Prediction of acute failure asphyxia · Predictionofacuterenal failure after birth asphyxia DS Roberts, GBHaycock, RNDalton, CTurner, PTomlinson, LStimmler, J WScopes Abstract...

Archives ofDisease in Childhood 1990; 65: 1021-1028

Prediction of acute renal failure after birth asphyxia

D S Roberts, G B Haycock, R N Dalton, C Turner, P Tomlinson, L Stimmler, J W Scopes

AbstractTwenty one babies of34-41 weeks' gestationalage with birth asphyxia (5 minute Apgar score65 or umbilical artery pH 67-2) were studiedduring the first two days of life to find outwhether the urinary excretion of tubularmarkers of renal function is of value in theearly diagnosis of acute renal failure. Urinaryretinol binding protein, myoglobin, andN-acetyl-,3-D-glucosaminidase (NAG), ex-pressed as a ratio with urinary creatinine,were measured and excretion profiles repeatedat 3-6 days in 15 infants and at 7-14 days in11 infants. Plasma creatinine concentration,creatinine clearance, plasma myoglobin con-centration, and fractional sodium excretionwere measured where possible in asphyxiatedinfants. Control data were obtained from 50healthy infants: 28 gave urine samples alone,17 urine and blood, and five blood alone.Normal urinary values were derived from 17,25, and three infants, respectively, for thethree time periods. The number of controlsamples was limited for ethical reasons.Four asphyxiated infants had acute renal

failure (group 1), four had tubular dysfunctionwithout glomerular disturbance (group 2) and13 had normal renal function (group 3). Group1 were clearly identified by greatly increasedurinary retinol binding protein (>27 000,ug/mmol creatinine) and myoglobin (>1500pg/mmol creatinine) excretion measured inthe first two days of life. In control infants therange of excretion of retinol binding proteinwithin the same time period was 3 to 967pg/mmol creatinine and urinary myoglobinwas undetectable. Excretion ofNAG failed todiscriminate between groups 1 and 2. Acuterenal failure occurred only in infants who hadheavy myoglobinaemia. Tubular dysfunctionin group 2 was transient and not accompaniedby plasma electrolyte disturbances.We conclude that measurement of urinary

excretion of retinol binding protein or myo-globin after birth is helpful in the earlydiagnosis of acute renal failure.

Acute renal failure is a recognised complicationof birth asphyxia; it carries a poor immediateprognosis and may result in permanent renaldamage in up to 40% of survivors. " The recentreports of increased urinary excretion of P2-microglobulin in infants with evidence ofintrauterine fetal distress (meconium stainedliquor) by Cole et al, and in sick infants by Tacket al, led these authors to suggest that subclinical,hypoxic, renal damage may be a relatively

common, underdiagnosed finding.7 8The early recognition of acute renal failure is

particularly important in asphyxiated infantswith hypoxic encephalopathy, in whom a stablebiochemical milieu is vital, because it facilitatesthe administration of appropriate fluid andelectrolyte replacement. The diagnosis is fre-quently difficult to make, however, becausemany of the clinical and biochemical findingsthat are helpful in establishing the diagnosis ofacute renal failure in adults and children areunreliable in neonates. The presence or absenceof oliguria may be misleading as 7% of healthybabies fail to pass urine until the second day oflife,9 and non-oliguric acute renal failure hasbeen reported in neonates. The plasma creatinineconcentration on day 1 is a poor guide to aninfant's renal function, because it mainly reflectsthe maternal creatinine concentration.11 12 Afurther difficulty in the interpretation of theplasma creatinine concentration is that themethods of measurement in common use (modi-fications of the colorimetric Jaffe reaction) aresubject to errors from interfering chromogenssuch as bilirubin and pyruvate,'3 substancesthat are likely to be present in increasedconcentrations in the blood of sick infants. Thevalue of indexes such as the urine:plasmaconcentration ratio of sodium, creatinine, andurea are limited by the overlap that existsbetween infants with functional (prerenal)failure, and those with intrinsic renal failure. 1416The fractional excretion of sodium or renalfailure index may be of diagnostic value, butboth need to be interpreted with caution inthe preterm,17 and in infants who have beengiven a saline challenge or treated with diureticsor aminophylline.

Myoglobinuria is a cause of renal failure inadults,'8 and a strong correlation between theduration of postnatal oliguria and the severity ofmyoglobinaemia has been reported in asphyxi-ated infants.'9 The presence of myoglobinuriamay therefore be a useful indicator of acuterenal failure.20

This study was designed to investigate thevalue of measurements of the excretion ofretinol binding protein, myoglobin, and N-acetyl-,B-D-glucosaminidase (NAG) during thefirst two days of life as predictors of clinicallyimportant acute renal failure after birthasphyxia, and to follow the excretion profilesinto the second week of life.

