long-term clinical effects of a peritoneal dialysis fluid with less glucose degradation products

Kidney International, Vol. 59 (2001), pp. 348–357

Long-term clinical effects of a peritoneal dialysis fluid withless glucose degradation products

BENGT RIPPE, OLE SIMONSEN, OLLE HEIMBURGER, ANDERS CHRISTENSSON, BORJE HARALDSSON,GUNNAR STELIN, LARS WEISS, FINN-DAVID NIELSEN, SUSANNE BRO, MICHAEL FRIEDBERG,and ANDERS WIESLANDER

University Hospital of Lund, Lund, Huddinge Hospital, Stockholm, Malmo University Hospital and Sahlgrenska UniversityHospital, Gothenburg, Lanssjukhuset Halmstad, Centralsjukhuset Karlstad, and Centrallasarettet, Boras, Sweden; RigshospitaletCopenhagen and Hvidovre Hospital, Copenhagen, Denmark; and Gambro AB R & D, Lund, Sweden

observed in the overnight effluent during treatment with theLong-term clinical effects of a peritoneal dialysis fluid with lessnew fluid.glucose degradation products.

Conclusions. We conclude that the new fluid with a higherBackground. Glucose degradation products (GDPs) are cy-pH and less GDPs is safe and easy to use and has no negativetotoxic in vitro and potentially toxic in vivo during peritonealeffects on either the frequency of peritonitis or peritoneal trans-dialysis (PD). We are presenting the results of a two-yearport characteristics as compared with conventional ones. Ourrandomized clinical trial of a new PD fluid, produced in aresults indicate that the new solution causes less mesothelialtwo-compartment bag and designed to minimize heat-inducedand interstitial damage than conventional ones; that is, it mayglucose degradation while producing a near neutral pH. Thebe considered more biocompatible than a number of conven-effects of the new fluid over two years of treatment on mem-tional PD solutions currently in use.brane transport characteristics, ultrafiltration (UF) capacity,

and effluent markers of peritoneal membrane integrity wereinvestigated and compared with those obtained during treat-ment with a standard solution. Conventional fluids for peritoneal dialysis (PD) areDesign. A two-group parallel design with 80 continuous am-

potentially bioincompatible because of their hyperosmo-bulatory peritoneal dialysis patients was used. The patientslality, low pH, and high lactate and glucose concentra-were randomly assigned to either the new fluid (N 5 40) or

to a conventional one (N 5 40), and were stratified with respect tions [1, 2]. Especially the high glucose concentrationsto age, diabetes, and time on PD. Peritoneal transport charac- in conventional fluids have been incriminated as beingteristics were assessed by the Personal Dialysis Capacity (PDCe) detrimental to the peritoneum, not the least by providingtest at 1, 6, 12, 18, and 24 months after inclusion and by weighing

the substrate for nonenzymatic glycosylation of tissuethe overnight bag daily. Infusion pain and handling were evalu-proteins [2, 3]. Furthermore, during heat sterilization,ated using a questionnaire. Peritoneal mesothelial and intersti-

tial integrity were evaluated by analyzing overnight effluent some of the glucose in conventional solutions is affecteddialysate concentrations of CA 125, hyaluronan (HA), pro- by the added heat and converted to fructose throughcollagen-1-C-terminal peptide (PICP), and procollagen-3-N- isomerization and degraded to biologically reactive sub-terminal peptide (PIIINP) at 1, 6, 12, 18, and 24 months.

stances [4–7]. Chemically identified glucose degradationResults. The handling of the new two-compartment bag wasproducts (GDPs) commonly found in PD fluids are methyl-considered easy, and there were no indications of increased

discomfort with the new system. Furthermore, no changes in glyoxal, glyoxal, formaldehyde, 3-deoxyglucosone (3-DG),peritoneal fluid or solute transport characteristics were ob- acetaldehyde, 2-furaldehyde, and 5-hydroxymethyl-fur-served during the study period for either fluid, and neither aldehyde [6, 7].were there any differences with regard to peritonitis incidence.

These substances have been demonstrated to inhibitHowever, significantly higher dialysate CA 125 (73 6 41 vs.cell proliferation and to cause necrosis of fibroblasts,25 6 18 U/mL), PICP (387 6 163 vs. 244 6 81 ng/mL), and

PIIINP (50 6 24 vs. 29 6 13 ng/mL) and significantly lower macrophages, and human mesothelial cells in vitro [8–13].concentrations of HA (395 6 185 vs. 530 6 298 ng/mL) were There are indeed indications that GDPs, by causing pro-

tein denaturation or cross-linking, are basically cytotoxic[14]. More specific detrimental effects of GDP in PDKey words: peritonitis, interstitial damage, dialysate solutions, biocom-

patible PD solution. fluids in vitro include impairment of cytokine releaseand superoxide radical generation by peritoneal macro-Received for publication December 6, 1999phages and by blood mononuclear and polymorphonu-and in revised form July 10, 2000

Accepted for publication July 25, 2000 clear cells [9, 15]. In animal models, GDPs also seem tocause impairment of leukocyte rolling and an increased 2001 by the International Society of Nephrology

348

Rippe et al: Glucose degradation products 349

venular flow velocity [16], even though in acute rat exper- fluids are demonstrated in Table 1. To ensure a balancedcomposition of the patient material in each group, aiments no negative effects of GDP were observed with

respect to peritoneal solute transport or ultrafiltration stratified randomization was performed with respect topatient age (,55, .55 years), diabetes (using insulin or(UF) [17]. Clinical studies indicate that GDP, at least in

combination with a low solution pH, may bring about not), and time on PD (,9 months, .9 months; Table 2).Initially, the investigation was intended as a one-yearinfusion pain and impaired UF capacity [18, 19].

