1855

ClinicalChemist,y42:111855-1860 (1996)

Development of human renal function: referenceintervals for 10 biochemical markers in fetal urine

F1&NcoIsE MULLER,l* MARC DOMMERGUES,2 LAURENCE BUSSIjRES,1 STEPHEN LORTAT-JACOB,3CHANTAL Loiixr,4 J-FRcoIs OURY,5 YVES AIGRAIN,6 PATRICK NIAUDET,7

PHILIPE AEGERTER,8 and YVES DUMEZ2

Evaluation of fetal renal function by analysis of fetal urinesampled in utero may improve perinatal care after a prena-tal diagnosis of bilateral obstructive uropathy. We providereference intervals for 10 fetal urinary compounds andexamine their variation with gestational age. Forty-onefetuses with bilateral obstructive uropathy (urine sampledbetween 20 and 38 weeks of gestational age) had normal,healthy values for serum creatmnine (50 gzmoVL) at ages1-2 years. These cases were thus assumed to represent areasonable approximation to healthy values. Sodium andi2 -microglobulin concentrations significantly decreasedwith gestational age; calcium, ammonia, and creatininesignificantly increased; glucose, phosphorus, chloride, urea,and total protein concentrations did not vary. Our resultsprovide reference values for prenatal evaluation of fetalrenal function and suggest that glomerular filtration ofmacromolecules and tubular reabsorption of glucose andphosphorus are mature by 20 weeks of gestation, whereastubular reabsorption of sodium and 132-microglobulin in-creases progressively during the second half of gestation.

INDEXING TERMS: prenatal diagnosis #{149}renal function #{149}bilateralobstructive uropathy #{149}fetal status #{149}sodium . 132-microglobulmn.calcium #{149}ammonia #{149}creatinine #{149}glucose #{149}phosphorus #{149}chloride#{149}urea . protein #{149}age-related effects

Obstructive uropathies are the most frequent fetal anomaliesdiagnosed by prenatal ultrasound. In cases such as those involv-ing posterior urethral valves, the obstruction is potentially

Biochianie, H#{244}pitalAmbroise Pare, 92104, Boulogne, Universit#{233}Paris V,France.

2 Maternit#{233}Port-Royal Baudeloque, Hopital Cochin, Paris.Chirurgie Infantile and N#{233}phrologiePediatrique, Hopital Necker-Enfants

Malades, Paris.N#{233}phrologiePCdiatrique, Gynecologie-Obstetrique, and 6 Chirurgie In-

fantile, H#{244}pitalRobert DebrC, Paris.S Informatique M#{233}dicale,H#{244}pitalAmbroise Pare, Boulogne.Author for correspondence. Fax 331 49 09 58 63.Received August 7, 1995; accepted June 11, 1996.

curable but may induce a wide spectrum of renal damage. In themost severe cases, major renal dysplasia results in perinataldeath, or survival with renal failure. In the mildest forms ofobstructive uropathy, the infants survive with normal renalfunction.

On the basis of their experimental work with fetal lambs,Adzick et al. [1] studied human renal function by analysis of fetalurine sampled in utero and showed that fetal urine sodium waspredictive of short-term perinatal outcome. More recently, weand our colleagues have shown that prenatal assessment ofvarious fetal urinary analytes, including sodium and f32-micro-globulin, can predict postnatal renal function later in childhood[2, 3]. However, the clinical value of these prognostic tests hasbeen questioned [4, 5] because of the lack of precise referencevalues, which may hamper the interpretation of fetal urinalysis,and discrepancies in the definitions of normal renal function.

In this study, we have derived reference intervals (mean SD) for fetal urinalysis of 10 compounds from results for a groupof 41 fetuses with isolated bilateral uropathy, in whom renalfunction was subsequently (postnatal age 1-2 years) found to benormal according to stringent criteria.

Materials and MethodsBetween 1986 and 1992, 167 pregnant women underwent fetalurine sampling because of diagnosis of bilateral dilatation of thefetal urinary tract at routine ultrasonography. In all cases, thefollowing investigations were performed: (a) detailed sonogra-phy specifying the location of obstruction, the amniotic fluidvolume, and the renal parenchymal structure; (b) fetal karyotyp-ing by amniocentesis or fetal blood sampling; and (c) transab-dominal fetal urine sampling through an ultrasound-guided20-gauge needle.

Urine was obtained from a renal pelvis or a ureter wheneverpossible, or from the fetal bladder when pelves or ureters wereinaccessible. When pelvic dilatations were not symmetrical,urine was obtained either from the least-enlarged pelvis or fromboth sides; in the case of bilateral sampling, results from theleast-enlarged pelvis were used in the analysis. To keep theprocedure minimally invasive, we made no attempt to decom-

1856 Muller et al.: Reference intervals for analytes in fetal urine

press the urinary tract, and repeat urine samplings were notperformed routinely.