Patients and methodsTwenty one infants who had had an episode ofbirth asphyxia, defined as a 5 minute Apgar

Department ofPaediatrics,United Medical andDental Schools ofSt Thomas's andGuy's Hospitals, London,St Thomas's CampusD S RobertsJ W ScopesGuy's Hospital CampusG B HaycockR N DaltonC TurnerP TomlinsonL StimmierCorrespondence to:Dr D S Roberts,Jenny LindChildren's Department,Norfolk and NorwichHospital, Brunswick Road,Norwich NRI 3SR.Accepted 23 March 1990

1021

on August 27, 2020 by guest. P

rotected by copyright.http://adc.bm

j.com/

Arch D

is Child: first published as 10.1136/adc.65.10_S

pec_No.1021 on 1 O

ctober 1990. Dow

nloaded from

Roberts, Haycock, Dalton, Turner, Tomlinson, Stimmler, Scopes

score of 5 or below, or an umbilical artery pH of7-2 or below at delivery, were studied in theneonatal units of these hospitals. The study wasapproved by the ethics committees of bothhospitals and parental consent was obtained inall cases. Infants were managed according to thestandard protocols used in the two units, nomodifications being made as a result of partici-pation in the study.

Causes of birth asphyxia included severeshoulder dystocia, intrapartum cord traction,and haemorrhage. Sepsis was implicated in thecause in only one infant, delivered in poorcondition at 36 weeks, whose cultures yieldedListeria monocytogenes after 48 hours' incubation.

All infants were of 34 weeks' gestational ageor above. This lower limit was chosen becausenephronogenesis is complete by this time andcomparisons are not usually complicated bydevelopmentally determined differences inrenal function.2' Gestational age was assessedfrom the menstrual history, ultrasonography, orby the clinical criteria of Dubowitz et al.22Wherever possible, infants were studied on

several occasions during the first two weeks oflife and results grouped into three postnataltime periods: 1-2 days, 3-6 days, and 7-14

Table I Number of infants in each group providing samples for each analysis at each timeperiod

Analysis Asphyxiated infants Control infants*(n=21) (n=50)

Time period (days) Time period (days)

1-2 3-6 7-14 1-2 3-6 7-14

Retinol binding protein 21 15 11 17 27 3Urinary myoglobin 21 15 11 17 27 3N-acetyl-fl-D-glucosaminidase 21 15 11 17 27 3Urine output 19 0 0 0 0 0Plasma creatinine 21 14 10 4 13 3Creatinine clearance 19 14 9 0 0 0Plasma myoglobulin 20 14 8 4 14 3Fractional excretion of sodium 19 14 9 4 11 2

*Of the 50 control infants, 28 gave urine samples, five gave blood samples, and 17 gave both.

Table 2 Clinical data of infants studied on days 1-2. Values are expressed as geometricmean (range)

Asphyxiated Control p Value*infants infants (days 1-2)(n=21) (n=17)

Gestational age (weeks) 39-3 (34-41) 37-7 (34-41) NSBirth weight (g) 3400 (2200-4800) 3400 (1900-4700) NSApgar score:At 1 minute 2-3 (0-7) 8-6 (6-10) <0 01At 5 minutes 4-1 (0-9) 9-6 (8-10) <0 01

*Wilcoxon rank sum test with Bonferroni correction.

days. All infants were studied within the firsttwo days of life and measurements were repeatedbetween 3-6 days in 15, and in the second weekof life in 11 infants. It was not possible to repeatmeasurements in some infants because of deathor discharge.A total of 50 non-asphyxiated infants served

as controls for the study; 28 provided urinesamples, 17 both urine and blood samples, andfive urine samples alone. Seventeen controlurine specimens were available for analysisduring the first two days of life, 27 during the3-6 day period, and three during the period7-14 days. With the exception of two infantswho provided urine samples on different days,non-asphyxiated infants were studied on a

single occasion.Blood samples from all infants were obtained

when blood was drawn for clinical purposes,and data are limited for this reason. Thenumbers of individual analyses are shown intable 1. Plasma was separated immediately andstored at -20°C until analysis. Urine was

collected by the method of Tarlow and alsostored at -20C.2s None of the infants studiedwas treated with aminoglycosides, which mayaffect tubular function.2425

Glomerular function was estimated by theplasma creatinine concentration and endogenouscreatinine clearance (calculated as urine crea-

tinine concentrationxurine flow rate (ml/min)/plasma creatinine concentration). Tubularperformance was assessed by the urinary con-

centrations of retinol binding protein, NAG,and myoglobin, divided by the urinary creatinineconcentration. Fractional excretion of sodium(calculated as the urine:plasma sodium ratio/urine:plasma creatinine ratio) was alsomeasured, and results are reported whendiuretics were not administered.Sodium was estimated in plasma and urine by

flame photometry. Creatinine was measured inplasma by high performance liquid chromato-graphy with cation exchange (HPLC),26 and inurine by an automated, reaction rate version ofthe Jaffe reaction.27 Retinol binding protein andmyoglobin were measured in urine by enzymelinked immunosorbent assays (ELISA) usingrabbit antisera (Dako Ltd).28 (Retinol bindingprotein ELISA: P Tomlinson, personal com-

munication, myoglobin ELISA: D Roberts,unpublished observations). Myoglobin in plasmawas also estimated by ELISA. The limit ofdetection for both assays was 2 [tg/l. NAG

Table 3 Results of biochemical tests done during first two days of life. Values expressed as geometric mean (range)

Control Asphyxiated infantsinfants(n=17) Group I Group 2 Group 3

(n=4) (n=4) (n= 13)

Urinary excretion of:Retinol binding protein(tg/mmol creatinine) 75 (3-%7) 42 906 (27 759-81 112) 7570 (3721-15 243) 154 (6-1418)

