Clinical use of estimating glomerular filtration rate equ-ations during pregnancy

doi:10.4236/health.2011.31006

Paper Menu >>

Journal Menu >>

Vol.3, No.1, 32-36 (2011) Health

doi:10.4236/health.2011.31006

Clinical use of estimating glomerular filtration rate equ-

ations during pregnancy

Luiz Paulo José Marques1*, Regina Rocco2, Maria Helena Victor1,

Benedita Calheiros de Novaes2, Ana Luiza Batista de Carvalho1, Omar da Rosa Santos1

1Renal Unit of Gaffrèe and Guinle University Hospital-Department of Medicine, Federal University of Rio de Janeiro State, Rio de

Janeiro, Brazil; [email protected]

2Obstetric Unit of Gaffrèe and Guinle University Hospital-Department of Medicine, Federal University of Rio de Janeiro State, Rio

de Janeiro, Brazil

Received3 November 2010; revised 5 November 2010; accepted 8 November 2010.

ABSTRACT

Background: Kidney disease, even when mild,

was once considered so major an impediment

to successful pregnancy and so dangerous to

the mother’s wellbeing. High-risk pregnancy

mainly associated to renal impairment may oc-

cur in 10-20% of gestations and it is very im-

portant that renal function is closely monitored

to prevent or minimize maternal and fetal com-

plications. This study was designed to investi-

gate the performance of Cockcroft-Gault CGeq

and the simplified MDRDeq equations in healthy

pregnant women to assess renal function. Me-

thods: We studied 167 normal ambulatory

pregnant women and kidney function was con-

temporaneously estimated through the CGeq

and the simplified MDRDeq and calculated

through the creatinine clearance (Ccr). Serum

and urinary creatinine were assa yed using Jaffé

reaction method in the same AutoAnalyser.

Results: When we compared calculated and es-

timated clearences for measurement of kidney

function we observed that CGeq overestimated

renal function (CGeq = 168.41 ± 38.80 ml/

min/1.73 m2, Ccr = 146.27 ± 30.49 ml/min / 1.73 m2,

p < 0.001), MDRDeq underestimated renal func-

tion (Ccr = 146.27 ± 30.49 ml/min / 1.73 m2,

MDRDeq = 129.15 ± 29.28 ml/min / 1.73m2, p <

0.001). Conclusions: Our results demonstrated

that CGeq overestimated, MDRDeq underesti-

mated significantly kidney function during ges-

tation in healthy women and cannot be recom-

mended to assess renal function in obstetric

practice. Ccr remains a useful clinical tool in

pregnant women until the development of a

specific equation that considers the several

important maternal renal physiological altera-

tions and provides the measure of GFR the

most unbiased and precise as possible.

Keywords: Pregnancy, Kidney function,

Glomerular Fi l tration Rate, MDRD equation,

Cockcroft-Gault equation.

1. INTRODUCTION

The pregnancy promotes several important changes in

renal hemodynamics. There are increases in cardiac

output and plasma volume that appear to parallel those

of Glomerular filtration rate (GFR) and Renal Plasma

Flow (RPF), but vascular resistance and blood pressure

fall at the same time [1]. Maternal renal adaptation is

characterized markedly increase in GFR, approximately

50% greater than pre-pregnant value [2]. High-risk

pregnancies, mainly associated to renal impairment, may

occur in 10-20% of pregnant women and increase ma-

ternal and fetal morbidity and mortality [3]. Furthermore,

in the last years, the number of gestations in kidney dis-

eases patients increased markedly. So, it is very impor-

tant that renal function is closely monitored during

pregnancy to prevent or minimize maternal and fetal

complications.

Glomerular filtration rate (GFR) is considered to be

the key marker of kidney function, critical for detection,

evaluation and management of renal impairment [4].

However, the precise measurement of GFR is invasive,

time consuming, expensive and technically difficult.

More recently, calculation of estimated GFR (eGFR)

using an empirical mathematical prediction formula

based on single point endogenous serum markers have

been encouraged as a simple, rapid and reliable means o f

assessing kidney function with varying degrees of suc-

cess. These methods are generally accepted as a better

L. P. J. Marques et al. / Health 3 (2011) 32-36

tool for estimating renal function than serum creatinine

alone. National and international organizations recom-

mend that clinical labor atories report est imated GFR and

that clinicians use estimated GFR to evaluate kidney

function for all patients. [5] However, the guidelines

specifically exclude interpretation in pregnant women in

the absence of validation.

