. 2004). There have been a number of other radiopharmaceuticals and

foods devised to measure gastric emptying and these have been found to be less

stable than in vivo 99mTc-sulphur colloid chicken liver; examples include Technegas-

labelled rice (Kwiatek et al. 1999), 99mTc-sulphur colloid mashed potato (Gentilcore

et al. 2006a, O'Donovan et al. 2004a), 99mTc-sulphur colloid pancakes (Borovicka et

al. 1996, Kreiss et al. 1998), 99mTc-sulphur colloid mushrooms (Frier and Perkins

1994), 99mTc(V)Thiocyanate-labelled oil (Carney et al. 1995, Cunningham et al.

1991, Edelbroek et al. 1992, Horowitz et al. 1993a, Pilichiewicz et al. 2003), 99mTc-

albumin colloid scrambled eggs (Taillefer et al. 1987) and 99mTc-labelled resin-

oatmeal (Domstad et al. 1980). While there is now available a radiolabel for almost

all food types (e.g. solids, liquids, semisolids, fats), careful consideration must be

given to the radioisotope used and its binding characteristics to the chosen meal.

More specifically, the radioisotope may separate from the solid component of the

meal and instead bind with the liquid phase, leading to falsely faster rates of solid

emptying (Kim et al. 2000). For example, it is known that 99mTc-sulphur colloid binds

Chapter 2

25

with greater affinity to egg than does 99mTc-DTPA or pertechnetate (Kim et al. 2000,

Thomforde et al. 1995).

The liquid phase of the test meal may be labelled with 67Gallium which has been

chelated with either ethylenediaminetetraacetic acid (67Ga-EDTA) or

diethylenetriamine penta-acetic acid (67Ga-DTPA) (Bellen et al. 1995, Collins et al.

1983). Alternative isotopes include 111Indium (as 111In-DTPA or 111In-oxine)

(Simonian et al. 2004) and 113mIn-DTPA (Collins et al. 1991, Edelbroek et al. 1993,

Jones et al. 1995b). Solid and liquid emptying may, therefore, be evaluated

concurrently via the use of two different radioisotopes of differing energy (Tc for the

solid; Ga/In for the liquid). For single isotope studies of liquid emptying, 99mTc-

sulphur colloid (Bellen et al. 1995, Collins et al. 1983) and 99mTc-DTPA (Lin et al.

2005, Malagelada et al. 1984) have been used as the radiolabel. Concurrent solid and

liquid emptying via dual isotope imaging is made possible by the differing physical

characteristics of the radioisotopes used. For example, 99mTc has a half-life of 6 hours

and a gamma energy peak of 140 keV; 67Ga has a half-life of 78 hours and three

gamma energy peaks of 90 keV, 185 keV and 296 keV. Thus, it is paramount to use a

low energy collimator when imaging 99mTc and a medium energy collimator when

imaging 67Ga alone or in conjunction with 99mTc.

While scintigraphy has several advantages over other methods, its major

disadvantages include its associated radiation burden (the radiation dose

approximates that received from a single abdominal radiograph), which precludes

studies during pregnancy and lactation and poses relevant risks to children requiring

the scan. Accordingly, there are limits to the dose of radiation, which, in turn, restricts

the number of scans permitted, posing diagnostic difficulty for patients requiring

Chapter 2

26

multiple scans, although, in most cases, the benefit of clinical diagnosis outweighs

the potential associated risk. Scintigraphy is inconvenient (taking two to six hours to

perform), requires access to a gamma camera, is relatively expensive, labour-

intensive and is time-consuming for both patients and technicians (Horowitz and

Dent 1991). Images must be corrected for tissue attenuation, isotope decay, patient

movement and down-scatter of energy from isotopes during dual isotope studies.

Tissue attenuation may be corrected by calculating geometric means (square root

[anterior x posterior counts]) (Moore et al. 1985), which represents the ‘gold

standard’ method, or by the acquisition of a left lateral image at the end of the study if

acquiring with only anterior or posterior dynamic images (Collins et al. 1983). Left

anterior oblique images have been used to reduce tissue attenuation (Maurer et al.

1991), however, this may represent the least preferred method due to large

interindividual differences in the anatomic position of the stomach.

There was, until recently, a lack of standardisation across different centres,

particularly with respect to the composition and volume of the test meal, the type of

gamma camera employed, the posture of the subject, the duration of data acquisition

and correction factors used, and resultantly, the calculation of gastric emptying rates,

thereby hindering comparisons of studies performed between different centres and

limiting interpretation of results (Kim et al. 2000). Tougas et al. (2000) reported a

simplified scintigraphic technique in the assessment of gastric emptying that used

four static scans over 4 hours, following consumption of a 99mTc-sulphur colloid-

labelled low-fat Egg-Beater meal, in a multicentre setting. Results were found to

compare favourably with conventional, more labour-intensive, dynamic scintigraphic

imaging. Normal values of gastric retention (95th percentile) were defined as < 90 %

at 1 hour, < 60 % at 2 hours and < 10 % at 4 hours. Gastric retention > 10 % at 4

Chapter 2

27

hours was reported to be indicative of gastroparesis (Tougas et al. 2000). A number

of groups (Lin et al. 2005, Mariani et al. 2004, Stanghellini et al. 2000, Tougas et al.

2000, Ziessman et al. 2004) have published acceptable standards for the measurement

of gastric emptying, including the joint consensus reached by the American Motility

Society and the European Society of Neurogastroenterology and Motility (Lin et al.

2005). More recently, members of the American Neurogastroenterology and Motility

Society and the Society of Nuclear Medicine published a consensus statement of

recommendations for a standardised method of measuring gastric emptying by

scintigraphy, which are in accordance with those described by Tougas et al. (2000)

(Abell et al. 2008b, Abell et al. 2008a).

2.2.2 Ultrasonography

Transabdominal ultrasonography has been developed relatively recently as an

alternative technique in the measurement of gastric emptying (Bateman and

Whittingham 1982, Bolondi et al. 1985, Duan et al. 1993, Holt et al. 1986).

Ultrasonography represents a safe, non-radioactive, non-invasive, inexpensive,

widely available, portable method for measuring gastric emptying (Bateman and

Whittingham 1982, Bolondi et al. 1985, Duan et al. 1993, Gentilcore et al. 2006a,

Gilja et al. 1995a, Gilja et al. 1996, Gilja et al. 1997, Gilja et al. 1999, Hausken et al.

1991, Holt et al. 1986, Tefera et al. 2002).

Transabdominal two-dimensional (2D) ultrasonography of the stomach has been used

to study gastric emptying (Bateman and Whittingham 1982, Bolondi et al. 1985, Holt

et al. 1986, Hveem et al. 1996, Ricci et al. 1993), as well as gastric contractions and

propagation of waves (Ahluwalia et al. 1994, Bateman et al. 1977, Hausken et al.

1991, Hausken and Berstad 1992, Holt et al. 1980), gastroduodenal flow and velocity

Chapter 2

28

(Hausken et al. 1992, King et al. 1984), strain rate imaging (Uematsu et al. 1997) and

gastric accommodation (Gilja et al. 1995b). In order to measure the gastric emptying

rate, parasagittal sections of the antrum with the aorta and mesenteric vein in the field

of view, serving as landmarks, are recorded (Bolondi et al. 1985). A region-of-

interest drawn around the circumference of the antrum at various time points enables

calculation of antral area at each time point. The fasting antral area is subtracted from

the antral area at each time point post-meal ingestion. Gastric emptying (“percent

retention”) at any time point is calculated as the antral area at any time point as a

percentage of the maximum antral area (Hveem et al. 1996). 2D ultrasonography has

been validated against the ‘gold standard’, scintigraphy, in the measurement of

gastric emptying of both low- and high-nutrient drinks in healthy young subjects

(Hveem et al. 1996) (Figure 2.2). However, the technique is associated with some

limitations, including the associated difficulty in imaging obese patients and in those

patients with a significant amount of bowel gas, it is operator-dependent, technically

demanding and can only be used to quantify the gastric emptying rate of specific

meals, predominantly liquids. Error may also be introduced when drawing regions-of-

interest around the assumed shape of the antrum (Gilja et al. 1996).

Figure 2.2: Parasagittal 2D ultrasonographic image of the antrum (indicated by arrow).

Chapter 2

29

More recently, three-dimensional (3D) ultrasonography has been developed to

measure gastric emptying rate (Gentilcore et al. 2006a, Gilja et al. 1995a, Gilja et al.

1997, Tefera et al. 2002) and has demonstrated greater accuracy compared with 2D

ultrasonography (Gilja et al. 1997). Transverse sections are recorded throughout the

entire length of the stomach, imaging both the proximal and distal regions.

Customised software provides area and volume measurements of each slice. Each of

the 2D slices is then reconstructed using specialised software to form a reconstructed

3D image of the total stomach. Gastric emptying of the stomach (percent retention) at

any time point is calculated as the volume at a particular time point relative to the

maximum volume (i.e. immediately following meal ingestion) (Gilja et al. 1997)

(Figure 2.3).

Figure 2.3: Ultrasonographic image of the stomach demonstrating region-of-interest (a) and 3D reconstructed volumetric image of the total stomach (b).

3D ultrasonography has recently been validated against the ‘gold standard’ method

for quantifying gastric emptying, scintigraphy, in healthy humans (Gentilcore et al.

2006a), however, it has hitherto not been applied to patients with gastroparesis.

Chapter 2

30

Chapter 6 reports an evaluation of 3D ultrasonography compared with scintigraphy,

in the measurement of gastric emptying in patients with diabetic gastroparesis.

2.2.3 Stable isotope breath tests

Stable isotope breath tests involve the use of the nonradioactive isotope 13C, which is

usually bound to a medium chain triglyceride (octanoic acid) or a proteinaceous algae

(Spirulina) and incorporated into a test meal (Lee et al. 2000). Stable isotope breath

tests usually measure gastric emptying of solids. The 13C-labelled substrate is then

ingested, emptied from the stomach into the duodenum, absorbed across the proximal

small intestine, metabolised by the liver and ultimately oxidised to 13CO2, which can

be detected in expired air (Camilleri et al. 1998, Kim et al. 2000, Lee et al. 2000).

Breath samples are collected at specific time intervals for up to 3 – 6 hours and

measured subsequently for 13CO2 enrichment by isotope ratio mass spectrometry or

laser infrared spectroscopy, thereby giving a measure of the gastric emptying rate.

This technique relies on the assumption that the gastric emptying rate is the sole rate-

limiting step in the delivery of 13CO2 to expired air (Hellmig et al. 2006). This may

prove unreliable in certain patients, for example, those with malabsorption states or

pancreatic, or hepatic, impairment; in this circumstance, substrate metabolism rather

than gastric emptying will preside as the rate-limiting step (Camilleri et al. 1998,

Mariani et al. 2004). Stable isotope breath tests cannot provide information about

intragastric distribution of a meal, as they are not an imaging technique (Mariani et

al. 2004). These tests are cheaper and simpler compared with traditional scintigraphy

such that studies may be performed in the office, at the bedside or in the community

(Hellmig et al. 2006). There is no associated radiation exposure, thereby allowing

studies in children and pregnant women. Studies (Bluck and Coward 2006, Chew et

al. 2003, Choi et al. 1997, Ghoos et al. 1993, Lee et al. 2000) have reported a high

Chapter 2

31

correlation between gastric emptying results from 13C breath tests and scintigraphy

performed simultaneously, and more recently, simultaneous dual isotopic breath tests

using 14C to label the solid, and 13C, to label the liquid, component of the meal has

been validated against the ‘gold standard’ scintigraphy in the measurement of gastric

emptying in health and diabetes (Chew et al. 2003). Further data in patients with

abnormally delayed and rapid gastric emptying are required to more clearly elucidate

the precision of this technique compared with scintigraphy. Owing to its simplicity

and ease of performance, isotopic breath tests may provide a useful screening tool in

patients with possible gastroparesis.

2.2.4 Magnetic resonance imaging

Magnetic resonance imaging (MRI) represents a non-invasive, yet detailed, method to

quantify the emptying rate, intragastric distribution, volume and contractile activity of

the stomach (Feinle et al. 1999, Schwizer et al. 1992). MRI can also be used to

measure gastric accommodation (Mearadji et al. 2001) and provide information about

intragastric distribution (Faas et al. 2002). Solid and liquid gastric emptying rates

have been shown to correlate well with simultaneously performed scintigraphy

(Feinle et al. 1999). MRI does not involve radiation exposure and enables repeated

and prolonged measurements. A useful research tool, its clinical application is,

however, restricted by its high cost, limited availability and long procedure time (Kim

et al. 2000). Furthermore, the technique requires patients to be imaged in the supine

position which may influence gastric emptying rates and intragastric meal

distribution. Patients are also required to hold their breath during acquisition (so as to

minimise motion artefacts) and may find confinement in the scanner uncomfortable

(Schwizer et al. 1996). The benefit of MRI as a valid diagnostic tool in the clinical

setting remains to be elucidated.

Chapter 2

32

2.2.5 Single photon emission computed tomography

Single photon emission computed tomography (SPECT) has been used to measure

the gastric accommodation response to a meal (Bouras et al. 2002, Kuiken et al.

1999, Simonian et al. 2004). This technique is based on the ability of the parietal

(oxyntic) and nonparietal (mucous) gastric cells to take up 99mTc-pertechnetate,

following intravenous administration. Tomographic images are acquired with a dual-

headed gamma camera with the patient in the supine position, and 3D images are

subsequently reconstructed and total gastric volumes calculated. SPECT, unlike the

barostat, is non-invasive, does not require intubation and has the ability to measure

accommodation of the total stomach (Mariani et al. 2004). SPECT has been validated

in vitro and in vivo, wherein, changes in total gastric volume in response to a meal

measured via SPECT were comparable to changes measured by barostat (Bouras et

al. 2002). More recently, simultaneous measurements of gastric emptying and

volume following ingestion of solid and liquid meals, have been performed

(Simonian et al. 2004), allowing accommodation and emptying to be quantified in a

single test. While this study (Simonian et al. 2004) was performed in healthy

subjects, the technique has the potential to be applied to investigations of the

pathophysiology of the stomach in disease. Disadvantages of SPECT include its

associated radiation burden, the need to image patients in the (non-physiological)

supine position and the inability to measure gastric sensory responses (Bouras et al.

2002).

2.2.6 Radiopaque marker techniques

Radiological (X-ray) tracking of ingested radiopaque markers has been used over the

past several decades for measurement of gastric emptying of nondigestible solids

(Feldman et al. 1984, Horikawa 1998, Loreno et al. 2004). Somewhat surprisingly,

Chapter 2

33

studies have reported that this technique correlates well with emptying rates measured

via scintigraphy (Stotzer et al. 1999) and two-dimensional ultrasonography (Loreno

et al. 2004). A major pitfall is the high radiation burden. Furthermore, only emptying

of nondigestible solids can be quantified, which, requires phase III activity (Coupe et

al. 1991, Hinder and Kelly 1977). Further studies are required to determine the

sensitivity and specificity and to standardise the method (Park and Camilleri 2006).

2.2.7 Applied potential tomography / Impedance

epigastrography

Applied potential tomography (APT) and impedance epigastrography (IE) use

changes in electrical resistivity or impedance to measure gastric emptying. APT

involves the application of a current through sixteen electrodes that are placed in a

circular array on the skin’s surface around the trunk at the level of the eighth costal

cartilage, which corresponds to the gastric fundus or body (Avill et al. 1987). Gastric

emptying is quantified by measuring the change in resistivity when a meal of fixed

composition empties from the stomach and an impedance image is produced

(Mushambi et al. 1992). Although applied potential tomography has been reported to

correlate with dye dilution and scintigraphic techniques (Avill et al. 1987, Mangnall

et al. 1991, Mushambi et al. 1992), this method is restricted to either solid (Mangnall

et al. 1991) or liquid (Avill et al. 1987) meals only, and cannot measure the gastric

emptying rate of mixed solid/liquid meals. APT requires pharmacological

suppression of gastric acid secretion by H2 receptor antagonists (e.g. ranitidine,

cimetidine) as the presence of gastric acid affects resistivity recordings (Mushambi et

al. 1992, Soulsby et al. 2006).

Chapter 2

34

Impedance epigastrography measures the impedance in a cross-section through the

body and involves the application of four electrodes, two anteriorly and two

posteriorly, over the gastric area, through which a current is directed, and an

impedance trace is produced (Mangnall et al. 1988). IE is also dependent on

pharmacological gastric acid suppression and is sensitive to movement, for example,

whole body movement by the patient, or small intestinal transit or gall bladder

emptying. APT has been reported to be more reproducible and reliable than IE, and

has shown better correlation with scintigraphy (Mangnall et al. 1988).

2.2.8 Absorption kinetics of orally administered drugs

An alternative, and relatively inexpensive, technique to measure gastric emptying

involves the oral administration of a drug and measurement of its blood or salivary

concentrations (Horowitz et al. 2002b). This method rests on the principle that the

rate of drug absorption is a measure of the rate of gastric emptying.

2.2.8.1 Paracetamol (acetaminophen) absorption

Paracetamol (acetaminophen) has been used in this way since its pharmacological

properties are ideal: absorption of paracetamol from the stomach is negligible, but it

is rapidly and completely absorbed across the small intestine by passive diffusion

(Clissold 1986, Willems et al. 2001), so that gastric emptying is the rate-determining

step for absorption (Clements et al. 1978, Heading et al. 1973, Nimmo 1976, Willems

et al. 2001). The procedure involves ingestion of a drink containing paracetamol

(usually 20 mg/kg or 1.5 g) and repeated venous blood, or saliva, sampling at regular

time intervals for measurement of paracetamol concentrations (Maddern et al. 1985).

There are close correlations between salivary and plasma/serum concentrations of

paracetamol (Gandia et al. 2003, Maddern et al. 1985, Sanaka et al. 2000, Smith et

Chapter 2

35

al. 1991). Based on the premise that paracetamol absorption exclusively reflects the

gastric emptying rate, it has been used to test the gastric emptying rate of liquids and

serum paracetamol has been shown to correlate with the rate of gastric emptying of

liquids (Clements et al. 1978). However, the concentration of paracetamol in serum

also reflects factors unrelated to gastric emptying, such as individual differences in

first-pass metabolism, unequal distribution and individual rates of elimination, as the

pharmacokinetics of paracetamol vary between and within individuals (Johansson et

al. 2003, Medhus et al. 2001). This method is, therefore, not considered to be very

accurate. To circumvent these pitfalls, Medhus et al. (2001) developed an algorithm

for the paracetamol absorption test, adjusting for individual first-pass metabolism,

unequal distribution of paracetamol in body fluids during periods of rapid changes in

its concentration and individual rates of elimination. The algorithm thus converts the

serum paracetamol concentrations into a percentage of the meal emptied over time

(Medhus et al. 2001), although literature reports of its validation against the ‘gold

standard’, scintigraphy, are lacking. The paracetamol absorption test is, however,

relatively popular due mainly to its safety and simplicity, it is well tolerated by

patients, reproducible and it is widely available. Its disadvantages include the

potential confounding influence of the first-pass effect, the inability to measure solid

emptying, the requirement for frequent blood sampling and the provision of only an

indirect measure of the gastric emptying rate (Gandia et al. 2003, Johansson et al.

2003, Kim et al. 2000, Medhus et al. 1999, Medhus et al. 2001, Sanaka et al. 1998,

Sanaka et al. 2000, Willems et al. 2001).

2.2.8.2 Oral glucose absorption

3-ortho-methyl-D-glucose (3-OMG) is a synthetic glucose analogue that is actively

absorbed, but not metabolised by the liver or intestinal mucosa and is renally cleared.

Chapter 2

36

Measurements of serum 3-OMG concentrations may, therefore, reflect the rate at

which glucose is actively absorbed across the small intestine (Fordtran et al. 1962,

Jones et al. 2001a, Rayner et al. 2002). Accordingly, serum 3-OMG levels may

provide an index of the rate at which 3-OMG empties from the stomach, since the

rate of oral drug absorption is determined, at least to some extent, by the rate of

gastric emptying (Jones et al. 2001a, Rayner et al. 2002). Much like the paracetamol

absorption test, the 3-OMG absorption method is safe, easy to perform and does not

require exposure to ioising radiation. While several studies (Jones et al. 2001a,

Rayner et al. 2002, Schwartz et al. 2002) have used 3-OMG concentration as a

reliable marker of glucose absorption, there has also been some evidence to suggest

that measurement of serum 3-OMG levels following oral administration of the

compound may reflect the rate of gastric emptying (Jones et al. 2001a). Jones et al.

(2001) demonstrated a correlation between oral glucose absorption (reflected by 3-

OMG concentration) and both the gastric emptying rate and blood glucose

concentration following ingestion of a high-nutrient drink, thereby accounting for 41

% to 48 % of the variance in glucose concentration.

2.3 Measur e m e n t o f i n t r a g a s t r i c p r e s s u r e a n d

contractions

Measurement of intragastric pressures and contractions can provide a considerable

amount of information about normal and disordered gastrointestinal motility. There

are a number of methods to measure intragastric pressures and contractions, including

manometry, the barostat, scintigraphy, ultrasonography, fluoroscopy and MRI. The

following discusses the barostat, ultrasonography and strain rate imaging, with a

primary focus on manometry, as this methodology is employed within the context of

this thesis.

Chapter 2

37

2.3.1 Manometry

Manometry, through the measurement of intragastric pressure, can quantify antral,

pyloric and duodenal motility (Hansen 2002, Heddle et al. 1988c). Motility patterns

in these various regions can be measured by a catheter comprising an adapted pyloric

sleeve sensor, in conjunction with multiple perfused side-holes positioned in the

antrum and duodenum (Figure 2.4). An infusion port positioned near the distal end of

the catheter allows for infusion of various test nutrients, such as dextrose (Heddle et

al. 1988b) and lipid (Heddle et al. 1988a). The distal end of the manometric catheter

is passed through an anaesthetised nostril and its movement through the upper

gastrointestinal tract is measured by manometric patterns. Patients undergoing this

procedure are required to fast overnight from both solids and liquids. The catheter

passes through the stomach and into the small intestine by peristalsis (Jones et al.

1995a, Little et al. 2005). Correct positioning of the sleeve sensor across the pylorus

is maintained by measuring the transmucosal potential difference (TMPD) between

sideholes located on either end of the sleeve, i.e. between the most distal antral

sidehole (~ -40 mV) and the most proximal duodenal sidehole (~ 0 mV) (Heddle et

al. 1988a, Heddle et al. 1988c, Heddle et al. 1988b, Heddle et al. 1989). An

intravenous cannula filled with sterile, isotonic saline is inserted into the

subcutaneous tissue of the forearm as a reference electrode, for this purpose.

Manometry provides accurate measurements of contractions which occlude the lumen

(Hansen 2002). Accordingly, contractions which do not completely occlude the

lumen may not be captured by manometry (Fone et al. 1990). Manometric pressure

recordings are digitised and recorded on a computer running custom software, and

subsequently analysed for a number of parameters including the number and

amplitude of isolated pyloric pressure waves (IPPWs) and the number and amplitude

Chapter 2

38

of pressure waves in the antrum and duodenum (Heddle et al. 1988a, Heddle et al.

1988c, Heddle et al. 1988b, Heddle et al. 1989, Little et al. 2005). A valuable

research tool, manometry is also used in the clinical setting, although, not routinely in

many centres. It is a specialised technique that is technically demanding and may be

uncomfortable for the patient, particularly when the catheter must remain in situ over

prolonged periods (Parkman et al. 2004).

Figure 2.4: Schematic representation of manometric catheter with 16 sideholes (channels) spaced at 1.5 cm intervals, comprising six antral sideholes, two TMPD sideholes on either side of the pyloric sleeve sensor, seven duodenal sideholes and one infusion port.

2.3.2 Barostat

The gastric barostat allows measurement of proximal gastric tone, and thereby,

gastric accommodation and compliance, which cannot be captured by manometry.

The technique involves the introduction of a thin-walled plastic balloon into the

gastric fundus (Azpiroz and Malagelada 1985a, Azpiroz and Malagelada 1987,

Hebbard et al. 1995b, Hebbard et al. 1996a, Hebbard et al. 1996b). A catheter

Chapter 2

39

connects the balloon to a barostat machine, which modulates expansion of the balloon

to constant pressure, or volume (up to a maximum of 1.2 L) (Schwizer et al. 2002).

Changes in intrabag pressure or volume, thus, reflect changes in gastric tone, in

response to various interventions (Hebbard et al. 1995b, Hebbard et al. 1996a,

Hebbard et al. 1996b, van der Schaar et al. 1999). The barostat method is invasive,

technically demanding, uncomfortable for the patient and difficult to use with solids

(Schwizer et al. 2002). The barostat balloon may affect gastric emptying or

intragastric distribution, and can be associated with artefacts, such as air leak or

inappropriately low volumes (Tutuian et al. 2008). Despite its limitations, the barostat

technique represents the only method which currently enables simultaneous

measurement of intragastric pressure and volume.

2.3.3 Ultrasonography

Ultrasonography has been used to measure gastric contractions and propagation of

waves (Ahluwalia et al. 1994, Bateman et al. 1977, Hausken et al. 1991, Hausken

and Berstad 1992, Holt et al. 1980, Hveem et al. 2001). The distal stomach is imaged

in the sagittal plane relative to other anatomic landmarks, such as the aorta and

superior mesenteric artery, and measurement of changes in antral area induced by a

contraction, relative to the relaxed area, enables measurement of frequency and

amplitude of contractions. Ultrasonography is more sensitive than manometry in the

detection of antral contractions, particularly non lumen-occlusive contractions, when

compared with manometry, which, as discussed, is better suited to the detection of

lumen-occlusive contractions (Hveem et al. 1995, Hveem et al. 2001).

Chapter 2

40

2.3.4 Strain rate imaging

Strain rate imaging (SRI) is a non-invasive, relatively novel, Doppler

ultrasonographic technique which enables measurement of the rate of mechanical

deformation (strain) (Ahmed et al. 2006, Gilja et al. 2002, Gilja et al. 2007). The

accuracy of SRI in measuring strain has been assessed in vitro in the porcine stomach

and shown to measure radial strain accurately, but to be less accurate in measuring

circumferential strain (Ahmed et al. 2006). SRI has also been shown to enable

detailed mapping of radial strain distribution of the gastric wall in vivo in healthy

subjects (Gilja et al. 2002). Purely a research tool at present, future studies in disease

states, such as functional dyspepsia, may further elucidate the biomechanical factors

underlying such conditions.

2.4 Measurement of gastric electrical activity

Gastric slow waves are controlled by a gastric pacemaker, located on the greater

curvature of the stomach, which depolarises at a frequency of 3 cycles per minute

(Chapter 1.2.2). Gastric electrical activity recordings may be made from the

gastrointestinal mucosa, serosa or the skin surface (Parkman et al 2003).

2.4.1 Electrogastrography

Electrogastrography (EGG) is a method for recording gastric electrical activity.

External EGG, which involves the placement of electrodes on the skin overlying the

stomach for the measurement of gastric electrical activity, is by far the most common

method used in humans, due to its non-invasive nature (Parkman et al. 2003). By

contrast, intraluminal, or surgically implanted, electrodes are usually reserved for

animal studies (Toporowska-Kowalska et al. 2006).

Chapter 2

41

Some disorders of gastric motor function are associated with abnormalities of EGG

frequency or amplitude (Chang 2005). Patients with diabetic gastroparesis, for

example, have been reported to demonstrate tachygastria (an increase in the number

of slow waves) and bradygastria (a decrease in the number of slow waves) (Koch et

al. 1989). At present, EGG is primarily considered a research tool and its clinical

usefulness has been questioned (Abid and Lindberg 2007).

2.5 Measurement of transpyloric flow

Transpyloric flow, the movement of fluid across the pylorus, has been measured by a

number of rechniques, including real-time ultrasonography (King et al. 1984),

Doppler ultrasonography (Hausken et al. 1992, Pallotta et al. 1998), scintigraphy

(Jones et al. 1995a), fluoroscopy (Rao et al. 1996), MRI (Marciani et al. 2001) and

impedance (Savoye et al. 2003, Savoye-Collet et al. 2003). These techniques have

provided valuable insights into transpyloric flow events, however, each has

limitations. All methods (excluding impedance) require that the stomach be filled

with a test meal or contrast agent, preventing studies during the interdigestive state

(Savoye et al. 2003, Savoye-Collet et al. 2003). Scintigraphy and fluoroscopy are

associated with a radiation burden, while other methods (Doppler ultrasonography,

real-time ultrasonography, impedance) require access to specialised equipment and a

high level of expertise (Gilja et al. 2007).

Chapter 2

42

2.6 Conclusions

Since the first measurements of gastric emptying performed in the 13th Century by the

Holy Roman Emperor Frederick II, techniques to evaluate gastric motility have come

a long way. While a number of such techniques to measure gastric emptying have

limited clinical utility at present, they have great application in the research setting.

Their continued growth, concomitant with the discovery of new technologies, holds

great promise for both clinical and research purposes in the future. Chapter 6 reports a

validation study comparing 3D ultrasonography with the ‘gold standard’,

scintigraphy, in the measurement of gastric emptying in patients with diabetic

gastroparesis.