N-acetyl-p3-D-glucosaminidase(smol substrate/hour/mmolcreatinine) 134 (32-362) 4707 (1135-13 484) 1520 (692-7736) 153 (24-310)

Myoglobin (og/mmol creatinine) Not detectable 6004 (1526-13 195) 4 (0-4-430) 1 (Not detectable-17)Fractional sodium excretion (%) 0-2 (0-1-0 5) 32 (13-71) 1-7 (0-9-3 4) 0 3 (0 1-2 0)Creatinine clearance (ml/min/kg) Not measured 0 01 (0U003-0U04) 0-8 (06-1-0) 1-2 (0-43-1)Plasma concentrations of:

Creatiine (ILmol/l) 75 (59-93) 257 (188-344) 101 (77-116) 78 (51-116)Myoglobin (tigIl) 48 (32-73) 2071 (855-7625) 104 (38-262) 95 (21-390)

Urine output (ml/kg/hour) Not measured 0-3 (0-1-08) 2-1 (1-5-2-7) 2-0 (0-5-57)

1022

on August 27, 2020 by guest. P

rotected by copyright.http://adc.bm

j.com/

Arch D

is Child: first published as 10.1136/adc.65.10_S

pec_No.1021 on 1 O

ctober 1990. Dow

nloaded from

Prediction ofacute renalfailure after birth asphyxia

100000

10000

r-

amc

-a)

EE

0-.(DaapCD

-a_C

Figure I Urinaryconcentration ofretinolbinding protein measuredduring the first two days oflife in control (open symbols)and asphyxiated infants(closed symbols).Asphyxiated infants weredivided into three subgroupsaccording to the degree ofretinol binding protcinexcretion. The scale islogarithmic, and the barindicates the mean andtolerance limits ofnormal.

1000

100

10.

* Groupi1**

.

0

0

0

0000

0

00

0000

0

0

00

gO0

0

400 -

300 -

0

E

.C

E

200

100 -

0

0

0

0

0

0

A

CIO A~* a-~~~M

1-2

0

00

-00doo

A*fAU

I_

3-6

Postnatal age (days)

Figure 2 Plasma creatinine concentrations with increasing postnai(open symbols) and the three subgroups ofasphyxiated infants (closecircles, group 2: triangles, and group 3: squares). The horizontal baieach group.

S

0

I0

0

I Group 2

Group 3

10

-C

E

CDX 0-1 -

CD

0_._1

l- 0-01-

U

win-2-' 4

0

10-

0-c

1-

0

C 0.1-

0

AA U

:~J0

00

0

Figure 3 Creatinine clearance according to birth weight,and urine output, measured during the first two days oflife inthe three subgroups ofasphyxiated infants (group 1: circles,group 2: triangles, and group 3: squares). The scak islogarithmic, and the horizontal bars indicate the mean foreach group.

activity was measured by an automated colori-metric method using p-nitrophenyl-N-acetyl-j3-D-glucosaminide as substrate (Sigma ChemicalCo).29When several estimations were made in any

infant for each study period, the arithmeticmean was calculated and used in furtheranalyses.Grouped results are presented in the figures

as individual data points and in the tablesas geometric means and ranges.

ResultsTable 2 shows the clinical data of the studyinfants on days 1-2. There were no significantdifferences in gestational age or birth weightbetween the asphyxiated infants and controls.Four asphyxiated infants (19%) developed

* acute renal failure (defined as a plasma creatinineconcentration of over 130 ,umol/l for at least twoconsecutive days). One infant required treat-ment with peritoneal dialysis for hyperkalaemicdysrhythmia.Nine infants (43%) had asphyxial encephalo-

pathy with seizures and seven of these had somedegree of renal dysfunction (three had acuterenal failure, and four had tubular proteinuria).A further six infants (29%) had transientcerebral irritability and feeding problems, butonly one of these infants developed acute renalfailure. Six infants who had no cerebralsymptoms had normal renal function. Threeinfants died (14%), two during the course of thestudy, at 2 and 10 days of age, respectively.The relevant biochemical data of the study

infants on days 1-2 are shown in table 3. TheO

A upper tolerance limits of normal for retinolA : binding protein and NAG were defined as the

___ geometric mean plus two logarithmic SDs for7-14 the respective set of control data. These were

2968 ,ug/mmol creatinine and 465 ,umol sub-strate/hour/mmol creatinine, respectively. Myo-

tal age in control subjects globin was undetectable in the urine of healthydsymbols-group 1:rs indicate the mean for newborns on days 1-2.

When the urinary retinol binding protein:

1023

.

on August 27, 2020 by guest. P

rotected by copyright.http://adc.bm

j.com/

Arch D

is Child: first published as 10.1136/adc.65.10_S

pec_No.1021 on 1 O

ctober 1990. Dow

nloaded from

Roberts, Haycock, Dalton, Turner, Tomlinson, Stimmler, Scopes

100-

10-

1 -

0.1 -

0

00 I

0

.

AI

am.Uf

0

0

0

as

taI0

1-2 3-6Postnatal age (days)

Figure 4 Fractional excretion ofsodium with increasing postnat(open symbols) and asphyxiated infants (closed symbols-group 1and group 3: squares). The scale is logarithmic and the horizontaleach group.