It has been demonstrated that GDPs are very potent clinical study, but it was later prolonged to include astudy period of two years, without adding any new pa-agents of nonenzymatic cross-linking of proteins, leading

to the formation of advanced glycation end products tients.(AGEs). In particular, dicarbonyl compounds such as

Study population3-DG and methylglyoxal participate in these reactions[7, 20–22]. A recent report has demonstrated extremely Patients were recruited if they were at least 18 years

old, gave their fully informed consent, and were able tohigh concentrations of 3-DG in conventional PD fluids[7]. The role of GDP in AGE formation is of great rele- use 2 L bags with a calcium concentration of 1.35 mmol/L.

Patients were excluded from the study if they had avance, since AGEs are considered to participate in theremodeling and fibrosis of the peritoneal membrane that positive screen for hepatitis B or HIV, or had an active

malignancy or were pregnant. Most of the patients (58%)may occur during long-term continuous ambulatory peri-toneal dialysis (CAPD) treatment and that may be linked previously had not been treated on CAPD. Once found

eligible, patients were reported to a central randomiza-to an increased risk of UF failure [23, 24].To avoid the generation of GDP, the PD fluid may tion office (at Gambro, Lund, Sweden) and were as-

signed to one of the two groups or strata (according tobe sterilized by filtration instead of being autoclaved.Some generation of GDP will still occur during storage, age, diabetes, and time on PD) using a premade random-

ization schedule. The mean age at entry was 57 yearshowever. Another way to avoid the generation of GDPis to dispense the fluid into a two-compartment bag, where for the control group and 58 years for the study group.

Diabetic patients represented 43% of the control groupconcentrated glucose is kept at a low pH (in a small com-partment), separate from the electrolyte-buffer solution and 35% of the study group, respectively (Table 2).(in a larger compartment). When this system is subjected

Study procedureto conventional heat sterilization and subsequent stor-age, the generation of GDP will be very limited. Before Patients attended the participating hospital on five

occasions during the study period, that is, at 1, 6, 12, 18,use, the two compartments, however, have to be mixed.This is the concept of the test solution in the present and 24 months after inclusion. Study parameters were

sampled within two weeks from the scheduled point ofstudy. Toxicological and chemical analyses of a fluidproduced and delivered in this way revealed that, besides time. Sampling did not take place within an episode of

peritonitis, but was postponed for four weeks after fullhaving a pH close to neutral, it showed very low levelsof GDP, essentially below the detection limit for 3-DG, recovery from peritoneal infection.formaldehyde, and acetaldehyde [25]. With this composi-

Assessment of peritoneal membranetion, the solution’s in vitro cytotoxicity as well as itsexchange characteristicsability to induce AGE formation was clearly reduced

[3, 25]. The aim of our current study was to investigate Personal Dialysis Capacity (PDCe) was used as a toolto assess the peritoneal transport characteristics [26].the long-term impact of this new fluid of PD on perito-

neal solute and fluid transport characteristics, as well The PDC procedure involves urine, blood, and dialysatesampling during a standardized CAPD day carried outas on some tentative markers of peritoneal membrane

integrity. according to a special exchange schedule. The programis based on the three-pore model of peritoneal transport[27, 28] and describes peritoneal exchange using the fol-

METHODSlowing parameters: (1) area parameter (A0/Dx), determin-

Study description ing the diffusion capacity of small solutes, and indirectly,the hydraulic conductance of the membrane (LpS); (2)The study had a prospective, randomized, parallel de-

sign, including 80 patients randomly assigned to either the reabsorption rate of fluid from the peritoneal cavityto the blood after peak time, when the glucose gradientGambrosol 40 CAPD solution [control group (C); N 5

40] or to the new, differently produced and nearly neutral has largely dissipated (JvAR); and (3) the large pore fluidflow (JvL), which determines the loss of proteins to thefluid that was low in GDP [study group (S); N 5 40].

Both groups were using the Gemini 10 design (with drain PD fluid. Residual renal function (RRF) was also evalu-ated in the PDC procedure and standardized to standardbag) or a slight modification thereof. The concentrations

of GDP, pH, and the in vitro cytotoxicity of the different body surface area (1.73 m2).

Rippe et al: Glucose degradation products350

Table 1. Chemical and biological characteristics of the two different fluids tested in the clinical study

Control group Study groupOne compartment Two compartments

Characteristics(contents of fluid bag) 1.5% 4% 1.5% 4%

pH 5.4360.01 5.3060.01 6.4360.01 6.2660.01Cytotoxicity % 52.760.3 75.360.3 3.061.5 20.761.2UV-abs 228 nm 0.5260.00 0.8360.02 0.1360.00 0.2860.03UV-abs 284 nm 0.2060.00 0.4960.01 0.1960.00 0.4860.03Acetaldehyde lmol/L 22665 26462 ,1 160Formaldehyde lmol/L 4.660.1 9.160.4 ,1.7 ,1.7Methylglyoxal lmol/L 22.760.5 33.361.4 ,2.8 ,2.8Glyoxal lmol/L 5.560.3 7.560.1 5.860.3 6.660.35-HMF lmol/L 2.260.1 6.960.1 8.360.2 22.561.42-Furaldehyde lmol/L ,0.2 0.660.2 ,0.2 ,0.23-Deoxyglucosone 11865 413619 12.361.2 65610

Each sample represents the mean value from three separate bags 6 SEM. Data are transcribed from references 30 and 11. Abbreviations are: abs, absorption,5-HMF, 5-hydroxymethyl-furaldehyde.