Chapter 3

43

3Chapter 3DISORDERED GASTRIC EMPTYING IN DIABETES MELLITUS

3.1 Introduction

Disordered gastric motility, associated with delayed gastric emptying, occurs

frequently in patients with diabetes mellitus and represents an important clinical

problem (Haans and Masclee 2007, Horowitz et al. 2001, Horowitz et al. 2002b, Park

and Camilleri 2006, Patrick and Epstein 2008). ‘Gastroparesis’ is literally defined as

a form of gastric ‘paralysis’ and has been classified as delayed gastric emptying in the

absence of mechanical obstruction (Horowitz et al . 2002b). Gastroparesis is

frequently defined as a rate of emptying which is greater than, or equal to, two

standard deviations outside a normal, control range (Horowitz and Dent 1991,

Horowitz and Fraser 1995, Horowitz et al. 2001, Horowitz et al. 2002b). It has been

Chapter 3

44

suggested that a distinction be made between gastroparesis and delayed gastric

emptying, with a diagnosis of gastroparesis restricted to those patients in whom

gastric emptying is grossly delayed. More recently, the term ‘diabetic gastropathy’

has been devised to integrate both delayed gastric emptying and the presence of upper

gastrointestinal symptoms, such as nausea, vomiting, bloating and discomfort, which

all commonly accompany the condition and contribute to morbidity (Smith and Ferris

2003, Talley 2003). Gastroparesis may be associated with a variety of underlying

disorders; furthermore, the specific pathophysiologies contributing to disordered

gastric emptying in diabetes mellitus (autonomic neuropathy, enteric neuropathy and

glycaemic control) are interrelated (Horowitz et al. 2002b, Smith and Ferris 2003).

3.2 Prevalence

Disordered gastric function in patients with diabetes mellitus has been recognised for

more than 50 years (Kassander 1958, Rundles 1945) although its prevalence has until

recently been underestimated. Kassander, in 1958, coined the term ‘gastroparesis

diabeticorum’, describing the condition as abnormal gastric retention in

asymptomatic, insulin-treated patients with diabetes (Kassander 1958). Gastroparesis

was historically regarded an infrequent complication of diabetes, however, there is

now unequivocal evidence that delayed gastric emptying occurs frequently and

represents a common clinical problem (Horowitz et al. 1996b, Horowitz et al. 2001).

At present, while there is a lack of true population-based studies, data from cross-

sectional studies have revealed that gastric emptying of solid and/or nutrient liquid,

measured by scintigraphic radionuclide studies, is delayed in 30 – 50 % of adult

outpatients with type 1 (Horowitz et al. 1986) or type 2 (Horowitz et al. 1989b)

diabetes (Figure 3.1). It has been argued that as the majority of data on the prevalence

of diabetic gastroparesis are from tertiary medical centres, the prevalence of

Chapter 3

45

gastroparesis may have been overestimated (Camilleri 2006, Haans and Masclee

2007, Smith and Ferris 2003). While delayed gastric emptying occurs frequently in

longstanding diabetes, in many cases, the magnitude of the delay in gastric emptying

is relatively modest (Figure 3.1). Solid and liquid gastric emptying rates correlate

poorly in patients with diabetes, which serves to explain why the prevalence of

delayed gastric emptying is highest when both solid and nutrient-containing liquids

are measured, usually simultaneously (Horowitz et al. 1991, Jones et al. 1996).

Intragastric distribution of solid and liquid meal components is also frequently

abnormal in patients with diabetes and often associated with delayed gastric emptying

(Jones et al. 1995b, Urbain et al. 1993).

Figure 3.1: Gastric emptying of solid (100 g minced beef) and liquid (10 % dextrose) in 87 patients with longstanding diabetes (67 type 1, 20 type 2) and 25 healthy subjects. Shaded areas represent normal ranges; horizontal lines reflect median values. Reproduced from Horowitz et al. (1991).

While it has been reported that gastric emptying is frequently delayed in children

(approximately 30 %) and adolescents with type 1 diabetes (Cucchiara et al. 1998,

Reid et al. 1992), gastric emptying is occasionally abnormally rapid in diabetes

Chapter 3

46

(Horowitz et al. 1991, Lipp et al. 1997, Phillips et al. 1992, Schwartz et al. 1996,

Weytjens et al. 1998). This may particularly be the case in recently diagnosed

asymptomatic “early” type 2 diabetes (Phillips et al. 1992, Schwartz et al. 1996). It

has been postulated that this may contribute to poor glycaemic control and

progression of diabetes in such patients (Phillips et al. 1992).

There has been only one long-term, longitudinal study of gastric emptying in type 1

and type 2 diabetes mellitus (Jones et al. 2002), which demonstrated minimal change

in solid and liquid gastric emptying over a mean follow-up of twelve years,

suggesting that gastric emptying is relatively stable over time and that diabetic

gastroparesis is not a rapidly progressive disorder associated with a poor prognosis

(Jones et al. 2002). It must be noted, however, that glycaemic control, as assessed by

glycated haemoglobin, improved at follow-up in this study (Jones et al. 2002), an

observation that probably reflects heightened therapeutic attempts to normalise

glycaemia in diabetes following the findings of the Diabetes Control and

Complications Trial (DCCT) (1993) and the United Kingdom Prospective Diabetes

Study (UKPDS) (1998).

3.3 Aetiology and pathophysiology

Gastroparesis can occur as a complication of virtually any disease that has the

capacity to cause neuromuscular dysfunction of the gastrointestinal tract (Camilleri

2006, Smith and Ferris 2003, Tack 2007). Diabetes mellitus, however, represents a

predominant cause of gastroparesis (Horowitz and Fraser 1994, Horowitz et al.

1996b, Horowitz et al. 2001). Other aetiologies of delayed gastric emptying,

including medication, idiopathic causes and surgery, are listed in Table 3.1 (Horowitz

Chapter 3

47

and Dent 1991, Horowitz et al. 2002b, Patrick and Epstein 2008, Smith and Ferris

2003).

Table 3.1 Common aetiologies of delayed gastric emptying

MedicationsOpiates, anticholinergics, glucagon, -adrenergic agonists, calcium channel antagonists

SurgeryVagotomy and gastric resection, fundoplication, oesophagotomy, gastric bypass, Whipple procedure, heart/lung transplantation

InfectionsEpstein-Barr virus, varicella, parvo-like viruses, Chagas disease, Clostridium botulinum

Central nervous system disordersCerebrovascular accidents/trauma, malignancy, seizures

Peripheral nervous system disordersParkinson’s disease, Guillain-Barre disease, multiple sclerosis, dysautonomias

Neuropsychiatric disordersAnorexia nervosa/bulimia, rumination syndrome

Rheumatological diseasesProgressive systemic sclerosis, systemic lupus erythematosus, polymyositis/dermatomyositis

Endocrine and metabolic disordersDiabetes mellitus, hypothyroidism, Addison’s disease, hypercalcaemia, hypokalaemia, renal failure, pregnancy

Paraneoplastic disordersAssociated with breast cancer, small cell lung cancer, pancreatic cancer, others

Neuromuscular disordersIdiopathic gastroparesis, amyloidosis, chronic intestinal pseudo-obstruction, myotonic dystrophy

Critical illnessAdapted from Smith and Ferris (2003).

Chapter 3

48

Several pathophysiological abnormalities contribute to the gastroparesis syndrome

observed in patients with diabetes mellitus, the most important of which include

fundic hypomotility, antral hypomotility, gastric dysrhythmia and disordered

antropyloroduodenal coordination (Park and Camilleri 2006). Enteric neuropathy has

been implicated in the pathogenesis of diabetic gastroparesis, more specifically, a loss

of interstitial cells of Cajal (Forster et al. 2005) and a reduction in the number of

nitric oxide synthase neurons (He et al. 2001). However, the main pathogenetic

factors of diabetic gastroparesis resulting in motor dysfunction of the stomach are

autonomic neuropathy and glycaemic control (Horowitz and Dent 1991, Horowitz et

al. 1994, Horowitz and Fraser 1995, Horowitz et al. 2001, Horowitz et al. 2002b).

3.3.1 Autonomic neuropathy

Gastric emptying is dependent on the function of the vagus nerve. Rundles, in 1945,

attributed delayed gastric emptying in diabetes to irreversible vagal damage (Rundles

1945), however, it has subsequently been established that this is not the case. There is

currently no direct quantitative test to measure gastrointestinal autonomic nerve

function, and as such, cardiovascular autonomic nerve function is used as a substitute,

indirect, measure of the function of the abdominal vagus (Asakawa et al. 2005).

Ewing et a l . (1985) developed five cardiovascular reflex tests to measure

cardiovascular autonomic nerve function and these have been the heart rate response

to deep breathing (R-R interval), the heart rate response to standing up (“30:15”

ratio), the blood pressure response to standing up, and the systolic blood pressure

response to sustained hand grip and Valsalva manoeuvre (Asakawa et al. 2005,

Ewing et al. 1985). In the author’s studies, heart rate response to deep breathing (R-R

interval) and heart rate response to standing up (“30:15” ratio) were used to evaluate

Chapter 3

49

parasympathetic function, while systolic blood pressure response to standing up was

used to assess sympathetic function (Chapters 7, 9, 10).

There is a higher prevalence of delayed gastric emptying in diabetic patients with

cardiovascular autonomic neuropathy than in those without (Asakawa et al. 2005,

Buysschaert et al. 1987, Darwiche et al. 2001, Horowitz et al. 1991, Merio et al.

1997, Ziegler et al. 1996), however, the correlation between autonomic nerve

function (assessed by cardiovascular nerve function) and gastric emptying rate is only

weak (Buysschaert et al. 1987, Campbell et al. 1977, Horowitz et al. 1991, Koçkar et

al. 2002, Merio et al. 1997, Ziegler et al. 1996) (Figure 3.2).

Figure 3.2: Relationship between solid gastric emptying and cardiovascular autonomic nerve function in diabetes mellitus. The shaded area represents the normal range. Reproduced from Horowitz et al. (1991).

Chapter 3

50

3.3.2 Blood glucose concentration

While neuropathic changes may contribute to disordered gut function in diabetes, it is

established that acute changes in the blood glucose concentration have a major,

reversible, effect on gastrointestinal motility in all regions of the gastrointestinal tract

in both healthy subjects (MacGregor et al. 1976, Schvarcz et al. 1995a, Schvarcz et

al. 1997) and diabetic patients (Fraser et al. 1990, Samsom et al. 1997, Schvarcz et

al. 1993, Schvarcz et al. 1997). In healthy subjects, hyperglycaemia slows gastric

emptying (MacGregor et al. 1976, Schvarcz et al. 1997). In patients with type 1

diabetes, the rate of gastric emptying of both solids (Fraser et al. 1990, Samsom et al.

1997) and nutrient-containing liquids (Fraser et al. 1990) has also been found to be

substantially slower during hyperglycaemia (~ 16 – 20 mmol/L) than during

euglycaemia (~ 4 – 8 mmol/L). Similar effects are likely to be evident in patients with

type 2 diabetes in that a higher blood glucose level at baseline is associated with a

more prolonged liquid emptying (Horowitz et al. 1989b). Changes in the blood

glucose concentration that are within the normal postprandial range have also been

shown to influence gastric emptying. Emptying of both solids and liquids is slower at

a blood glucose concentration of 8 vs. 4 mmol/L in both healthy subjects and patients

with uncomplicated type 1 diabetes (Schvarcz et al. 1997). The motor correlates of

the slowing of gastric emptying induced by acute hyperglycaemia include proximal

gastric relaxation (Hebbard et al. 1996a), suppression of antral contractions (Hasler et

al. 1995) and stimulation of localised pyloric pressure waves (Fraser et al. 1991b).

In contrast, insulin-induced hypoglycaemia has been shown to accelerate gastric

emptying of both solid and nutrient-containing liquids in both healthy subjects

(Schvarcz et al. 1995a) and in patients with uncomplicated type 1 diabetes (Schvarcz

et al. 1993). Hypoglycaemia (~ 1.9 mmol/L) was reported to accelerate the gastric

Chapter 3

51

emptying rate approximately twofold when compared to euglycaemia (~ 4 – 7

mmol/L) in patients with type 1 diabetes mellitus (Schvarcz et al. 1993) (Figure 3.3).

It has been proposed that this may represent an important mechanism in the

counterregulation of hypoglycaemia (Horowitz et al. 2002b). There is no information

about the effects of hypoglycaemia in patients with longstanding, complicated

diabetes mellitus, including those with gastroparesis and autonomic neuropathy.

These issues form the focus of the study reported in Chapter 7.

Figure 3.3: The effect of hypoglycaemia (~ 1.9 mmol/L) on solid and liquid gastric emptying in 8 uncomplicated type 1 diabetic patients. Reproduced from Schvarcz et al. (1993).

While numerous studies have evaluated the effects of intravenous glucose on gastric

motility, information about the effects of other monosaccharides, such as fructose, are

limited. Fructose is used widely in the diabetic diet as it is sweeter than equienergetic

Chapter 3

52

glucose (thereby conferring the same sweetness for a lower energy burden) and has

been demonstrated to exert lower glycaemic responses following oral administration

in type 2 diabetes (Crapo et al. 1980, Vozzo et al. 2002) and healthy subjects (Crapo

et al. 1980, Horowitz et al. 1996a, Kong et al. 1999a, Vozzo et al. 2002). Oral

fructose is also associated with higher plasma insulin concentrations in patients with

type 2 diabetes compared with healthy subjects (Vozzo et al. 2002). Fructose has

been shown to stimulate GLP-1 secretion, albeit to a much lesser extent than an

equicaloric load of glucose in normal subjects (Kong et al. 1999a), although there is

no difference in GLP-1 concentrations after fructose ingestion in people with and

without diabetes (Toft-Nielsen et al. 2001, Vozzo et al. 2002). Oral fructose has been

shown to empty from the stomach at a slightly faster rate than glucose (Elias et al.

1968, Guss et al. 1994), however, when administered intraduodenally, both glucose

and fructose produce comparable effects on antropyloroduodenal motility (Rayner et

al. 2000a). Further studies to assess the effect of intravenous fructose compared with

glucose on glycaemia, gastric emptying and antropyloroduodenal motility are

indicated (Chapter 8).

3.4 Clinical features of diabetic gastroparesis

Disordered gastric emptying in diabetes may be associated with upper gastrointestinal

symptoms, malnutrition, anorexia, altered oral drug absorption and poor control of

blood glucose concentrations, which is now recognised as a major contributor to

micro- and macro-vascular disease (Horowitz and Dent 1991). The gastric emptying

rate is also a determinant of postprandial hypotension, which occurs frequently in

diabetes and in the elderly and represents an important clinical problem, as it can

predispose to syncope and falls, and in more severe cases, stroke and angina

(Horowitz et al. 1989b).

Chapter 3

53

3.4.1 Upper gastrointestinal symptoms

There have been few comprehensive studies detailing the prevalence, determinants or

importance of upper gastrointestinal symptoms in diabetic patients. However, it is

apparent that the prevalence of upper gastrointestinal symptoms is increased in type 1

and type 2 diabetes and is linked to psychological distress, which influences quality

of life adversely (Horowitz et al. 2002b). Schvarcz et al. (1996) reported a higher

prevalence of upper gastrointestinal symptoms including loss of appetite, early

satiety, nausea, vomiting, postprandial fullness and abdominal distension among

patients with type 1 diabetes compared with age- and sex-matched controls (Schvarcz

et al. 1996). A relatively recent population-based study of 15 000 Australian adults

found that all upper and lower gastrointestinal symptoms were increased in patients

with type 2 diabetes compared with controls, and that there was a higher prevalence

of these symptoms among patients in whom self-reported glycaemic control was poor

compared with those who self-reported average or good glycaemic control (Bytzer et

al. 2001). Possible determinants of gastrointestinal symptoms in diabetes include

disordered motility, poor glycaemic control, autonomic nervous dysfunction, social

aspects (psychology and demographics), medication, visceral hypersensitivity and

Helicobacter pylori infection (Horowitz et al. 2002b). Horowitz et al. (1991)

examined the relationship between upper gastrointestinal symptoms and gastric

emptying rate in patients with type 1 and type 2 diabetes and found that the

correlation between these, while statistically significant, was weak (Horowitz et al.

1991) (Figure 3.4). The potential predictors of solid and liquid gastric emptying rates

have been evaluated in patients with type 1 and type 2 diabetes and abdominal

bloating/fullness and the female sex were found to be associated with delayed gastric

emptying (Jones et al. 2001b). Similarly, a study conducted in hospitalised type 2

patients showed gastric emptying to be slower in females and in those patients

Chapter 3

54

reporting nausea and early satiety (Kojecky et al. 2008). Accordingly, gastroparesis

should be regarded as a marker of gastrointestinal motor abnormality, rather than as

the direct cause of symptoms (Horowitz et al. 1991).

Figure 3.4: The relationship between upper gastrointestinal symptoms and gastric emptying of a solid meal in 87 type 1 and type 2 diabetic patients. The shaded area represents the normal range. Reproduced from Horowitz et al. (1991).

3.4.2 Oral drug absorption

Most drugs are absorbed more rapidly from the small intestine than from the stomach

(Horowitz et al. 1989a). Consistent with the pH-partition hypothesis, weakly basic

lipid-soluble drugs are absorbed more rapidly from the small intestine than from the

acidic stomach, hence, the rate of absorption will be limited by the rate of gastric

emptying. Yet, acidic drugs, too, are absorbed more rapidly from the intestine than

the stomach, presumably owing to the greater surface area of the small intestine

Chapter 3

55

(Heading et al. 1973). As discussed in Chapter 4, the rate of gastric emptying, both

during the fasting and fed state, is an important determinant of the rate of oral drug

absorption (Hebbard et al. 1995a, Horowitz et al. 1989a, Nimmo 1976). In patients

with severe gastroparesis, clinically relevant aberrant plasma drug concentrations

may result, particularly when drugs (especially tablets or capsules that do not readily

undergo disintegration and dissolution in the stomach) are ingested with or after a

meal. This poses clinical significance when rapid onset of a drug is necessary, for

example, some oral hypoglycaemic agents in diabetic patients (Groop et al. 1989,

Horowitz et al. 2002b).

3.4.3 Impact of gastric emptying on glycaemic control

It is now apparent that in both normal subjects and diabetic patients, the rate of

gastric emptying plays a major role in the regulation of postprandial blood glucose

homeostasis by controlling the delivery of carbohydrate to the small intestine (Jones

et al. 1995b). Horowitz et al. (1993) reported that the rate of gastric emptying

accounts for approximately 34 % of the variance in peak plasma glucose after a 75 g

oral glucose load in normal subjects, so that a faster rate of emptying is associated

with higher initial plasma glucose concentrations (Horowitz et al. 1993b). Similar

data were evident in patients with type 2 diabetes managed by diet, although the

magnitude of the rise in blood glucose concentration in response to emptying of

glucose was greater (Jones et al. 1996). In type 1 patients with gastroparesis, the

immediate postprandial insulin requirement to maintain eulgycaemia is less (Ishii et

al. 1994). Accordingly, delayed gastric emptying has the potential to result in poor

control of blood glucose concentrations in diabetic patients, particularly those with

type 1 and insulin-treated type 2 diabetes, by causing a mismatch between the onset

of action of insulin administered exogenously or oral hypoglycaemic agents, and

Chapter 3

56

absorption of nutrients from the small intestine (Chaikomin et al. 2006, Horowitz and

Fraser 1995), and gastroparesis is a recognised cause of hypoglycaemia in this group

(Horowitz et al. 2002b). In contrast, type 2 patients who are not taking insulin may

benefit by strategies that slow gastric emptying, even if the latter is delayed, given the

defect in phase 1 insulin release (Gonlachanvit et al. 2003). In light of this, and

together with the fact that poor glycaemic control is a major contributor to micro-

and, probably, macro-vascular disease, modulation of the gastric emptying rate in

order to improve glycaemic control is pertinent to the management of diabetes.

Furthermore, it is now well established that postprandial, as opposed to fasting,

glycaemia (HbA1c) represents a specific target for treatment, owing to the fact that

humans spend the majority of their time in the postabsorptive or postprandial, rather

than the fasting, state (Chaikomin et al. 2006, Rayner and Horowitz 2006).

Postprandial glycaemia may also represent an independent risk factor of

cardiovascular disease, even in healthy humans (Gerich 2003).

In patients with type 1 diabetes, prokinetic agents taken before a meal to accelerate

gastric emptying may potentially offset the mismatch between exogenous insulin and

nutrient delivery, potentially improving glycaemic control (Chaikomin et al. 2006).

In patients with type 2 diabetes who are not treated with insulin improved

postprandial glycaemic response may result from slowing the gastric emptying rate

by the addition of guar gum to a meal (Russo et al. 2003) and by ingesting an oil

“preload” immediately prior to a carbohydrate-containing meal (Gentilcore et al.

2006b).

Chapter 3

57

3.4.4 Therapeutic approaches to the management of

glycaemic control

The ‘incretin effect’, which is characterised by the higher insulin secretory response

observed following an oral compared with an intravenous glucose load despite

eliciting similar rises in glycaemia (Perley and Kipnis 1967), is reduced or

diminished in patients with type 2 diabetes (Nauck et al. 2004). Since the discovery

that one of the incretin hormones glucagon-like peptide 1 (GLP-1) stimulates insulin

release, suppresses glucagon secretion, slows gastric emptying, suppresses energy

intake and may increase –cell mass in type 2 diabetes (Chapter 1.4.5.1), a number of

novel therapeutic classes of ‘incretin-based’ drugs have emerged for the treatment of

type 2 diabetes. Because intact, biologically active GLP-1 is rapidly degraded (within

minutes) by the protease dipeptidyl peptidase-IV (DPP-IV), ‘incretin-based’ therapies

follow either of two main approaches: GLP-1 receptor agonists (‘incretin mimetics’)

represent one approach and include exenatide and liraglutide, which mimic the effect

of GLP-1 but are DPP-IV-resistant. An alternative approach is that of the DPP-IV

inhibitors (‘incretin enhancers’), such as sitagliptin and vildagliptin, which potentiate

incretin action by inhibiting DPP-IV and thereby increase the half-life of endogenous

GLP-1.

3.4.4.1 GLP-1 receptor agonists

Exenatide, which is synthesised from exendin-4 (itself originally derived from the

venom of the Gila monster (Heloderma suspectum)), shares 53 % amino acid

sequence identity to GLP-1 (Baggio and Drucker 2007). Exenatide has a circulating

half-life of 60 – 90 minutes and plasma levels remain elevated approximately 4 – 6

hours after subcutaneous injection. In clinical trials, exenatide, given by twice-daily

subcutaneous injections, caused significant reductions in HbA1c and glycaemia (both

Chapter 3

58

pre- and post-prandially), improved parameters of –cell function, promoted weight

loss and slowed gastric emptying (Buse et al. 2004, DeFronzo et al. 2005, Fineman et

al. 2003, Kendall et al. 2005, Kolterman et al. 2003, Linnebjerg et al. 2008).

Gastrointestinal adverse effects (nausea) are common (Buse et al. 2004, DeFronzo et

al. 2005, Kendall et al. 2005), but do not usually compromise long-term use. The

dominant effect of exenatide to reduce postprandial glycaemic excursions may relate

to slowing of gastric emptying, even in patients with autonomic neuropathy

(Linnebjerg et al. 2008).

Because exenatide is administered by twice-daily injections, attempts were directed to

develop an agent with less frequent parenteral administration. Liraglutide has a half-

life of 10 – 14 hours following subcutaneous injection and may be given once daily

(Drucker and Nauck 2006). Liraglutide has been shown to reduce pre- and post-

prandial glycaemia and HbA1c, in addition to promoting weight loss or preventing

weight gain, however, data on its effect on gastric emptying are conflicting (Degn et

al. 2004, Juhl et al. 2002). Mild nausea, vomiting and diarrhoea are the most common

side effects (Degn et al. 2004, Nauck et al. 2006), but again do not compromise long-

term usage (Horowitz et al. 2008). More recently, exenatide has been formulated as a

sustained-release once-weekly injection (Kim et al. 2007) and when administered,

has been shown to result in greater reductions in HbA1c, when compared with

exenatide given twice daily, with no increased risk of hypoglycaemia (Drucker et al.

2008).

3.4.4.2 DPP-IV inhibitors

The DPP-IV inhibitors, through their potentiation of endogenous GLP-1, have

demonstrated similar effects to the GLP-1 receptor agonists in preclinical and clinical

Chapter 3

59

studies, including stimulation of insulin and suppression of glucagon secretion,

reductions in HbA1c, decreases in postprandial glycaemic excursions, and

preservation of –cell mass by stimulating cell proliferation and inhibiting apoptosis

(Drucker and Nauck 2006). However, DPP-IV inhibitors are generally not associated

with gastrointestinal adverse effects or weight loss, and their effect on gastric

emptying has not yet been fully elucidated. Vildagliptin has been reported to

decelerate gastric emptying in one study in type 2 diabetes reported in abstract form

(Woerle et al. 2007), while another study (Vella et al. 2007) reported no effect on

gastric emptying in type 2 diabetes. Effects of DPP-IV inhibition on gastric emptying

may well be less than those of exogenous GLP-1 (Little et al. 2006a) or GLP-1

analogues (Linnebjerg et al. 2008) because the resulting plasma concentrations of

GLP-1 are physiological, rather than pharmacological. Recently, sitagliptin was in

fact found to be less effective than exenatide at slowing gastric emptying (DeFronzo

et al. 2008).

Within this class, sitagliptin is the only agent to have been approved in Australia and

USA (vildagliptin is approved for use within the European Union), however, a

number of other agents including saxagliptin and denagliptin are currently under

clinical trial investigation (Barnett 2008).

Chapter 3

60

3.5 Conclusions

Disordered gastric motility occurs frequently in diabetes and represents an important

clinical problem. The main pathogenetic factors are autonomic neuropathy and

glycaemic control. The effects of hypoglycaemia on gastric emptying in longstanding

type 1 diabetes are investigated in Chapter 7, while the comparative effects of

fructose, glucose and saline on gastric emptying and antropyloroduodenal motility are

evaluated in the study reported in Chapter 8.

Chapter 4

61

4Chapter 4TREATMENT OF GASTROPARESIS

4.1 Introduction

The key principles in the management of gastroparesis include relieving

gastrointestinal symptoms, correction of hydration and malnutrition, optimisation of

glycaemic control (in diabetic patients) and judicious use of prokinetic and antiemetic

therapies where appropriate. While non-pharmacological interventions, including

dietary and lifestyle modifications, are important in the management of gastroparesis,

the majority of patients with gastroparesis require pharmacological therapy. Current

treatment options are suboptimal and new therapeutic alternatives are emerging.

Chapter 4

62

4.2 Non-pharmacological interventions

4.2.1 Dietary and lifestyle modifications

Patients with gastroparesis are usually advised to adhere to a diet that is low in both

fat and fibre, as fat empties from the stomach slowly (Chapter 1.3.3) and indigestible

fibres can be difficult to empty and lead to bezoar formation (Abell et al. 2006,

Rayner and Horowitz 2005). These concepts have, however, not be shown to be

effective. It is also often recommended that patients consume several, small, frequent

meals rather than few, large meals per day (Park and Camilleri 2006). Alcohol and

nicotine delay gastric emptying and their consumption should be discouraged. Liquid

supplements, or pureed food may be better tolerated than solid meals due to their

faster rate of emptying from the stomach (Abell et al. 2006, Patrick and Epstein

2008). These measures aim to promote the emptying of ingesta from the stomach at a

rate that is not exceedingly delayed.

It is well established, as previously discussed (Chapter 3.3.2), that hyperglycaemia

adversely affects gastric motor function (Fraser et al. 1990, Hebbard et al. 1996a,

Hebbard et al. 1996b, MacGregor et al. 1976, Samsom et al. 1997, Schvarcz et al.

1997), as do changes in the blood glucose concentration that are within the normal,

postprandial range (Schvarcz et al. 1997). Moreover, the gastroprokinetic effect of

erythromycin (Chapter 4.3.3) is attenuated during hyperglycaemia in both healthy

subjects (Jones et al. 1999b) and patients with type 1 diabetes (Petrakis et al. 1999a).

Accordingly, increased blood glucose monitoring and intensive therapeutic attempts

to nomalise glycaemia in patients with diabetes may further assist in the management

of diabetic gastroparesis.

Chapter 4

63

4.3 Pharmacological interventions

The majority of currently available pharmacological agents for the treatment of

gastroparesis are prokinetics that aim to increase the frequency and amplitude of

contractile activity within the stomach and, thereby, accelerate the gastric emptying

rate (Camilleri 2007, Parkman et al. 2004). While prokinetics are often used to

alleviate symptoms, there is, as previously discussed (Chapter 3.4.1), a poor

correlation between the presence of symptoms and the rate of gastric emptying

(Horowitz et al. 1991). Established prokinetic agents in most common use include

metoclopramide, domperidone and erythromycin and, now much less frequently,

cisapride (Table 4.1). There are a number of novel therapeutic alternatives under

investigation.

Table 4.1 Commonly used prokinetic agents

Drug Actions Dose Route Adverse effects

Metoclopramide D2 antagonist 10 mg three to four times daily IV, SC, Dystonia, tardive dyskinesia,5-HT3 antagonist IM, Oral sedation, hyperprolactinaemia5-HT4 agonist

Domperidone D2 antagonist 10 – 20 mg two to four times daily Oral Hyperprolactinaemia

Erythromycin Motilin agonist 50 – 250 mg three to four times daily IV, oral Nausea, vomiting, abdominal pain, dysrhythmia

Cisapride 5-HT4 agonist 10 – 20 mg three to four times daily Oral Dysrhythmia, abdominal pain,5-HT3 antagonist diarrhoea

Chapter 4

65

4.3.1 Metoclopramide

Metoclopramide is a substituted benzamide with both prokinetic and antiemetic

properties. Metoclopramide is a central and peripheral dopamine (D2) receptor

antagonist, a 5-hydroxyriptamine-4 (5-HT4) agonist, a 5-HT3 antagonist (at high

doses) and a cholinesterase inhibitor (Syed et al. 2005). The prokinetic effects, which

are limited primarily to the proximal gut, occur as a result of D2 receptor antagonism

in the myenteric plexus, facilitated acetylcholine release from enteric cholinergic

neurons (5-HT4 receptors) and muscarinic receptor sensitisation and, therefore,

smooth muscle contraction (Tonini et al. 2004).