0

0

0

0

0

o0

A--

4-

00

o80

0

00o

OC00

c0

a

Uan

s.

0

* 000

8

0

0

I -2

1 -2I I L

3-6Postnatal age (days)

asphyxiated infants are considered in thesethree subgroups.Plasma creatinine values in control infants

and in the asphyxiated infants studied seriallyare shown in fig 2. Infants in group 1 alreadyhad greatly increased plasma creatinine concen-trations during the first two days of life, whichincreased further towards the end of the firstweek and then declined during the secondweek. Infants in groups 2 and 3 showed the

*___ expected decline in plasma creatinine concen-8 * . tration with increasing postnatal age.

* Creatinine clearance during the first two daysof life, expressed according to birth weight, wascorrespondingly reduced in group 1; all theseinfants were oliguric at the time of initialassessment with urinary flow rates of less than1 ml/kg/hour (fig 3).Tubular dysfunction was confirmed in

L*................. 0 groups 1 and 2 by the fractional excretion of7-14 sodium (fig 4).

In control infants the mean urinary excretiontal age in control subjects of retinol binding protein and myoglobin, and: circles, group 2: triangles, the fractional excretion of sodium were constantbars indicate the mean for during the first two weeks of life; NAG excretion

showed a slight rise with increasing postnatalage. Myoglobin was detected in the urine ofonly one control infant, between 3 and 6 days:the concentration was at the limit of the assayand the infant did not show the highest concen-tration of retinol binding protein or of NAG in

* the control group. In infants in group 1 themean excretion of all markers declined towardstheir respective normal ranges with increasingtime after the episode of asphyxia, the retinolbinding protein most slowly. The same was true

° of group 2 with the exception ofNAG, the meanexcretion of which remained unchanged. Excre-

Al tion of all markers in group 3 was comparablewith control values throughout the first two

- weeks of life (fig 5, 6, and 7).0 Free myoglobin seems to be handled by the

kidney as a low molecular weight protein, in asimilar way to retinol binding protein. In thepresence of an increased filtered load (a raised

* plasma concentration) it appears in the urine indetectable quantities in the presence of tubulardysfunction, as indicated by an increased urinaryretinol binding protein concentration.

__________ Plasma myoglobin concentration was greatly7-14 increased in infants in group 1 (fig 8), but

concentrations decreased in all groups withincreasing postnatal age.rigurze. unn4ary excrew-ojl(re.tnoJ onaLng proietn wtm tncreasing posna&taagen conLro

subjects (open symbols) and asphyxiated infants (closed symbols-group 1: circles, group 2:triangles, and group 3: squares). The scale is logarithmic and the horizontal bars indicate themeanfor each group.

creatinine ratio of asphyxiated infants (measuredin the first two days of life) was displayedgraphically, the values were noted to be dividedinto three distinct subgroups (fig 1). Group(n=4) had concentrations that were 15-44 timesabove the upper tolerance limit of normal.Group 2 (n=4) had concentrations that were2-8 times above the upper tolerance limit ofnormal, and group 3 (n= 13) had concentrationsthat were within the normal range.

All infants in group had acute renal failure.Glomerular function was normal in groups 2and 3. Throughout the rest of this paper the

DiscussionThe measurement of various tubular markers ofrenal function has proved useful in the diagnosisand surveillance of a number of renal dis-orders.3"33 Low molecular weight proteins,such as P32-microglobulin and retinol bindingprotein, are freely filtered at the glomerulus andalmost completely reabsorbed by the proximaltubular cells in health, so that when tubulardysfunction is present, increased quantitiesappear in the urine. The report of raised urinaryj32-microglobulin concentrations in infants withmeconium stained liquor who were otherwisewell led to the suggestion that subclinicaltubular damage was present in a proportion ofthese infants, and that measurement of low

ELB0

0C0

a,x

04)

aL

100000 -

a).Cr

c 10000--aEcECD-s 1000 -Ca,

-B 100-C

-5C

a)> 10-

C

D

1024

,0~o0

j7;,"jv,o c r A;,N,4;*"T 'hirntoq*l 0114 th

on August 27, 2020 by guest. P

rotected by copyright.http://adc.bm

j.com/

Arch D

is Child: first published as 10.1136/adc.65.10_S

pec_No.1021 on 1 O

ctober 1990. Dow

nloaded from

Prediction ofacute renalfailure after birth asphyxia

0

o *o 0o A -00 A

000 IeQet -M

A0

A

0

1-2 3-6 7-14Postnatal age (days)

Figure 6 Urinary excretion ofN-acetyl-,fD-glucosaminidase (NAG) with increasingpostnatal age in control subjects (open symbols) and asphyxiated infants (closed symbols-group 1: circles, group 2: triangles, andgroup 3: squares). The scale is logarithmic and thehorizontal bars indicate the meanfor each group.