Serum creatinine has been pivotal for renal function

evaluation because it is an inexpensive, common test in

clinical practice. The factors associated with creatinine

measure are well understood: Creatinine is derived from

the metabolism of creatine in skeletal muscle and from

dietary protein intake; it is released into the circulation at

a relative ly constant rate and has a stable plasma concen-

tration. Creatinine is freely filtered across the glomerulus

and is neither reabsorbed nor metabolized by the kidney.

However, approximately 10 to 40 percent of urinary crea-

tinine is derived from tubular secretion by the organic

cation sec retory pat hways in the proxi m al tubule [6].

There are no fewer than 47 predictions equations cur-

rently available that use creatinine as endogenous mark-

er of renal function, although the most common equa-

tions used to estimate renal function in adults are the

Cockcroft-Gault (CG eq) and the simplified equation from

the Modification of Diet in Renal Disease Study

(MDRDeq) [7]. These two equations differ in many as-

pects, including the predicted index (creatinine clearance

for CGeq and GFR for MDRDeq), units of predictions

(ml/mim for CGeq and ml/mim/1,73 m2 for MDRDeq)

and variables used for the prediction (serum creatinine,

gender, age and body weight for CGeq and serum creati-

nine , gender, age, and race for MDRDeq) [8,9].

Recently, other endogenous proteins easily measured

in clinical laboratories have been suggested as better

markers of GFR. Cystatin C, considered superior to

creatinine in non pregnant women, has been recom-

mended as an alternative and possibly superior marker in

pregnancy. However, others factors such as diabetes, fat

mass and inflammation may alter serum Cystatin C level

and some reports have shown substantial variability in

the relationship between GFR and Cystatin C among

different populations [10], and more r ecent data serious-

ly question its use in pregnant populations [11].

Inulin clearance may be considered the gold standard

method to measure GFR in pregnancy, but it is costly,

cumbersome and not practical outside a clinical research

setting. The endogenous creatinine clearance (Ccr), the

primary tool to assess kidney function in non pregnant

subjects, is equally useful in clinical practice for eva-

luating renal function during the gestational period. Cli-

nicians caring for obstetrics patients are still waiting data

if GFR estimating equations are accurate in pregnancy to

use in clinical practice.

In the present study, we investigated the performance

of CGeq and MDRDeq equations while using ClCr as the

best approximation for GFR in healthy pregnant women.

We aimed to analyse the performance of these estimating

equations to predict renal function in pre gnancy.

2. MATERIALS AND METHODS

One hundred sixty seven normal ambulatory pregnant

women accepted to participate in the study that was

conducted from January to December 2008 and renal

function was contemporaneously estimated and calcu-

lated. All subjects were normoten sive with no history of

renal, cardiac, diabetic or vascular disease, and individu-

als having treatment with drugs affecting creatinine se-

cretion were excluded. The normal pregnant women

selected were between 20th and 30th weeks of gestation.

All subjects gave written informed co nsent for creatin ine

clearance studies and the clinical research was approved

by the local Human Research Ethics Committee.

Each pregnant woman on whom the study was carried

out on an outpatient basis was carefully instru cted in the

collection of a 24 hour urine specimen. Blood samples

were drawn at the morning before their usual morning

meal when weight and height were obtained. They

stayed on a free diet and on their normal physical activi-

ty. Creatinine was measured using a kinetic colourime-

tric assay (Jaffé method) on the same Roche Hitachi 917

equipment (Roche Diagnostic GmbH, Mannheim, Ger-

many) and the assay was calibrated with an IDMS ref-

erence laboratory. Two samples were collected with in-

terval of one week and the mean of the results was used

to calculate renal function by creatinine clearance (Ccr),

Cockcroft-Gault eq uation (CGeq) and the simplified equ-

ation from the Modification of Diet in Renal Disease

Study (MDRDeq).

Ccr was calculated (ml/min):

Ccr = urinary creatinine mg/dl X 24 hour urinary vo lume

ml/min / serum creatinine mg/dl

The prediction equations were calculated:

CGeq (ml/min) = 140 – age years X weightkg/72 X se-

rum creatinine mg/dl with use of the 0,850 multiplier for

female gender.