Metoclopramide accelerates gastric emptying and intestinal transit by increasing the

tone and amplitude of antral contractions, relaxing the pylorus and duodenal bulb,

coordinating antroduodenal motility, and enhancing peristalsis of the duodenum and

jejunum (Parkman et al. 2004). Metoclopramide also normalises gastric slow-wave

dysrhythmias via inhibition of dopamine-induced gastric smooth muscle relaxation

(Haans and Masclee 2007). Metoclopramide has also been shown to improve

symptoms of postprandial fullness and nausea (Shakil et al. 2008). The antiemetic

effects of metoclopramide are related to antagonism of D2 and 5-HT3 receptors on

vagal and brainstem pathways in the area postrema and in the vomiting centre. Long-

term efficacy of metoclopramide has not been substantiated and the prokinetic effects

may potentially diminish over time (Lata and Pigarelli 2003).

Metoclopramide is available in several formulations. It may be given orally,

intravenously and subcutaneously; parenteral administration proves useful in cases

where symptoms are severe. Time to onset is 3 minutes when administered

intravenously, 10 15 minutes following intramuscular injection and 30 60

Chapter 4

66

minutes when taken orally, with pharmacological effects lasting 1 2 hours.

Metoclopramide has a 4 hour half-life and is primarily renally cleared. The usual oral

dose is 10 mg four times daily, taken 30 minutes before meals and at bedtime

(Parkman et al. 2004, Rayner and Horowitz 2005).

The use of metoclopramide is somewhat restricted by a high prevalence of central

nervous system adverse effects. Mild central neurological effects, including

restlessness, agitation, dizziness and drowsiness, occur in up to 40 % of patients,

while dystonic reactions, characterised by trismus, torticollis and occulogyric crisis,

occur in approximately 1 % of patients (Lata and Pigarelli 2003). During more

prolonged use, tardive dyskinesia may develop, which is potentially irreversible.

Metoclopramide may also cause hyperprolactinaemia and, thereby, galactorrhoea due

to antidopaminergic effects on the central nervous system (Abell et al. 2006, Smith

and Ferris 2003).

4.3.2 Domperidone

Domperidone is a benzamide derivative with specific antagonist activity on D2

receptors (Parkman et al. 2004). Its mechanism of action, much like metoclopramide,

is via dopaminergic antagonism of the myenteric plexus with a resultant increase in

the duration of antral and duodenal contractions (Smith and Ferris 2003). It is

relatively impermeable to the blood-brain barrier and, therefore, causes less central

neurological adverse effects. Domperidone also antagonises D2 receptors in the

chemoreceptor trigger zone in the area postrema (outside of the blood-brain barrier)

and, as such, imparts antiemetic activity (Tonini et al. 2004).

Chapter 4

67

Domperidone is available only in tablet form (in Australia) and the usual dose is 10

mg three times daily before meals and an additional dose at bedtime (Rayner and

Horowitz 2005). Owing to its more favourable central adverse effect profile, doses of

domperidone may be increased more readily than with metoclopramide, although

antidopaminergic adverse effects may still occur (Tonini et al. 2004). The plasma

half-life after a single oral dose is 7 hours in healthy subjects and is more prolonged

in patients with severe renal insufficiency.

Domperidone appears to be effective in the management of symptomatic

gastroparesis, including that associated with diabetes mellitus (Sugumar et al. 2008),

with fewer side effects than metoclopramide (Patterson et al . 1999). Acute

administration of domperidone has been reported to accelerate gastric emptying of

solids and liquids in both healthy subjects and in diabetic patients with autonomic

neuropathy, with the greatest effect being observed in those patients with the greatest

delay in gastric emptying (Horowitz et al. 1985). Chronic administration (35 – 51

days) has, however, been reported to accelerate liquid, but not solid, emptying

(Horowitz et al. 1985). Domperidone has also been shown to alleviate upper

gastrointestinal symptoms and improve quality of life in patients with diabetes

(Silvers et al. 1998, Sugumar et al. 2008).

4.3.3 Erythromycin

Erythromycin is a macrolide antibiotic that also has the ability to act as a motilin

agonist through its interaction with motilin receptors, which are abundant in the

enteric nervous system, particularly in the stomach and upper gastrointestinal tract.

Motilin is a polypeptide hormone which resides in endocrine cells of the distal

stomach and duodenum (Abell et al. 2006). Erythromycin accelerates gastric

Chapter 4

68

emptying, increases the frequency and amplitude of antral and duodenal contractions

and induces phase III MMC contractions (Haans and Masclee 2007). The effect of

erythromycin on gastrointestinal symptoms is controversial, although there is

evidence to suggest that it significantly improves bloating in patients with functional

dyspepsia and gastroparesis (Arts et al. 2005).

Erythromycin, which has a half-life of 60 – 150 minutes, is available in oral and

intravenous formulations. Intravenous erythromycin is thought to be the most potent

prokinetic drug when given intravenously and may serve a role acutely, in the initial

management of gastroparesis (Park and Camilleri 2006). The typical oral dose is 50 –

250 mg three to four times per day and intravenous dosing is usually 1 – 2 mg/kg

(max 6 mg/kg) three times per day. Intravenous administration should be restricted to

short-term use for an acute exacerbation of diabetic gastroparesis, while oral

administration should be the chosen route for chronic use, with the liquid formulation

preferred over capsules as the need for dissolution and disintegration to occur in the

stomach is eliminated (Parkman et al. 2004). Chronic use is restricted, however, due

to the potential risk of bacterial resistance, in addition to the development of

tachyphylaxis due to down-regulation of motilin receptors. The major adverse effects

of erythromycin include abdominal pain and cramping, nausea, diarrhoea, vomiting

and rash. Erythromycin may also cause adverse cardiac effects, including Q-T

interval prolongation and torsades de pointes (Rayner and Horowitz 2005). The

gastroprokinetic effect of erythromycin has been shown to be attenuated during

hyperglycaemia in both healthy subjects (Jones et al. 1999b) and patients with type 1

diabetes (Petrakis et al. 1999a). Even variations in the blood glucose concentration

that are within the normal postprandial range diminish the gastrokinetic effect of

erythromycin (Jones et al. 1999a).

Chapter 4

69

4.3.4 Cisapride

Cisapride was arguably the drug of first choice for the treatment of gastroparesis,

until the discovery of its capacity to cause serious adverse cardiac events, including

prolongat ion of the Q-T interval in a dose-dependent manner and associated

potentially fatal cardiac dysrhythmias, including torsades de pointes (Parkman et al.

2004). Consequently, its use has been restricted in many countries, including

Australia, where it is available for certain patients under special access from the

manufacturer. Cisapride acts as a partial 5-HT4 agonist and causes release of

acetylcholine from the myenteric plexus. Cisapride accelerates gastric emptying,

improves antral, jejunal and colonic motility and increases the number of antral,

duodenal and pyloric pressure waves to promote expulsion of stomach contents

(Fraser et al. 1994). It also increases lower oesophageal sphincter pressure and as

such, has efficacy in the treatment of gastro-oesophageal reflux disease. Cisapride

also exerts antiemetic effects, which are thought to be due to its weak antagonism of

5-HT3 receptors (Smith and Ferris 2003).

Cisapride is formulated as tablets and oral liquid, and the recommended dose, which

is to be kept as low as possible, is 10 20 mg two to four times daily, at least 30

minutes before food. Time to maximal concentration is 2 hours and the half-life is 7

10 hours. Cisapride is metabolised by cytochrome P450 3A4 (CYP3A4) and,

therefore, has the potential to interact with a substantial number of drugs which

inhibit this enzyme (Abell et al. 2006). Additional contraindications to the use of

cisapride include pre-existing cardiac dysrhythmias, family history of Q-T interval

prolongation and concomitant use of medications which have the ability to prolong

the Q-T interval (e.g. sotalol, amiodarone) (Parkman et al. 2004). Combination of

cisapride with erythromycin is contraindicated, due not only to CYP3A4 interaction,

Chapter 4

70

but also to their independent ability to prolong the Q-T interval (Rayner and

Horowitz 2005). In addition to dysrhythmias, other adverse effects of cisapride

include abdominal pain and cramping, diarrhoea and headache. Long-term efficacy of

cisapride is uncertain and its use is generally discouraged.

A systematic analysis comparing the effects of the conventional prokinetic agents

metoclopramide, domperidone, erythromycin and cisapride has been performed

(Sturm et al. 1999) with the following conclusions: (i) erythromycin was found to be

superior with respect to acceleration of gastric emptying when compared to other

prokinetics, and (ii) both erythromycin and domperidone were apparently the most

effective in improving gastrointestinal symptoms. A consistent finding in all studies

included in the systematic analysis was a lack of association between changes in

gastric emptying rate and improvement in symptoms (Sturm et al. 1999).

4.3.5 Antiemetic agents

While some prokinetic agents may offer symptomatic relief, antiemetic agents are

frequently required to alleviate nausea and vomiting. The phenothiazine antiemetics,

including promethazine and prochlorperazine, act as dopamine antagonists in the area

postrema and can be administered orally, rectally and parenterally. The recommended

dose of promethazine is 12.5 – 50 mg every 4 - 6 hours, while prochlorperazine is

given at a dose of 5 – 10 mg every 6 hours. Sedation and extrapyramidal side effects

are not uncommon with both drugs (Rabine and Barnett 2001). Selective 5-HT3

receptor antagonists, such as ondansetron and tropisetron, have been used

successfully for chemotherapy-induced nausea and vomiting, however, their efficacy

in the treatment of gastroparesis has not yet been substantiated (Rayner and Horowitz

2005).

Chapter 4

71

4.3.6 Novel therapies

There is a clear need for new treatment alternatives and recent research has focussed

on the development of agents that possess prokinetic and antiemetic activity, but

which have fewer side effects than existing, conventional prokinetic agents (Table

4.2).

Table 4.2 Novel prokinetic agents

Drug class Drug name

Motilin receptor agonist Alemcinal (ABT-229)

KC11458

Mitemcinal

Motilin-related peptide Ghrelin

5-HT4 receptor agonist Tegaserod

D2 receptor antagonist Levosulpiride

D2 receptor antagonist/ Itoprideacetylcholinesterase inhibitor

Synthetic erythromycin analogues, termed ‘motilides’ and ‘motilactides’, that possess

prokinetic activity, but which are devoid of antibiotic properties, and may have

overcome the tachyphylaxis issue, are in development. Alemcinal (ABT-229), one of

the first of this class of motilin receptor agonists, was tested in patients with

functional dyspepsia (with normal and delayed gastric emptying) (Talley et al. 2000),

and in diabetic gastroparesis (Talley et al. 2001). These studies demonstrated a lack

of efficacy with no relief (Talley et al. 2000), or a worsening (Talley et al. 2001), of

Chapter 4

72

symptoms compared with placebo. Another motilin agonist, KC11458, was studied in

patients with diabetic gastroparesis and shown not to accelerate gastric emptying or

improve symptoms (Russo et al. 2004). More recently, mitemcinal, another motilin

receptor agonist, was found to have prokinetic activity and improve upper

gastrointestinal symptoms in patients with type 1 and 2 diabetes (McCallum and

Cynshi 2007a, McCallum and Cynshi 2007b).

Ghrelin (Chapter 1.4.5.5), bearing structural resemblance to motilin, has received

much attention recently due to its important role in appetite and body weight

regulation. As discussed (Chapter 1.4.5.5), ghrelin has been shown to exert

gastroprokinetic effects in healthy volunteers (Levin et al. 2006) and patients with

diabetic gastroparesis (Murray et al. 2005). In patients with idiopathic gastroparesis,

ghrelin accelerated gastric emptying and also reduced meal-related symptoms (Tack

et al. 2005). Accordingly, there is interest in the development of ghrelin-like drugs

which can be used to treat gastroparesis, such as the potent and specific ghrelin

agonist, TZP-101, recently developed by Tranzyme Pharma (Fraser et al. 2008).

Tegaserod, a partial agonist at 5-HT4 receptors, is indicated primarily in the treatment

of constipation-predominant irritable bowel syndrome (Galligan and Vanner 2005).

There was inconsistent evidence that it accelerated gastric emptying (Degen et al.

2005, Talley et al. 2006), but clinical trials demonstrating gastroprokinetic effects of

tegaserod in patients with gastroparesis are lacking. Moreover, tegaserod was

removed from the market in early 2007, as pooled clinical trial data revealed a higher

incidence of myocardial infarction, stroke and unstable angina in patients taking

tegaserod compared with those taking placebo (Hammerle and Surawicz 2008).

Chapter 4

73

The D2 receptor antagonist, levosulpiride, has been studied in patients with

gastroparesis and dyspeptic symptoms and there is evidence to indicate that it

possesses gastroprokinetic properties (Mansi et al. 1995, Mansi et al. 2000, Mearin et

al. 2004). Although a relatively old compound, and available only in some countries

(e.g. Belgium, Spain) (Tack 2008), it is thought that levosulpiride may confer

advantages over other conventional D2 receptor antagonists on the market, namely,

metoclopramide and domperidone, and further studies are needed.

More recently, substantial research has focused on itopride, a derivative of

metoclopramide, which stimulates gastrointestinal motor activity through synergistic

effects of D2 receptor blockade and acetylcholinesterase inhibition (Iwanaga et al.

1994, Iwanaga et al. 1996). Itopride has a number of potential advantages over

conventional prokinetics, including its inability to traverse the blood-brain barrier,

and therefore, lacks central neurological effects. Moreover, it is not metabolised by

cytochrome P450 enymes, thereby, reducing the opportunity for drug-drug

interactions (Mushiroda et al. 2000). Studies in dogs have demonstrated

gastroprokinetic activity (Iwanaga et al. 1996) and moderate antiemetic action

(Iwanaga et al. 1996). Human studies in functional dyspepsia have shown conflicting

results; significant improvements in symptoms have been reported in one study

(Holtmann et al. 2006), but this was not confirmed in subsequent studies (Choung et

al. 2007, Talley et al. 2008). Itopride had no effect on gastric emptying in healthy

humans (Choung et al. 2007) but this does not exclude the possibility of an effect in

gastroparesis. Chapter 9 reports a study which has evaluated the effects of itopride on

gastric emptying in diabetic gastroparesis.

Chapter 4

74

Since the pathogenesis of diabetic gastroparesis has been attributed, in part, to vagal

nerve dysfunction occurring as part of a generalised autonomic neuropathy, recent

research has focussed on C-peptide, or ‘connecting-peptide’. There is evidence that

C-peptide has the capacity to improve autonomic nerve function in type 1 diabetes

(Ido et al. 1997, Johansson et al. 1996, Johansson et al. 2000, Sima et al. 2001,

Steiner 1978). C-peptide links the A and B chains of proinsulin and, during insulin

synthesis, is cleaved from proinsulin and released in amounts equimolar with those of

insulin (Steiner 1978). C-peptide has traditionally been considered biologically

inactive, however, recent studies indicate that C-peptide has preventive and

ameliorating effects on the chronic complications of type 1 diabetes, including

autonomic nerve dysfunction, in experimental animal models (Ido et al. 1997, Sima

et al. 2001) and in humans (Johansson et al. 1996, Johansson et al. 2000), following

both acute and chronic administration.

While a relationship between autonomic nerve dysfunction and delayed gastric

emptying does exist, albeit weak (Buysschaert et al. 1987, Horowitz et al. 1991),

there have been no studies to date that have evaluated the effect of C-peptide on

gastric emptying in type 1 diabetes. This is addressed in the study reported in Chapter

10.

4.4 Other medical therapies

4.4.1 Intrapyloric botulinum toxin

Pylorospasm, that is prolonged periods of pyloric tone and phasic contractions

localised to the pylorus, has been reported to occur in patients with diabetic

gastroparesis (Mearin et al. 1986). Botulinum toxin is a bacterial toxin which inhibits

Chapter 4

75

the release of acetylcholine and resultantly causes muscle paralysis (Friedenberg et

al. 2008). Intrapyloric injection of botulinum toxin via endoscopy (25 IU into each of

four quadrants of the pylorus) has been reported to accelerate gastric emptying and

improve symptoms in a number of earlier studies (Bromer et al. 2005, Ezzeddine et

al. 2002, Lacy et al. 2002, Lacy et al. 2004, Miller et al. 2002), although none of

these was controlled. Two, small controlled trials have been conducted (Arts et al.

2006, Arts et al. 2007), one of which demonstrated an improvement in solid, but not

liquid, gastric emptying and an improvement in several meal-related symptoms (Arts

et al. 2006). The other study (Arts et al. 2007) failed to show any benefit of

botulinum toxin over placebo with respect to gastric emptying and symptoms.

Consistent with these findings, a recent, large, randomised, double-blind, placebo-

controlled trial demonstrated that botulinum toxin, relative to placebo, did not

accelerate gastric emptying nor improve symptoms (Friedenberg et al. 2008).

Accordingly, intrapyloric botulinum toxin should not be recommended unless in the

context of a clinical trial, or possibly when all other therapies have failed

(Friedenberg et al. 2008).

4.4.2 Gastric electrical stimulation

Gastric electrical stimulation, or gastric pacing, has been used to treat drug-refractory

gastroparesis via implantable gastric stimulators. There are three main methods in

use: (i) high energy, low frequency pulses, (ii) low energy, high frequency pulses and

(iii) sequential gastric neural stimulation (Maranki and Parkman 2007). Gastric

stimulation by high energy, low frequency pulses entrains a regular slow wave

rhythm of 3 cycles/min and has been reported to improve both gastric emptying and

symptoms in patients with gastroparesis (McCallum et al. 1998). Low energy, high

frequency stimulation involves the implantation of two electrodes onto the serosal

Chapter 4

76

surface overlying the pacemaker area on the greater curvature of the stomach by

laparoscopy or laparotomy (Park and Camilleri 2006). The electrodes are connected

to a neurotransmitter implanted subcutaneously in the abdominal wall, which

resembles a cardiac pacemaker, and stimulates the smooth muscle layer by high

frequency (12 cycles/min), low energy pulses of short duration. High frequency

gastric electrical stimulation has been reported to reduce vomiting frequency,

improve symptoms, improve quality of life and reduce hospitalisations, with no, or

only modest improvement, in gastric emptying (Abell et al. 2002, de Csepel et al.

2006, Familoni et al. 1997, Lin et al. 2004, Lin et al. 2006, Mason et al. 2005,

McCallum et al. 2005). Only one study has been controlled (Abell et al. 2002), which

demonstrated marginal improvement in gastric emptying at 12 months, although there

was no change in gastric emptying initially. The high frequency method does not

entrain the gastric slow wave. Infection poses the main complication associated with

implantation of the neurotransmitter and 5 – 20 % of patients require removal of the

device. The third method of gastric electrical stimulation involves the placement of

ring-shaped electrodes encircling the antrum (Mintchev et al. 2000). Sequential

stimulation of these electrodes induces propagated antral contractions with the aim of

forceful gastric emptying. This method has shown promising results in dogs

(Mintchev et al. 2000), but is yet to be tested in humans.

4.4.3 Surgery

Surgery is rarely routinely recommended. The majority of patients suffering

gastroparesis will respond to medical treatment, however, 2 – 5 % of patients have

severe gastroparesis refractory to drug therapy and require multiple hospitalisations

(Syed et al. 2005). Surgical interventions in these patients are aimed at symptom

palliation, stomach decompression, provision of enteral access and acceleration of

Chapter 4

77

gastric emptying. A venting gastrostomy may be inserted by endoscopy, surgery or

via radiological techniques and can be opened periodically to relieve meal-associated

symptoms including nausea, bloating and abdominal discomfort (Smith and Ferris

2003). A feeding jejunostomy tube may be inserted surgically or endoscopically, so

as to provide a means by which to deliver hydration and nutrition when patients can

no longer tolerate oral or gastrostomy feeding. Insertion of percutaneous tubes,

however, is associated with significant complications (Syed et al. 2005). Major

surgical resections, including near-total or partial gastrectomies with Roux-en-Y

reconstructions, have been associated with a reduction in symptoms (Eckhauser et al.

1988) and have been used with a degree of success in diabetic gastroparesis (Bell and

Ovalle 1999, Reardon et al. 1989).

In patients with diabetes, pancreatic transplantation (Kennedy et al. 1990) and

pancreatic islet transplantation (Lee et al. 2005) have been reported to halt, or

reverse, diabetic neuropathy. Furthermore, significant improvements in gastric

function (as assessed by electrogastrography) and gastrointestinal symptoms have

been reported in pat ients who have undergone combined pancreas-kidney

transplantation, compared with kidney-alone transplantation (Hathaway et al. 1994),

although these are uncontrolled studies.

Chapter 4

78

4.5 Conclusions

Treatment of gastroparesis poses substantial clinical challenges and current

therapeutic options are suboptimal. Conventional prokinetic agents, including

metoclopramide, domperidone and erythromycin, predominate the clinical treatment

of gastroparesis, and although newer treatment options are emerging, no single agent

has proven to be superior to the conventional agents to date. Accordingly, further

well-designed and well-controlled research is imperative in the development of new,

effective approaches in the treatment of this challenging condition. Chapter 9

evaluates a novel prokinetic agent, itopride, in longstanding diabetic gastroparesis

and Chapter 10 reports the effects of C-peptide on gastric emptying in longstanding

type 1 diabetes.

Chapter 5

79

5Chapter 5EFFECT OF EXENDIN(9-39), A GLUCAGON-LIKE PEPTIDE-1

(GLP-1) ANTAGONIST, ON GASTRIC EMPTYING AND

GLYCAEMIA IN HEALTHY HUMANS

5.1 Summary

The ‘incretin’ hormone, glucagon-like peptide-1(7-36) amide (GLP-1), when given

exogenously in pharmacological doses, lowers fasting and postprandial glycaemia

through stimulation of insulin release, inhibition of glucagon secretion and slowing of

gastric emptying. Studies using the specific GLP-1 antagonist, exendin(9-39), have

established that endogenous GLP-1 modulates insulin and glucagon secretion. The

role of endogenous GLP-1 in the regulation of gastric emptying is uncertain. The

aims of this study were to determine the effects of endogenous GLP-1, using

Chapter 5

80

exendin(9-39), on gastric emptying of a solid meal and glycaemia, in healthy

subjects. Eleven healthy subjects (8 male, 3 female; age: 48.6 ± 6.2 yr (range 19 71

yr), body mass index: 26.7 ± 1.2 kg/m2) received exendin(9-39) (300 pmol/kg/min) or

placebo (isotonic saline) intravenously (from t = 30 180 min) on two separate days

in a double-blind, randomised, crossover design. Gastric emptying (scintigraphy)

(from t = 0 180 min) and glycaemia (glucometer) (from t = 30 240 min) were

measured following ingestion of a mashed potato meal (~ 2600 kJ) labelled with 20

MBq 99mTc-sulphur colloid. When compared with placebo, exendin(9-39) accelerated

gastric emptying (P = 0.0001) from t = 45 180 min (P < 0.005 for all), and reduced

the 50 % emptying time (T50) (exendin(9-39): 67.7 ± 7.6 min vs. placebo: 83.4 ± 7.1

min; P = 0.0003). Postprandial blood glucose levels were greater (P = 0.003) during

infusion with exendin(9-39) compared with placebo (i.e. t = 30 180 min), and the

peak blood glucose was also greater on exendin(9-39): 10.8 ± 0.6 mmol/L vs.

placebo: 9.5 ± 0.7 mmol/L; P = 0.03. The magnitude of the rise in blood glucose at 60

min was inversely related to the T50 when data from both studies were combined (r =

0.46, P = 0.04). It is concluded that GLP-1 plays a physiological role to slow gastric

emptying in healthy subjects, which impacts positively on postprandial glycaemic

excursions.

5.2 Introduction

Glucagon-like peptide-1(7-36) amide (GLP-1), a proglucagon-derived peptide, is an

‘incretin’ hormone which is released predominantly from mucosal L-cells of the

distal ileum and colon in response to the presence of luminal nutrients, including

carbohydrate (Nauck et al. 1993), fat (Feinle et al. 2003), and protein (Herrmann et

al. 1995). Plasma levels of GLP-1 increase some three-fold following a meal in

humans (Edwards et al. 1999, Kreymann et al. 1987), the magnitude of this increase

Chapter 5

81

being dependent on the small intestinal nutrient load (Pilichiewicz et al. 2007) and

length of small intestine exposed to nutrient (Little et al. 2006b). When given

exogenously in pharmacological concentrations, GLP-1 increases insulin, and

suppresses glucagon, secretion in humans with and without type 2 diabetes during

hyperglycaemia (Nauck et al. 1993). Exogenous GLP-1 has also been shown to

stimulate insulin gene expression and biosynthesis (Fehmann and Habener 1992),

enhance peripheral glucose disposal (D'Alessio et al. 1994) and suppress energy

intake (Brennan et al. 2005, Flint et al. 1998) in humans. Pharmacological doses of

GLP-1 also affect gastrointestinal function, including a dose-related slowing of

gastric emptying (Delgado-Aros et al. 2002, Flint et al. 2001, Little et al. 2006a,

Naslund et al. 1999, Nauck et al. 1997a), occurring as a result of relaxation of the

proximal stomach (Delgado-Aros et al. 2002, Schirra et al. 2002), inhibition of

antropyloric pressure waves and antroduodenal motility (Schirra et al. 2000, Schirra

et al. 2006) and stimulation of isolated pyloric pressure waves and pyloric motility

(Schirra et al. 2000, Schirra et al. 2006). These diverse effects of exogenous GLP-1

act in concert to reduce fasting, and particularly, postprandial, glucose concentrations,

and have provided the impetus for the development of GLP-1 analogues and GLP-1

receptor agonists for the management of type 2 diabetes. The dominant mechanism

by which exogenous GLP-1 (Little et al. 2006a, Meier et al. 2003) and its analogues

(Linnebjerg et al. 2008) reduce postprandial glycaemic excursions in health and type

2 diabetes appears to be slowing of gastric emptying (Linnebjerg et al. 2008, Little et

al. 2006a, Meier et al. 2003).

In contrast to the pharmacological effects of GLP-1, the physiological role(s) of GLP-

1 remain poorly defined. These can be explored using the specific, competitive, GLP-

1 receptor antagonist, exendin(9-39) (Edwards et al. 1999, Salehi et al. 2008, Schirra

Chapter 5

82

et al. 1998a). Blockade of GLP-1 receptors with exendin(9-39) (Schirra et al. 1998a),

during intravenous glucose, has been shown to abolish the insulinotropic effect, and

attenuate the glucose-lowering effect, an apparently ‘physiological’ dose of GLP-1 in

both animals (D'Alessio et al. 1996, Kolligs et al. 1995, Wang et al. 1995) and

humans (Edwards et al. 1999, Schirra et al. 1998a), in a dose-dependent manner

(Schirra et al. 1998a). Furthermore, in humans, exendin(9-39) increases fasting blood

glucose (Edwards et al. 1999, Schirra et al. 2006) and potentiates the glycaemic

response (by ~ 35 %) to an oral glucose load (Edwards et al. 1999) and a duodenal

glucose infusion (Schirra et al. 2006). In these studies, exendin(9-39) has been shown

to increase both fasting (Schirra et al. 2006) and postprandial glucagon

concentrations (Edwards et al. 1999), with inconsistent effects on insulin (D'Alessio

et al. 1996, Edwards et al. 1999, Kolligs et al. 1995, Schirra et al. 2006).

It has been recognised that the adverse effects of exendin(9-39) on postprandial

glycaemia may potentially relate to acceleration of gastric emptying (Chelikani et al.

2005, Edwards et al. 1999, Kumar et al. 2008, Salehi et al. 2008, Schirra et al. 2006),

but there is little information about this. In both healthy subjects (Horowitz et al.

1993b) and patients with type 2 diabetes (Horowitz et al. 1991, Jones et al. 1995b),

even modest differences in the rate of gastric emptying can have a substantial effect

on the glycaemic response to a meal. Endogenous GLP-1 has been shown to inhibit

antroduodenal motility and stimulate pyloric motility during intraduodenal glucose

infusion (Schirra et al. 2006), as well as increase proximal gastric accommodation

and compliance, and stimulate pyloric motility during duodenal infusion of a mixed

liquid meal at a rate designed to mimic normal gastric emptying (Schirra et al. 2009).

The effect of intravenous infusion of exendin(9-39) on gastric emptying has been

evaluated in rats, and exendin(9-39) was shown to completely block the inhibition of

Chapter 5

83

gastric emptying induced by intravenous GLP-1 (Chelikani et al. 2005). In mice,

intraperitoneal exendin(9-39) has been reported to accelerate gastric emptying by 10

% (Kumar et al. 2008). To date, only one study (Salehi et al. 2008) has evaluated the

effects of endogenous GLP-1 on gastric emptying in humans; in this study, following

ingestion of a 75 g oral glucose load in healthy subjects, there was no apparent

difference in gastric emptying between exendin(9-39) and control infusions, however,

the method used to assess gastric emptying (absorption of xylose) was insensitive

(Salehi et al. 2008). - The ‘gold standard’ technique for measurement of gastric

emptying is scintigraphy (Horowitz et al. 1991).

The aims of this study were to determine the effect of exendin(9-39) on gastric

emptying, as measured by scintigraphy, of a high-carbohydrate, semisolid meal in

healthy humans, and to relate any change in gastric emptying to that on glycaemia.