Table 2. Characteristics of the study patients or no?” then the degree of the pain sensation was as-sessed by a nine-point visual analog scale (VAS).Characteristics Control group (N 5 40) Study group (N 5 40)

The breaking of the frangible pin and the mixing ofAge (years)a 57 (26–82) 58 (28–80)Malesb 30 (75%) 25 (63%) the two compartments of the new bag were each evalu-Diabeticsb 17 (43%) 14 (35%) ated by four grade scales using scores from 1 to 4 (1,Time on CAPDc

simple; 4, difficult).0–9 months 34 35.9 months 6 5

Participating centersa Mean age (range)b Number of patients (percent) The patients were recruited from nine participatingc Number of patients in different categories (percent)

centers. The number of study patients (S) and controlpatients (C) is given within parenthesis: University Hos-pital of Lund (7 S; 8 C), Malmo University Hospital (4 S;3 C), Sahlgrenska University Hospital, Gothenburg (12 S;Ultrafiltration was evaluated during the study by daily

registration by the patient of the weight of the overnight 9 C), Lanssjukhuset Halmstad (0 S; 1 C), CentralsjukhusetKarlstad (1 S; 3 C), Centrallasarettet Boras (2 S; 1 C),bag using an electronic scale. The concentration of glu-

cose in the night bag was 2.5%. Time for start and end Huddinge Hospital, Stockholm (3 S; 6 C; all in Sweden),Rigshospitalet Copenhagen (4 S; 1 C), Hvidovre Hospi-of the overnight dwell was recorded, and the dwell time

was set at 10 hours. tal, Copenhagen (7 S; 8 C; in Denmark). The study proto-col was approved by the ethics committee of each partici-

Dialysate markers pating center.Samples of spent dialysate from the overnight bag were

Statistical analysiscollected and stored without any further treatment andfrozen at 2708C for later analysis. Analysis was performed Values are presented as box plots in the figures, with

the largest and smallest observed values (if not an outlierusing commercially available kits for CA 125 [enzyme-linked immunosorbent assay (ELISA); Boehringer Mann- or extreme) as bars on the top or at the bottom of the

box. The median value is indicated within the box, eachheim, Mannheim, Germany], hyaluronan (HA; RIA;Pharmacia, Uppsala, Sweden), procollagen-1-C-terminal box including 50% of the values. Values of more than

1.5 box lengths from the 25th percentile are presentedpeptide (PICP). and procollagen-3-N-terminal peptide(PIINP; RIA from Orion Diagnostics, Espoo, Finland). as outliers (o), and values of more than 3 box lengths

from the 25th percentile, are presented as extremes (*).A separate study showed that the level of CA 125 wasunaffected by not subjecting the PD fluid to centrifugation The dropout rate as well as the peritonitis incidence

(with respect to first episode only) in each treatmentbefore freezing and storage (unpublished observation).group was compared by the log rank test [29]. In addition,

Assessment of infusion pain and handling the Mann–Whitney test was employed to compare thedistribution of the patient-specific numbers of peritonitisThe patients completed a questionnaire covering sub-

jective symptoms in connection with infusion of the PD episodes during the one-year follow-up. Peritonitis inci-dence, with respect to all episodes, was calculated andsolution. If the answer was yes on the first question of

“Do you feel any discomfort or pain during infusion, yes compared according to Rothman [30]. Infusion pain was


Table 3. Distribution of the patients with respect to number ofperitonitis episodes

Control group Study groupNumber of episodes (N 5 40) (N 5 40)

0 26 271 12 82 1 13 1 14 0 3

7 dropped out within the first month (4 in the controlgroup and 3 in the study group), and 16 chose (for per-sonal reasons) not to continue a second year. Fifty-twopatients completed the first year, and 13 patients com-pleted both years in the study, which represents a drop-out rate during the first year of 35% (including the 7patients who dropped out at the very start) and a dropoutrate during the second year of 64% (75%, including the16 patients who chose not to continue a second year).Except for the 7 initial dropouts and the 16 patients whofor various reasons decided not to continue the study,Fig. 1. Proportion of patients in the study over time, illustrating the

dropout rate. There was no significant difference between the group- the major dropout causes were transfer to hemodialysisspecific functions (P 5 0.25; log-rank test). The darker line is the study (N 5 6), transfer to automatic peritoneal dialysis (APD)group and the lighter line is the control group.

or 2.5 L bags (N 5 8), transplantation (N 5 9), death(N 5 7), or various other reasons (N 5 14). No significantdifference was observed in the dropout frequency be-

compared between the treatment groups using Fisher’s tween the control group or the study group (P 5 0.25;exact test as well as the Mann–Whitney test. log-rank test).