Simplified MDRDeq (ml/min/1.73 m2) = 175 [serum

creatinine mg/dl X 0.011312]-1.154 X Ageyears-0.205 X

0.742.

To make allowance for comparison with Ccr, MDRDeq

and CGeq results were adjusted for 1.73 m2 of body sur-

face area (BSA) according to the formula:

BSA (m2) = 0.007184 X height cm0.725 X weight kg0.425.

Statistical procedures included analysis of mean and

SD for categorical and numerical data, analysis of 95%

confidence intervals (95%CIs) and analysis of variance

(ANOVA) associated with Bonferroni multiple compar-

isons test.

L. P. J. Marques et al. / Health 3 (2011) 32-36

3. RESULTS

A total of 167 normal ambulatory pregnant women;

mean age 32.06 ± 5.04 years were included in this study.

Calculated creatinine clearance, CGeq and MDRDeq

Tabl e 1 . Clinical and laboratorial data at the time of creatinine

measurement.

Number of Pregnant women 167

Age (year) 32.06 ± 5.04

Time of Pregnancy (week) 25.08 ± 3.03

Weight gain (Kg) 12.04 ± 3.11

Serun creatinine (mg%) 0.59 ± 0.12

Creatinine Clearance (ml/min/1.73 m2) 146.27 ± 30.49

Cockcroft-Gaulteq (ml/ mi n/1.73 m2) 168.41 ± 38.80

MDRDeq ( ml /min/1.73 m2) 129.15 ± 29.28

Tabl e 2. Renal function (ml/min/1.73 m2) measured by creati-

nine clearance (Ccr), Cockcroft-Gault (CGeq) and MDRDeq and

comparison among the measurements.

Ccr CGeq MDRDeq Bias 95% con-

fidence

interval

CGeq x

Ccr 146.27

± 30.49 168.41

± 38.80 - 22.13 16.69 To

-27.50

Ccr x

MDRDeq 146.27

± 30.49 - 129.15 ±

29.28 17.12

11.68 To

22.56

CGeq x

MDRDeq - 168.41

± 38.80 129.15 ±

29.28 39.26

33.82 To

44.70

equations were performed from 20 to 29 weeks of preg-

nancy (25.80 ± 3.03 weeks) and the weight gain during

gestation until laboratory tests was 12.04 ± 3.11 kg. ( Ta-

ble 1).

The main analyses for indices of kidney function were

with data normalized per 1.73 m2 and the results ob-

served were Ccr = 146.27 ± 30.49 ml/min / 1.73 m2, CGeq

= 168.41 ± 38.80 ml/min/1.73 m2, MDRDeq = 129.15 ±

29.28 ml/min/1.73 m2. (Figure 1).

When we compared calculated and estimated clear-

ances for measurement of kidney function during preg-

nancy we observed that CGeq overestimated renal func-

tion (168.41 ± 38. 80 ml/min/1.73 m2 and 146.27 ± 30.49

ml/min/1.73 m2, CGeq versus Ccr respectively, p < 0.001

and the bias 22.13 ml/min) and MDRDeq underestimated

renal function (129.15 ± 29.28 ml/min/1.73 m2 and

146.27 ± 30.49 ml/min/1.73 m2, MDRDeq versus Ccr

respectively, p < 0.001 and the bias 17,12 ml/min). (Ta-

ble 2, Figure 2 and Figure 3).

4. DISCUSSIO N

Effective renal plasma flow increase and Glomerular

filtration rate change in pregnancy. GFR increase mar-

kedly during gestation, approximately 50% greater than

non pregnant values during the second trimester, these

high levels are maintained through gestational period

until the 36th week, after which a decrease of 15% to

20% may occur (2). We studied kidney function of 167

normal ambulatory pregnant women from 20 to 29

Figure 1. Renal function measured by creatinine clearance (a),

Cockcroft-Gaulteq (b) and MDRDeq (c) during pregnancy.

Figure 2. Poor correlation between renal function measured by

Creatinine Clearance (ClCr) and Cockcroft-Gaulteq (CG) in

normal pregnancy.