5.3 Materials and Methods

5.3.1 Subjects

Eleven healthy subjects (8 male, 3 female; age: 48.6 ± 6.2 yr (range 19 – 71 yr), body

mass index: 26.7 ± 1.2 kg/m2) were studied. Subjects were randomly selected from

volunteers responsive to advertisements placed on hospital and university notice

boards. All subjects were non-smokers and none were taking medication known to

influence gastrointestinal function. None had a history of diabetes, gastrointestinal

disease or surgery, significant respiratory, cardiac or hepatic disease, chronic alcohol

abuse or epilepsy. All subjects were screened for hepatic disease and none had

evidence of significantly elevated (greater than twice the upper limit of normal) liver

Chapter 5

84

function enzymes (AST, ALT, LDH, GGT). Female subjects were not pregnant or

lactating.

5.3.2 Experimental protocol

Each subject underwent two randomised, double-blind, placebo-controlled studies,

separated by an interval of 5 – 9 days. On each study day, subjects received an

intravenous (iv) infusion of either (i) exendin(9-39) (RP10872, GenScript Corp.,

Piscataway, NJ, USA) at 300 pmol/kg/min, or (ii) placebo (saline 0.9 %w/v) at a rate

of 1 mL/min, commencing 30 min before ingestion of a meal (i.e. at t = 30 min) and

continuing for 210 min, i.e. until t = 180 min.

On each study day, subjects attended the Department of Nuclear Medicine, Positron

Emission Tomography and Bone Densitometry at 09:00 h following an overnight fast

(14 h for solids and 12 h for liquids). Intravenous cannulae were inserted into the

medial antecubital vein of each forearm; one for blood sampling and the other for

intravenous infusion of exendin(9-39) or placebo.

Written, informed consent was obtained from each subject prior to their enrolment.

The protocol and advertisements were approved by the Human Research Ethics

Committee of the Royal Adelaide Hospital and all studies were performed in

accordance with the Declaration of Helsinki.

5.3.3 Measurement of gastric emptying

Gastric emptying was measured for 3 hours using a standardised, single-isotope,

scintigraphic test (Collins et al. 1983). The test meal comprised 65 g powdered potato

(Deb Instant Mashed Potato, Continental, Epping, NSW, Australia) and 20 g glucose,

Chapter 5

85

reconstituted with 200 mL water labelled with 20 MBq 99mTc-sulphur colloid, and

mixed with 45 g melted margarine (Fairy Cooking Margarine, Peerless Foods,

Braybook, Vic, Australia). The energy content of the meal was ~ 2600 kJ (total

carbohydrate 62 g, total fat 40 g). The time for meal ingestion was standardised at

between 5 – 8 min, and t = 0 min was defined as the time of meal completion.

Radioisotopic data were acquired with the subject seated with their back against a

gamma camera (GEnie; GE Healthcare Technologies, Milwaukee, WI, USA) at 1-

minute intervals for the first hour and at 3-minute intervals thereafter, for a total of

180 minutes. Data were corrected for subject movement, radionuclide decay and –

ray attenuation, the latter using correction factors derived from a lateral image of the

stomach (Collins et al. 1983). The lag phase (Tlag), determined visually as the time

between meal completion and the appearance of radioactivity in the proximal small

intestine (Collins et al. 1983), was quantified. Regions-of-interest were drawn around

the total stomach, which was further divided into proximal and distal stomach

regions, by dividing the long axis of the stomach into two equal halves (Jones et al.

1995b). Gastric emptying curves (expressed as “percentage retention” over “time”)

were thus derived for total, proximal and distal stomach (Jones et al. 1995b). The

intragastric retention at t = 0, 15, 30, 45, 60, 75, 90, 105, 120, 150 and 180 min was

calculated; the time taken for 50 % of the meal to empty (T50) was also quantified.

5.3.4 Measurements of blood glucose and plasma GLP-1,

GIP, insulin and glucagon

During each gastric emptying measurement, venous blood samples (~ 20 mL) were

obtained immediately before (-30 min) commencement of iv infusion, at t = -15 min,

immediately after meal completion (t = 0 min), and then at t = 15, 30, 45, 60, 90, 120,

150, 180, 210 and 240 min, for subsequent measurement of blood glucose and plasma

Chapter 5

86

GLP-1 (total and intact), GIP (total and intact), insulin and glucagon. At the time of

submission of this thesis, hormone data (GLP-1, GIP, insulin and glucagon) were

unavailable.

Blood glucose concentrations were determined immediately using a portable blood

glucose meter (Medisense Precision QID, Abbott Laboratories, Bedford, MA, USA).

In each study, peak blood glucose, defined as the greatest increment above baseline,

and time to peak, were also determined.

5.3.5 Statistical analysis

Data from ten subjects were included in the statistical analyses, as one subject had

markedly delayed gastric emptying (> 3 standard deviations above the mean) on

placebo, and was excluded on the basis of statistical advice. Data (gastric emptying

and blood glucose) were evaluated using repeated measures analysis of variance

(ANOVA) with “treatment” and “time” as factors. Area under the curve (AUC) was

calculated using the trapezoidal rule. Blood glucose data were evaluated both during

the period of infusion of exendin(9-39) (t = 30 180 min), and separately, as the

postprandial rise in blood glucose (t = 0 90 min). Gastric emptying was evaluated

from t = 0 180 min. Student’s t-tests (two-tailed) for paired comparisons were used

to compare sample means (i.e. T50, Tlag, blood glucose AUC, peak blood glucose

and time to peak). Relationships between variables were analysed using linear

regression analysis. All analyses, unless stated otherwise, were performed using

Statview (version 5.0; Abacus Concepts, Berkeley, CA, USA) and SuperANOVA

(version 1.11, Abacus Concepts, Berkeley, CA, USA). Data are presented as mean

values ± standard error of the mean (SEM). An error probability of P < 0.05 was

considered significant in all analyses.

Chapter 5

87

5.4 Results

The studies were well tolerated and no significant adverse events were reported. In

one subject, gastric emptying was markedly delayed on placebo (T50: 229 min), and

all data in this subject were excluded from analyses, on the basis of statistical advice.

5.4.1 Gastric emptying and intragastric distribution

During both treatments, gastric emptying approximated an overall linear pattern after

a very short lag phase (exendin(9-39): 1.3 ± 0.2 min vs. placebo: 1.6 ± 0.3 min; P =

0.34). Gastric emptying was faster (P = 0.0001) with exendin(9-39) compared with

placebo from t = 45 – 150 min (P < 0.0001) and at 180 min (P < 0.005) (Figure 5.1a).

The T50 was also less (exendin(9-39): 67.7 ± 7.6 min vs. placebo: 83.4 ± 7.1 min; P =

0.0003). There was no significant difference in either proximal (P = 0.18; Figure

5.1b), or distal (P = 0.34; Figure 5.1c), gastric emptying between treatments, although

mean proximal gastric retention was less with exendin(9-39) (Figure 5.1b).

Figure 5.1: Retention of a mashed potato meal in the (a) total, (b) proximal and (c) distal, stomach during intravenous infusion of exendin(9-39) (300 pmol/kg/min) and placebo (0.9 %w/v saline at 1mL/min). Data are mean values SEM; n = 10.

Chapter 5

88

5.4.2 Blood glucose concentration

There was no significant difference in blood glucose between treatments at either

baseline, i.e. t = 30 min, (exendin(9-39): 6.18 ± 0.35 mmol/L vs. placebo: 6.11 ±

0.33 mmol/L; P = 0.73), or t = 0 min (exendin(9-39): 6.51 ± 0.41 mmol/L vs. placebo

6.11 ± 0.25 mmol/L; P = 0.22). After the meal, there was a rise (P = 0.0001 for both)

in blood glucose, from t = 0 90 min, which was significant from t = 15 min during

exendin(9-39) (P = 0.0006), and from t = 15 min during placebo (P = 0.007), infusion

(Figure 5.2). At t = 240 min, blood glucose levels were less than baseline (P < 0.05

for both). Blood glucose concentrations were greater (P = 0.03) during exendin(9-39)

compared with placebo (i.e. from t = -30 – 180 min). Peak blood glucose

concentration was also greater on exendin(9-39) compared with placebo (exendin(9-

39): 10.8 ± 0.6 mmol/L vs. placebo: 9.5 ± 0.7 mmol/L; P = 0.003), without any

difference in the time to peak between treatments (exendin(9-39): 81.0 ± 11.9 min vs.

placebo: 79.5 ± 9.2 min; P = 0.83). Total area under the blood glucose curve was also

greater during exendin(9-39) compared with placebo from t = 0 – 90 min (exendin(9-

39): 799.4 ± 126.1 mmol.min/L vs. 703.4 ± 33.2 mmol.min/L; P = 0.003). When

calculated as the change from baseline, area under the blood glucose curve was also

greater during exendin(9-39) compared with placebo from t = 0 – 90 min (exendin(9-

39): 243.2 ± 14.7 mmol.min/L vs. 153.5 ± 17.8 mmol.min/L; P = 0.001).

Chapter 5

89

Figure 5.2: Blood glucose concentrations during intravenous infusion of exendin(9-39) (300 pmol/kg/min) and placebo (0.9 %w/v saline at 1mL/min). Data are mean values SEM; n = 10.

5.4.3 Relationships between blood glucose and gastric

emptying

When data from both studies were combined, there was a significant inverse

relationship between the rise in blood glucose concentration from baseline (t = -30

min) and gastric emptying (T50), i.e. the rise in blood glucose was greater when

gastric emptying was relatively more rapid (e.g. at t = 60 min; r = -0.46, P = 0.04)

(Figure 5.3).

Chapter 5

90

Figure 5.3: Relationship between the magnitude of the rise in blood glucose at 60 min and the T50 during intravenous infusion of exendin(9-39) (300 pmol/kg/min) and placebo (0.9 %w/v saline at 1mL/min). Data are from both treatment visits; n = 20.

5.5 Discussion

This study provides novel insights into the effects of endogenous GLP-1 on gastric

emptying in humans. Intravenous administration of the GLP-1 receptor antagonist,

exendin(9-39), accelerated gastric emptying and, this was associated with an increase

in the glycaemic response to a high-carbohydrate, mashed potato meal in healthy

subjects. These observations establish that GLP-1 plays a physiological role to slow

gastric emptying in health and indicates that this impacts on glycaemia.

In this study, the effect of exendin(9-39) on gastric emptying was evaluated using

scintigraphy, the ‘gold standard’ measure of gastric emptying in humans (Horowitz et

al. 1991), unlike a previous study, which employed an insensitive technique (Salehi

et al. 2008). The magnitude of the observed acceleration of gastric emptying induced

Chapter 5

91

by exendin(9-39) was substantial, with a reduction in the mean 50 % emptying time

of some 15 %. It has been reported that exendin(9-39) affects antropyloroduodenal

motility in both the fasted state and in response to duodenal nutrient infusion (Schirra

et al. 2000, Schirra et al. 2006), however, gastric emptying was not measured in these

studies. There has, hitherto, been only one study (Salehi et al. 2008) which has

evaluated the effects of exendin(9-39) on gastric emptying in humans, which reported

no effect on gastric emptying of a 75 g oral glucose load in 300 mL water, in healthy

subjects. However, in this study, an indirect and relatively insensitive technique

(absorption of xylose) was used to quantify gastric emptying, a test which is based on

the incorrect premise that xylose absorption is exclusively dependent on the rate of

gastric emptying, so that inter-individual differences in metabolism, distribution and

elimination are ignored. In rats, absorption of xylose across the gastric mucosa is

known to be substantial (Stradley et al. 1986). The use of scintigraphy in the present

study also enabled meaurement of intragastric distribution, which is not possible with

the xlyose absorption technique. While there was no significant difference in either

proximal or distal gastric emptying, the pattern of proximal emptying reflected that of

total gastric emptying and was, on average, faster with exendin(9-39) compared with

placebo. Hence, the absence of any difference may well reflect a type 2 statistical

error. The dose of exendin(9-39) in this study was based on that used in two previous

studies in humans (Edwards et al. 1999, Schirra et al. 2006), where 300 – 500

pmol/kg/min was administered intravenously without any adverse effect.

That exendin(9-39) accelerated gastric emptying is consistent with observations in

animal studies, which have assessed the effect of exendin(9-39) on gastric emptying

(Chelikani et al. 2005, Imeryüz et al. 1997, Kumar et al. 2008, Tolessa et al. 1998).

For example, in rats, intravenous exendin(9-39) has been shown to completely block

Chapter 5

92

the GLP-1-induced inhibition of gastric emptying of saline (Chelikani et al. 2005,

Imeryüz et al. 1997). Intracerebroventricular administration of exendin(9-39) to rats

at high (1000 fmol), but not at lower doses (750 fmol and 75 fmol), attenuates the

inhibition of gastric emptying induced by intracerebroventricular administration of

GLP-1 (Imeryüz et a l . 1997). Subcutanous (but not intracerebroventricular)

exendin(9-39) at the same dose (6 pmol/kg) that reversed the delay induced by GLP-

1, has also been shown to completely reverse the inhibitory effect of glucose on

gastric emptying (Imeryüz et al. 1997). In mice, intraperitoneal exendin(9-39) has

been shown to accelerate gastric emptying of a high-carbohydrate, high-fat, semisolid

meal by 10 % (Kumar et al. 2008). However, exendin(9-39) alone, after both

subcutaneous and intracerebroventricular administration, had no effect on gastric

emptying of saline in the absence of GLP-1 (Imeryüz et al. 1997). The effects of

exendin(9-39) are likely to reflect the effects of the test meal on GLP-1 secretion; as

suggested by studies in which gastric emptying of saline was quantified (Imeryüz et

al. 1997). The mashed potato meal was high in both carbohydrate and fat, and has

been shown to be a potent stimulant of GLP-1 secretion (O'Donovan et al. 2004a).

An increase in both fasting and postprandial glucose concentrations with exendin(9-

39) is well documented (Edwards et al. 1999, Schirra et al. 1998a, Schirra et al.

2006), and the observations of the present study are, in general, consistent with this,

although there was no significant difference in baseline blood glucose. The increased

glycaemic response to the meal after exendin(9-39) is likely to, at least in part, reflect

the acceleration of gastric emptying as indicated by the observed inverse relationship

between the rise in blood glucose concentration with the rate of gastric emptying

(T50). Gastric emptying is known to account for approximately 34 % of the variance

in peak plasma glucose after a 75 g oral glucose load in healthy subjects and patients

Chapter 5

93

with diet-controlled type 2 diabetes, so that a more rapid rate of emptying is

associated with higher plasma glucose (Horowitz et al. 1993b). In healthy subjects,

the relationship between duodenal carbohydrate delivery with both glycaemia and

GLP-1 secretion is non-linear (Chaikomin et al. 2005, O'Donovan et al. 2004b), so

that there is little difference in the glycaemic response to intraduodenal glucose at a

rate of 2 kcal/min when compared to 4 kcal/min, but a much greater GLP-1 response

to 4 kcal/min (Pilichiewicz et al. 2007). Relative stimulation of glucagon is likely to

contribute to the observed effects of exendin(9-39) on glycaemia (Schirra et al.

1998a, Schirra et al. 2006). The more rapid gastric emptying induced by exendin(9-

39) would favour an increase, rather than a decrease, in the postprandial insulin

response (Edwards et al. 1999). Hence, plasma hormone data (GLP-1, GIP, insulin

and glucagon), which are unavailable at the time of submission of this thesis, would

allow more definitive interpretation of the observations.

It should also be recognised that as the acceleration of gastric emptying induced by

exendin(9-39) was associated with relative hyperglycaemia, the magnitude of the

effect of exendin(9-39) may have been underestimated, i.e. acute hyperglycaemia,

including variations in blood glucose within the normal postprandial range (Schvarcz

et al. 1997) is known to slow gastric emptying and attenuate the effects of prokinetic

drugs (Jones et al. 1999a, Jones et al. 1999b). For these reasons, the study by Salehi

et al. (2008) employed a glucose clamp with subjects studied during physiological

glycaemia (~ 8.9 mmol/L), making their negative observations even more surprising

in retrospect.

The mechanisms by which GLP-1 inhibits gastric emptying are thought to be vagally-

dependent (Delgado-Aros et al. 2003, Imeryüz et al. 1997, Schirra et al. 1998b);

Chapter 5

94

certainly cholinergic mechanisms mediate the effects of GLP-1 on gastric

accommodation (Schirra et al. 2009). It is therefore possible that the GLP-1-induced

slowing of gastric emptying will be attenuated in patients with autonomic neuropathy,

including patients with diabetes (Delgado-Aros et al. 2003) and this warrants study.

Recently, a study assessing the effect of exendin(9-39) on gastric emptying in mice

have identified a genetic variation in the expression of the Glp1r gene that may be

responsible for differences in gastric emptying (Kumar et al. 2008). This issue would

be of interest to evaluate in humans.

In conclusion, exendin(9-39) accelerates gastric emptying in healthy subjects and this

contributes to potentiation of the glycaemic response; these findings establish that

endogenous GLP-1 is an enterogastrone in humans.

Chapter 6

95

6Chapter 6MEASUREMENT OF GASTRIC EMPTYING OF A HIGH -

NUTRIENT LIQUID IN DIABETIC GASTROPARESIS BY 3D

ULTRASONOGRAPHY

6.1 Summary

Gastric emptying is delayed in 30 – 50 % of patients with longstanding type 1 or type

2 diabetes and may be associated with gastrointestinal symptoms and poor glycaemic

control. Scintigraphy represents the ‘gold standard’ for measurement of gastric

emptying, but is associated with a radiation burden. Three-dimensional (3D)

ultrasonography has recently been demonstrated to provide a valid measure of liquid

gastric emptying in healthy subjects, however, the technique has not been validated in

patients with gastroparesis. The primary aim of this study was to compare

Chapter 6

96

measurements of gastric emptying of a high-nutrient glucose drink by 3D

ultrasonography and scintigraphy in diabetic gastroparesis. Ten patients (8 type 1, 2

type 2, 6 M, 4 F, aged 46.1 ± 4.5 years, body mass index 29.1 ± 1.6 kg/m2, duration

of diabetes 19.6 ± 3.3 years) with diabetic gastroparesis (defined as the retention at

100 min of solid (100 g minced beef) ≥ 61 % and/or 50 % emptying time for liquid

(150 mL 10 % dextrose) ≥ 31 min), were studied. Concurrent measurements of

gastric emptying by scintigraphy and 3D ultrasonography were performed following

ingestion of 75 g glucose in 300 mL water (255 kcal) labelled with 20 MBq 99mTc-

sulphur colloid. There was no significant difference in gastric emptying between the

two techniques (50% emptying times (T50s): scintigraphy – 103.3 ± 10.0 min vs.

ultrasonography - 98.8 ± 10.4 min; P = 0.60). There was a significant correlation

between the scintigraphic and ultrasonographic T50s (r = 0.67, P = 0.03). The limits

of agreement for the T50s were acceptable at -57.22 min and +48.22 min (mean

difference –4.50 min). Blood glucose concentrations after the drink were greater

when gastric emptying was relatively more rapid (e.g. at t = 60 min; scintigraphy: r =

-0.65, P = 0.04; 3D ultrasonography: r = -0.78, P = 0.008). In conclusion, 3D

ultrasonography appears to provide a valid, and non-invasive, measure of gastric

emptying of a high-nutrient liquid in diabetic gastroparesis.

6.2 Introduction

Gastroparesis, defined as delayed gastric emptying (characteristically, a rate of

emptying which is more than two standard deviations outside of a normal range

(Horowitz et al. 2002b)) resulting from disordered gastric motility, is now recognised

to affect 30 – 50 % of patients with longstanding type 1 or type 2 diabetes mellitus

and to be associated with upper gastrointestinal symptoms, impaired oral nutrient and

drug absorption, and poor glycaemic control (Horowitz and Dent 1991, Horowitz et

Chapter 6

97

al. 2006, Rayner et al. 2001, Samsom et al. 2003). The gastroduodenal motor

dysfunctions underlying abnormally slow gastric emptying in diabetes include

decreased antral contractions, impaired coordination between the antrum and

duodenum, increased pyloric contractions and impaired proximal gastric relaxation

(Horowitz et al. 2002b, Rayner et al. 2001). Predictably, intragastric meal distribution

is also frequently abnormal in diabetes (Jones et al. 1995b). The relationship between

gastric emptying of solid and liquid meal components in diabetes is, however,

relatively weak (Horowitz et al. 2006, Jones et al. 1995b). While gastric emptying of

low-nutrient liquids is frequently normal, the prevalence of delayed emptying of

solids and high-nutrient liquids appears to be comparable (Horowitz et al. 2002b,

Horowitz et al. 2006).

Scintigraphy is currently the ‘gold standard’ technique for measurement of gastric

emptying in both clinical and research settings (Horowitz et al. 2002b). With

scintigraphy intragastric meal distribution can also be evaluated, usually by dividing

the total stomach ‘region-of-interest’ half way along the long axis of the stomach into

proximal and distal regions (Jones et al. 1995b). The application of scintigraphy,

particularly in children and pregnant women, is restricted by necessity for exposure to

ionising radiation. Scintigraphy also requires specialised, expensive, equipment and

is, therefore, relatively costly and not always readily available. Alternative techniques

to measure gastric emptying have been developed (Samsom et al. 2003), including

transabdominal ultrasonography (Bolondi et al. 1985, Gentilcore et al. 2006a, Gilja et

al. 1997, Hausken et al. 1992, Holt et al. 1980, Hveem et al. 1996), which does not

expose patients to ionising radiation, is safe, non-invasive, widely available and

allows bedside monitoring due to the portability of the equipment (Bolondi et al.

Chapter 6

98

1985, Gentilcore et al. 2006a, Gilja et al. 1997, Hausken et al. 1992, Holt et al.

1980).

Transabdominal two-dimensional (2D) ultrasonography is used widely to quantify

gastric emptying. This technique measures gastric emptying indirectly by quantifying

changes in antral cross-sectional area, or diameter (Bolondi et al. 1985, Hausken et

al. 1993, Hveem et al. 1996), an approach which necessitates assumptions about the

geometric shape of the antrum prior to volume calculation and is poorly suited to

evaluation of intragastric meal distribution. More recently, three-dimensional (3D)

ultrasonographic techniques have been developed to measure gastric emptying

(Gentilcore et al. 2006a, Gilja et al. 1994, Gilja et al. 1995a, Gilja et al. 1997, Tefera

et al. 2002). Gilja et al. described a 3D ultrasonographic technique based on magnetic

scanhead tracking and demonstrated high accuracy with this technique in vitro (using

a porcine stomach) (Gilja et al. 1998) and in vivo (in healthy humans) (Tefera et al.

2002). Gastric half-emptying times were shown to be more accurate, and less

variable, based on 3D, compared with 2D, data (Gilja et al. 1997). 3D

ultrasonography offers a significant advantage over its 2D counterpart in that

assumptions regarding the geometric shape of the stomach are not required prior to

volume calculation and, by imaging both the proximal and distal stomach, intragastric

meal distribution can also be evaluated by determining the antral area visually (Tefera

et al. 2002). More recently, we have validated 3D ultrasonographic measurement of

gastric emptying against the ‘gold standard’ technique, scintigraphy, in healthy,

normal subjects (Gentilcore et al. 2006a). - Gastric half-emptying times as measured

by 3D ultrasonography and scintigraphy, were not significantly different and

correlated closely following ingestion of both low- and high-nutrient liquids

(Gentilcore et al. 2006a). Whilst 3D ultrasonography may provide a valid measure of

gastric emptying rate in healthy humans, it has hitherto not been applied to patients

Chapter 6

99

with gastroparesis. Furthermore, measurements of intragastric meal distribution by

scintigraphy and 3D ultrasonography have not been compared.

Gastric emptying is pivotal to postprandial blood glucose homeostasis (Horowitz et

al. 1993b, Horowitz et al. 2006, Rayner et al. 2001). Hence, after ingestion of an oral

glucose load (such as that employed in an oral glucose tolerance test) the initial

glycaemic response is greater when the rate of gastric emptying is relatively more

rapid in both healthy subjects, and in patients with diabetes (Horowitz et al. 1993b,

Jones et al. 1995b, Jones et al. 1996).

The primary aim of this study was to compare measurements of gastric emptying and

intragastric distribution of a high-nutrient drink using 3D ultrasonography and

scintigraphy in patients with diabetic gastroparesis. By using a drink containing 75 g

dextrose in 300 mL water, the relationship between glycaemia and the rate of gastric

emptying assessed by 3D ultrasonography could also be evaluated.

6.3 Materials and Methods

6.3.1 Subjects

Ten patients with diabetes mellitus (8 type 1, 2 type 2; 6 male, 4 female), aged 46.1 ±

4.5 years, body mass index 29.1 ± 1.6 kg/m2 and duration of known diabetes 19.6 ±

3.3 years, with known gastroparesis, were studied. All patients had participated in a

recent study which evaluated the effects of the putative prokinetic drug, itopride, on

gastric emptying in diabetes (Stevens et al. 2008). Gastroparesis was defined as ≥ 61

% gastric retention of a 100 g minced beef patty (labelled with 20 MBq 99mTc-sulphur

colloid chicken liver) at 100 min and/or a 50 % emptying time of 150 mL 10 %

Chapter 6

100

dextrose (labelled with 6 MBq 67Ga-EDTA) ≥ 31 min on the placebo test, as assessed

by scintigraphy (Horowitz and Dent 1991). Of the ten patients, four had delayed solid

emptying, five had delayed liquid emptying and one had delayed emptying of both

solid and liquid. Autonomic nerve function had been evaluated using three, non-

invasive, cardiovascular reflex tests, the results of each were scored as 0 = normal, 1

= borderline and 2 = abnormal, for a maximum score of 6. The mean score for

autonomic neuropathy was 3.4 ± 0.3 (range 2 – 5); seven of the ten patients had

evidence of autonomic neuropathy (i.e. total score ≥ 3) and in the remaining three, the

result was ‘borderline’ (i.e. total score ≥ 1 and < 3) (Horowitz and Dent 1991,

Stevens et al. 2008). Patients were not selected on the basis of the presence, or

absence of gastrointestinal symptoms. None had a history of significant hepatic,

cardiac, renal or respiratory disease, chronic alcohol abuse or epilepsy or

gastrointestinal surgery, apart from uncomplicated appendicectomy. All patients were

using insulin, albeit one with type 2 diabetes who took gliclazide only. Patients taking

medication known to influence gastrointestinal motility were excluded and smoking

was prohibited for 24 h prior to, and during, the gastric emptying measurement.

Upper gastrointestinal symptoms were assessed by questionnaire on the morning of

the first study day, prior to consumption of the drink. ‘Gastric’ (anorexia, nausea,

early satiation, abdominal bloating/fullness, vomiting, abdominal pain) and

‘oesophageal’ (dysphagia, heartburn and acid regurgitation) symptoms were graded

as 0 = none, 1 = mild (the symptom could be ignored), 2 = moderate (the symptom

could not be ignored, but did not influence daily activities) and 3 = severe (the

symptom influenced daily activities). As there were nine symptoms, the maximum

possible total score was 27 (Horowitz and Dent 1991, Jones et al. 1995b). Written,

informed consent was obtained from each subject prior to their enrollment in the

study. The protocol was approved by the Human Research Ethics Committee of the

Chapter 6

101

Royal Adelaide Hospital and all studies were performed in accordance with the

Declaration of Helsinki.

6.3.2 Experimental protocol

Patients attended the Department of Nuclear Medicine, Positron Emission

Tomography and Bone Densitometry at 08.30 h after an overnight fast of at least 12 h

for solids and liquids and an intravenous cannula was inserted into a medial

antecubital vein for subsequent blood sampling. Each patient underwent concurrent

measurements of gastric emptying by scintigraphy, 2D ultrasonography and 3D

ultrasonosgraphy (Gentilcore et al. 2006a). The test drink comprised 300 mL water

containing 75 g dextrose (25 %w/v) (255 kcal) labelled with 20 MBq 99mTc-sulphur

colloid. Patients were informed that they would be given glucose, and the nine

patients on insulin adjusted their dose accordingly.

6.3.3 Measurement of gastric emptying

6.3.3.1 Scintigraphy

The drink was administered at t = -2 min and consumed within 2 min. Time zero (t =

0 min) was defined as the time of drink completion and gastric emptying was

monitored for 180 min. Radioisotopic data were acquired with the subject seated with

their back against a gamma camera (GEnie; GE Healthcare Technologies,

Milwaukee, WI, USA) at 1-min intervals for the first hour and at 3-min intervals

thereafter. Data were corrected for patient movement, radionuclide decay and –ray

attenuation, using previously described methods (Gentilcore et al. 2006a, Horowitz

and Dent 1991). The lag phase was determined visually as the time between drink

completion and the appearance of radioactivity in the proximal small intestine.

Chapter 6

102

Regions-of-interest were drawn around the total stomach, which was further divided

into proximal and distal stomach regions, by dividing the long axis of the stomach

into two equal halves (Jones et al. 1995b). Gastric emptying curves (expressed as the

percentage retention over time) were thus derived for total, proximal and distal

stomach (Jones et al. 1995b). The intragastric retention at t = 0, 15, 30, 45, 60, 90,

120, 150 and 180 min was calculated and the time taken for 50 % of the drink to

empty (SCT50) was also quantified.

6.3.3.2 Ultrasonography

Ultrasonography measurements were performed using a Logiq™ 9 ultrasonography

system (GE Medical Systems, Milwaukee, WI, USA) with TruScan Architecture,

including built-in magnetically sensored 3D. For 3D positioning and orientation

measurement (POM), a transmitter was placed close to the subject and a snap-on

sensor attached to a 3.5 C broad spectrum 2.5 - 4 MHz convex transducer (Tefera et

al. 2002). As the transmitter produces a spatially varying magnetic field, and ferrous

and conductive metals distort the magnetic field, all metal objects were removed from

the patient and from the area directly between the POM transmitter and sensor (Liao

et al. 2004). The POM transmitter was placed close (approximately 20 - 30 cm) (Gilja

et al. 1997) to the left side of the subject, at the level of the stomach, so that the

subject was positioned between the ultrasonography scanner and the transmitter.