U

A A

A

* 0

00

a

AW

I 13-6

Postnatal age (days)

I

7-14

Figure 7 Urinary excretion ofmyoglobin with increasing postnatal age in control subjects(open symbols) and asphyxiated infants (closed symbols-group 1: circles, group 2: triangles,and group 3: squares). The scale is logarithmic and the horzontal bars indicate the meanforeach group.

molecular proteinuria might be a useful adjunctin the assessment of their renal function.7Comparison data from studies in adults have

shown that urinary retinol binding proteinconcentration correlates well with that of 2-microglobulin, and that it is of equal sensitivityin the detection of tubular dysfunction.34 Wechose to measure retinol binding proteinbecause it is stable at lower urinary pH values atwhich degradation of 032-microglobulin occurs,and the need for manipulation of urinary pH onvoided specimens is eliminated.35

Excretion of retinol binding protein in neo-nates is higher than in adults. Our controlinfants achieved a mean retinol binding proteinexcretion in the first two days of life of 75pig/mmol creatinine, with similar values in thesecond week of life, in contrast to a reportedmean concentration of 77 [tg/mmol creatininein adult men.28 This tenfold difference is not aspurious increase in the retinol binding proteinconcentration as a result of the lower urinarycreatinine concentrations found in infants; theabsolute urinary retinol binding protein concen-trations in the infants were greater, with amaximal recorded value of 4000 ,ug/l, comparedwith the maximal normal adult value reportedin the above study of 540 jig/l. This presumablyreflects reduced proximal tubular reabsorptionin neonates.The mean retinol binding protein excretion

reported here for control infants with gestationalages of more than 34 weeks is similar to thevalue obtained by Clark et al,36 but out uppertolerance limit of normal is roughly five timeshigher and reflects a greater range of retinolbinding protein excretion values in the groupstudied.We identified three groups of asphyxiated

infants according to their urinary retinol bind-ing protein:creatinine ratio. The four infantswith acute renal failure had the highest excretion(range 27 759-81 112 pg/mmol creatinine) andthere was clear discrimination between theseinfants and the others. In turn, infants in group2 (3721-15 243) were clearly differentiated fromthose in group 3 (6-1418). Infants with acuterenal failure also had the heaviest myoglobinuria(range 1526-13 195 [ig/mmol creatinine) and, aswith retinol binding protein, were sharplydistinguished from the others. Those in group 2had detectable, but lesser, degrees of myo-globinuria (0 4-430). Gross pigmenturia, givinga brown colour to the urine, was seen in allinfants in group 1, and in the infant from group2 who had the highest myoglobin concentrationof that group. In three infants from group 3myoglobin was detected in the urine, but thevalues were at the limit of detection. Our resultsare in agreement with those of Kojima et al.20Tubular dysfunction was confirmed by thefractional excretion ofsodium, which was grosslyabnormal in group 1 (mean 31 9%) mildlyincreased in group 2 (1-7%), and normal ingroup 3 (0-3%). That in control infants wasbelow 1% and in sodium balance studies thefractional excretion of sodium in healthy infantsof similar gestational age with a sodium intakeof s,2-5 mmol/kg/day has been reported to benormally less than 1%.17

100000

0

*0

C.)

c

E0-CE

a,

E-

a,

0

10000 -

*1000-

100-

10-

*a

-I.-a

0.

100000

10000-a,

C

C.).-aE

cnC"r-.a0

0

E

C

0S

0

.

1000-

100-

10-

1*

1-2

1025

a

A

000 *-- __

on August 27, 2020 by guest. P

rotected by copyright.http://adc.bm

j.com/

Arch D

is Child: first published as 10.1136/adc.65.10_S

pec_No.1021 on 1 O

ctober 1990. Dow

nloaded from

Roberts, Haycock, Dalton, Turner, Tomlinson, Stimmler, Scopes

100000-

z--

r.-

80

Co

a,

0

C:.a_0

0

E

Co

E

Co

10000-

1000.

100-

10-

-v-0

AuA I

A*

A a 00oo

-OOo

,40

0

-IF0

a

AU

0

*-

a

1-2I

3-6 7-14Postnatal age (days)

Figure 8 Plasma myoglobin concentrations with increasing postnatal age in control subjects(open symbols) and asphyxiated infants (closed symbols-group 1: circles, group 2: triangles,and group 3: squares). The scale is logarithmic and the horizontal bars indicate the meanforeach group.

NAG is a lysosomal enzyme of a molecularweight that precludes its filtration at theglomerulus, it is also rapidly cleared from thecirculation by the liver so that increased urinaryexcretion is a consequence of intrinsic renaldamage.37 NAG excretion was considerablyincreased in infants with acute renal failure(range 1135-13 484 [tmol substrate/hour/mmolcreatinine) and in infants with tubular dysfunc-tion and no glomerular disturbance (692-7736).This presumably reflects a similar initial degreeof tubular cell damage in the two groups. NAGis clearly a sensitive measure of tubular damage,but not specific enough to discriminate theinfants with acute renal failure.