From the daily assessments of the weight of the over-Peritonitisnight bags, mean values were calculated for each subject

for defined time intervals. The “initial” mean was based Table 3 demonstrates the distribution of patients withon data from weeks two through five of the study period, respect to number of peritonitis episodes. No significantand the following mean values were based on data col- difference with respect to number of episodes or timelected during each of the following consecutive five-week to the first peritonitis was observed (P 5 0.9 and P 5periods. Subgroup analyses were performed with respect 0.35, respectively). The peritonitis incidences were 17to diabetes and time on PD before inclusion (0 vs. .0 out of 36.41 years (0.467 per year) and 25 out of 45.07month). The Kruskal–Wallis test was used to compare years (0.555 per year) in control and study groups, re-treatment subgroups. The Mann–Whitney test was used spectively (P 5 0.6). Total peritonitis incidence for theto compare overnight drained weights, area parameters, study was 1 out of 23.5 months.and dialysate markers for the treatment groups with re-

Peritoneal transport characteristicsspect to initial values and (patient specific) changes dur-ing the study. A P value of less than 0.05 is referred to In Figure 2, the normalized “unrestricted pore areaas significant. Bonferroni’s correction was applied when- over unit diffusion path length” (“area parameter” forever multiple comparisons were performed, and the P small solute transport; A0/Dx) is shown as a functionvalue obtained was less than 0.05. Exact P values were of study time in each group. There were no significantcalculated (StatXact 3; Cytel Software Corporation, differences among control and study groups initiallyCambridge, MA, USA) in situations in which the num- (after 1 month) nor in assessments at months 6, 12, 18,bers of observations were small. All P values reported and 24, respectively. This also holds true for the hydraulicare two tailed. conductance (LpS), the final reabsorption rate (JvAR),

the large pore fluid flux (JvL), and the renal residualfunction (data not shown).RESULTS

Dropout rates Overnight bag weightThe overnight bag weight, assessed as a function ofThe dropout rate in the study is demonstrated in Fig-

ure 1. Out of the 80 patients randomized to the study, treatment time in each group, is shown for consecutive


Fig. 2. Box plot of normalized (%) area pa-rameter (A 0 /Dx) as a function of treatmenttime in the (A) control group and (B) studygroup. Each box includes 50% of all measuredvalues, with the largest and smallest observedvalue as bars on the box top or at the boxbottom and the median value indicated withinthe box. Values of more than 1.5 box lengthsfrom the 25th percentile are presented as out-liers (o), and values of more than three boxlengths from the 25th percentile are presentedas extremes (*). N is the number of patients.There were no significant differences amongstudy and control groups at any time withrespect to A 0 /Dx.

five-week periods in Figure 3. There were no significant cantly higher in the study group (P , 0.001), and HAwas significantly lower (P , 0.04). The use of individualdifferences whatsoever between the groups (P 5 0.21).

Moreover, the (patient specific) changes from the “ini- mean values was justified by the observation that thepatient-specific values did not change noticeably duringtial” value (mean value of weeks 2 through 5) to each

of the following five-week periods did not differ signifi- months 1 through 24.cantly among the groups.

Infusion pain and handlingDialysate markers The relative as well as the total number of patients

who experienced pain or discomfort during infusion de-Figures 4 and 5 demonstrate the levels of CA 125 andHA among the two groups, respectively, at months 1, 6, creased from 31% (21 patients) at 1 month and 33% (19

patients) at 6 months to 13% (6 patients) at 12 months,12, 18, and 24. CA 125 levels were highly significantlyincreased for the study group at all points of measure- to 13% (2 patients) at 18 months, and finally, to 20%

(2 patients) at 24 months. Since a sufficient number ofment, whereas HA showed significantly lower valuesat month 12, and reductions that were of borderline patients reported pain at one and six months, these data

were chosen for further evaluation. At one month of thesignificance at months 1 and 18 (P 5 0.05). At 24 months(and partly at 18 months), however, no powerful statisti- study, 40% (12 of 30) of the patients on conventional

fluid and 24% (9 of 37) of the patients on the new fluidcal analysis could be applied because of the low numberof patients. The levels of PICP and PIIINP were signifi- noted some kind of pain or discomfort. After six months,

these figures were 46% (11 of 24) and 24% (8 of 34),cantly higher in the test group at months 1, 6, 12, 18,and 24. Thus, for PICP, P equaled 0.001 at 1 month, P , respectively. Although the new fluid tended to yield

fewer patients with discomfort or pain, this was not found0.001 at 6 months, P , 0.001 at 12 months, P 5 0.02 at18 months, and P 5 0.9 at 24 months. For PIIINP, P to be statistically significant (P 5 0.19 and P 5 0.09,

respectively, using Fisher’s exact test for the dichoto-was 0.002 at 1 month, P , 0.001 at 6 months, P , 0.001at 12 months, P 5 0.3 at 18 months, and P 5 0.07 at mous pain variable). For the patients reporting pain, no

significant differences were seen in the distribution of24 months. A power calculation demonstrated that aminimum of 14 patients was needed in each group at the degree of pain using the nine-point scale rate (at 1

month P 5 0.12, at 6 months P 5 0.13, Mann–Whitneyany time to reveal a 20% significant difference (P ,0.05; 80% power). test).

Evaluation of the study questionnaire indicated thatTo use all of the collected data in the statistical evalua-tion, mean values for all the individual patient marker the handling of the new bag was easy. On a four-grade

scale with scores from 1 to 4 (1, simple; 4, difficult), thedata for months 1 through 24 were calculated and com-pared (Table 4). CA 125, PICP, and PIIINP were signifi- new fluid scored a mean value (declining over time)


Fig. 3. Box plots of drained weight in the overnight bag (2.5% glucose concentration) as a function of treatment time in the (A) control and (B)study groups. N is the number of patients. A patient had a valid weekly observation, if at least one value was reported during each week. Thenumber of valid observations decreased more markedly in the control group than in the study group. No statistical difference among control andstudy groups was obtained.

Fig. 4. Box plot of the concentration of CA125 (U/mL) in overnight peritoneal effluentfor the (A) control and (B) study groups. Thestudy group showed significantly higher valuesat all points of measurement (P , 0.001 at 1month, P , 0.001 at 6 months, P 5 0.001 at12 months, P 5 0.003 at 18 months, and P 50.007 at 24 months).