None of the fetuses underwent intrauterine uroamnioticshunting, and the results of fetal urinalysis were not taken intoaccount to alter postnatal management. Complete follow-up wasachieved for 157 of the 167 cases. There were 63 terminations ofpregnancy: 25 for associated anomalies (7 chromosome anom-alies, 18 multiple malformations with normal karyotype) and 38for marked oligohydramnios with severe bilateral renal damage.Two fetuses with multiple organ malformations died in utero.There were also 21 neonatal deaths: 3 from associated malfor-mations not diagnosed prenatally (2 VACTEL associations, 1severe hydrocephalus), 2 from neonatal sepsis, and 16 fromacute respiratory distress, pulmonary hypoplasia, and renalfailure. Two children died before age 6 months, one fromsudden infant death syndrome and one from megacystis-micro-colon. By 1 year postpartum, 69 children were still alive. Two ofthese (1 male, 1 female) had megacystis-microcolon; the remain-ing 67 survivors all had isolated obstructive uropathies, withoutany associated malformation. In all survivors, follow-up includedclinical examination in a pediatric nephrology unit, radiologyand ultrasonography of the urinary tract, serial biochemicalanalysis, and bacteriological urinalysis. All these infants werefollowed until at least age 1 year (range of follow-up, 1-7 years).Postnatal renal function was not considered normal in the 26cases with postnatal serum creatinine concentrations >50jimollL (5.6 mg/L) at 1-2 years. In contrast, 41 children wereconsidered to have normal renal function, and these 41 form thedatabase of this study. Twenty-two of these 41 cases werepreviously included in a study investigating the ability of fetalurinalysis to predict the renal function of children with bilateralobstructive uropathy [2].

The criterion we used to define normal postnatal renalfunction was serum creatinine S0 p.mol/L (5.6 mg/L) at age1-2 years. This threshold represents 2SD above the mean inhealthy children [6, 7]. The height and weight of all childrenwhose serum creatinine was 50 mol/L were within therespective normal ranges. In addition, none of these childrenhad any clinical symptom related to renal fimction. Renalfunction also remained normal in those followed for >2 years.

In these 41 cases, bilateral urinary tract dilatation wasdiagnosed between 20 and 38 weeks since the mothers lastmenstrual period. Fetal urine was sampled from a renal pelvis in29 cases and from the bladder in 12. When several samples werecollected from the same fetus at different gestational ages, onlythe first sample was used in the calculations.

Before performing biochemical urinalysis, we checked for theabsence of contamination with blood by making red and whiteblood cell counts. Lack of contamination with amniotic fluid wasnoted by the absence of intestinal enzymes, namely, y-glutamyl-transferase, aminopeptidase M, and alkaline phosphatase isoen-zymes. The following analytes were studied: sodium, chloride,calcium, phosphorus, ammonium, urea, creatinine, glucose,total proteins, and $32-microglobulin. Sodium was determinedby a spectrophotometric method, urea by glutamic dehydroge-nase enzymatic assay (Abbott, N. Chicago, IL), creatimne by theJaffe method (Beckman, Galway, Ireland), proteins by sulfosal-

icylic acid precipitation, chloride by a direct thiocyanate color-imetric method, calcium by o-cresolphthalein colorimetric assay(Abbott), phosphorus by molybdate colorimetry (Abbott), am-monia by an L-glutamyl dehydrogenase enzymatic method(Boehringer, Mannheim, Germany), glucose by enzymatic col-orimetry (Merck, Paris, France), and 132-microglobulin by RIA(Pharmacia, Uppsala, Sweden).

The urinary concentration of each compound was expressedas mean SD and was studied as a function of gestational agethrough the construction of age-related reference centiles byusing absolute residuals [8].

ResultsReference intervals were established from results for 41 fetuseswith bilateral obstructive uropathy whose postnatal renal func-tion at age 1 year was normal (defined as postnatal serumcreatimne

Reference values

Gestational age8Protein, g/Lp2-Microgbo1, mg/LUrea, mmol/LCreatinine, .tmol/LAmmonia, mol/LSodium, mmol/LChloride, mmol/LGlucose, mmol/LCalcium, mmol/LPhosphate, mmol/L

Weeks of amenorrhea.

Mean SD32 4

0.04 0.070.96 1.2

8.7 2.9216 68695 410

50 950 7

0.15 0.190.65 0.360.15 0.25

95% confidenceInterval

0.00-0.220.00 3.3

4.4-15

105-363225-1900

31-64

36-660.00-0.74

0.10-1.520.00-0.88

Clinical Chemi.rtiy 42,No. 11, 1996 1857

Table 1. Reference values for analytes In fetal urine(41 cases).