In the infants who were studied serially,retinol binding protein, NAG, and myoglobinexcretion declined in group 1, retinol bindingprotein more slowly than either NAG or myo-

globin, and all remained above the normal range

into the second week of life. In one infant,measurements were repeated at the age of sixweeks and urinary retinol binding protein andNAG were still greatly increased, suggestingthat tubular recovery, if it occurs at all, is slow.The increased excretion of NAG in the secondweek of life is probably not the result ofcontinued tubular cell exfoliation, but ofincreased lysosomal activity in surviving cells;tubular reabsorption of low molecular weightproteins occurs by endocytosis, the proteinsthen fusing with lysosomes within which theyare digested. Urinary NAG concentrations were

highest in those infants with the greatest excre-tion of retinol binding protein and myoglobin,suggesting that this persistent use is the result ofa stimulated exocytosis similar to that shown inanimal experiments.38 The similar continuedrise in urinary NAG, with more or less

constant mean values throughout the two weekpostnatal period in infants in group 2, supportsthis suggestion. In these infants retinol bindingprotein and myoglobin excretion steadily de-clined, and by the second week remained highin only one infant; measurements were repeatedin the third week in this infant and all resultswere then normal. This indicates that tubulardysfunction in these infants is transient. Infantsin group 3 had concentrations comparable withthose in control infants throughout the study.Our single criterion for the diagnosis of acute

renal failure was a plasma creatinine concen-tration of over 130 [tmol/l for at least two days.The presence of oliguria was not required so wedid not overlook infants with non-oliguric acuterenal failure. Similar plasma creatinine concen-trations have been used as diagnostic criteria inother studies.' 2 39 We measured plasmacreatinine using high performance liquidchromatography because the method is specific;with modified Jaffe techniques various chromo-gens present in plasma interfere with thereaction resulting in overestimations or under-estimations of the creatinine concentration thatcannot be predicted for any individual measure-ment and may produce errors in excess of100%.40 Within two days, the plasma creatinineconcentrations in infants in group 1 wereconsiderably raised (mean 257 [tmol/l). Themean values were similar in groups 2 and 3,comparable to our limited control data, andwithin the expected range for infants of similarpostnatal age reported by other workers." 12Creatinine clearance expressed according tobirth weight,4' was correspondingly reduced(mean 0-01 ml/min/kg) in group 1. The clear-ances measured in groups 2 (0-8), and 3 (1-2),are similar to those described by Coulthard in areview of published data on glomerular filtrationrates expressed by birth weight in healthynewborns.42

All infants in group 1 were oliguric in the firsttwo days of life with urinary flow rates of lessthan 1 ml/kg/hour. One infant from group 3 hada reduced urine output, the remainder had flowrates in excess of 1 ml/kg/hour.

Severe muscle damage associated with subse-quent renal failure is well documented inadults,'8 and the same phenomenon probablyoccurs in neonates.43 Birth asphyxia is associatedwith myoglobinaemia in the first few days oflife, the severity of which is-in turn-relatedto the duration of postnatal oliguria.'9 Plasmamyoglobin concentrations, measured in thefirst 2 days of life in our asphyxiated infants,ranged from 21 to 7625 ,tg/l. The infants withacute renal failure had the highest concentra-tions, and the infant who required dialysis hadthe highest of all. These concentrations aresimilar to those reported by Kasik et al,'9 andsimilar to those seen in adults with acuterhabdomyolysis and acute renal failure."4 Atbirth, infant muscle contains roughly 10% ofthe adult myoglobin content45; the plasmaconcentrations achieved by several of theasphyxiated infants therefore indicate consider-able muscle breakdown.

If the normal range of plasma creatinine atbirth and an expected daily rise above 18 timol/l

1026

00

00 0

0

on August 27, 2020 by guest. P

rotected by copyright.http://adc.bm

j.com/

Arch D

is Child: first published as 10.1136/adc.65.10_S

pec_No.1021 on 1 O

ctober 1990. Dow

nloaded from

Prediction ofacute renalfailure after birth asphyxia 1027

in the presence of acute renal failure areconsidered,46 the plasma creatinine concentra-tions in our infants represent an extremely rapidrise. Such a rise has been reported in myo-globinuric renal failure in adults, and is theresult of the release of large quantities ofintracellular creatine from damaged muscle.47Our results show that an asphyxial insult maybe associated with tubular dysfunction, but thatacute renal failure supervenes only in thepresence of high plasma concentrations of myo-globin: the exaggerated rise in plasma creatininein our infants, coupled with the heavy myo-globinaemia and myoglobinuria, suggests thatthese infants had rhabdomyolysis induced renalfailure.Our identification of infants with no glomeru-

lar disturbance but with detectable tubulardysfunction confirms previous reports, althoughthis was a less common occurrence in our study(19%) than in others.7 8 Plasma myoglobinconcentrations in these infants were lower thanin those with acute renal failure, evidence thatthey sustained less severe asphyxial damage.These infants undoubtedly have a degree oftubular dysfunction that is clinically unsuspectedbut which can be shown with sensitive tests.The higher mean fractional excretion of sodiumin this group of infants has implications forelectrolyte balance, although we found no signi-ficant differences in the plasma sodium andpotassium concentrations between these infantsand those without tubular dysfunction duringthe time periods 1-2 and 3-6 days (Wilcoxonrank sum test with correction factor).

In conclusion, 62% of the infants studied,who were over 34 weeks' gestational age andhad had an episode of birth asphyxia, showedno apparent impairment of renal function, 19%had overt acute renal failure, and a further 19%had tubular dysfunction indicated by an in-creased excretion of tubular proteins, but with-out any associated glomerular disturbance.These figures do not represent the absoluteincidence of renal damage in our units becausethe numbers presented here include infants notborn in the hospital and exclude those whoreceived treatment with aminoglycoside anti-biotics.