For 3D data acquisition, subjects were scanned immediately before (t = -5 min), and

after (t = 0 min), drink ingestion, followed by images at t = 0, 15, 30, 45, 60, 90, 120,

150 and 180 min. Subjects were instructed to hold their breath at the end of

inspiration (Liao et al. 2004) and the stomach was scanned by a continuous

translational movement along its long axis, starting proximally at the left subcostal

Chapter 6

103

margin, where the transducer was tilted cranially to image the superior part of the

stomach (Gilja et al. 1997), and moving distally to the gastroduodenal junction (Gilja

et al. 1997, Liao et al. 2004) to produce transverse sections of the entire stomach. On

each occasion, the scanning time approximated 10 seconds. When gastric

contractions were observed, acquisition was interrupted until after the contraction

wave had passed. The stored images were copied to CD-ROM and then transferred to

a Windows workstation. Data processing and volume estimation were performed with

the use of EchoPAC-3D software® (GE Vingmed Sound, Horten, Norway), as

described (Tefera et al. 2002). Regions-of-interest were drawn around the stomach on

sagittal sections, which were subsequently interpolated to produce a 3D image of the

total stomach. The volume of the drink in the total stomach was derived and

expressed as a percentage of the volume at t = 0 min immediately following ingestion

of the drink (i.e. 100 %). The 50 % emptying time (UST50) was also determined

(Gentilcore et al. 2006a). Following 3D reconstruction, the division line between the

proximal and distal stomach was determined as the vertical section at the angular

incisure at the lesser gastric curvature. The volume of drink in the distal stomach was

calculated and the volume of the proximal stomach was then determined by

subtracting antral volume from the total volume (Tefera et al. 2002). Gastric

emptying curves (expressed as the percentage retention over time) were thus derived

for total, proximal and distal stomach at t = 0, 15, 30, 45, 60, 90, 120, 150 and 180

min. Images were assessed for visible amounts of air in the stomach and patients were

excluded from the study if the amount was considered significant.

For 2D data acquisition, the transducer was positioned in the epigastrium by the left

subcostal margin and titled cranially. Sagittal sections were acquired with the left

renal pelvis in longitudinal projection and the left lobe of the liver and tail of the

Chapter 6

104

pancreas serving as internal landmarks. Regions-of-interest were drawn around the

proximal gastric area in the sagittal section by tracing from the top margin of the

fundus and 7 cm downwards along the axis of the stomach, and proximal sagittal area

was recorded (Gilja et al. 1997).

6.3.4 Measurement of blood glucose

Venous blood samples (~ 10 ml) were obtained at t = -2, 15, 30, 45, 60, 90, 120, 150

and 180 min. Blood glucose concentrations were determined immediately using a

portable blood glucose meter (MediSense Optium meter, MediSense Products, Abbott

Laboratories, Bedford, MA, USA).

6.3.5 Statistical analysis

The sample size of ten patients was based on the number of subjects included in our

previous study (Gentilcore et al. 2006a), which demonstrated good agreement

between scintigraphic and 3D ultrasonographic measurements of gastric emptying of

both low-nutrient and high-nutrient drinks in healthy subjects. Data were evaluated

using repeated measures analysis of variance (ANOVA) with ‘treatment’ and ‘time’

as factors. Student’s t-test (two-tailed) for paired comparisons was used to compare

sample means. Relationships between variables were analysed using linear regression

analysis. Limits of agreement analyses were performed according to Bland and

Altman (Bland and Altman 1986), such that the difference between SCT50 and

UST50 was plotted against the mean of the two methods (difference plot) (Bland and

Altman 1986) and limits of agreement were defined as acceptable if within the mean

(2 SD) difference (Bland and Altman 1986). All analyses, unless stated otherwise,

were performed using Statview (version 5.0; Abacus Concepts, Berkeley, CA, USA)

and SuperANOVA (version 1.11, Abacus Concepts, Berkeley, CA, USA). Limits of

Chapter 6

105

agreement analyses were performed by a professional statistician using SAS 9.1 (SAS

Institute Inc., Cary, NC, USA). Data are presented as mean values standard error of

the mean (SEM). A P-value < 0.05 was considered significant in all analyses.

6.4 Results

All patients tolerated the study well and none had visible amounts of intragastric air

significant enough to warrant exclusion from the study. Five patients reported

gastrointestinal symptoms; three of these rated at least one symptom as ‘severe’. The

mean score for upper gastrointestinal symptoms was 1.4 ± 0.5; scores for ‘gastric’

and ‘oesophageal’ symptoms were 1.0 ± 0.5 and 0.4 ± 0.3, respectively.

6.4.1 Gastric emptying and intragastric distribution

With both techniques gastric emptying approximated an overall linear pattern after a

short lag phase (the latter measured scintigraphically as 2.3 ± 0.7 min). There was no

significant difference between the T50s (SCT50: 103.0 10.0 min vs. UST50: 98.8

10.4 min, P = 0.60), nor any difference in the overall curves for total stomach

emptying between the techniques (P = 0.79) (Figure 6.1a). There was a significant

correlation between the scintigraphic and 3D ultrasonographic T50s (r = 0.67, P =

0.03) (Figure 6.2). The limits of agreement for the T50s were -57.22 min and +48.22

min (mean difference –4.5 min) (Figure 6.3).

There was a trend (P = 0.06) for greater retention of the drink in the proximal

stomach using 3D ultrasonography compared with scintigraphy (Figure 6.1b). In

contrast, there was a significant method*time interaction (P = 0.0001) between

scintigraphic and 3D ultrasonographic measurements of distal gastric emptying; from

Chapter 6

106

t = 0 – 45 min and at t = 120 min (P < 0.05 for all), distal gastric retention was greater

using scintigraphy when compared with 3D ultrasonography (Figure 6.1c).

Figure 6.1: Retention of dextrose (75 g / 300 ml) in (a) total, (b) proximal and (c) distal stomach, quantified by scintigraphy and 3D ultrasonography. Data are mean values ± SEM; *** P < 0.0001, ** P < 0.01, * P < 0.05.

Figure 6.2: Relationship between scintigraphic (SCT50) and 3D ultrasonographic (UST50) 50 % emptying times for the drink (75 g dextrose in 300 mL water).

Chapter 6

107

Figure 6.3: Limits of agreement for scintigraphic (SCT50) and 3D ultrasonographic (UST50) 50 % emptying times (T50s) for the drink (75 g dextrose in 300 mL water).

There was a significant relationship between proximal area (measured by 2D

ultrasonography) and proximal volume (measured by 3D ultrasonography) (r = 0.63,

P < 0.0001) (Figure 6.4).

Figure 6.4: Relationship between proximal stomach sagittal area (quantified by 2D ultrasonography) and proximal volume (quantified by 3D ultrasonography) in all patients across all time points.

Chapter 6

108

6.4.2 Blood glucose concentration

There was an increase (P < 0.0001) in blood glucose after the drink from t = 15 min

(P = 0.045). At t = 180 min (15.7 0.9 mmol/L), the blood glucose was still greater

(P = 0.0004) than at baseline (t = -2 min; 10.0 0.8 mmol/L) (Figure 6.5).

Figure 6.5: Blood glucose concentrations following ingestion of the drink (75 g dextrose in 300 mL water). Data are mean values ± SEM.

6.4.3 Relationships between blood glucose and gastric

emptying

With both techniques there was a significant inverse relationship between the blood

glucose concentration and total gastric retention, i.e. the blood glucose was greater

when gastric emptying was relatively more rapid (e.g. at t = 60 min; scintigraphy: r =

-0.65, P = 0.04; 3D ultrasonography: r = -0.78, P = 0.008).

Chapter 6

109

6.5 Discussion

This study establishes that 3D ultrasonography is a valid, and non-invasive, measure

of gastric emptying of a high-nutrient drink in patients with diabetic gastroparesis, as

demonstrated by the close agreement between ultrasound with scintigraphy. The

validity of both techniques is supported by the observed correlation between the

glycaemic responses to the drink with the rate of gastric emptying.

The observed concordance between 3D ultrasonography with scintigraphy is

consistent with our previous study in healthy volunteers (Gentilcore et al. 2006a),

where gastric emptying of both low- and high-nutrient drinks were measured by both

3D ultrasonography and scintigraphy in healthy young subjects and a good correlaton

and agreement between both techniques was evident (Gentilcore et al. 2006a). All our

patients had diabetic gastroparesis, as determined previously by scintigraphy (Stevens

et al. 2008), and, as would be expected, there was a poor correlation between gastric

emptying of solids and liquids in this group (Horowitz and Dent 1991, Horowitz et al.

2002b). – Of the 10 patients, five had delayed solid emptying, four had delayed liquid

emptying, and one had delayed emptying of both solid and liquid meal components

(Stevens et al. 2008). In all patients the diabetes was longstanding and seven of the

ten had evidence of autonomic neuropathy, which is not surprising given that the

incidence of autonomic neuropathy increases with the duration of diabetes (Jones et

al. 2002). While there is a high prevalence of upper gastrointestinal symptoms in

patients with type 1 and type 2 diabetes with gastroparesis, it is well recognised that

some may be asymptomatic (Bytzer et al. 2001, Horowitz and Dent 1991, Schvarcz et

al. 1996).

Chapter 6

110

The high-nutrient drink used in this study was selected to effectively stimulate small

intestinal feedback inhibition on gastric emptying – such liquids are known to empty

from the stomach at comparable rates to digestible solids, with the exception that the

lag phase is substantially shorter (Akkermans et al. 1984). For this reason, although

evaluation of gastric emptying of solids is likely to be more sensitive than that of low-

nutrient liquids in the diagnosis of gastroparesis (Wright et al. 1985), the sensitivity

of high-nutrient liquids is probably comparable to that of solids (Horowitz and Dent

1991, Jones et al. 1995b). The dextrose load in the drink (75 g) is that which is

traditionally employed in the oral glucose tolerance test, as used in the clinical

setting. The glycaemic response to this has been demonstrated to correlate with

gastric emptying in both healthy subjects and type 2 diabetes (Jones et al. 1996), so

that gastric emptying accounts for 30 - 40 % of the variance in peak plasma glucose

(Horowitz et al. 1993b, Jones et al. 1996). Hence, the observed correlations between

the glycaemic response and gastric emptying as assessed by both 3D ultrasonography

and scintigraphy were anticipated and serve to further validate the ultrasonographic

technique.

Intragastric distribution is frequently abnormal in diabetic gastroparesis (Jones et al.

1995b) as well as other conditions associated with disordered gastric motility,

including functional dyspepsia (Hausken et al. 1993) and gastro-oesosophageal reflux

disease (Tefera et al. 2002), and may potentially modulate gastrointestinal symptoms.

3D ultrasonography, unlike its 2D counterpart, enables measurement of intragastric

meal distribution (Gilja et al. 1997) as is the case with scintigraphy (Jones et al.

1995b). Our observation of a significant difference between scintigraphic and 3D

ultrasonographic measurements of proximal and distal stomach emptying is

congruous with the discrepant methods of dividing the stomach into proximal and

Chapter 6

111

distal regions. Given that the ‘distal stomach’ is generally larger, and the ‘proximal

stomach’ smaller, with scintigraphy, our observations are expected. It may, therefore,

be argued, that scintigraphy will be relatively more sensitive in detecting changes in

the distal stomach emptying. Not surprisingly, there was a significant relationship

between proximal area, as measured by 2D ultrasonography, and the corresponding

proximal volume quantified by 3D ultrasonography. While previous reports (Gilja et

al. 1997) have demonstrated a significant relationship between 2D antral area and the

corresponding 3D distal volume, this represents the first demonstration of a

correlation pertaining to the proximal stomach.

In conclusion, this study establishes that 3D ultrasonography provides a valid, and

non-invasive, measure of gastric emptying of a high-nutrient liquid in patients with

diabetic gastroparesis.

Chapter 7

112

7Chapter 7INSULIN – INDUCED HYPOGLYCAEMIA ACCELERATES

GASTRIC EMPTYING OF SOLIDS AND LIQUIDS IN

LONGSTANDING TYPE 1 DIABETES

7.1 Summary

The rate of gastric emptying of carbohydrate is a major determinant of postprandial

glycaemia. In healthy subjects and patients with uncomplicated type 1 diabetes, there

is evidence that gastric emptying may be accelerated by insulin-induced

hypoglycaemia. The objective of this study was to determine the effects of acute

hypoglycaemia on gastric emptying in longstanding type 1 diabetes, and to evaluate

whether the response to hypoglycaemia is influenced by the rate of gastric emptying

during euglycaemia and/or autonomic dysfunction. Gastric emptying of a solid/liquid

Chapter 7

113

meal (100 g 99mTc–minced beef and 150 mL 67Ga-EDTA labelled water) was

measured by scintigraphy on two separate days, during hypoglycaemia and

euglycaemia. The studies took place at the Department of Nuclear Medicine, PET and

Bone Densitometry at the Royal Adelaide Hospital. Twenty type 1 patients (4 female,

16 male, age 45.9 ± 2.3 years and duration of known diabetes 18.0 ± 2.7 years) were

recruited from outpatient clinics and the Diabetes Centre at the Royal Adelaide

Hospital. Hypoglycaemia ( 2.6 mmol/L) was established 15 min prior to, and

maintained for 45 min after, meal consumption. On one of the days, autonomic nerve

function was evaluated using cardiovascular reflex tests. Twelve of the 20 subjects

had autonomic neuropathy. Gastric emptying of both solid (P < 0.001) and liquid (P <

0.05) was faster during hypoglycaemia. The magnitude of this acceleration was

greater when the rate of gastric emptying during euglycaemia was slower (solid: %

retention at 100 min, r = -0.52, P < 0.05 and liquid: 50 % emptying time, r = -0.82, P

< 0.0001, but not influenced by autonomic nerve function. Insulin-induced

hypoglycaemia accelerates gastric emptying of solids and liquids in longstanding type

1 diabetes even in those patients with delayed emptying, and is likely to be an

important mechanism in the counter-regulation of hypoglycaemia.

7.2 Introduction

It is now recognised that the rate of gastric emptying is a major determinant of the

glycaemic response to carbohydrate-containing meals in both type 1 and type 2

diabetes (Ishii et al. 1994, Jones et al. 1996, Rayner et al. 2001) and that gastric

emptying is abnormally slow in 30 – 50 % of patients with longstanding diabetes

(Horowitz et al. 1989b, Horowitz et al. 1991, Horowitz et al. 1996b, Keshavarzian et

al. 1987, Merio et al. 1997, Samsom et al. 2003, Ziegler et al. 1996). The risk of

gastroparesis is known to be greater in those patients with autonomic neuropathy

Chapter 7

114

(usually assessed by cardiovascular reflex tests), but the relationship between the

delay in gastric emptying and the presence of autonomic neuropathy is not strong

(Horowitz et al. 1991, Merio et al. 1997, Migdalis et al. 2001, Stacher et al. 2003,

Ziegler et al. 1996). Disordered gastric emptying in diabetes may also result from

acute changes in the blood glucose concentration (Cucchiara et al. 1998, Fraser et al.

1990, MacGregor et al. 1976, Rayner et al. 2001, Samsom et al. 1997, Schvarcz et al.

1993, Schvarcz et al. 1995a, Schvarcz et al. 1997).

While it is well established that acute hyperglycaemia slows gastric emptying in both

normal subjects (MacGregor et al. 1976, Rayner et al. 2001) and type 1 patients

(Cucchiara et al. 1998, Fraser et al. 1990, Samsom et al. 1997), there is relatively

little information about the effects of hypoglycaemia on gastric emptying (Schvarcz

et al. 1993, Schvarcz et al. 1995a) and only one study has hitherto evaluated patients

with diabetes (Schvarcz et al. 1993). The initial report by Schvarcz et al. (1993)

involved 8 young adult patients with uncomplicated type 1 diabetes of short duration;

during acute hypoglycaemia gastric emptying was apparently much faster (Schvarcz

et al. 1993). The observations were subsequently confirmed in a study of 8 healthy

young adults (Schvarcz et al. 1995a). A substantial methodological limitation of both

studies (Schvarcz et al. 1993, Schvarcz et al. 1995a) is that they were not randomised

– gastric emptying was always measured initially during euglycaemia and,

subsequently, during hypoglycaemia. Furthermore, although a dual isotope technique

was used to measure gastric emptying of solids and liquids concurrently, the labelling

was demonstrably imprecise, because solid and liquid meals were reported to empty

from the stomach at about the same rate during euglycaemia, whereas solids are

known to empty from the stomach much more slowly than low-nutrient liquids

(Collins et al. 1983). Hence, the conclusions derived from these studies (Schvarcz et

Chapter 7

115

al. 1993, Schvarcz et al. 1995a) may not be valid. There is no information about the

effects of acute hypoglycaemia on gastric emptying in patients with longstanding

type 1 diabetes, nor is it known whether the response to hypoglycaemia is influenced

by the rate of gastric emptying during euglycaemia, or the presence of gastroparesis.

Schvarcz et al. (1995) reported in healthy subjects that the acceleration of gastric

emptying induced by hypoglycaemia is blocked by concurrent administration of

atropine, indicating that cholinergic stimulation is important in mediating the effect

(Schvarcz et al. 1995b). It is, however, not known whether the gastric emptying

response to hypoglycaemia is modified by the presence of autonomic neuropathy.

The aims of this study were to determine the effects of acute hypoglycaemia on

gastric emptying, and evaluate whether the response to hypoglycaemia is influenced

by the rate of gastric emptying during euglycaemia or autonomic dysfunction, in

longstanding type 1 patients.

7.3 Materials and Methods

7.3.1 Subjects

Twenty type 1 patients; 4 female, 16 male, aged 45.9 ± 2.3 years, body mass index

26.3 ± 0.7 kg/m2 and duration of known diabetes 18.0 ± 2.7 years, were recruited

from outpatient clinics and the Diabetes Centre at the Royal Adelaide Hospital.

Glycated haemoglobin was 8.1 ± 0.3 % (normal < 6 %). Severe cardiac or respiratory

disease, previous gastrointestinal surgery (apart from uncomplicated appendectomy)

and the use of medication known to affect gastrointestinal motility, represented

exclusion criteria. The plasma creatinine was also required to be within the normal

Chapter 7

116

range (≤ 0.12 mmol/L). Patients were not selected on the basis of gastrointestinal

symptoms, gastric emptying status, or autonomic nerve function.

A physical examination was performed to assess diabetic microvascular

complications. Retinopathy was graded on a recent ophthalmological assessment.

Peripheral neuropathy was diagnosed when absent ankle reflexes were associated

with either motor or sensory changes (Horowitz et al. 1991). Written, informed

consent was obtained from each patient prior to enrolment in the study in accordance

with the Declaration of Helsinki and the protocol was approved by the Human Ethics

Committee of the Royal Adelaide Hospital.

7.3.2 Experimental protocol

Each subject attended the Department of Nuclear Medicine, PET and Bone

Densitometry at about 0900 h after an overnight fast (12 h for solids, 10 h for liquids)

on two separate occasions for measurement of gastric emptying. Smoking was

prohibited for 24 h prior to each gastric emptying measurement. On one day, the

blood glucose concentration was maintained in the euglycaemic range (~ 6 mmol/L)

for the duration of the gastric emptying measurement, while on the other day,

hypoglycaemia (~ 2.6 mmol/L) was induced and maintained for 60 min, followed by

euglycaemia. The studies were performed in a single-blind, randomised, fashion and

the two study days were separated by a minimum of 4 days. Two intravenous

cannulae were inserted, one in an antecubital vein of the right arm for infusion of

glucose and insulin, and the other retrogradely on the dorsum of the left hand for

blood sampling. The left hand was heated with an electric pad to “arterialise” the

venous blood. A blood pressure cuff (DINAMAP; Johnson & Johnson, Tampa, FL,

Chapter 7

117

USA) was placed around the left arm for measurements of systolic blood pressure

(SBP), diastolic blood pressure (DBP) and heart rate (HR).

The blood glucose concentration was stabilised at the desired level using a glucose-

insulin clamp (Jones et al. 1998). The rate of insulin infusion (Actrapid; Novo

Nordisk Pharmaceuticals, Auckland, New Zealand) based on a standard rate of 80

mU/m2.min, was initially variable (0 – 10 min) according to the subject’s body

surface area, and then constant for the remainder of the study (Jones et al. 1998,

Simonson et al. 1985), whereas the glucose (25 %w/v) infusion was varied to

maintain the blood glucose at the desired level (Jones et al. 1998). When venous

blood glucose was stabilised at either 6 mmol/L (euglycaemia) or 2.6 mmol/L

(hypoglycaemia) for 15 min, subjects ingested the meal. On the hypoglycaemic day,

the blood glucose concentration was maintained at approximately 2.6 mmol/L for 60

min (i.e. t = -15 – 45 min), returned to euglycaemic levels (6 mmol/L) between t = 45

- 75 min, and maintained at that level for the remainder of the study (i.e. t = 75 – 120

min). On the euglycaemic day, the blood glucose was maintained at approximately

6.0 mmol/L from t = -15 – 120 min. Blood glucose concentrations were measured

every 5 min throughout each study using a portable glucose meter (Medisense

Precision QID; Abbott Laboratories, Bedford, MA, USA). Plasma glucose was

measured using the hexokinase method on the venous blood samples (~20 mL)

obtained at t = -15, 0, 15, 30, 45, 60, 75, 90, 105, and 120 min.

Chapter 7

118

7.3.3 Measurement of gastric emptying

Gastric emptying was measured using a dual isotope scintigraphic technique (Collins

et al. 1983, Jones et al. 1995b). The test meal comprised 100 g lean minced beef

labelled with 20 MBq 99mTc-sulphur colloid chicken liver, and 150 mL water labelled

with 7 MBq 67Ga-EDTA (ethylenediaminetetraacetic acid) (Jones et al. 1996). The

solid meal was eaten within 5 minutes, followed by the water, which was consumed

within 1 minute. Radioisotopic data were acquired with the subject seated with their

back against a gamma camera (Siemens, Chicago, IL, USA) at 1-minute frames for

the first hour and 3-minute frames thereafter. Time zero (t = 0) was defined as the

time of meal completion and gastric emptying was monitored for 120 minutes. Data

were corrected for radionuclide decay, gamma-ray attenuation and subject movement,

using previously described methods (Collins et al. 1983, Jones et al. 1997). Regions-

of-interest were drawn for the total stomach, which was subsequently divided into

proximal and distal stomach regions. Gastric emptying curves for total, proximal and

distal stomach regions, expressed as percent retention over time, were then derived

(Jones et al. 1997). For the solid component of the meal, the lag phase (Tlag, defined

as the time at which activity was first seen in the proximal small intestine) and the

percentage of the solid meal remaining in the stomach at 100 min (T100), were

determined; for the liquid component, the time for 50 % emptying (T50) was derived

(Horowitz et al. 1987, Jones et al. 1995b). Delayed gastric emptying was defined as a

T100 of > 66 %, and a T50 > 35 min, based on an established normal range (Collins

et al. 1983).

Chapter 7

119

7.3.4 A s s e s s m e n t o f upper gastrointestinal and

hypoglycaemic symptoms

Prior to the commencement of the first glucose clamp, the following upper

gastrointestinal symptoms were assessed: lack of appetite, nausea, early satiation,

vomiting, upper abdominal discomfort or distension, abdominal pain (“gastric

symptoms”), dysphagia, heartburn and acid regurgitation (“oesophageal symptoms”)

using a validated questionnaire (Horowitz et al. 1991, Jones et al. 1996). The severity

of each symptom was graded as 0 = none, 1 = mild, 2 = moderate, 3 = severe for a

maximum total score of 27.

Hypoglycaemic symptoms were evaluated each day at t = -30, -15, -10, -5, 0, 15, 30,

45, 60, 75, 90 and 120 min, and subjects were asked to score the following

symptoms: pounding heart, shakiness, sweating, headache, difficulty thinking and

slowed thinking, on a scale of 1 7, where 1 indicated that the subject did not have

the symptom, and 7 indicated that the symptom was experienced in the extreme; the

maximum possible score was 42 (Jones et al. 1998).

7.3.5 Measurement of blood pressure and heart rate

Systolic, diastolic blood pressure and heart rate were measured at t = -30, -15, 0, 15,

30, 45, 60, 75, 90, 105 and 120 min using an automated device (DINAMAP; Johnson

& Johnson, Tampa, FL, USA).

7.3.6 Assessment of autonomic nerve function

Autonomic nerve function was measured at the end of the second visit, approximately

two hours after the completion of the gastric emptying measurement, using

standardised cardiovascular reflex tests (Ewing and Clarke 1982, Ewing et al. 1985,

Chapter 7

120

Horowitz et al. 1991, Jones et al. 1995b). Parasympathetic function was evaluated by

the variation (R-R interval) of the heart rate during deep breathing and in response to

standing ("30:15" ratio). Sympathetic function was assessed by the fall in systolic

blood pressure in response to standing. The result of each test was scored as 0 =

normal, 1 = borderline and 2 = abnormal. A score > 3 was considered to indicate

definite autonomic dysfunction (Ewing and Clarke 1982, Ewing et al. 1985, Horowitz

et al. 1991, Jones et al. 1995b).

7.3.7 Statistical analysis

Individual comparisons between the two “treatment” groups (hypoglycaemia vs

euglycaemia) were performed using Student’s t-tests. Data were analysed using

repeated measures analysis of variance (ANOVA) with “treatment” and “time” as

variables. In the case of a “treatment by time” interaction, contrasts were used to

compare individual time points between the two treatment groups to examine pre-

planned hypotheses. Relationships between gastric emptying and other parameters

were assessed by linear regression analysis. Data are shown as mean ± SEM unless

stated otherwise. A P-value of < 0.05 was considered significant in all analyses.

7.4 Results

No serious untoward effects occurred on either of the test days. Although some

subjects experienced mild neuroglycopaenic symptoms, all were able to cooperate

and eat the test meal. Twelve of the 20 patients had definite evidence of autonomic

neuropathy, with a mean score of 3.0 0.3. Five subjects had proliferative

retinopathy and five (25 %) had evidence of peripheral neuropathy. Blood glucose

concentrations closely approximated the desired levels; during hypoglycaemia, the

Chapter 7

121

blood glucose between t= -15 and 45 min was 2.6 0.07 mmol/L (Figure 7.1). There

was no difference in the blood glucose concentration between the 2 days after 75 min.

Figure 7.1: Blood glucose concentrations in studies conducted during hypoglycaemia and euglycaemia. Gastric emptying was measured between t = 0 and 120 min. Data are mean values SEM; * P < 0.05 and # P < 0.001 compared with euglycaemia.

7.4.1 Gastric emptying

Five of the 20 subjects had delayed solid, and 7 delayed liquid, emptying during

euglycaemia. In two subjects, gastric emptying of solid was markedly delayed (solid

T100 ≥ 3 standard deviations outside the normal range). There was no significant

difference in gastric emptying of solids (solid T100) or liquids (liquid T50) during

euglycaemia in patients with autonomic neuropathy when compared to the remainder

of the group (data not shown).

There was a significant “treatment by time” interaction (P < 0.0001) for solid gastric

emptying; gastric emptying was faster (P < 0.001) during hypoglycaemia from 60

Chapter 7

122

min (Figure 7.2a). The lag phase (hypoglycaemia vs. euglycaemia) was 16.8 4.3

min vs. 22.7 6.0 min, (P = 0.20) and the solid T100 was 35.7 5.4 % vs. 48.6 5.2

%, (P = 0.04). There was also a “treatment by time” interaction for liquid emptying (P

< 0.05); gastric emptying was faster (P < 0.05) during hypoglycaemia when

compared to euglycaemia from 15 min (Figure 7.2b). The liquid T50 also tended to

be less during hypoglycaemia (26.4 4.8 min vs. 41.2 7.6 min, P = 0.09). Of the 20

patients, three had delayed emptying for solid, and two for liquid, on the

hypoglycaemic day.

There was no significant difference between euglycaemia and hypoglycaemia in the

retention of solid in the proximal stomach (Figure 7.2c). However, there was a

“treatment by time” interaction for the retention of liquid in the proximal stomach (P

< 0.01), so that retention was less (P < 0.05) on the hypoglycaemic day (Figure 7.2d).

There was a significant “treatment by time” interaction (P < 0.01) for gastric

emptying of solid from the distal stomach, so that retention in the distal stomach was

less (P < 0.05) during hypoglycaemia when compared to euglycaemia (Figure 7.2e).

There was no significant difference in distal stomach retention of liquid between the

two days (Figure 7.2f).

Chapter 7

123

Figure 7.2: Gastric emptying and intragastric distribution of solid and liquid meal components during hypoglycaemia and euglycaemia. Data are mean mean values SEM; * P < 0.05 and # P < 0.01 compared with euglycaemia.

Chapter 7

124

When the cohort was divided according to whether their gastric emptying was normal

or delayed during euglycaemia (i.e. solid and/or liquid emptying), the magnitude of

the acceleration of gastric emptying of liquids during hypoglycaemia was greater

(change in T50: 38.9 21.2 min vs. 1.8 2.6 min, P < 0.05) in those with delayed

emptying of liquid; for solid emptying, this was not quite significant (change in T100:

30.0 19.8 % vs. 7.3 3.6 %, P = 0.09). In the two subjects with markedly delayed

gastric emptying of solids, gastric emptying was faster during hypoglycaemia. When

the cohort was divided into those with and without cardiovascular autonomic

neuropathy, there was no significant difference in the magnitude of the change in

gastric emptying of either solid (change in T100: 8.9 4.1 % vs. 19.0 13.3 %, P =

0.40) or liquid (change in T50: 4.6 9.5 min vs. 30.0 14.1 min, P = 0.14), nor were

there any relationships between either the total score for autonomic neuropathy or the

“30:15” ratio for the heart rate response to standing between the change in gastric

emptying of solids or liquids during euglycaemia and hypoglycaemia.