DiscussionThis study establishes reference intervals for the concentrationsof 10 urinary compounds in fetuses with bilateral uropathy whosurvived free of clinical symptoms of renal failure in infancy andwhose serum creatinine was

r = 0,49; P 50 .emol/L at ages 1-2 years.

1858 Muller et al.: Reference intervals for analytes in fetal urine

survivors with normal postnatal renal function are adequatecontrols. For sodium, /32-microglobulin, and calcium, datapoints from fetuses with poor outcome overlapped little withour reference intervals, underscoring the potential clinical use-fulness of these markers.

In addition to this potential clinical application, our resultsprovide new insights into human kidney development. Fetalurine production can be documented as early as 12 weeks afterconception. The morphological development of human kidney

has been extensively described. Briefly, the branching process ofthe collecting system is completed by 20 weeks. The inductionof the metanephric blastema by the distal portion of thebranches results in the formation of 8-12 generations ofnephrons. This process of nephrogenesis is complete by 32-36weeks of gestation, at which time each kidney has -800 000nephrons [19]. Little, however, is known about the functionalaspects of renal development. Our results suggest that thevarious functions of the kidney do not develop at the same time.

r = 0.48; P

0 aU

#{149}. . aU

#{149}

46

-J 3oE

22

!

U000

500

-j

300C

aE2000

100

a

0 A

aa

I

.

a

a#{149} U: #{149}. #{149}#{149}#{149}aU UU U

aa

a A AU

#{149}:- 0 0

a --------__._-__ aN-ft ---

160

140

F20E 100az80

60

40

EaC,

I 20 22 24 20 28 38 32 34 36 35 40

3

A UU.

2 I #{149}U UU #{149}

I #{149}#{149}.#{149}A 0I #{149} 0 a A 0

0 a

nI 0. 0 #{149}e

I 8 20 22 24 26 25 30 32 3 4 36 3 5 4 0

-J

0EE0a.

0

-J

0EEa0U

0

I 8 20 22 24 26 20 30 32 34 36 35 40 I 8 20 22 24 26 20 30 32 3 4 3 6 3 8 4 0

10UU. aA

#{149}U 0U U #{149}#{149}I aaao

#{149} LU #{149}#{149}u#{149}a DU#{149}#{149}: I a 02U

A#{176}

:U

3000

A

0*

-J

E

I

12

.

lb .

#{149}#{149}

2 U. U #{149} Aa

0 1 *III.u*.*o.

1 5 20 22 24 26 28 30 32 34 36 30 40

Gestatlonal age, weeks

Clinical Chemistry 42, No. 11, 1996 1859

:.#{163}#{149} #{149} a a. . 0

.. II 0#{149}#{149} #{149}. DoI__ __

0 0

00

I 20 22 24 26 28 30 32 34 36 3S 40

UU

#{149}#{149}1 #{149} a. U

a #{149} aa#{149} a

a. .#{149}#{149} 00 0

*

U ffri 1r_20 22 24 26 28 30 32 34 36 38 40

2000

-l0,I000

z

IS 20 22 24 20 28 30 32 34 36 38 40

Gestatlonal age, weeks

Fig. 2. Fetal urinalysis: abnormal cases.Data points for 70 cases with poor outcome are superimposed on the 5th, 50th, and 95th centiles referenceregressioncurves.A, Termination of pregnancy becauseof renaldamage with severe ollgohydramnios (n = 38).#{149},Neonataldeath caused by renal failure (n = 16). El. Survivors with serum creatinine >50 mol/L at ages1-2 years (n - 26). In some cases, urine sample volumewas too small to allow measurement of all analytes.

1860 Muller et al.: Reference intervals for analytes in fetal urine

For instance, at 20 weeks of gestation, the low protein contentof fetal urine suggests that glomerular renal function, analyzedin terms of protein filtration, is mature. The virtual absence ofurinary glucose and phosphorus at this gestational age suggeststhat the maturation of tubular reabsorption of these compoundsis also achieved by then. Earlier reports of higher fetal urinaryconcentrations of phosphate (0.91 mmollL) at 16 weeks [14]suggest immature tubular reabsorbtion of phosphate at thisstage of development. In contrast, our results are consistent withprogressive maturation of tubular reabsorption of sodium and132-microglobulin and tubular secretion of calcium after 20weeks. These changes in biological markers of tubular matura-tion parallel the increase in kidney mass and the number ofnephrons and confirm previous reports (based on smaller series)showing a decreased sodium concentration throughout gesta-tion [16, 18]. In addition, our data indicate that fetal urinarycalcium increases significantly (P