Fig. 5. Box plot of the concentration of hya-luronan (HA) (ng/mL) in overnight peritonealeffluent for the (A) control and (B) studygroups. The study group showed lower (ormarginally lower) values at 1, 12, and 18months. (P 5 0.05 at 1 month, P 5 0.6 at 6months, P 5 0.04 at 12 months, P 5 0.05 at18 months, and P 5 0.26 at 24 months). Over-all, the HA concentration was lower in thestudy group (Table 4).

Table 4. Dialysate concentration of CA 125, HA, that a new glucose-based solution of more or less identi-PICP and PIIINP

cal composition as conventional ones, but having a nearControl group Study group neutral pH and being largely free of GDP, significantly

Dialysate marker N 5 34 N 5 36 affects some effluent markers of peritoneal membraneCA 125 U/mL 25618 73641 integrity. Thus, CA 125, PICP, and PIIINP were foundHA ng/mL 5306298 3956185

to be significantly higher and HA lower with the newPICP ng/mL 244681 3876163PIIINP ng/mL 29613 50624 fluid after one month and during up to 18 months of

treatment. However, there were no measurable effectsThe individual patient mean values for months 1 to 24 were calculated foreach treatment group. Data are shown as mean 6 SD. CA 125, PICP and PIIINP on peritoneal exchange parameters, including UF, withvalues were higher in the study group (P , 0.001) and HA was lower (P , 0.04).

either the new or with the conventional control solutionduring the two years of observation. A much longerstudy period would have been required to reveal any

between 1.1 and 1.4 (frangible pin) and 1.0 to 1.2 (fluid changes in these parameters, since substantial increasesmixing). The study represents approximately 257 treat- in small solute (glucose) transport and reductions in UFment months and close to 65,000 exchanges with the new tend to first appear—if at all—after three or four yearsbag. Thus, the handling of the new bag can be considered of treatment using conventional solutions [32].easy and safe. Mesothelial cells lining the entire peritoneum produce

lubricants and surface tension-lowering substances, suchSubgroup analysesas phospholipids, but also fibrinolytic and antithrombo-Subgroup analyses did not reveal any differencesgenic substances, as well as cytokines and chemokines,among diabetics versus nondiabetics or among newlyand a number of other compounds, including HA, CArecruited patients versus patients who had been on PD125, and collagen peptides [33, 34]. In this context, CApreviously with regard to any of the parameters discussed125 and phospholipids can be regarded as more or lesspreviously in this article. Among the first-year dropouts,specific for the mesothelium, whereas the other com-21 patients completed at least the first month of thepounds also are produced by other cells in the perito-study, and of these, 13 belonged to the control groupneum. In fact, CA 125 effluent levels have been claimedand 8 to the study group. In neither group the dropoutsto reflect the mass of viable mesothelial cells in stablediffered significantly with respect to A0/Dx (transportPD patients [35–37]. During PD, the mesothelial cellstype) or CA 125 dialysate level from the non-dropoutsusually become cubic and multinucleate, tend to loseof the corresponding group.their microvilli, and express more rough endoplasmaticreticulum, at least over longer observation periods [38].

DISCUSSION The expression of more rough endoplasmatic reticulumshould, at least partly, reflect an increased mesothelialThe present study supports and extends the results of a

previous preliminary crossover study [31], demonstrating metabolic activity in response to the continuous loss of


secretion products to the dialysate that occurs in PD rise in CA 125 may reflect a rapid regeneration of themesothelium during treatment with the new solution. Ac-[33]. Thus, the mesothelium is apparently “up-regulated”

to produce more secretion during CAPD in response to tually, ex vivo studies have recently demonstrated thatmesothelial cells (from the overnight effluent) grow tothe repeated exchanges of dialysate. Conceivably, up-

regulation of the mesothelium in this respect would be confluence in culture in a shorter time when harvestedfrom patients who have been on PD fluids with less GDPmore pronounced for a more physiological solution than

for a conventional one. Hence, the production of phos- than when harvested from patients previously exposed toconventional solutions (abstract; Jarkelid et al, Perit Dialpholipids, CA 125, collagen peptides, etc., can be ex-

pected to be higher with the new, more physiological Int 19:S54, 1999). These findings indicate that the mesothe-lial cells seem more viable when exposed to the new fluid.PD fluid. On the other hand, inflammatory stimulation

of the mesothelium using proinflammatory cytokines has Hyaluronan is a major component of the interstitialground substance in most tissues. It is secreted into thenot been found to affect the release of CA 125 above

baseline levels, in contrast to the situation for HA (dis- peritoneal cavity by fibroblasts in the interstitium andlocally by the mesothelial cells [41]. HA is involved incussed later in this article) [35].

Mesothelial cells are frequently lost during the course several physiological processes, such as tissue repair andremodeling, and also in primary morphogenesis [42]. Theof CAPD. Thus, along with an impaired remesotheliali-

zation, a frequent finding in long-term PD, partial meso- local HA production by tissue fibroblasts (and by themesothelium) increases very markedly, for example, dur-thelial denudation occurs [38]. Impaired survival of meso-

thelial cells might thus serve as an early indication of a ing peritonitis [41]. On the other hand, in edema statesor during conditions of high fluid flows through an unin-more or less severe peritoneal tissue damage or irritation,

eventually terminating in fibrosis. Such changes seem flamed interstitium, HA is easily washed out from thetissue via the lymphatics [43]. The fraction of HA that isto occur during long-term treatment concomitant with

microvascular alterations, such as capillary neovasculari- produced by mesothelial cells forms a mesothelial surfacecoat together with other glycosaminoglycans (GAGs) andzation, and possibly, increases in vascular permeability.