There was a significant relationship between the magnitude of the acceleration of

gastric emptying in response to hypoglycaemia and the rate of gastric emptying

during euglycaemia for both the solid T100 (r = 0.52, P < 0.05) and liquid T50 (r =

0.82, P < 0.0001) (Figure 7.3), i.e. hypoglycaemia had a greater effect when gastric

emptying was relatively more slow.

Chapter 7

125

Figure 7.3: The relationship between the magnitude of the change in gastric emptying for the solid retention at 100 min (T100) and liquid 50 % emptying time (T50) between hypoglycaemia and euglycaemia and the rate of gastric emptying during euglycaemia. Individual data for the 20 subjects are shown.

7.4.2 Upper gastrointestinal and hypoglycaemic symptoms

Upper gastrointestinal symptoms were present in 13 of the 20 subjects; the total score

was 3.0 0.83 (“gastric” symptoms: 2.2 0.7; “oesophageal” symptoms: 0.9 0.3).

There was no significant relationship between the total symptom score and either the

solid T100 (r = -0.17, P = 0.47) or the liquid T50 (r = -0.31, P = 0.18) during

euglycaemia.

There was a significant “treatment by time” interaction for “pounding heart” (P <

0.0001), “shakiness” (P < 0.0001) and “sweating” (P < 0.0001). The symptom of

“pounding heart” was greater on the hypoglycaemic day between t = –30 and 15 min

(P < 0.01); “shakiness” was greater during hypoglycaemia from t = –30 to 45 min (P

< 0.01); “sweating” was greater during hypoglycaemia between t = –15 and 15 min

(P < 0.01), when compared to euglycaemia (data not shown). There was a “treatment

by time” interaction for total symptoms (P < 0.05); symptoms of hypoglycaemia were

Chapter 7

126

greater from t= –15 to 15 min (P < 0.01) when compared to the euglycaemic day

(Figure 7.4). When the cohort was divided into those with and without cardiovascular

autonomic neuropathy, there was no significant difference in the total score for

hypoglycaemic symptoms during the hypoglycaemic study day (data not shown).

Figure 7.4: Symptoms of hypoglycaemia during hypoglycaemia and euglycaemia. Data are mean values SEM. #P < 0.01 compared with euglycaemia.

7.4.3 Blood pressure and heart rate

There was no difference in baseline systolic or diastolic blood pressure, or heart rate

between the two days: (SBP: hypoglycaemia 122.1 3.4 mmHg vs. euglycaemia

121.1 3.6 mmHg, DBP: 68.2 2.2 mmHg vs. 69.8 2.2 mmHg and HR: 77.4 3.1

beats/min vs. 73.6 2.5 beats/min). There was also no significant difference in

systolic or diastolic blood pressure during the gastric emptying measurements

between the two treatments. There was, however, a “treatment by time” interaction

Chapter 7

127

for heart rate (P < 0.05), so that heart rate was greater (P < 0.05) from t= –15 to 15

min during hypoglycaemia (data not shown).

7.5 Discussion

The present study evaluated the acute effects of insulin-induced hypoglycaemia on

gastric emptying in a heterogeneous group of patients with longstanding type 1

diabetes. Only one study has hitherto evaluated the effects of hypoglycaemia on

gastric emptying in type 1 diabetes (Schvarcz et al. 1993); this study only included 8

subjects, all of whom had uncomplicated diabetes, and interpretation is hampered by

methodological limitations. The present study was randomised, and gastric emptying

of both solid and liquid meal components was evaluated with a precise scintigraphic

technique (Collins et al. 1983, Jones et al. 1995b). This study demonstrated that in

longstanding type 1 diabetes: (i) hypoglycaemia accelerates gastric emptying of both

solid and liquid meal components and (ii) the magnitude of this acceleration is greater

in those patients who have slower gastric emptying during euglycaemia.

It is now well established that acute elevations in blood glucose concentrations have a

major, reversible, effect on gastric emptying (Fraser et al. 1990, MacGregor et al.

1976, Rayner et al. 2001, Samsom et al. 1997), as well as motility in other regions of

the gastrointestinal tract (de Boer et al. 1992, Russo et al. 1996, Zhang et al. 2004), in

both healthy subjects and diabetic patients. Even elevations in the blood glucose

concentration that are within the normal postprandial range influence gastric

emptying; in both healthy subjects and uncomplicated type 1 patients; emptying of

solids and liquids is slower at a blood glucose concentration of 8 vs. 4 mmol/L

(Schvarcz et al. 1997). Aylett (1962) was the first to report that gastric emptying of

water was accelerated by insulin-induced hypoglycaemia, in a group of patients with

Chapter 7

128

duodenal ulceration (Aylett 1962). Schvarcz et al. subsequently performed studies

relating to the effect of acute hypoglycaemia (blood glucose ~ 1.9 mmol/L) on gastric

emptying in uncomplicated type 1 patients (Schvarcz et al. 1993) and healthy

subjects (Schvarcz et al. 1995a). Although the radioisotopic labelling was imprecise,

and the studies were not randomised, the apparent acceleration of gastric emptying

was substantial, with a reduction in the 50 % emptying times of solids and liquids of

~ 50 % (Schvarcz et al. 1993, Schvarcz et al. 1995a). These studies, like the present

study, were not designed to discriminate between the potential effects of

hypoglycaemia and hyperinsulinaemia on gastric emptying. However, subsequent

studies indicate that hyperinsulinaemia per se has no effect on either gastric motility

or gastric emptying. In particular, euglycaemic hyperinsulinaemia has no effect on

postprandial antral motility in healthy subjects (Hasler et al. 1995) or on gastric

emptying of solids or liquids in patients with uncomplicated type 1 and type 2

diabetes (Kong et al. 1999b). Hence, the effects of hyperglycaemia on gastric

emptying are most unlikely to be attributable to hyperinsulinaemia.

We used a double-isotope radionuclide technique, which has been used extensively

by our group to evaluate gastric emptying in diabetes (Horowitz et al. 1991, Jones et

al. 1995b): both solid and liquid meal components were labelled precisely (Collins et

al. 1983). The cohort studied had longstanding type 1 diabetes (mean duration 18 yr).

Five had delayed solid emptying, seven of 20 had delayed liquid emptying during

euglycaemia, and 60 % had autonomic neuropathy; these prevalences are comparable

with those reported previously (Horowitz et al. 1991, Keshavarzian et al. 1987,

Samsom et al. 2003, Ziegler et al. 1996). The magnitude of the observed acceleration

of gastric emptying of solids and liquids during hypoglycaemia was substantial, and

consistent with the reports by Schvarcz et al. (Schvarcz et al. 1993, Schvarcz et al.

Chapter 7

129

1995a). Because it is now established that even minor changes in the rate of small

intestinal delivery of carbohydrate may have a major effect on glycaemia (O'Donovan

et al. 2004b, Rayner et al. 2001), the observed acceleration of gastric emptying is

likely to represent an important counter-regulatory mechanism to hypoglycaemia. It is

of interest that the acceleration of gastric emptying of liquids was evident very soon

after meal ingestion, whereas the acceleration of solid emptying was not apparent

until approximately 60 min, and hypoglycaemia had no effect on the lag phase for

solid emptying. During the lag phase, solids are ground into small particles by the

antrum before emptying commences (Collins et al. 1983, Horowitz and Dent 1991).

Hence, it appears that hypoglycaemia does not affect this component of gastric

mechanics and that solids must be “liquefied” for their emptying to be accelerated.

The acceleration of gastric emptying by hypoglycaemia was associated with changes

in intragastric meal distribution, as would be expected: during hypoglycaemia, the

retention of liquid in the proximal stomach, and that of solid in the distal stomach,

were decreased.

Issues of clinical relevance that could be addressed by our study because of

heterogeneity of the cohort were whether the effect of hypoglycaemia on gastric

emptying was influenced by the rate of emptying during euglycaemia and/or

autonomic nerve function. The magnitude of the acceleration of gastric emptying was

shown to be inversely related to the rate of gastric emptying during euglycaemia for

both solid and liquid meal components; when gastric emptying was relatively slower,

the response was greater. Hence, there is no evidence to suggest that the response to

hypoglycaemia may be impaired by gastroparesis; rather, as has been shown to be the

case with many gastrokinetic drugs, the response is, in general, more marked when

gastric emptying is delayed (Collins et al. 1983, Horowitz and Dent 1991). It should,

Chapter 7

130

however, be recognised that solid gastric emptying was markedly delayed in only two

subjects. Although acceleration of gastric emptying by hypoglycaemia was observed

in both of these subjects, this group may potentially respond differently. Schvarcz et

al. (1995) reported that the acceleration of gastric emptying induced by

hypoglycaemia in healthy subjects was blocked by atropine (Schvarcz et al. 1995b).

Hence, it would not have been surprising if the response to hypoglycaemia proved to

be dependent on autonomic (particularly parasympathetic) function. There was,

however, no clear evidence to support this concept, although there was perhaps a

trend (P = 0.14) for the acceleration of liquid emptying to be less in those patients

with autonomic neuropathy, and the possibility of a type 2 statistical error must be

acknowledged. Autonomic nerve function was evaluated, as in previous studies using

standardised cardiovascular reflex tests (Ewing et al. 1985, Jones et al. 1996, Merio

et al. 1997, Stacher et al. 2003, Ziegler et al. 1996), which are probably a reasonable

surrogate for (Buysschaert et al. 1987), but certainly not a direct measure of,

gastrointestinal autonomic function. Hence, the obervations relating to the effect of

autonomic function should be viewed circumspectly. It remains possible that drugs

with anticholinergic activity could influence the gastric emptying response to

hypoglycaemia, as suggested by Schvarcz et al. (1995).

This study does not provide additional insights into either the mechanism(s) by which

hypoglycaemia accelerates gastric emptying, or the gastroduodenal motor correlates

of this effect. In relation to the former, stimulation of both sympathetic and

parasympathetic activity (Berne and Fagius 1986, Schvarcz et al. 1995b) may be

important. The rate of gastric emptying is dependent on the integration of motor

activity in the proximal stomach, antrum, pylorus and proximal small intestine

(Horowitz et al. 1994), and the motor dysfunctions in type 1 patients with

Chapter 7

131

gastroparesis are heterogeneous (Fraser et al. 1994, Samsom et al. 1997). Studies are

indicated to evaluate the effects of insulin–induced hypoglycaemia on postprandial

gastric motility in healthy subjects and type 1 patients. The effects of hypoglycaemia

on fasting antropyloric motility in normals are apparently unremarkable (Fellows et

al. 1987, Fraser et al. 1991a). The present study has also not determined whether

there is a “threshold” at which hypoglycaemia accelerates emptying, or whether the

response is continuous (there appears to be a direct relationship between changes in

gastric emptying/gastric motility and the magnitude of acute elevations in blood

glucose concentrations (Groop et al. 1989, Hasler et al. 1995)); the degree of

hypoglycaemia induced in this study (~ 2.6 mmol/L) was less than that used by

Schvarcz et al. (Schvarcz et al. 1993, Schvarcz et al. 1995a) (~ 1.9 mmol/L).

Chapter 8

132

8Chapter 8EFFECTS OF INTRAVENOUS FRUCTOSE ON GASTRIC

EMPTYING AND ANTROPYLORODUODENAL MOTILITY IN

HEALTHY SUBJECTS

8.1 Summary

Gastric emptying of glucose is regulated closely, not only as a result of inhibitory

feedback arising from the small intestine, but also because of the resulting

hyperglycaemia. Fructose is used widely in the diabetic diet and is known to empty

from the stomach slightly faster than glucose, but substantially slower than water.

The aims of this study were to determine whether intravenous fructose affects gastric

emptying and antropyloroduodenal motility and, how any effects compare to those

induced by intravenous glucose. Six healthy males (mean age 26.7 ± 3.8 yr)

Chapter 8

133

underwent concurrent measurements of gastric emptying of a solid meal (100 g

minced beef labelled with 20 MBq 99mTc-sulphur colloid) and antropyloroduodenal

motility on three separate, randomised days during intravenous infusion of either

fructose (0.5 g/kg), glucose (0.5 g/kg), or isotonic saline for 20 min. Gastric emptying

(scintigraphy), antropyloroduodenal motility (manometry) and blood glucose

(glucometer) were measured for 120 min. There was a rise in blood glucose (P <

0.001) after glucose (peak 16.4 ± 0.6 mmol/L), but not after fructose or saline.

Intravenous glucose and fructose both slowed gastric emptying substantially (P <

0.005 for both), without any significant difference between them. Between t = 0 – 30

min; the number of antral pressure waves was less after both glucose and fructose (P

< 0.002 for both) than saline, and there were more isolated pyloric pressure waves

during glucose (P = 0.003) compared with fructose and saline (P = NS for both)

infusions. It is concluded that intravenous fructose slows gastric emptying and

modulates gastric motility in healthy subjects, and that the magnitude of slowing of

gastric emptying is comparable to that induced by intravenous glucose.

8.2 Introduction

It is generally assumed that the delivery of nutrients to the small intestine is regulated

tightly, primarily as a result of feedback inhibition generated by small intestinal

luminal receptors, the magnitude of which is dependent on the length and region of

small intestine exposed to nutrient (Lin et al. 1989, Lin et al. 1990), so that the

overall rate of entry of nutrients into the small intestine approximates 2 kcal/min in

healthy subjects (Brener et al. 1983, Hunt et al. 1985, Lin et al. 1989, Moran and

McHugh 1981). The motor correlates of the slowing of gastric emptying induced by

the presence of nutrients in the small intestine include relaxation of the proximal

stomach (Azpiroz and Malagelada 1985b), suppression of antral motility (Heddle et

Chapter 8

134

al. 1988a) and stimulation of phasic and tonic pyloric contractions (Heddle et al.

1988b). Monosaccharides empty from the stomach more slowly than water or

isotonic saline because of small intestinal feedback (Guss et al. 1994), but there may

be subtle differences between them. In particular, fructose empties more rapidly than

glucose when given as intragastric loads to monkeys (Moran and McHugh 1981). The

slightly more rapid rate of emptying of oral fructose compared to oral glucose, has

also been established in humans (Elias et al. 1968, Guss et al. 1994, Horowitz et al.

1996a, Sole and Noakes 1989).

In addition to intraluminal mechanisms, there is evidence that plasma

monosaccharide concentrations may also affect gastric motility/emptying (Hebbard et

al. 1996a, MacGregor et al. 1976, Samsom et al. 1997, Schvarcz et al. 1997). This

could potentially contribute to the discrepant effects of monosaccharides on gastric

emptying. In particular, it is well established that acute hyperglycaemia, induced by

intravenous glucose has major, reversible, effects on gastrointestinal motor function.

Marked hyperglycaemia (~ 16 – 20 mmol/L) slows gastric emptying in healthy

subjects (MacGregor et al. 1976, Schvarcz et al. 1997) and patients with type 1

(Fraser et al. 1990, Samsom et al. 1997) and type 2 (Horowitz et al. 1989b) diabetes,

when compared with euglycaemia (~ 4 mmol/L). Even blood glucose concentrations

that are within the normal postprandial range (i.e. 4 - 8 mmol/L), have been shown to

affect gastric emptying in both healthy volunteers and uncomplicated type 1 patients

(Schvarcz et al . 1997). The slowing of gastric emptying induced by acute

hyperglycaemia is associated with suppression of antral pressure waves (Barnett and

Owyang 1988, Samsom et al. 1997) and their propagation (Samsom et al. 1997),

stimulation of basal and phasic pyloric pressure waves (Fraser et al. 1991b), an

increase in proximal gastric compliance (Hebbard et al. 1996a, Hebbard et al. 1996b),

Chapter 8

135

and suppression of duodenal motor activity (Russo et al. 1996). Acute

hyperglycaemia also affects the perception of sensations arising from the gut,

including fullness and nausea (Hebbard et al. 1996a, Hebbard et al. 1996b, Jones et

al. 1997, Lingenfelser et al. 1999). In view of the well documented effects of

hyperglycaemia, it is surprising that there is no information about the potential effects

of intravenous fructose on gastric emptying or gastric motility.

Fructose consumption has increased markedly in the last 20 - 30 years (Guss et al.

1994, Teff et al. 2004). Fructose is sweeter than isoenergetic glucose and, as such,

presents the advantage of offering the same level of sweetness for a lower energy

burden (Gerrits and Tsalikian 1993, Hallfrisch et al. 1983). For this reason, fructose

is used widely in the diabetic diet and has largely replaced sucrose in a number of

processed foods, particularly beverages (Gerrits and Tsalikian 1993, Vozzo et al.

2002). The glycaemic response to fructose is also substantially less than to glucose

(Bowen et al. 2007, Crapo et al. 1980, Horowitz et al. 1996a, Kong et al. 1999a,

Vozzo et al. 2002). There are also substantial differences in the effects of oral glucose

and fructose on the release of gastrointestinal hormones, including insulin (Vozzo et

al. 2002), and the ‘incretin’ hormones GLP-1 (Kong et al. 1999a, Toft-Nielsen et al.

2001, Vozzo et al. 2002) and GIP (Vozzo et al. 2002). Fructose-induced insulin

release is glucose-dependent in that insulin secretion following intravenous

(Dunnigan and Ford 1975), and oral (Reiser et al. 1987), fructose is greater during

hyper- than during eu-glycaemia.

The primary aim of this study was to determine whether intravenous fructose affects

gastric emptying and antropyloroduodenal motility, and if so, how these effects

compare to those induced by intravenous glucose.

Chapter 8

136

8.3 Materials and Methods

8.3.1 Subjects

Six healthy males (mean age: 26.7 ± 3.8 yr, body mass index: 26.4 ± 1.4 kg/m2) were

studied. Subjects were randomly selected from volunteers who responded to

advertisements posted on university notice boards. Subjects were asked to maintain a

normal diet for three days prior to each study and smokers were required to abstain

from tobacco for at least 12 h prior to each study day. No subject had a history of

diabetes mellitus, gastrointestinal disease or surgery, significant respiratory, cardiac

or hepatic disease, chronic alcohol abuse, gout or epilepsy. No subject was taking

medication known to influence gastrointestinal function.

8.3.2 Experimental protocol

Each subject underwent three randomised, single-bind studies, separated by at least

three days. During each study, subjects received an intravenous infusion of either

fructose or glucose (0.5 g/kg body weight dissolved in sterile water as 0.2 g/mL, for

both monosaccharides) or placebo (saline 0.9 %w/v), administered over 20 min i.e.

the total volume infused varied between subjects depending on body weight. The

dose of the monosaccharides was selected on the basis of previous studies (Aitken

and Dunnigan 1969, Dunnigan and Ford 1975, Elliott et al. 1967).

On each study day, subjects attended the Department of Nuclear Medicine, PET and

Bone Densitometry at approximately 08:00 h following an overnight fast (14 h for

solids and 12 h for liquids). A silicone-manometric assembly (~4 mm diameter)

(Dentsleeve, Adelaide, SA, Australia) was introduced into the stomach via an

anaesthetised nostril and allowed to pass through the stomach and into the duodenum

Chapter 8

137

by peristalsis (Rayner et al. 2000a). The manometric catheter consisted of 16

sideholes (channels) spaced at 1.5 cm intervals, comprising: six antral sideholes

(channels 1 – 6), a 4.5 cm sleeve sensor (channel 7), two sideholes on the back of the

sleeve sensor (channels 8 and 9), seven duodenal sideholes (channels 10 – 16), and an

infusion port (the latter was not used) (Rayner et al. 2000a). The correct position of

the catheter, i.e. with the sleeve sensor straddling the pylorus, was monitored by

continuous measurement of the transmucosal potential difference (TMPD) between

the most distal antral channel (channel 6; ~ -40 mV) and the most proximal duodenal

channel (channel 10; ~ 0 mV). For this purpose, an intravenous cannula filled with

sterile saline was placed subcutaneously in the posterior aspect of the forearm and

used as a reference (Rayner et al. 2000a). All channels were perfused with degassed,

distilled water, except for the two TMPD channels, which were perfused with

degassed, isotonic saline, at 0.15 mL/min. Two intravenous cannulae were inserted

into antecubital veins on opposing arms, one for blood sampling and the other for

intravenous infusion of fructose, glucose or placebo.

Subjects were positioned supine until the catheter was in the correct position, when

they were seated with their back against a gamma camera. They then consumed the

test meal, which comprised 100 g minced beef, labelled with 20 MBq 99mTc-sulphur

colloid chicken liver, followed immediately by 25 mL water (to clear the oesophagus

of food). The meal was consumed within 5 minutes and the time of meal completion

was defined as t = 0 min. Immediately after the meal, the intravenous infusion

(glucose, fructose or placebo) was initiated in single-blind, randomised order, and

ceased at t = 20 min. Gastric emptying and antropyloroduodenal motility were

monitored between t = 0 - 120 min.

Chapter 8

138

Each subject provided written, informed consent prior to their involvement. The

protocol was approved by the Human Ethics Committee of the Royal Adelaide

Hospital and all studies were performed in accordance with the Declaration of

Helsinki.

8.3.3 Measurement of blood glucose concentrations

Venous blood samples were obtained immediately prior to the intravenous infusion (t

= -5 min) and then at 0, 5, 10, 15, 30, 45, 60, 90 and 120 min. Blood glucose

concentrations were determined immediately using a portable blood glucose meter

(Medisense Companion 2 meter, Medisense Inc., Waltham, MA, USA).

8.3.4 Measurement of gastric emptying

Gastric emptying was measured scintigraphically. Radioisotopic data were acquired

at 1-minute intervals for the first hour and at 3-minute intervals thereafter. Data were

corrected for subject movement, radionuclide decay and –ray attenuation (Collins et

al. 1983). Regions-of-interest were drawn for the total stomach, which was

subsequently divided into proximal and distal stomach regions. Gastric emptying

curves for total, proximal and distal stomach regions, expressed as “percent retention”

over “time”, were then derived. From the gastric emptying curves, the intragastric

retention at 0, 15, 30, 45, 60, 75, 90, 105 and 120 min was derived. The amounts

remaining in the stomach at 45 min were compared as blood glucose concentrations

were expected to normalise soon after that time (Horowitz et al. 1996a). The lag

phase was determined visually as the time before radioactivity appeared in the

duodenum (Collins et al. 1983, Horowitz et al. 1991).

Chapter 8

139

8.3.5 Measurement of antropyloroduodenal motility

Pressure waves were analysed only when the sleeve sensor was positioned correctly

across the pylorus, according to previously defined TMPD criteria, and their

amplitude was 10 mmHg. Intraluminal pressures were recorded at 10 Hz using

custom software (HAD, written by A/Prof GS Hebbard, Melbourne, Vic, Australia),

written in LabVIEW 3.1.1 (National Instruments, Austin, TX, USA) and stored for

subsequent analysis. Data were converted for analysis (MAD, written by Prof CH

Malbert, Rennes, France) and artefacts eliminated by visual inspection of each

recording. Variables assessed were: (i) number and amplitude of antral pressure

waves (waves in any of the last 3 antral side holes of amplitude > 10 mmHg), (ii)

number and amplitude of isolated pyloric pressure waves (IPPWs) (IPPWs recorded

by the sleeve sensor in the absence of a pressure wave of onset within 5 sec of the

pyloric wave, occurring in the antral or duodenal side holes, of amplitude > 10

mmHg), and (iii) number and amplitude of duodenal pressure waves (waves in any of

the first 3 duodenal side holes of amplitude > 10 mmHg), using custom software

(written by Prof A Smout, Utrecht, The Netherlands). The number and amplitude of

pressure waves was determined manually by visual inspection of manometric

recordings, and measurements of amplitude were corrected by subtraction of baseline

pressure recordings.

8.3.6 Statistical analysis

Data (blood glucose, gastric emptying, IPPWs) were evaluated using repeated

measures analysis of variance (ANOVA) and are shown as mean ± standard error of

the mean (SEM). The number and amplitude of both antral and duodenal pressure

waves were analysed using Wald statistics with generalised estimating equations

(GEE) (Liang and Zeger 1986, Zeger et al. 1988) based on chi-square distribution.

Chapter 8

140

Manometric data (IPPWs, antral and duodenal pressure waves) were analysed in 30-

minute time intervals from t = 0 – 120 min, with a 15-minute baseline (i.e. t = -15 – 0

min). For the mean pressure analyses, the data were log-transformed, and a normal

distribution was used for the GEE. For the mean number of waves analyses, a poisson

distribution with log link was used, however, the data were overdispersed, so a

negative binomial distribution with log link was used instead. A compound symmetry

covariance structure was assumed for all analyses. Measurements from the last three

sideholes in the antrum were grouped for the analysis of number and amplitude of

antral pressure waves. Likewise, measurements from the first three sideholes in the

duodenum were grouped for the analysis of number and amplitude of duodenal

pressure waves. For pressure wave analysis (number and amplitude) of the antrum

and duodenum, data are expressed as means with upper and lower 95 % confidence

interval (95%CI) limits. Student’s t-tests were used to assess paired comparisons. A

P-value < 0.05 was considered significant in all analyses.

8.4 Results

All subjects tolerated the saline and glucose studies well. Of the 6 subjects, two

reported nausea, and four volunteered epigastric discomfort (mild and transient)

during fructose infusion.

8.4.1 Blood glucose

Baseline blood glucose did not differ between study days (t = -5 min: saline 5.4 ± 0.3

mmol/L vs. glucose 5.5 ± 0.3 mmol/L vs. fructose 5.2 ± 0.2 mmol/L) (Figure 8.1).

Mean blood glucose (calculated using 30-minute blood glucose levels from t = 0 –

120 min) was greater following glucose (7.5 ± 0.3 mmol/L) compared with both

saline (5.5 ± 0.2 mmol/L; P = 0.0008) and fructose (5.3 ± 0.2 mmol/L; P = 0.0002)

Chapter 8

141

infusions, with no difference between saline and fructose (P = 0.47). During glucose

infusion, the peak blood glucose concentration was 16.4 ± 0.6 mmol/L, and the time

to peak was 20.0 ± 3.2 min. At t = 120 min, blood glucose was less after glucose

compared to both saline (P = 0.0005) and fructose (P = 0.009) infusions, with no

difference between saline and fructose (P = 0.35).

Figure 8.1: Effects of intravenous fructose, glucose and saline(infused between t = 0 – 20 min) on the blood glucose concentration following ingestion of 100 g minced beef. Data are mean values ± SEM.

8.4.2 Gastric emptying and intragastric distribution

The overall emptying pattern of the minced beef approximated a linear function after

an initial lag phase (Figure 8.2a). The lag phases were 18.7 ± 2.9 min, 34.7 ± 7.3 min

and 35.3 ± 9.8 min for saline, glucose and fructose studies, respectively (P = NS for

all comparisons, although there were trends for the lag phase on the saline day to be

Chapter 8

142

shorter than the glucose (P = 0.07) and fructose (P = 0.09) days. There was a

treatment-by-time interaction between the three treatments (P= 0.0001) for total

gastric emptying; gastric emptying from the total stomach was slower with both

glucose and fructose compared with saline between t = 45 – 120 min (P < 0.005 for

all) with no significant difference between glucose and fructose. From t = 0 – 45 min,

there was a treatment-by-time interaction between the three treatments (P = 0.0003)

for total gastric emptying; at t = 45 min, gastric emptying was slowest with glucose

compared with both fructose (P < 0.05) and saline (P < 0.0001), although fructose

was slower than saline (P < 0.0001). In five of the 6 subjects, gastric emptying was

substantially slower after intravenous fructose than after intravenous saline, and in the

remaining one, gastric emptying was comparable. Between t = 45 – 120 min, there

was no difference in total gastric emptying between the three treatments: saline 55.7

± 2.7 % vs. glucose 58.2 ± 6.3 % vs. fructose 53.2 ± 6.3 % (P = NS for all).

There was a treatment-by-time interaction between the three treatments (P = 0.0001)

for proximal gastric emptying (Figure 8.2b); retention in the proximal stomach was

greater with glucose compared to saline (from t = 15 – 90 min; P < 0.05 for all), and

greater with fructose compared with saline (from t = 15 – 105 min; P < 0.01), with no

difference between glucose and fructose.

Chapter 8

143

Figure 8.2: Effects of intravenous fructose, glucose and saline (infused between t = 0 – 20 min) on (a) total, (b) proximal, and (c) distal, gastric emptying of 100 g minced beef. Data are mean values ± SEM.

Chapter 8

144

There was a treatment-by-time interaction between the three treatments (P = 0.0001)

for distal gastric emptying (Figure 8.2c); retention in the distal stomach was initially

greater after saline than both glucose (from t = 15 – 30 min; P < 0.002) and fructose

(from t = 15 – 45 min; P < 0.01), with no difference between glucose and fructose

(from t = 15 – 45 min). Retention in the distal stomach was subsequently less with

saline compared with both glucose (from t = 75 – 120 min; P < 0.001) and fructose

(from t = 90 – 120 min; P < 0.05), with no difference between glucose and fructose

(between t = 105 – 120 min).