Infants with acute renal failure were clearlyidentified by their increased excretion of retinolbinding protein and myoglobin. Urinary retinolbinding protein:creatinine ratios in excess of27 000 ,ug/mmol creatinine, and urinarymyoglobin:creatinine in excess of 1500 [tg/mmolcreatinine seem to be of diagnostic value. NAGexcretion, although a sensitive indicator oftubular injury, is not specific for acute renalfailure.

Plasma creatinine concentration, creatinineclearance, and fractional excretion of sodiumdefined the group with acute renal failure, butall have some disadvantage that affects theirpotential diagnostic value. The HPLC methodwe used, which measures 'true' plasma creatinineis of restricted availability; creatinine clearance*is subject to the added inaccuracies incumbentupon timed urine collections and incompletebladder emptying,48 and the measurement ofthe fractional excretion of sodium requires

simultaneous blood and urine samples andcautious interpretation when diuretics or asaline load have been administered.The measurement of retinol binding protein

or myoglobin is simple, inexpensive, and can becarried out on random samples of urine. Wesuggest that the measurement of retinol bindingprotein or myoglobin should be incorporatedinto the routine assessment of asphyxiatedinfants at birth, as they give an early indicationof the adequacy of renal function.

We thank the parents and participating infants for their help andunderstanding. We thank sisters B Clarkson-White and B Nicholand all staff of the neonatal units of St Thomas's and Guy'sHospitals for their help and patience.

This study was supported in part by the Children NationwideMedical Research Fund. DS Roberts was supported by StThomas's Hospital Research Endowment Fund.

1 Chevalier RL, Campbell F, Brenbridge ANAG. Prognosticfactors in neonatal acute renal failure. Pediatrics 1984;74:265-72.

2 Anand SK, Northway JD, Crussi FG. Acute renal failure innewborn infants. J Pediatr 1978;92:985-8.

3 Anand SK. Acute renal failure in the neonate. Pediatr ClinNorth Am 1982;29:791-800.

4 Dauber IM, Krauss AN, Symchych PS, Auld PA. Renalfailure following perinatal anoxia. J Pediatr 1976;88:851-5.

5 Meeks ACG, Sims DG. Treatment of renal failure inneonates. Arch Dis Child 1988;63:1372-6.

6 Stapleton FB, Jones DP, Green RS. Acute renal failure inneonates: incidence, etiology and outcome. PediatricNephrology 1987;1:314-20.

7 Cole JW, Portman RJ, Lim Y, Perlman JM, Robson AM.Urinary ,32-microglobulin in full-term newborns: evidencefor proximal tubular dysfunction in infants with meconium-stained liquor. Pediatrics 1985;76:958-64.

8 Tack ED, Perlman JM, Robson AM. Renal injury in sicknewborn infants: a prospective evaluation using urinary 112-microglobulin concentrations. Pediatrics 1988;81:432-40.

9 Sherry SN, Kramer I. The time of passage of first stool andfirst urine by the newborn infant. J Pediatr 1955;46:158.

10 Grylack L, Medani C, Hultzen C, Sivasubramanian K,Davitt MK, Scanlon JW. Nonoliguric acute renal failure inthe newborn. A prospective evaluation ofdiagnostic indexes.AmJ3 Dis Child 1982; 136: 518-20.

11 Trompeter RS, Al-Dahhan J, Haycock GB, Chik G, ChantlerC. Normal values for plasma creatinine concentrationrelated to maturity in normal term and preterm infants. IntJ Pediatr Nephrol 1983;4:145-8 .

12 Rudd GD, Hughes EA, Placzek MM, Hodes DT. Referenceranges for plasma creatinine during the first month of life.Arch Dis Child 1983;58:212-5.

13 Spencer K. Analytical reviews in clinical biochemistry: theestimation of creatinine. Ann Clin Biochem 1986;23:1-25.

14 Matthew OP, Jones AS, James E, Bland H, Groshong T.Neonatal renal failure: usefulness of diagnostic indices.Pediatrics 1980;65:57-60.

15 Norman ME, Asadi FK. A prospective study of acute renalfailure in the newborn infant. Pediatrics 1979;63:475-9.

16 Ellis EN, Arnold WC. Use of urinary indexes in renal failurein the newborn. AmJ3 Dis Child 1982;136:615-7.

17 Al-Dahhan J, Haycock GB, Chantler C, Stimmler L. Sodiumhomeostasis in term and preterm neonates: 1 Renal aspects.Arch Dis Child 1983;58:335-45.

18 Bywaters EGL, Beall D. Crush injuries with impairment ofrenal function. Br Med J 1941;i:427-32.

19 Kasik JW, Leuschen MP, Bolam DL, Nelson RM. Rhabdo-myloysis and myoglobinaemia in neonates. Pediatrics 1985;76:255-8.

20 Kojima T, Kobayashi T, Matsuzaki S, lwase S, Kobayashi Y.Effects of perinatal asphyxia and myoglobinuria ondevelopment of acute renal failure. Arch Dis Child 1985;60:908-12.

21 Potter EL, Thierstein ST. Glomerular development in thekidney as an index of fetal maturity. J Pediatr 1943;22:695.

22 Dubowitz LMS, Dubowitz V, Goldberg C. Clinical assess-ment of gestational age in the newborn infant. J Pediatr1970;77:1-10.