Figure 3. Poor correlation between renal function measur e d by

Creatinine Clearance (ClCr) X MDRDeq in normal pregnancy.

weeks of pregnancy (25.80 ± 3.03 weeks), in the period

L. P. J. Marques et al. / Health 3 (2011) 32-36

of pregnancy where the renal function is more stable.

The Clearance of endogenous creatinine (Ccr) was the

best approximation of GFR in pregnancy to assess renal

function in clinical practice. Because there is, at most, a

small increase in the production and excretion of creati-

nine during normal gestation, the large increments

in Ccr lead to reduction in its plasma level. Lower seru m

creatinine levels observed in pregnancy are a reflection

of haemodilution due to the increase of plasma volume

as well as hyperfiltration. Thus, values considered nor-

mal in nonpregnant women may reflect abnormal renal

function in preg nancy [12].

Accurate GFR measurements using inulin infusion are

impractical for large-scale application in pregnant popu-

lation, while Ccr is made complex by the need of obtain-

ing accurate 24 hour urine collection. To minimize this

last limitation we instructed women to take note of the

start and end time of urine collection to allow for com-

putation of the length of collection, and used the mean of

two creatinine samples collected with interval of one

week to calculate renal function. All the creatinine as-

says were performed in the same Technican AutoAna-

lyser to prevent the high inter-laboratory variations of

creatinine [13] and the kidney function was contempo-

raneously estimated and calculated.

When we compared the results of estimating renal

function using CGeq and MDRD eq equations with Ccr, we

observed that CGeq overestimated (168.41 ± 38.80

ml/min/1.73 m2 and 146.27 ± 30.49 ml/min/1.73 m2, p <

0.001 ) and MDRDeq underestimated renal function

(129.15 ± 29.28 ml/min/1.73 m2 and 146.27 ± 30.49

ml/min/1.73 m2, p < 0.001), which demonstrated that

CGeq and simplified MDRDeq equations are not useful

measure of GFR in normal pregnancy to estimate renal

function. Some studies have also demonstrated in normal,

hypertensive and preeclamptic pregnancy, that both the

CGeq and MDRDeq were also inaccurate in predicting

kidney function when compared with Ccr [14-16].

Bias (difference between measured (Ccr) and esti-

mated (CGeq and MDRDeq) kidney function) was used as

a measure of performance. Bias of CGeq prediction in

pregnant women (22.13 from 16.69 to 27.57 ml/min)

may be explained by the rule that weight coefficients of

the CGeq do not differentiate between muscular mass

(relevant to creatinine generation) and non-muscular

mass (not relevant to creatinine generation). Thus, the

CGeq during gestation transforms any weight gain dif-

ference into a difference in predicted kidney function

that tends to be overestimated in overweight. Generally

in obstetric population, healthy women gain approx-

imately 12.5 kg in the first pregnancy and 1 kg less in

subsequent gestation and increase the body surface area.

In our pregnant women group weight gain was 12.04 ±

3.11 kg during gestation until laboratory tests to assess

renal function and may have contributed to overesti-

mated renal function by CGeq. However, the correction

of clearance results to 1.73 m2 of body surface area was

not capable to correct th is overesti mated resu lt [17].

In normal gestation, our results demonstrated that

MDRDeq underestimated renal function (Bias of 17.12

from 11.68 to 22.56 ml/min) and is not us eful measur e in

pregnant women on whom GFR increases markedly.

Some large studies have recently concluded that using

the MDRDeq in healthy subjects is problematic. MDRD eq

was derived from an adult population with chronic kid-

ney disease most of whom had a GFR less than 60

ml/min/1.73 m2. Perhaps more important, GFR was sys-

tematically underestimated by MDRDeq when GFR rises

above 60 ml/ min/1.73 m2 in healthy subjects and clini-

cal laboratory that estimate GFR by MDRDeq usually

report the result as GFR > 60 ml/min/1.73 m2 [18,19 ].

Similarly in pregnancy, recent studies comparing

MDRDeq with inulin clearance in a small sample of

healthy pregnant women have also demonstrated that

MDRDeq underestimated GFR during gestation [20,21].