Compared with the capillary barrier, mesothelial cells and phospholipids [33, 42]. The continuous “rinsing” of theperitoneal membrane by conventional PD solutions mostinterstitial tissue are, however, unlikely to play an impor-

tant role in the peritoneal transport of fluid and solutes likely causes mesothelial HA production to increaseslightly, since the concentration of HA is known to be[2, 27, 28, 36, 39]. Thus, early mesothelial and interstitial

changes would not be reflected by alterations in peritoneal considerably lower in fluid obtained at the first washoutin patients entering CAPD [41]. However, the markedtransport parameters [36].

Longitudinal examination of the dialysate during the increases in HA production by both fibroblasts andmesothelial cells that occurs in peritonitis reflect in-course of CAPD, in accordance with the morphological

changes, has demonstrated declining CA 125 dialysate flammatory stimulation. In fact, HA seems to act as aperipheral inflammation marker [42]. In this perspective,levels with increasing treatment time [37]. Our results of

higher dialysate CA 125 levels during the use of the new the lower HA production in the test group is consistentwith a more physiological situation for the new solution.fluid strongly indicate that the mesothelial cells are not

as negatively affected throughout the treatment as by con- Procollagen-1-C-terminal peptide and PIIINP are re-leased during the synthesis of collagen I and collagen III,ventional PD fluids. Elevations of dialysate in CA 125

were observed early, that is, after a treatment period of respectively, and are used as markers for overall collagensynthesis [33, 34, 44]. Dialysate concentrations reflectonly one month [40], which was similar to observations

in our previous cross-over study [31]. Here, significant locally produced PICP and PIIINP by fibroblasts andmesothelial cells [33, 44]. Our results of a higher concen-increases in CA 125 were observed as soon as 6 and 12

weeks of treatment for a solution largely free of GDP. tration of PICP and PIIINP for the new fluid most proba-bly reflect an improved cell function of the mesotheliumFurthermore, the CA 125 levels again dropped to low

values after switching the patients back to conventional and thus more collagen release to the effluent. The datamay, of course, also indicate an ongoing fibrosis with adialysis fluids [31, 40]. Whether these changes reflect alter-

ations of the mesothelial cell mass, or alternatively, were higher collagen turnover. In this context, however, it isnoticeable that Graff et al did not show a correlationdue to changes in CA 125 production per cell is not known.

One should keep in mind that approximately 40% of our between the production of collagen peptides and long-term CAPD or peritonitis incidence, conditions thatpatients had been previously exposed to conventional PD

fluids. This group of patients also showed an elevation of would be expected to be associated with more peritonealfibrosis [44].CA 125 during treatment with the new fluid; that is, they

did not differ from the group of patients who had never The mechanism by which GDP alone or in combina-tion with high glucose concentrations (and other toxicbeen exposed to PD fluids before. Thus, if mesothelial

cell mass is positively influenced by the new fluid, the influences) affects the mesothelial cells remains to be


Hospital Gothenburg, Sweden), and Christina Landgren, Christinaelucidated. One possibility is that there is a direct toxicSchonborg, Anette Larsen, Helen Andersson, Annika Larsson, Helle

action of GDP through carbonyl-induced cross-linking Larsson, Thierry Crost, Eva Svensson, Lena Jarkelid, Per Kjellstrand,of surface proteins and of intracellular proteins or nucleic Mary-Ann Ernstsson, and Lena Olsson (Gambro AB). We also thank

Ulf Stromberg for statistical analysis and Kerstin Wihlborg for skillfulacids. For instance, formaldehyde, a well-known fixative,typing of the manuscript.easily penetrates the cell membrane and immediately

cross-links several intracellular components. Further- Reprint requests to Dr. Bengt Rippe, Department of Nephrology,University Hospital of Lund, S-221 85 Lund, Sweden.more, GDP can promote AGE formation in an environ-E-mail: [email protected] of high glucose levels, which is of great relevance

since AGE seem to reduce the viability of mesothelialREFERENCEScells [45, 46]. It is conceivable that the major detrimental

1. Holmes CJ: Biocompatibility of peritoneal dialysis solutions. Peritaction of GDP over longer periods of PD treatment isDial Int 13:88–94, 1993linked to such a promotion of AGE formation in the

2. Krediet RT, Zweers MM, Ho-dac-Pannekeet MM, et al: Thepresence of high glucose levels [23, 46]. effect of various dialysis solutions on peritoneal membrane viabil-

ity. Perit Dial Int 19(Suppl 2):S257–S266, 1999One cannot rule out the possibility that the higher pH3. Dawnay A, Wieslander AP, Millar DJ: Role of glucose degrada-in the new fluid is a major factor in improving its biocom-

tion products in the generation of characteristic AGE fluorescencepatibility, at least in the short term. Actually, in a recent in peritoneal dialysis fluid. Proceedings of the 6th International

Symposium on the Maillard Reaction, 1997, pp 332–337rat study, GDP-free solutions seemed to reduce the ini-4. Kjellstrand P, Martinson E, Wieslander A, et al: Developmenttial peritoneal vasodilatation significantly, and hence, the

of toxic degradation products during heat-sterilization of glucose-initial inflation of peritoneal glucose transport produced containing fluids for peritoneal dialysis: Influence of time and tem-by conventional (acidic) solutions. However, this oc- perature. Perit Dial Int 15:26–32, 1995