8.4.3 Antropyloroduodenal manometry

8.4.3.1 Antral pressure waves

There was no significant difference in the number of antral waves at baseline between

the three studies (saline: 4.95 [95%CI: 1.68, 14.64] vs. glucose: 2.84 [95%CI: 2.43,

3.32] vs. fructose: 1.96 [95%CI: 0.98, 3.91]) (Figure 8.3a). There was a significant

treatment-by-time interaction between the three infusions (P < 0.0001). There was a

significant increase in the number of antral waves following the meal during saline,

glucose and fructose (P < 0.0001 for all) infusions. The rise in the number of antral

waves from baseline occurred promptly following meal ingestion (i.e. t = 0 – 30 min)

during infusion with saline (P = 0.02), but not during glucose (P = 0.21) or fructose

(P = 0.65), i.e. the stimulation of antral waves was attenuated by infusion of glucose

and fructose. The rise in the number of antral waves was not evident until t = 30 – 60

min after fructose (P < 0.0001) and t = 90 – 120 min after glucose (P < 0.0001). Over

the 120 min period, the number of antral waves was greater after saline compared to

both fructose (P = 0.002) and glucose (P < 0.0001) infusion, and greater for fructose

compared to glucose (P = 0.0004).

Chapter 8

145

The amplitude of antral pressure waves at baseline was higher for saline (33.92

mmHg [95%CI: 24.79, 46.41]) than glucose (17.53 mmHg [95%CI: 11.73, 26.21]) (P

= 0.01), and tended to be greater compared with fructose (20.57 mmHg [95%CI:

16.41, 25.78]) (P = 0.09), with no significant difference between fructose and glucose

(P = 0.49). Following meal ingestion, there was no significant difference in the

amplitude of antral pressure waves between the three infusions (data not shown).

8.4.3.2 Isolated pyloric pressure waves

There was no difference in the number of isolated pyloric pressure waves (IPPWs) at

baseline between the three studies (saline: 2.00 ± 0.58 vs. glucose: 3.67 ± 1.84 vs.

fructose: 1.33 ± 0.99) (Figure 8.3b). There was a significant (P = 0.0001) time effect

for the three treatments for the duration of the study. There was an increase in the

number of IPPWs from baseline to 30 min immediately following meal ingestion (i.e.

t = 0 30 min) during infusion with glucose (P = 0.003), but not with saline (P =

0.21) or fructose (P = 0.06), i.e. the number of IPPWs was increased during infusion

of glucose, but not fructose or saline. The rise in the number of IPPWs was not

evident until t = 30 60 min for both saline (P = 0.018) and fructose (P = 0.0002).

For the period t = 0 120 min, there was no treatment effect on the number of

IPPWs.

There was no significant difference in the amplitude of IPPWs at baseline between

the three treatments (data not shown). There was a significant (P = 0.0001) time

effect for the amplitude of IPPWs with the three treatments for the duration of the

study. There was a significant increase in the amplitude of IPPWs from baseline

immediately following meal ingestion (i.e. t = 0 30 min) during infusion with saline

(P = 0.02), glucose (P = 0.0005) and fructose (P = 0.04). The magnitude of the

Chapter 8

146

increase in amplitude at t = 0 – 30 min was greater with glucose compared with

fructose (P = 0.05), but not saline (P = 0.37) and there was no difference in the

amplitude of IPPWs at t = 0 – 30 min between saline and fructose (P = 0.28) (data not

shown).

8.4.3.3 Duodenal pressure waves

There was no significant difference in the number of duodenal waves at baseline

between the three infusions (saline: 11.09 [95%CI: 5.43, 22.66] vs. glucose: 7.81

[95%CI: 4.44, 13.74] vs. fructose: 6.83 [95%CI: 3.35, 13.95]) (Figure 8.3c). There

was a significant increase in the number of duodenal waves from baseline following

meal ingestion (i.e. t = 0 – 30 min) during infusion with saline (P < 0.0001) and

glucose (P < 0.0001), but not fructose (P = 0.27). However, for the 120 min period,

the number of duodenal waves was not significantly different between the three

infusions.

There was no significant difference in the amplitude of duodenal pressure waves at

baseline or after the meal between the three infusions (data not shown).

Chapter 8

147

Figure 8.3: Effects of intravenous fructose, glucose and saline (infused between t = 0 – 20 min) on number of (a) antral pressure waves (recorded by the last three antral channels), (b) isolated pyloric pressure waves, and (c) duodenal pressure waves (recorded by the first three duodenal channels), following ingestion of 100 g minced beef. Data are mean values with upper and lower 95%CI (a, c) and mean values ± SEM (b).

Chapter 8

148

8.5 Discussion

This study establishes that when given intravenously, fructose has the capacity to

slow gastric emptying and modulate antropyloroduodenal motility in healthy subjects

and that the magnitude of these effects is comparable to those induced by an identical

intravenous glucose load. In particular, intravenous administration of fructose

resulted in a transient slowing of gastric emptying and intragastric distribution of a

solid meal, associated with suppression of antral and duodenal pressure waves and

(non-significant) stimulation of isolated pyloric pressure waves.

The demonstrated effects of intravenous glucose, resulting in peak plasma glucose

concentrations of ~ 16 mmol/L on gastric emptying (MacGregor et al. 1976, Samsom

et al. 1997, Schvarcz et al. 1997) and antropyloroduodenal motility (Fraser et al.

1991b, Russo et al. 1996, Samsom et al. 1997) are consistent with previous reports.

As has been shown, these effects were transient, indicating that they were secondary

to hyperglycaemia. The mechanism(s) mediating the effects of hyperglycaemia on

gastric motility are uncertain, but may be vagally-dependent (Schvarcz et al. 1995b).

Glucose-dependent neurons are also known to be present in the myenteric plexus (Liu

et al. 1999) and central nervous system (Mizuno and Oomura 1984). Although

plasma fructose was not quant i f ied, the effects of f ructose on gast r ic

emptying/motility are probably, like glucose, dependent on the plasma

concentrations. As expected, intravenous fructose had no effect on plasma glucose

(Aitken and Dunnigan 1969, Dunnigan and Ford 1975, Tounian et al. 1994), nor

would it be expected to affect insulin secretion (Dunnigan and Ford 1975). Hence, it

appears that both glucose and fructose have the capacity to slow gastric emptying

both as a result of their interaction with receptors in the small intestinal lumen (Elias

et al. 1968, Guss et al. 1994, Horowitz et al. 1996a, Sole and Noakes 1989) and also

Chapter 8

149

the elevation of plasma monosaccharide concentrations. It should be recognised that

the author studied only a relatively small number of subjects and the possibility that

there are modest differences in the effects of intravenous fructose and glucose, which

may potentially contribute to the slightly more rapid gastric emptying of oral fructose

when compared to oral glucose, cannot be excluded (Elias et al. 1968, Guss et al.

1994, Horowitz et al. 1996a, Moran and McHugh 1981, Sole and Noakes 1989),

although this may well be explicable by the differential effects on gut hormone

release, particularly GLP-1 (Kong et al. 1999a, Rayner et al. 2000a, Vozzo et al.

2002). Four subjects experienced transient, mild epigastric discomfort during

intravenous fructose infusion. This has been reported during rapid fructose infusion in

much higher doses (0.5 g/kg in 5 min to 1.5 g/kg in 60 min) (Elliott et al. 1967), and

its cause remains uncertain. In the current study, the severity and duration of

discomfort did not compromise the completion of the study and effects of fructose on

gastric emptying appeared consistent.

That fructose has the capacity to slow gastric emptying as a result of its plasma levels

is not surprising given that intravenous administration of high-caloric nutrients

(parenteral nutrition) has been reported to delay solid (MacGregor et al. 1979) and

liquid (Bursztein-De Myttenaere et al. 1994) gastric emptying in humans. It has been

suggested that this reflects the stimulation of gastric acid secretion by the amino acid

content of the parenteral feed (Isenberg and Maxwell 1978, McArthur et al. 1983),

with a concomitant suppression of pancreatic and biliary secretion and a reduction in

the buffering capacity of the duodenum (Bursztein-De Myttenaere et al. 1994).

Energy intake may also be suppressed during parenteral nutrition, possibly as a result

of slower gastric emptying (Hunt 1980). Intravenous administration of high-dose

Chapter 8

150

amino acids has also been shown to decrease antral motility (Gielkens et al. 1999).

As with fructose, the mechanisms mediating these effects remain to be determined.

Chapter 9

151

9Chapter 9E F F E C T O F I T O P R I D E O N G A S T R I C E M P T Y I N G I N

LONGSTANDING DIABETES MELLITUS

9.1 Summary

Delayed gastric emptying occurs in 30 50 % of patients with longstanding type 1 or

2 diabetes, and represents a major cause of morbidity. Current therapeutic options are

limited. We aimed to evaluate the effects of itopride on gastric emptying in

longstanding diabetes. Twenty-five patients (20 type 1, 5 type 2; 10 male, 15 female;

mean age 45.2 ± 2.7 yr; BMI 27.5 ± 0.9 kg/m2; duration diabetes 20.2 ± 2.4 yr) were

enrolled in a double-blind, placebo-controlled, randomised, crossover trial. Subjects

received both itopride (200 mg) and placebo tid for 7 days, with a washout of 7 14

days. Gastric emptying (scintigraphy), blood glucose (glucometer) and upper

Chapter 9

152

gastrointestinal symptoms (questionnaire) were measured following each treatment

period. The test meal comprised 100 g minced beef labelled with 99mTc-sulphur

colloid and 150 mL 10 % dextrose labelled with 67Ga-EDTA. There was a slight trend

for itopride to accelerate both solid (P = 0.09) and liquid (P = 0.09) gastric emptying.

The effect of itopride on emptying of both solids and liquids tended to be greater

when the emptying on placebo was slower (solids: r = 0.39, P = 0.057; liquids: r =

0.44, P < 0.03). Twelve (48 %) patients had delayed solid and/or liquid gastric

emptying on placebo and in this group, itopride modestly accelerated liquid (P <

0.05), but not solid (P = 0.39), emptying. Itopride had no effect on mean blood

glucose during the gastric emptying measurement (placebo: 9.8 ± 0.6 mmol/L vs.

itopride: 9.6 ± 0.6 mmol/L), or gastrointestinal symptoms (placebo: 1.4 ± 0.4 vs.

itopride: 1.8 ± 0.5). Itopride, in a dose of 200 mg tid for 7 days, tends to accelerate

gastric emptying of liquids and solids in longstanding diabetes. The magnitude of this

effect appears to be modest and possibly dependent on the rate of gastric emptying

without treatment.

9.2 Introduction

Delayed gastric emptying occurs in 30 – 50 % of patients with longstanding type 1 or

type 2 diabetes and may be associated with upper gastrointestinal symptoms,

impaired nutrient and drug absorption, and poor glycaemic control (Horowitz et al.

1991, Horowitz et al. 2006, Rayner et al. 2001, Samsom et al. 2003). Treatment with

prokinetic drugs, including metoclopramide, domperidone, cisapride and

erythromycin, forms the mainstay of therapy of symptomatic diabetic gastroparesis

(Talley 2003). While short-term administration of all of these drugs has been shown

to accelerate gastric emptying and improve symptoms, both the magnitude of

symptomatic improvement and the change in gastric emptying are variable (Talley

Chapter 9

153

2003). Moreover, all of the currently available drugs have significant limitations. The

use of metoclopramide is associated with a high incidence of adverse central nervous

system effects (Tonini et al. 2004). There is evidence that the prokinetic effects of

erythromycin, metoclopramide and domperidone are not sustained during chronic

administration (Horowitz et al. 1985, Horowitz et al. 1987); in the case of

erythromycin, tachyphylaxis probably reflects down-regulation of motilin receptors

(Talley 2003). Cisapride was arguably the therapy of first choice (Horowitz et al.

1987) but has had its use curtailed greatly due to its capacity to prolong the Q-T

interval and induce potentially fatal cardiac arrhythmias (Evans and Krentz 1999).

The prokinetic effect of some drugs, including erythromycin and cisapride, is also

attenuated during hyperglycaemia (Horowitz et al. 2002a, Jones et al. 1999a, Jones et

al. 1999b, Petrakis et a l . 1999a, Petrakis et al. 1999b). Hence, current

pharmacological therapy is suboptimal and there is a need for new treatment options.

Itopride is a benzamide derivative bearing a distinct structural resemblance to

prokinetic drugs including cisapride, metoclopramide and domperidone (Holtmann et

al. 2006, Iwanaga et al. 1990, Iwanaga et al. 1991, Iwanaga et al. 1996, Kakrani and

Madraki 2002, Tsubouchi et al. 2003). Itopride blocks dopamine (D2) receptors on

cholinergic motor neurons and inhibits acetylcholinesterase (AChE) to increase the

acetylcholine level, and, thereby, stimulate gastrointestinal motility (Iwanaga et al.

1991, Sakaguchi et al. 2001, Tsubouchi et al. 2003). Itopride is highly polar and,

therefore, does not cross the blood-brain barrier readily so that the risk of

extrapyramidal effects is low. Itopride also does not affect the Q-T interval and is

metabolised by flavine-dependent monooxygenases (FMO3), rather than the

cytochromes P450, so the potential for drug-drug interactions is low (Mushiroda et al.

2000).

Chapter 9

154

A recent study reported that itopride was effective in the treatment of functional

dyspepsia with a low risk of adverse effects (Holtmann et al. 2006). In the dog,

itopride has been shown to dose-dependently stimulate motility in the antrum,

duodenum and colon (Tsubouchi et al. 2003). In a Japanese study, itopride, when

administered as a single dose of 50 mg, was reported to accelerate gastric emptying,

as assessed by the relatively insensitive acetaminophen method, in 15 ‘chronic

gastritis’ patients, of whom 11 apparently had delayed gastric emptying (Harasawa

and Miwa 1993). There is only limited evidence that itopride is an effective

prokinetic agent in patients with diabetes, and there have hitherto been no

randomised, placebo-controlled trials. In a study reported in abstract form, oral

administration of itopride (150 mg/day for two weeks) improved both gastric

emptying and gastric myoelectrical activity in 12 type 2 diabetic patients with

peripheral neuropathy (Basque et al. 2005). However, the study was not randomised

and gastric emptying was quantified by the acetaminophen (Sanaka et al. 1998) and

radiopaque marker (Horowitz and Fraser 1995) methods, rather than scintigraphy,

which is considered to be the ‘gold standard’ (Horowitz and Fraser 1995).

The purpose of this study was to evaluate the effect of itopride on gastric emptying of

solids and liquids measured by scintigraphy in an unselected cohort of patients with

longstanding diabetes mellitus, using a randomised, placebo-controlled design.

9.3 Materials and Methods

9.3.1 Subjects

Twenty-five Caucasian patients with diabetes (20 type 1, 5 type 2; 10 male, 15

female), aged 45.2 ± 2.7 years, body mass index 27.5 ± 0.9 kg/m2, duration of known

Chapter 9

155

diabetes 20.2 ± 2.4 years and glycated haemoglobin 8.7 ± 0.4 %, were enrolled in the

study. Patients were randomly selected from those attending outpatient clinics at the

Royal Adelaide Hospital and through advertisements posted in local community

newspapers. Patients were not selected on the basis of the presence, or absence, of

gastrointestinal symptoms or known gastroparesis. None had a history of liver,

cardiac or respiratory disease or gastrointestinal surgery, apart from uncomplicated

appendicectomy. All patients had normal renal function (serum creatinine 0.05 – 0.12

mmol/L; calculated creatinine clearance > 60 mL/min). One of the 5 patients with

type 2 diabetes was using insulin. The oral hypoglycaemic agents used to treat

patients with type 2 diabetes included metformin, gliclazide, glimepiride,

rosiglitazone, pioglitazone and acarbose. Patients taking medication known to

influence gastrointestinal motility were excluded and smoking was prohibited for 24

h prior to, and during, each gastric emptying measurement. Written, informed consent

was obtained from each subject prior to their enrolment in the study. The protocol and

advertisements were approved by the Human Research Ethics Committee of the

Royal Adelaide Hospital and all studies were performed in accordance with the

Declaration of Helsinki.

9.3.2 Experimental protocol

Subjects were enrolled in a randomised, double-blind, placebo-controlled, crossover

trial consisting of a screening visit (V1) (performed within 14 days from treatment

start), followed by two 7 day treatment periods, during which each subject received

both itopride (200 mg) and placebo three times daily before meals. The two treatment

periods were randomised and separated by 7 14 days. Autonomic nerve function

was assessed during the screening visit (V1). Immediately following each treatment

period (i.e. V2 and V3), subjects attended the Department of Nuclear Medicine, PET

Chapter 9

156

and Bone Densitometry at 09.30 hours after an overnight fast (14 h for solids, 12 h

for liquids) and an intravenous cannula was inserted into an antecubital vein for

subsequent blood sampling. The venous blood glucose concentration was then

measured. - If the blood glucose concentration was < 12 mmol/L, the subject was

asked to administer their usual morning dose of insulin 15 minutes before the test

meal; if the blood glucose concentration was ≥ 12 mmol/L, the subject was instructed

to administer their usual dose of insulin immediately and the gastric emptying

measurement was not commenced until the blood glucose concentration was < 12

mmol/L. Subjects on oral hypoglycaemic agents took their usual morning dose

approximately 15 minutes before the test meal. Subjects took their last dose of trial

medication 60 minutes before commencement of the gastric emptying measurement.

Compliance was assessed by a count of returned tablets.

9.3.3 Measurement of gastric emptying

Gastric emptying was measured using a standardised, dual isotope scintigraphic test

(Collins et al. 1983). The test meal comprised 100 g lean minced beef, labelled with

20 MBq 99mTc-sulphur colloid chicken liver, followed immediately by 150 mL 10 %

dextrose, labelled with 6 MBq 67Ga-EDTA. The solid component of the meal was

consumed within 5 min, followed by the liquid within 1 min. Radioisotopic data were

acquired with the subject seated with their back against a gamma camera (GEnie; GE

Healthcare Technologies, Milwaukee, WI, USA) at 1-min intervals for the first hour

and at 3-min intervals thereafter. Time zero was defined as the time of meal

completion and gastric emptying was monitored for 120 min. Data were corrected for

subject movement, radionuclide decay and –ray attenuation, the latter by using

correction factors derived from a lateral image of the stomach (Collins et al. 1983).

The lag phase was determined visually as the time between meal completion and the

Chapter 9

157

appearance of radioactivity in the proximal small intestine (Collins et al. 1983). From

the gastric emptying curves (expressed as the percentage retention over time), the

intragastric retention at 0, 15, 30, 45, 60, 75, 90, 105 and 120 min was derived

(Collins et al. 1983). The amount (%) of solid remaining in the stomach at 100 min

(T100) and the time taken for 50 % of the liquid to empty (T50) were also quantified

(Horowitz et al. 1991); gastric emptying was considered to be delayed when the solid

T100 was > 61 % and/or the liquid T50 was > 31 min, based on an established normal

range (Horowitz et al. 1987, Horowitz et al. 1991).

9.3.4 Measurement of glycaemic control

During each gastric emptying measurement, venous blood samples (5 mL) were

obtained at t = –2, 30, 60, 90 and 120 min. Blood glucose concentrations were

determined immediately using a portable blood glucose meter (Medisense

Companion 2 meter, Medisense Inc., Waltham, MA, USA). The mean of these blood

glucose measurements was calculated.

9.3.5 Assessment of upper gastrointestinal symptoms

Upper gastrointestinal symptoms were assessed by questionnaire on the morning of

both study days, prior to consumption of the test meal. “Gastric” (anorexia, nausea,

early satiation, abdominal bloating/fullness, vomiting, abdominal pain) and

“oesophageal” (dysphagia, heartburn and acid regurgitation) symptoms were graded

as 0 = none, 1 = mild (the symptom could be ignored), 2 = moderate (the symptom

could not be ignored, but did not influence daily activities) and 3 = severe (the

symptom influenced daily activities). As there were nine symptoms, the maximum

possible total score was 27 (Jones et al. 1995b).

Chapter 9

158

9.3.6 Assessment of autonomic nerve function

Cardiovascular autonomic nerve function was assessed on V1 using standardised

cardiovascular reflex tests (Ewing and Clarke 1982). Parasympathetic function was

calculated by the variation (R-R interval) of the heart rate during deep breathing (E/I)

and the immediate heart rate response to standing ("30:15" ratio). Sympathetic

function was assessed by the fall in systolic blood pressure in response to standing.

Each of the tests was scored 0 = normal, 1 = borderline, 2 = abnormal, for a

maximum total score of 6. A score of 3 was considered to be indicative of

autonomic neuropathy (Horowitz et al. 1991, Jones et al. 1995b).

9.3.7 Statistical analysis

The sample size of 25 patients was based on the number of subjects included in our

previous studies, which demonstrated beneficial effects on gastric emptying after oral

prokinetic therapy (domperidone and cisapride) in patients with diabetes (Horowitz et

al. 1985, Horowitz et al. 1987), in order to provide statistical power of 80 % with P <

0.05. Data (gastric emptying and blood glucose concentration) were evaluated using

repeated measures analysis of variance (ANOVA) with “treatment” and “time” as

factors. Treatment order effect was assessed by repeated measures ANOVA, with

“treatment*order” as a between-subjects factor. A carryover effect of the drug was

deemed evident if there was significance for the “treatment*order” main effect or any

of its interactions. Student’s t-tests (two-tailed) for paired comparisons were used to

compare sample means (T100 and T50). Relationships between variables were

analysed using linear regression analysis. Data are presented as mean values ± SEM.

For the rejection of the null hypothesis, an error probability of P < 0.05 was

considered as significant.

Chapter 9

159

9.4 Results

No serious adverse events were reported. The total score for autonomic neuropathy

was 3.0 ± 0.2 (median score 3, range 0 4). Fifteen of the 25 patients had evidence

of autonomic neuropathy (i.e. total score ≥ 3) and in the remaining 10, the result was

‘borderline’ (i.e. total score ≥ 1 and < 3). Assessment of tablet counts revealed

adequate patient compliance with respect to medication; 3 patients missed 1 of the 22

doses and 1 patient missed 3 doses. There was no evidence of a treatment order effect

in any of the analyses.

9.4.1 Gastric emptying

On both days, solid emptying approximated an overall linear pattern after an initial

lag phase (itopride: 17.6 ± 2.5 min vs. placebo: 20.6 ± 3.1 min; P = 0.31), and liquid

gastric emptying an overall monoexponential pattern, after a short lag phase (itopride:

1.3 ± 0.2 min vs. placebo: 1.4 ± 0.2 min; P = 0.65). There was a slight trend for an

acceleration of gastric emptying by itopride for both solids (P = 0.09) and liquids (P =

0.09) (Figure 9.1). However, there was no significant difference in either the solid

T100 (itopride: 47.7 ± 4.3 % vs. placebo: 52.2 ± 4.1 %; P = 0.23) or liquid T50

(itopride: 25.5 ± 2.4 min vs. placebo: 27.4 ± 2.2 min; P = 0.43) between treatments. In

all 25 patients, the emptying of both solids and liquids tended to be more accelerated

when the emptying on placebo was slower (solids: r = 0.39, P = 0.057; liquids: r =

0.44, P < 0.03) (Figure 9.2). Of the 25 patients, about one third emptied solids and

liquids more rapidly on placebo compared with itopride (Figure 9.2).

Chapter 9

160

Figure 9.1: Gastric emptying of (a) solid and (b) liquid meal components following treatment with itopride (200 mg po tid) and placebo (n = 25, data are mean values ± SEM).

Figure 9.2: Relationship between the magnitude of the change in gastric emptying (placebo - itopride) for (a) solid (retention at 100 min) and (b) liquid (50 % emptying time) with gastric emptying on placebo (n = 25, data are mean values ± SEM).

Chapter 9

161

Of the 25 patients, 12 (48 %) had delayed solid and/or liquid gastric emptying on

placebo. In this group, there was no difference in solid (P = 0.39) emptying, however,

there was a treatment-by-time interaction for liquid (P = 0.02); itopride accelerated (P

<0.05) liquid gastric emptying in this group (Figure 9.3). There was no significant

difference in either the solid T100 (itopride: 60.2 ± 6.1 % vs. placebo: 62.7 ± 5.9 %; P

= 0.72) or liquid T50 (itopride: 30.3 ± 3.9 min vs. placebo: 34.1 ± 3.0 min; P = 0.29)

between treatments. Seven of the 12 patients with delayed solid and/or liquid gastric

emptying on placebo had evidence of autonomic neuropathy (i.e. total score ≥ 3).

Figure 9.3: Gastric emptying of (a) solid and (b) liquid meal components following treatment with itopride (200 mg po tid) and placebo in patients with delayed gastric emptying of solids and/or liquids on placebo (n = 12, data are mean values ± SEM).

9.4.2 Blood glucose concentration

There was no significant difference in baseline blood glucose concentrations between

treatments (itopride: 8.3 ± 0.6 mmol/L vs. placebo: 8.4 ± 0.7 mmol/L). On both days,

there was a rise (P = 0.0001 for both) in blood glucose after the meal (Figure 9.4),

without any significant difference between itopride and placebo. There was also no

difference in mean blood glucose concentrations (itopride: 9.6 ± 0.6 mmol/L vs.

Chapter 9

162

placebo: 9.8 ± 0.6 mmol/L) during the studies. The blood glucose concentration-time

profile did not differ significantly between treatments in the 12 patients with delayed

solid and/or liquid gastric emptying on placebo (data not shown). The effect of

itopride on gastric emptying was not significantly related to the blood glucose

concentration (data not shown).

Figure 9.4: Blood glucose concentrations during gastric emptying measurements following treatment with itopride (200 mg po tid) and placebo (n = 25, data are mean values ± SEM).

In the 8 patients in whom gastric emptying of solids was faster on placebo than

itopride, there was a relationship with the change in blood glucose concentration (r =

0.83, P = 0.01), i.e. gastric emptying was slower when the blood glucose

concentration at baseline was relatively higher.

Chapter 9

163

9.4.3 Upper gastrointestinal symptoms

Fourteen patients had gastrointestinal symptoms on placebo; two of these rated at

least one symptom as ‘severe’. There was no difference in the total score for upper

gastrointestinal symptoms between treatments (itopride: 1.8 ± 0.5 vs. placebo: 1.4 ±

0.4), nor was there any difference in either “gastric” (itopride: 1.6 ± 0.4 vs. placebo:

1.3 ± 0.4) or “oesophageal” (itopride: 0.2 ± 0.1 vs. placebo: 0.1 ± 0.1) symptoms.

9.5 Discussion

Our observations indicate that itopride hydrochloride, when administered in a dose of

200 mg tid for 7 days, has little effect on gastric emptying of solid and/or liquid meal

components in patients with longstanding diabetes mellitus. This is perhaps

unexpected in view of previous reports relating to the effects of itopride on gastric

motility/gastric emptying (Basque et al. 2005, Harasawa and Miwa 1993).

Itopride has been reported to have beneficial effects on gastric emptying and

gastroduodenal motility in both animals (Iwanaga et al. 1990, Iwanaga et al. 1991,

Iwanaga et al. 1996, Tsubouchi et al. 2003) and humans (Basque et al. 2005,

Harasawa and Miwa 1993). In conscious dogs, intravenous itopride (3 mg/kg)

increased antral and duodenal contractility and antagonised dopamine-induced

inhibition of gastric contractions (Iwanaga et al. 1990). When administered orally,

itopride (30 mg/kg) accelerated gastric emptying in dogs and antagonised dopamine-

induced delay in gastric emptying in rats (Iwanaga et al. 1991). At higher doses (10

mg/kg iv), itopride enhanced contractile activity in all regions of the gastrointestinal

tract from the stomach to the colon in the dog (Tsubouchi et al. 2003). In 15 Japanese

‘chronic gastritis’ patients, in whom itopride (50 mg po), or placebo, was

administered 30 minutes prior to ingestion of a drink containing 1.5 g acetaminophen,

Chapter 9

164

itopride was reported to accelerate gastric emptying in all of them, as reflected by an

increase in the serum acetaminophen concentration 45 min after the test meal

(Harasawa and Miwa 1993). While it was suggested that of these 15 patients, 11 had

delayed gastric emptying, this diagnosis was made on the basis of serum

acetaminophen levels. In another Japanese study, itopride (150 mg po daily) was

administered for two weeks in 12 patients with type 2 diabetes who had peripheral

neuropathy and been “diagnosed with gastroparesis”. Itopride was reported to

accelerate gastric emptying of a meal containing radiopaque capsules and

acetaminophen, however, the study was not placebo-controlled or randomised

(Basque et al. 2005). In a more recent study performed in the US, itopride in doses of

100 mg and 200 mg tid reduced total gastric volume without accelerating gastric

emptying in healthy volunteers (Choung et al. 2007). Thus, although both a single

dose of 50 mg po (Harasawa and Miwa 1993), and 150 mg po administered for two

weeks (Basque et al. 2005), were reported to accelerate gastric emptying in Japanese

patients with gastroparesis, doses of 100 mg and 200 mg tid proved ineffective in

healthy volunteers (Choung et al. 2007). The absence of an effect of itopride on

gastric emptying in healthy subjects (Choung et al. 2007) does not argue strongly

against an effect in patients with gastroparesis, as it is well recognised that the effect

of prokinetic drugs is usually more marked when gastric emptying is delayed (Talley

2003). It should be also noted that in Caucasians, the AUC and Cmax for itopride are

some 30 50 % less than that in Japanese subjects for the identical dose, so that 50

mg itopride in Japanese subjects resulted in a Cmax comparable to that achieved with

a 100 mg dose in Caucasians (Seiberling 2001).

We intentionally selected a heterogeneous cohort of patients with longstanding type 1

or type 2 diabetes, the majority of whom had autonomic neuropathy and less than

Chapter 9

165

optimal chronic glycaemic control. While itopride did not accelerate gastric emptying

of solids or liquids significantly, the observed relationship between the magnitude of

the improvement of gastric emptying by itopride with gastric emptying on placebo

suggests that an acceleration of gastric emptying may only be evident in patients with

gastroparesis, particularly those with markedly delayed gastric emptying. As

discussed above, this would not be surprising (Talley 2003). Twelve of the 25

patients we studied had delayed solid and/or liquid gastric emptying (as would be

expected given the selection criteria) (Horowitz et al. 1991, Samsom et al. 2003) and

in this group, itopride accelerated gastric emptying of liquids significantly, although

the magnitude of this acceleration was modest. These data should also be regarded

circumspectly as there was no significant difference in the T50 value (P = 0.29).