23 Tarlow M. Urine and stool collection for metabolic studies inthe newborn. Arch Dis Child 1974;49:490-2.

24 Elinder G, Aperia A. Development of glomerular filtrationrate and excretion of ,l2-microglobulin in neonates duringgentamicin treatment Acta Paediatr Scand 1983;72:219-24.

25 Osborne JP. Urinary enzymes in the detection of renaldamage in the newborn. London: University of London,1980. (MD thesis.)

26 Ambrose RT, Ketchum DF, Smith JW. Creatinine deter-mined by 'high-performance' liquid chromatography. ClinChem 1983;29:256-9.

27 Cook JGH. Creatinine assay in the presence of protein. ClinChim Acta 1972;47:613-7.

28 Topping MD, Forster HW, Dolman D, Luczynska CM,Bernard AM. Measurement of urinary retinol-bindingprotein by enzyme-linked immunosorbent assay and its

on August 27, 2020 by guest. P

rotected by copyright.http://adc.bm

j.com/

Arch D

is Child: first published as 10.1136/adc.65.10_S

pec_No.1021 on 1 O

ctober 1990. Dow

nloaded from

1028 Roberts, Haycock, Dalton, Turner, Tomlinson, Stimmler, Scopes

application to the detection of tubular proteinuria. ClinChem 1986;32:1863-6.

29 Marulin D. Rapid colorimetric assay of P-D-galactosidase andN-acetyl-fo-D-glucosaminidase in human urine. Clin ChimActa 1976;73:453-61.

30 Wellwood JM, Ellis BG, Price RG, Hammond K, ThompsonAE, Jones NF. Urinary N-acetyl-fi-D-glucosaminidaseactivities in patients with renal disease. Br Med3J 1975;iii:408-11.

31 Kunin CM, Chesney RW, Craig WA, England AC, DeAngelisC. Enzymuria as a marker of renal injury and disease:studies of N-acetyl-p-D-glucosarninidase in the generalpopulation and in patients with renal disease. Pediatrics1978;62:75 1-60.

32 Watanabe K, Kojima T, Fukuda Y, et al. Reliability ofurinary N-acetyl-,1-D-glucosaminidase as an indicator ofrenal tubular damage in neonates. Biol Neonate 1987;52:16-21.

33 Gibb DM, Tomlinson PA, Dalton NR, Turner C, Shah V,Barratt TM. Renal tubular proteinuria and micro-albuminuria in diabetic patients. Arch Dis Child 1989;64:129-34.

34 Bernard AM, Moreau D, Lauwerys R. Comparison ofretinol-binding protein and f2-microglobulin determinationin urine for the early detection of tubular proteinuria. ClinChim Acta 1982;24:477-82.

35 Davey PG, Gosling P. f12-microglobulin instability in patho-logical urine. Clin Chem 1982;28:1330-3.

36 Clark PMR, Bryant TN, Hall MA, Lowes JA, Rowe DJF.Neonatal renal function assessment. Arch Dis Child 1989;64:1264-9.

37 Schlesinger P, Rodman JS, Frey M, et al. Clearance oflysosomal hydrolases following intravenous infusion, therole of the liver in the clearance of f-glucuronidase and N-

acetyl-4-D-glucosaminidase. Arch Biochem Biophys 1976;177:606-14.

38 Bernard A, Amor AL, Viau C, Lauwerys R. The renal uptakeof proteins: a non-selective process in conscious rats.Kidney Int 1988;34:175-85.

39 Olavarria F, Krause S, Barranco L, et al. Renal function infull-term newborns following neonatal asphyxia. ClinPediatr 1987;26:334-8.

40 Huang Y, Chiou WL. Creatinine XII: comparison of assaysof low serum creatinine levels using high-performanceliquid chromatography and two picrate methods. ] PharmSci 1983;72:836-7.

41 Coulthard MG, Hay EN. Weight as the best standard forglomerular filtration in the newborn. Arch Dis Child1984;59:373-5.

42 Coulthard MG. The maturation of glomerular filtration inpreterm and mature babies. Early Hum Dev 1985;11:281-92.

43 Haftel AJ, Eichner J, Haling J, Wilson ML. Myoglobinuricrenal failure in a newborn infant. 7 Pediatr 1978;93:1015-6.

44 Demos MA, Gitin EL, Kagen LJ. Exercise myoglobinemiaand acute exertional rhabdomyolysis. Arch Intern Med1974;134:669-73.

45 Kagen LJ, Christian CL. Immunologic measurements ofmyogloblin in human adult and fetal skeletal muscle. AmPhysiol 1966;211:656-60.

46 Guignard JP. Neonatal nephrology. In: Holliday MA, BarrattTM, Vernier RL, eds. Pediatric nephrology. 2nd Ed.Baltimore: Williams and Wilkins, 1987: 921-44.

47 Grossman RA, Hamilton RW, Morse BM, Penn AS,Goldberg M. Nontraumatic rhabdomyolysis and acuterenal failure. N Engl J Med 1974;291:807-1 1.

48 Roberts DS, Rendell B. Postmicturition residual bladdervolumes in healthy babies. Arch Dis Child 1989;64:825-8.

on August 27, 2020 by guest. P

rotected by copyright.http://adc.bm

j.com/

Arch D

is Child: first published as 10.1136/adc.65.10_S

pec_No.1021 on 1 O

ctober 1990. Dow

nloaded from