5. Henderson IS, Couper IA, Lumsden A: Potentially irritant glu-curred only when the GDP-free solutions were bufferedcose metabolites in unused CAPD-fluid, in Frontiers in Peritonealto a neutral pH. Indeed, it was recently shown that a Dialysis, edited by Maher JF, Winchester JF, New York, Field,

neutral bicarbonate (25 mmol/L)/lactate (15 mmol/L)- Rich, 1985, pp 261–2646. Nilsson-Thorell C, Muscalu N, Andren AHG, et al: The heat-buffered PD solution reduced inflow pain compared with

sterilization of fluids for peritoneal dialysis gives rise to severalconventional (acidic) control lactate solutions [47]. In aldehydes. Perit Dial Int 13:208–213, 1993fact, the combination of bicarbonate/lactate was more 7. Linden T, Forsback G, Deppisch R, et al: 3-Deoxyglucosone, a

promoter of advanced glycation end-products in fluids for perito-effective than bicarbonate (38 mmol/L) alone in reducingneal dialysis. Perit Dial Int 18:290–293, 1998infusion pain. This may be ascribed to the usually high 8. Wieslander A, Nordin MK, Kjellstrand PT, et al: Toxicity of

extracellular pCO2 produced when the intraperitoneal peritoneal dialysis fluids on cultured fibroblasts L-929. Kidney Int40:77–79, 1991bicarbonate concentration is .30 mmol/L and there is a

9. Wieslander A, Nordin M, Martinson E, et al: Heat-sterilizedsubsequent rapid diffusion of CO2 intracellularly, causing PD-fluids impair growth and inflammatory responses of culturedintracellular acidosis [48, 49]. cell lines and human leukocytes. Clin Nephrol 39:343–348, 1993

10. Wieslander A, Andren AHG, Nilsson-Thorell C, et al: AreWe conclude that a new PD fluid with a higher pHaldehydes in heat-sterilized peritoneal dialysis fluids toxic in vitro?and less GDP than conventional ones is safe and easy Perit Dial Int 15:348–352, 1995

to use, with no negative effects on either the frequency 11. Wieslander A, Forsback G, Svensson E, et al: Cytotoxicity, pHand glucose degradation products in four different brands of PD-of peritonitis or on peritoneal membrane transport char-fluids. Adv Perit Dial 12:57–60, 1996acteristics. Moreover, it turns out to be more biocompati- 12. Wieslander A: Cytotoxicity of peritoneal dialysis fluid: Is it related

ble for the patient, since according to all available evi- to glucose breakdown products? Nephrol Dial Transplant 11:958–959, 1996dence, it tended to cause less mesothelial and interstitial

13. Cendoroglo M, Sundaram S, Groves C, et al: Necrosis and apo-damage, although peritoneal microvascular function ap- ptosis of polymorphonuclear cells exposed to peritoneal dialysispeared to be unaffected. In the long term, the use of this fluid in vitro. Kidney Int 52:1626–1634, 1997

14. Barile FA: Introduction to in vitro cytotoxicology. Florida, CRCsolution rather than conventional dialysate formulas willPress, Inc., 1994, pp 27–35hopefully lead to a more intact peritoneal membrane 15. Jonasson P, Braide M: Acute in vivo toxicity of heat-sterilized

with better preserved transport characteristics. glucose peritoneal dialysis fluids to rat peritoneal macrophages.Perit Dial Int 18:376–381, 1998

16. Jonasson P, Bagge U, Wieslander A, et al: Heat-sterilized PD-ACKNOWLEDGMENTSfluid blocks leukocyte adhesion and increases flow velocity in ratperitoneal venules. Perit Dial Int 16(Suppl 1):S137–S140, 1996This study was supported by Gambro Sweden AB and also in part

17. Musi B, Rippe A, Carlsson O, et al: Effects of acidity, glucoseby funding from the Swedish Medical Research Council (grant no. 8285).degradation products, and dialysis fluid buffer choice on peritonealThe support and assistance from the following persons are gratefullysolute and fluid transport in rats. Perit Dial Int 18:303–310, 1998acknowledged: Gerd Hallgren and Susanne Zetterman (Centrallasaret-

18. Henderson IS, Couper IA, Lumsden A: The effect of shelf lifetet Boras, Sweden), Peter Johansson (Lanssjukhuset Halmstad, Swe-of peritoneal dialysis fluid on ultrafiltration in CAPD, in Peritonealden), Ann Berglund-Back (Centralsjukhuset Karlstad, Sweden), VreniDialysis, edited by La Greca G, Milan, Wichtig, 1986, pp 85–86Frolich (Kungsholmsdialyzen Stockholm, Sweden), Elsebeth Hilde-

19. Martis L, Henderson LW: Impact of terminal heat-sterilizationbrand, Lene Moller (Rigshospitalet Copenhagen, Denmark), Pia Tro-on the quality of peritoneal dialysis solutions. Blood Purif 15:54–60,jahn (Hvidovre Hospital, Denmark), Lena Krutzen (University Hospital1997of Lund, Sweden), Ann-Britt Grahed and Elisabeth Rasanen (Malmo

University Hospital, Sweden), Ellinor Broms (Sahlgrenska University 20. Lamb EJ, Cattell WR, Dawnay AB, et al: In vitro formation of


advanced glycation end-products in peritoneal dialysis fluid. Kid- 36. Pannekeet MM, Koomen GCM, Struijk DG, et al: Dialysate CA125 in stable CAPD patients: No relation with transport parame-ney Int 47:1768–1774, 1995

21. Dawnay AB, St J, Millar DJ: Glycation and advanced glycation ters. Clin Nephrol 44:248–254, 199537. Ho-Dac-Pannekeet MM, Hiralall JK, Struijk DG, et al: Longi-end-products formation with icodextrin and dextrose. Perit Dial