Hence, particularly given the P-value (0.09), the use of a larger sample size may well

have detected a difference in gastric emptying on itopride compared with placebo,

although any effect would seem likely to be small. Previous studies have

demonstrated substantial effects of short-term administration of other prokinetics (e.g.

domperidone, erythromycin, metoclopramide) with comparable, or smaller, numbers

of subjects (Horowitz et al. 1985, Horowitz et al. 1987, Janssens et al. 1990). It

should also be recognised that the prokinetic effects of such agents may be dose-

dependent, as is the case for erythromycin (Coulie et al. 1998) and that higher doses

may induce a motor pattern which slows gastric emptying (Coulie et al. 1998).

Hence, further evaluation of the effects of different doses of itopride on gastric

emptying in patients with diabetic gastroparesis, particularly those with markedly

delayed gastric emptying are warranted.

It should also be recognised that acute changes in the blood glucose concentration

have a substantial, and reversible, effect on gastric emptying in both healthy subjects

Chapter 9

166

and patients with diabetes. Marked hyperglycaemia (blood glucose 16 20

mmol/L) (Fraser et al. 1990) and even blood glucose concentrations that are within

the normal postprandial range (4 8 mmol/L) (Schvarcz et al. 1997), slow gastric

emptying when compared to euglycaemia, while insulin-induced hypoglycaemia

accelerates emptying (Chapter 7). Acute hyperglycaemia, including changes in the

blood glucose concentration within the postprandial range, may also attenuate the

response to prokinetic drugs (Horowitz et al. 2002a, Jones et al. 1999b, Petrakis et al.

1999a, Petrakis et al. 1999b). Hence, while we ensured that the blood glucose

concentration immediately prior to the gastric emptying measurement was < 12

mmol/L and the mean blood glucose during the gastric emptying measurements were

< 10 mmol/L, we cannot discount the possibility that the effects of itopride on gastric

emptying may be more marked during euglycaemia. This can only be resolved by the

use of glucose clamps to stabilise blood glucose concentrations in the euglycaemic

range. It is, however, relevant to note that while the effects of erythromycin on gastric

emptying have been evaluated during euglycaemia (Janssens et al. 1990), blood

glucose concentrations have not been stabilised (or in most cases even monitored) in

essentially all studies relating to the effects of prokinetic drugs on gastric emptying in

diabetes and despite this, beneficial effects have been demonstrable (Horowitz et al.

1985, Horowitz et al. 1987).

Upper gastrointestinal symptoms occur frequently in patients with diabetes and affect

quality of life adversely (Horowitz et al. 1991). The relationship between symptoms

and disordered gastric emptying is, however, relatively weak (Jones et al. 1995b).

Similarly, there is a poor correlation between the effects of prokinetic drugs on

symptoms and gastric emptying (Talley 2003). A beneficial effect on symptoms may

also potentially be mediated by mechanisms unrelated to acceleration of gastric

Chapter 9

167

emptying (Talley 2003). In a recent study, itopride in a dose of 50 mg po tid has been

reported to improve gastrointestinal symptoms in patients with functional dyspepsia,

but gastric emptying was not measured (Holtmann et al. 2006). The majority of our

subjects did not have severe symptoms and our study was not designed to evaluate

the effect of itopride on upper gastrointestinal symptoms.

Chapter 10

168

10Chapter 10ACUTE EFFECTS OF C-PEPTIDE ON GASTRIC EMPTYING IN

LONGSTANDING TYPE 1 DIABETES

10.1 Summary

Gastric emptying is delayed in 30 50 % of patients with longstanding type 1

diabetes mellitus, particularly in those who have autonomic neuropathy. There is

evidence that C-peptide improves autonomic nerve function in type 1 diabetes. The

aim of the present study was to evaluate the effects of C-peptide on solid and liquid

gastric emptying in longstanding type 1 diabetes. Eight type 1 patients (5 male, 3

female), aged 47.3 ± 3.1 years, with diabetes of duration 24.0 ± 2.0 years, were

studied. Gastric emptying of a mixed solid (100 g minced beef) and liquid (150 mL

10 % dextrose) meal, was measured by scintigraphy on two days during intravenous

Chapter 10

169

infusion of either C-peptide (6 pmol/kg/min) or isotonic saline. Infusions commenced

30 minutes prior to meal ingestion and continued for 150 minutes. Autonomic

function was evaluated after the gastric emptying measurement by three standardised

cardiovascular reflex tests. During the saline infusion, gastric emptying was delayed

in three patients. Three patients had definite autonomic neuropathy; in three others, at

least one of the three tests was borderline. C-peptide infusion, which resulted in

physiological serum concentrations (1.4 ± 0.2 nmol/L), had no effect on either solid

or liquid gastric emptying. In patients with longstanding type 1 diabetes, acute

intravenous administration of C-peptide (6 pmol/kg/min) does not appear to

accelerate gastric emptying.

10.2 Introduction

Delayed gastric emptying occurs in 30 50 % of patients with longstanding type 1

diabetes and may be associated with upper gastrointestinal symptoms, poor glycaemic

control and impaired oral drug absorption (Horowitz et al. 1991, Horowitz et al.

2002b). The pathogenesis of diabetic gastroparesis is poorly defined. Delayed gastric

emptying occurs more frequently in those patients with autonomic neuropathy

(Buysschaert et al. 1987, Horowitz et al. 1991, Kawagishi et al. 1997, Merio et al.

1997, Ziegler et al. 1996) (usually assessed by standardised cardiovascular reflex

tests) (Ewing and Clarke 1982) and it has been suggested that autonomic neuropathy

is the major factor underlying delayed gastric emptying (Stacher et al. 2003).

However, it is also clear that acute hyperglycaemia may slow gastric emptying

(Fraser et al . 1990, Horowitz et al. 2002b). Current treatment of diabetic

gastroparesis is less than optimal (Horowitz et al. 2002b).

Chapter 10

170

There is now compelling evidence that C-peptide, a cleavage moiety released from

proinsulin during insulin biosynthesis, that has traditionally been thought to be

physiologically inert (Steiner 1978), may improve autonomic nerve function in type 1

diabetes (Johansson et al. 1996, Johansson et al. 2000, Sima et al. 2001, Zhang et al.

2001). There is a lso evidence that C-peptide deficiency contributes to the

microvascular complications of type 1 diabetes (Johansson et al. 1992, Johansson et

al. 1993). C-peptide is released in equimolar amounts with those of insulin

(Rubenstein et al. 1969) and is, accordingly, not present in longstanding type 1

patients. In animal models of diabetes (Sima et al. 2001, Zhang et al. 2001), and type

1 patients (Johansson et al. 2000), C-peptide may improve peripheral nerve function.

For example, in type 1 diabetic BB/Wor rats, administration of C-peptide in

“replacement” doses for 2 – 8 months improved nerve conduction velocity, paranodal

swelling and axoglial dysjunction (Sima et al. 2001), and in type 1 patients, treatment

with C-peptide for 3 months was associated with an improvement in sensory nerve

conduction velocity (Johansson et al. 2000). Studies by Johansson et al. have

provided evidence that both acute (Johansson et al. 1996) and chronic (Johansson et

al. 2000) administration of C-peptide improves cardiovascular autonomic function in

type 1 patients. In particular, intravenous administration of C-peptide over a period of

3 hours and resulting in physiological blood levels, improved cardiovascular

autonomic function, particularly parasympathetic tests, such as heart rate variability

(Johansson et al. 1996).

The aim of this study was to evaluate the acute effects of C-peptide on gastric

emptying in patients with longstanding type 1 diabetes.

Chapter 10

171

10.3 Materials and Methods

10.3.1 Subjects

Eight patients with type 1 diabetes (5 male, 3 female, mean age: 47.3 ± 3.1 yr, body

mass index (BMI): 24.6 ± 1.0 kg/m2, duration of known diabetes: 24.0 ± 2.0 yr) were

studied. Participants were randomly selected from ambulant outpatients attending the

Royal Adelaide Hospital. Smokers were required to abstain from cigarettes for at

least 12 h prior to each study day. No patient had a history of gastrointestinal disease

or surgery, significant respiratory or cardiac disease, chronic alcohol abuse or

epilepsy. No patient was taking medication known to influence gastrointestinal

function. Glycated haemoglobin was 8.7 ± 0.4 % (normal < 6.0 %).

10.3.2 Experimental protocol

Each patient underwent two randomised, single-bind studies, separated by at least

three days. During each study, patients received an intravenous infusion of either C-

peptide (6 pmol/kg/min) (Clinalfa, Läufelfingen, Switzerland) or placebo (saline 0.9

%w/v), administered at an identical rate (3 mL/min) for 150 min i.e. the total volume

infused was 450 mL. The dose of C-peptide was chosen on the basis of a previous

study (Johansson et al. 1996).

On each day, patients attended the Department of Nuclear Medicine, PET and Bone

Densitometry at approximately 09:00 h following an overnight fast (14 h for solids

and 12 h for liquids). Two cannulae were inserted into antecubital veins, one for

blood sampling and the other for intravenous infusion of C-peptide or placebo. The

venous blood glucose concentration was then measured. If the blood glucose

concentration was < 12 mmol/L, patients were asked to administer their usual

Chapter 10

172

morning dose of insulin 15 minutes before the test meal; if the blood glucose

concentration was ≥ 12 mmol/L, patients were instructed to administer their usual

dose of insulin immediately and the intravenous infusion was not commenced until

the blood glucose concentration was < 12 mmol/L. The intravenous infusion

commenced 30 min prior to commencement of meal ingestion, and when the blood

glucose concentration was < 12 mmol/L, with the patients seated with their back

against a gamma camera (Collins et al. 1983, Horowitz et al. 1987, Horowitz et al.

1991). The test meal comprised 100 g minced beef, labelled with 20 MBq 99mTc-

sulphur colloid chicken liver, followed immediately by 150 mL 10 % dextrose,

labelled with 6 MBq 67Ga-EDTA (Jones et al. 2002) and was consumed within 5

minutes. The time of meal completion was defined as t = 0 min. Gastric emptying

was then monitored for 120 minutes. About 30 minutes after the completion of the

gastric emptying measurement, autonomic nerve function was evaluated using

cardiovascular reflex tests (Ewing and Clarke 1982).

Each patient provided written, informed consent prior to their involvement in the

study. The protocol was approved by the Human Ethics Committee of the Royal

Adelaide Hospital and all studies were performed in accordance with the Declaration

of Helsinki.

10.3.3 Measurement of gastric emptying

Gastric emptying was measured scintigraphically. Radioisotopic data were acquired

at 1-minute intervals for the first hour and at 3-minute intervals thereafter. Data were

corrected for subject movement, radionuclide decay and –ray attenuation (Collins et

al. 1983). From the gastric emptying curves (expressed as the percentage retention

over time), the intragastric retention at 0, 15, 30, 45, 60, 75, 90, 105 and 120 min was

Chapter 10

173

derived (Collins et al. 1983, Jones et al. 2001a, Jones et al. 2002). The amount (%) of

solid remaining in the stomach at 100 min (T100) and the time taken for 50 % of the

liquid to empty (T50) were also quantified (Horowitz et al. 1991); gastric emptying

was considered to be delayed when the solid T100 was > 61 % and/or the liquid T50

was > 31 min, based on an established normal range (Horowitz et al. 1987, Horowitz

et al. 1991). The lag phase was determined visually as the time before radioactivity

appeared in the duodenum (Collins et al. 1983).

10.3.4 Assessment of autonomic nerve function

Cardiovascular autonomic nerve function was assessed on both days using

standardised cardiovascular reflex tests (Ewing and Clarke 1982). Parasympathetic

function was calculated by the variation (R-R interval) of the heart rate during deep

breathing (E/I) and the immediate heart rate response to standing ("30:15" ratio).

Sympathetic function was assessed by the fall in systolic blood pressure in response

to standing. Each of the tests was scored 0 = normal, 1 = borderline, 2 = abnormal,

for a maximum total score of 6. A score of 3 was considered to be indicative of

autonomic neuropathy (Horowitz et al. 1991, Jones et al. 1995b).

10.3.5 Assessment of upper gastrointestinal symptoms

Upper gastrointestinal symptoms were assessed by questionnaire at the patient’s first

visit before the gastric emptying measurement. “Gastric” (anorexia, nausea, early

satiation, abdominal bloating/fullness, vomiting, abdominal pain) and “oesophageal”

(dysphagia, heartburn and acid regurgitation) symptoms were graded as 0 = none, 1 =

mild (the symptom could be ignored), 2 = moderate (the symptom could not be

ignored, but did not influence daily activities) and 3 = severe (the symptom

Chapter 10

174

influenced daily activities). As there were nine symptoms, the maximum possible

total score was 27 (Jones et al. 1995b).

10.3.6 Measurement of blood glucose and serum C-peptide

concentrations

Venous blood samples (20 mL) were obtained immediately prior to the intravenous

infusion (t = -30 min) and then at –15, -2, 15, 30, 45, 60, 90 and 120 min. Blood

glucose concentrations were determined immediately using a portable blood glucose

meter (Medisense Companion 2 meter, Medisense Inc., Waltham, Massachusetts,

USA). The mean of the blood glucose concentrations at –2, 30, 60, 90 and 120 min

was calculated. Serum was stored at –70 C until analysis of C-peptide concentrations

by radioimmunoassay. Serum C-peptide was measured on blood samples obtained at t

= -30 min and t = 30 min; the latter time point was selected on the basis that steady-

state concentrations would be achieved within 60 min after commencement of the

infusion (Faber et al. 1978, Johansson et al. 1996).

10.3.7 Statistical Analysis

Data were evaluated using repeated measures analysis of variance (ANOVA) and are

shown as mean values ± SEM. Student’s t-tests were used to assess paired

comparisons. A P-value < 0.05 was considered significant in all analyses.

10.4 Results

The studies were well tolerated by all patients. The score for “gastric” symptoms was

1.4 ± 0.8, that for “oesophageal” symptoms 0.6 ± 0.3, and the total score was 2.0 ±

0.9. After the saline infusion, the total score for autonomic neuropathy was 2.0 ± 0.5.

Chapter 10

175

Three patients had unequivocal evidence of autonomic neuropathy (i.e. total score

≥ 3), two patients had no evidence of autonomic neuropathy and in the remaining

three, at least one result was borderline (i.e. total score ≥ 1 and < 3). After the C-

peptide infusion, there was no difference than after saline in autonomic function (total

score 1.8 ± 0.5).

Baseline and mean blood glucose concentrations did not differ between the study

days (t = -30 min; C-peptide: 9.1 ± 0.9 mmol/L vs. control: 9.3 ± 0.8 mmol/L; mean

blood glucose; 9.1 ± 1.0 mmol/L vs. 9.4 ± 0.8 mmol/L). On both days blood glucose

increased after the meal (e.g. from t = 2 45 min, P < 0.001 for both) and declined

from t = 45 120 min (P = 0.0001 for both).

Serum C-peptide levels rose (P = 0.001) from 0.3 ± 0.1 nmol/L at baseline to 1.4 ±

0.2 nmol/L at t = 30 min during C-peptide infusion. C-peptide levels remained low

during the saline infusion (baseline: 0.2 ± 0.0 nmol/L; at t = 30 min: 0.2 ± 0.0

nmol/L).

10.4.1 Gastric emptying

On both days, solid emptying approximated an overall linear pattern after an initial

lag phase, and liquid gastric emptying an overall monoexponential pattern, after a

short lag phase. The solid and liquid lag phases did not differ between study days

(solid; C-peptide: 13.0 ± 3.7 min vs. control: 16.8 ± 4.0 min; liquid; 1.1 ± 0.0 min vs.

1.3 ± 0.3 min). During saline infusion, three of the eight patients had delayed gastric

emptying. C-peptide had no effect on either solid or liquid gastric emptying (solid

T100; C-peptide: 39.6 ± 9.5 % vs. control: 35.5 ± 6.0 %; liquid T50; 27.0 ± 3.5 min

vs. 24.9 ± 3.2 min) (Figure 10.1).

Chapter 10

176

Figure 10.1: Gastric emptying of (a) solid (100 g minced beef) and (b) liquid (150 mL 10 % dextrose) meal components in 8 patients with type 1 diabetes mellitus. Data are mean values ± SEM.

10.5 Discussion

This preliminary study, which represents the first evaluation of the effects of C-

peptide on gastric emptying, indicates that acute administration of C-peptide has no

effect on gastric emptying of either solids or liquids in longstanding type 1 diabetes.

The dose of C-peptide (6 pmol/kg/min) that was evaluated in this study has been used

previously (Johansson et al. 1996) and resulted in physiological serum concentrations

comparable to those reported by others (Johansson et al. 1996). Johansson et al.

(1996) studied 12 type 1 patients in whom cardiovascular autonomic nerve function

was assessed before and during a 3-hour intravenous infusion of either C-peptide (6

pmol/kg/min) or saline (Johansson et al. 1996). C-peptide infusion was associated

with significant improvements in heart rate variability (from 13 ± 1 % to 20 ± 2 %)

and the brake index in response to tilting (from 4.6 ± 1.0 % to 10.3 ± 2.2 %). All the

patients studied by Johansson et al. (Johansson et al. 1996) had symptoms of

Chapter 10

177

peripheral neuropathy and 11 of the 12 had evidence of cardiovascular autonomic

neuropathy.

There is a higher prevalence of delayed gastric emptying in type 1 patients with

cardiovascular autonomic neuropathy than in those without (Buysschaert et al. 1987,

Horowitz et al. 1991, Kawagishi et al. 1997, Stacher et al. 2003, Ziegler et al. 1996).

However, the relationship between gastric emptying and autonomic neuropathy is not

strong (Buysschaert et al. 1987, Horowitz et al. 1991). This may potentially reflect

autonomic impairment isolated to the gastrointestinal tract (Hosking et al. 1975).

However, the outcome of histological studies of the myenteric plexus and abdominal

vagus nerve in human type 1 diabetes is inconsistent (Britland et al. 1990, Guy et al.

1984, He et al. 2001, Yoshida et al. 1988). There is evidence that acute changes in the

blood glucose concentration affect gastric emptying (Fraser et al. 1990, Schvarcz et

al. 1993, Schvarcz et al. 1997). Hyperglycaemia (~ 16 – 20 mmol/L) slows (Fraser et

al. 1990, Samsom et al. 1997), while hypoglycaemia (~ 2.5 mmol/L) accelerates

(Schvarcz et al . 1993), gastric emptying in type 1 diabetes. Blood glucose

concentrations that are within the normal postprandial range (~ 8 mmol/L) may also

slow gastric emptying (Schvarcz et al. 1997). Acute hyperglycaemia may also affect

autonomic nerve function (Lam et al. 1993, Yeap et al. 1996).

It should be recognised that a number of factors may potentially contribute to the

negative outcome of our study. The cohort studied (n = 8) was relatively small; while

a type 2 error cannot be excluded, there was no suggestion of any effect of C-peptide

on gastric emptying (the mean data are virtually superimposed) making this unlikely.

While only 3 of the 9 patients had definite autonomic neuropathy, autonomic tests

were “borderline” in a further 3. Hence, while we cannot exclude the possibility that

Chapter 10

178

an effect of C-peptide would only be evident in patients with cardiovascular

autonomic neuropathy this, again, appears unlikely. The observed prevalence of

delayed gastric emptying was as would be expected in the cohort studied (Horowitz et

al. 1991, Horowitz et al. 2002b); while it is possible that the effects of intravenous C-

peptide may only be evident in those patients with delayed gastric emptying, there

was no evidence to suggest this. While the blood glucose concentrations were

stabilised to below 12 mmol/L prior to commencement of each study, they were not

maintained in the euglycaemic range (mean ~ 9.3 mmol/L). Given the established

effect of acute hyperglycaemia, even at values within the normal postprandial range,

on gastric emptying (Fraser et al. 1990, Schvarcz et al. 1993, Schvarcz et al. 1997)

and attenuating the response to prokinetic drugs (Jones et al. 1999a), we cannot

exclude this factor, although in the positive study by Johansson et al. (Johansson et

al. 1996) blood glucose levels varied from 4.6 – 8.7 mmol/L during the

measurements. Finally, our study only evaluated the effect of acute administration of

C-peptide; an effect of chronic administration of C-peptide on gastric emptying

cannot be excluded.

Chapter 11

179

11Chapter 11CONCLUSIONS

Delayed gastric emptying occurs in 30 – 50 % of patients with longstanding diabetes

and may be associated with upper gastrointestinal symptoms such as nausea and

vomiting, impaired nutrient and drug absorption, and poor glycaemic control. The

studies presented in this thesis relate to normal and disordered gastrointestinal

motility in healthy humans and in patients with diabetes mellitus. A number of broad

areas have been addressed, and these include: (i) insights into normal gastric motor

function, (ii) novel methodological approaches in the quantification of gastric

emptying, (iii) investigations into the pathogenesis of diabetic gastroparesis, and (iv)

novel therapeutic approaches to the treatment of diabetic gastroparesis.

Chapter 11

180

A number of gastrointestinal peptides, including glucagon-like peptide-1 (GLP-1), are

released from the stomach and small intestine in response to the ingestion of

nutrients. Exogenous GLP-1 lowers fasting and postprandial glycaemia through

stimulation of insulin release and inhibition of glucagon secretion, and delays gastric

emptying, however, the physiological role of endogenous GLP-1 on gastric emptying

is less clear. In Chapter 5, the putative role of endogenous GLP-1 as an

enterogastrone in healthy subjects, is established. The effects of endogenous GLP-1,

using the intravenous GLP-1 receptor antagonist exendin(9-39) or saline, on gastric

emptying (scintigraphy) and glycaemia were measured in healthy subjects, following

ingestion of a mashed potato meal in double-blind, randomised, crossover design. In

healthy subjects, exendin(9-39) accelerated gastric emptying and increased

postprandial glycaemia and there was a relationship between the rise in blood glucose

and the gastric emptying rate. Although unavailable at the time of submission of this

thesis, analysis of plasma hormone data (GLP-1, GIP, insulin and glucagon) would

allow more definitive interpretation of the observations. It would also be of interest to

study the effects of exendin(9-39) on gastric emptying during euglycaemia using a

glucose clamp, as acute hyperglycaemia is known to slow gastric emptying and

attenuate the effects of prokinetic drugs. Furthermore, measurements of

antropyloroduodenal motility would have provided a more profound understanding of

the precise motor mechanisms attributable to the observed acceleration of gastric

emptying.

Gastric emptying is delayed in some 30 – 50 % of patients with longstanding type 1

and type 2 diabetes mellitus, and represents an important clinical problem. While

scintigraphy remains the ‘gold standard’ in the measurement of gastric emptying, it is

associated with a radiation burden. Three-dimensional ultrasonography has recently

Chapter 11

181

been demonstrated to provide a valid measure of gastric emptying in healthy subjects.

In Chapter 6 , two techniques (sc int igraphy and three-dimensional (3D)

ultrasonography) to assess gastric emptying of a high-nutrient glucose drink were

compared in patients with diabetic gastroparesis. Concurrent measurements of gastric

emptying of a high-nutrient glucose drink by 3D ultrasonography and scintigraphy

were compared in patients with known diabetic gastroparesis. There was good

correlation and agreement between the two techniques, suggesting that 3D

ultrasonography may provide a valid, and non-invasive, measure of gastric emptying

in patients with diabetic gastroparesis. The validity of both techniques is supported by

the observed correlation between the glycaemic responses to the drink with the rate of

gastric emptying.

Acute changes in the blood glucose concentration are known to influence the gastric

emptying rate. Hyperglycaemia slows, and hypoglycaemia accelerates, gastric

emptying in healthy subjects and in patients with uncomplicated type 1 diabetes. The

effects of hypoglycaemia in longstanding type 1 diabetes, however, are not known. In

Chapter 7, the effects of insulin-induced hypoglycaemia on gastric emptying in

longstanding type 1 diabetes is evaluated. Gastric emptying of a mixed solid/liquid

meal was measured by scintigraphy on two separate days, during hypoglycaemia and

euglycaemia. Hypoglycaemia accelerated gastric emptying of both solids and liquids

and the magnitude of this acceleration was greater when the rate of gastric emptying

during euglycaemia was slower. This is likely to be an important mechanism in the

counter-regulation of hypoglycaemia. Further studies to determine whether there is a

“threshold” at which hypoglycaemia accelerates gastric emptying, or whether the

response is continuous are now warranted. The mechanism(s) by which

hypoglycaemia accelerates gastric emptying, or the gastroduodenal motor correlates

Chapter 11

182

of this effect, were not evaluated in this study. Accordingly, studies to evaluate the

effects of insulin-induced hypoglycaemia on postprandial antropyloroduodenal

motility in both type 1 patients and healthy subjects are indicated.

The effects of glucose on gastric motility have been well characterised, however, the

effects of other monosaccharides, such as fructose, are less well understood. Fructose

is used widely in the diabetic diet and is known to empty from the stomach slightly

faster than glucose. Chapter 8 assesses the effects of intravenous fructose, glucose

a n d s a l i n e ( f o r 2 0 m i n ) on solid gastric emptying (scintigraphy) and

antropyloroduodenal motility in healthy males in randomised, placebo-controlled,

crossover design. Intravenous glucose and fructose both slowed gastric emptying

without any significant difference between them, although there was a trend for faster

gastric emptying with fructose compared with glucose at 45 min. Immediately after

meal ingestion (t = 0 – 30 min): there were more isolated pyloric pressure waves

during glucose infusion, compared with fructose or saline; the number of antral

pressure waves was attenuated by infusion of both glucose and fructose, but not

saline; and the number of duodenal waves increased during saline and glucose, but

not fructose (although the magnitude of the increase was greater for saline compared

with glucose). Fructose induces a motor pattern in healthy subjects with suppression

of antral and duodenal waves, and slowing of gastric emptying; the magnitude of this

effect appears comparable to glucose. A limitation of this study was that only six

healthy subjects were studied. Further studies to evaluate the comparative effects of

fructose, glucose and saline on gastric emptying and antropyloroduodenal motility in

a larger cohort of subjects, and patients with diabetes are warranted.

Chapter 11

183

Prokinetic agents form the mainstay of treatment in patients with diabetic

gastroparesis, however, all have significant limitations. Itopride has been shown to be

effective in improving symptoms in functional dyspepsia, and while there is only

limited evidence that itopride is an effective prokinetic in patients with diabetes, there

have, hitherto, been no randomised, placebo-controlled trials. In Chapter 9, the effects

of itopride (200 mg) versus placebo (three times daily for 7 days) on mixed

solid/liquid gastric emptying (scintigraphy), glycaemia and upper gastrointestinal

symptoms were evaluated in patients with longstanding type 1 and 2 diabetes mellitus

in double-blind, placebo-controlled, randomised, crossover design. There was a trend

for itopride to accelerate both solid and liquid gastric emptying. Itopride accelerated

liquid, but not solid, gastric emptying in the cohort (48 %) with delayed solid and/or

liquid gastric emptying on placebo. The magnitude of the improvement in gastric

emptying of liquids and solids was inversely related to gastric emptying on placebo.

Thus, itopride accelerates gastric emptying of liquids, and possibly solids, in diabetic

gastroparesis, however, the magnitude of this appears to be relatively modest and

dependent on the basal rate of gastric emptying. Given the positive, albeit modest,

effect of itopride on gastric emptying in the cohort with gastroparesis, the use of a

larger sample size may well have detected a significant difference in gastric emptying

on itopride compared with placebo, and further studies are warranted. Moreover,

further evaluation of the effects of different doses of itopride on gastric emptying in

patients with diabetic gastroparesis, particularly those with markedly delayed gastric

emptying are indicated. Future studies that employ glucose clamps to stabilise the

blood glucose concentration within the euglycaemic range, also, are warranted.

It is well recognised that delayed gastric emptying occurs more frequently in patients

with autonomic neuropathy. There is now evidence to suggest that C-peptide

Chapter 11

184

improves autonomic nerve function in patients with type 1 diabetes mellitus. In

Chapter 10, gastric emptying of a mixed solid/liquid meal was measured by

scintigraphy on two separate days (during intravenous C-peptide and saline infusion)

in a single-blind, placebo-controlled, randomised, crossover design. Autonomic nerve

function was assessed at the end of each gastric emptying measurement using three

standardised cardiovascular reflex tests. Acute intravenous administration of C-

peptide, resulting in physiological serum concentrations, had no effect on either solid

or liquid gastric emptying in patients with longstanding type 1 diabetes mellitus. Six

of the 8 patients had evidence of autonomic nervous dysfunction, however, C-peptide

had no effect on the total score for autonomic neuropathy. Since the cohort studied

was relatively small, it would be of interest to evaluate the effects of C-peptide in a

larger sample size. Given the established effect of acute hyperglycaemia on gastric

emptying and attenuating the response to prokinetic drugs, a future study under

conditions of euglycaemia via the use of a glucose clamp, is indicated. While the

observations relate to the effect of acute administration of C-peptide, the chronic

effects of C-peptide on gastric emptying are yet to be elucidated, and further studies

are warranted.

Gastroparesis, particularly in patients with diabetes, represents an important clinical

problem. The studies presented in this thesis have provided fundamental insights into

the patterns, determinants and measurement of normal and disordered gastric motor

function and postprandial glycaemia in healthy humans and patients with diabetes,

and novel therapies to more effectively treat gastroparesis have been explored. Future

studies to further assess the complex pathogenesis and pathophysiology of

gastroparesis, and which include larger cohorts of patients, are now warranted.