Int 17:52–58, 1997 tudinal follow-up of CA 125 in peritoneal effluent. Kidney Int51:888–893, 199722. Niwa T, Katsuzaki T, Miyazaki S, et al: Amyloid b2-microglobulin

is modified with imidazolone, a novel advanced glycation end- 38. Dobbie JW, Lloyd JK, Gall CA: Categorization of ultrastructuralchanges in the peritoneal mesothelium, stroma and blood vesselsproduct, in dialysis related amyloidosis. Kidney Int 51:187–194,

1997 in uremia and CAPD patients, in Advances in Peritoneal Dialysis,edited by Khanna R, Nolph KD, Prowant BF, et al, Toronto,23. Nakayama M, Kawaguchi Y, Yamada K, et al: Immunohistochem-

ical detection of advanced glycosylation end-products in the perito- University of Toronto Press, 1990, pp 3–1239. Rippe B, Krediet RT: Peritoneal physiology: Transport of solutes,neum and its possible pathophysiological role in CAPD. Kidney

Int 51:182–186, 1997 in The Textbook of Peritoneal Dialysis, edited by Gokal RC, NolphKD, Dordrecht, The Netherlands, Kluwer Academic Press, 1994,24. Mahiout A, Ehlerding G, Brunkhorst R: Advanced glycation

end-products in the peritoneal fluid and in the peritoneal mem- pp 68–13240. Cappelli G, Bandiani G, Cancarini GC, et al: Low concentrationsbrane of continuous ambulant peritoneal dialysis patients. Nephrol

Dial Transplant 11:2–6, 1996 of glucose degradation products in peritoneal dialysis fluids andtheir impact on biocompatibility parameters: Prospective cross-25. Wieslander AP, Deppisch R, Svensson E, et al: In vitro biocom-

patibility of a heat-sterilized low-toxic and less acidic fluid for over study with a three-compartment bag. Adv Perit Dial 15:238–242, 1999peritoneal dialysis. Perit Dial Int 15:158–164, 1994

26. Haraldsson B: Assessing the peritoneal dialysis capacities of indi- 41. Yung S, Coles GA, Williams JD, et al: The source and possiblesignificance of hyaluronan in the peritoneal cavity. Kidney Intvidual patients. Kidney Int 47:1187–1198, 1995

27. Rippe B, Stelin G: Simulations of peritoneal solute transport dur- 46:527–533, 199442. Laurent TC, Laurent UBG, Fraser JRE: The structure and func-ing continuous ambulatory peritoneal dialysis (CAPD): Applica-

tion of two-pore formalism. Kidney Int 35:1234–1244, 1989 tion of hyaluronan: An overview. Immunol Cell Biol 74:A1–A7,199628. Rippe B, Stelin G, Haraldsson B: Computer simulations of peri-

toneal fluid transport in CAPD. Kidney Int 40:315–325, 1991 43. Lebel L, Smith L, Risberg B, et al: Effect of increased hydrostaticpressure on lymphatic elimination of hyaluronan from sheep lung.29. Altman DG: Practical Statistics for Medical Research. London,

Chapman & Hall, 1991 J Appl Physiol 64:1327–1332, 198844. Graff J, Joffe P, Fugleberg S, et al: Collagen markers in peritoneal30. Rothman KJ: Modern Epidemiology. Boston, Little Brown, 1986,

pp 153–156 dialysis patients. Adv Perit Dial 11:24–27, 199545. Nishida Y, Chang Shao J, Kiribayashi K, et al: Advanced glyca-31. Simonsen O, Wieslander A, Landgren C, et al: Less infusion

pain and elevated level of CA 125 by the use of a new and more tion end-products reduce the viability of human peritoneal meso-thelial cells. Nephron 80:477–478, 1998biocompatible PD-fluid. Adv Perit Dial 12:156–160, 1996

32. Heimburger O, Wang T, Lindholm B: Alteration in water and 46. Kawaguchi Y, Hasegawa T, Nakayama M, et al: Issues affectingthe longevity of the continuous peritoneal dialysis therapy. Kidneysolute transport with time on peritoneal dialysis. Perit Dial Int

19(Suppl 2):S83–S90, 1999 Int 52(Suppl 62):S105–S107, 199747. Mactier RA, Sprosen TS, Gokal R, et al: Bicarbonate and bicar-33. Hjelle JT, Miller-Hjelle MA, Dobbie J: The biology of the

mesothelium during peritoneal dialysis. Perit Dial Int 15(Suppl):S13– bonate/lactate peritoneal dialysis solutions for the treatment ofinfusion pain. Kidney Int 53:1061–1067, 1998S21, 1995

34. Pannekeet MM, Zemel D, Koomen GCM, et al: Dialysate markers 48. Ritter JM, Doktor HS, Benjamin N: Paradoxical effect of bicar-bonate on cytoplasmic pH. Lancet 335:1243–1246, 1990of peritoneal tissue during peritonitis and in stable CAPD. Perit

Dial Int 15:217–225, 1995 49. Schambye HT, Pedersen FB, Wang P: Bicarbonate is not theultimate answer to the biocompatibility problems of CAPD solu-35. Visser CE, Brouwer-Steenbergen JJE, Betjes MGH, et al: Can-

cer antigen 125: A bulk marker for the mesothelial mass in stable tions: A cytotoxicity test of CAPD solutions and effluents. AdvPerit Dial 8:42–46, 1992peritoneal dialysis patient. Nephrol Dial Transplant 10:64–69, 1995

long-term clinical effects of a peritoneal dialysis fluid with less glucose degradation products

Documents