In conclusion: CGeq overestimated and MDRDeq un-

derestimated significantly kidney function during gesta-

tion in healthy women and are not accurate enough to be

used as screening tests to assess renal function in obste-

tric population. At present, 24-h urine collection for en-

dogenous Ccr, and that may also be used for proteinuria

measurement, remains as the best method to measure

renal function during pregnancy in clinical practice, and

there is no estimating equation that provides accurate

eGFR during pregnancy. So, it is necessary to develop

an accurate equation to pregnant women that considers

the several important changes in maternal renal hemo-

dynamics and provides an eGFR the most unbiased and

precise as possible. Our results clearly demonstrated that

CGeq and MDRDeq cannot be recommended as a useful

clinical tool in obstetric pra ctice and c linicians caring for

obstetric patients may soon see eGFR routinely reported

alongside the traditional biochemistry.

REFERENCES

[1] Umans, J.A., Lindheimer, M.D. (1995) The renal adapta-

tions to pregnancy is now nostalgic. Journal of Clinical

Investigation, 96, 482-490. doi:10.1172/JCI118048

[2] Conrad, K. (2004) Mechanisms of renal vasodilation and

hiperfiltration during pregnancy. Journal of the Society

for Gynecologic Investigation, 7, 438-443.

doi:10.1016/j.jsgi.2004.05.002

[3] ACOG practice bulletin. (2002) Diagnosis and manage-

ment of preeclampsia and eclampsia. O b stetrics Gyne-

cology, 99, 159-166.

[4] Levey, A.S., Eckardt, K.U., Ts ukamoto, Y., et al. (2005)

Definition and classification of chronic kidney disease

(KDIGO). Kidney international, 67, 2089-2010.

doi:10.1111/j.1523-1755.2005.00365.x

L. P. J. Marques et al. / Health 3 (2011) 32-36

[5] National Kidney disease education program: Information

for health professionals (2004) National Institutes of

Health, National Institutes of Diabete s and Digestive and

Kidney Diseases, Bethesda, Maryland, USA.

[6] Shemesh, O., Golbetz, H., Kriss, J.P. and Myers , B.D.

(1985) Limitations of creatinine as a filtration marker in

glomerulopathic patients. Kidney Internation, 28,

830-838. doi:10.1038/ki.1985.205

[7] Johnson, D.W. (2005) Use of serum creatinine concentra-

tion to assess level of kidney function. Nephrology, 10,

s133-s139. doi:10.1111/j.1440-1797.2005.00487_1.x

[8] Cockcroft, D.W., Gaul,t M.H. (1976) Prediction of crea-

tinine clearance from serum creatinine. Nephron, 16,

31-41. doi:10.1159/000180580

[9] Levey, A.S., Green, T., Kusek, J.W., et al. (2000) A si m-

plified equation to predict glomerular filtration rate from

serum creatinine. Journal of the American Society Neph-

rology, 11, 151A.

[10] Stevens, L.A., Schmid, C.H., Greene, T., et al. (2009)

Factors other than GFR affect serum cystatin levels.

Kidney international, 75, 652-660.

doi:10.1038/ki.2008.638

[11] Akbari, A., Lepage , N., Keely, E., Clark, H.D., Jaffey, M.

and Filler, G. (2005) Cystatin C and beta trace protein as

markers of renal function in pregnancy. British Journal

of Obstetri cs Gynecology, 112, 575 -578.

doi:10.1111/j.1471-0528.2004.00492.x

[12] Jeyabalan, A. and Conrad, K.P. (2007) Renal function

during normal pregnancy and preeclampsia. Front Bios-

cience, 12, 2425-2437. doi:10.2741/2244

[13] Miller, W.G., Myers, G.L., Ashwood, E.R., et al. (2005)

Creatinine measure: state of the art in accuracy and in-

ter-laboratory harmonization. Archives Pathology & La-

boratory Medicine, 129, 297-304.

[14] Delemarre, F.M. and Schoenmakers, C.H. (2008) The

MDRD formula in pregnancy. BJOG, 115, 1192.

[15] Marques, L.P.J., Rocco, R., Victor, M.H., Novaes, B.C.,

Carvalho, A.L.B. and Santos, O.R. (2009) Clinical using

of estimating glomerular filtration rate equations during

pregnancy. World Congress of Nephrology, Milan Italy,

22-26 May 2009, poster Su. 197.

[16] Côté, A.M., Lam, E.M., von Dadelszen, P., Mattman, A.,

Magee, L.A. (2010) Monitoring renal function in hyper-

tensive pregnancy. Hypertens Pregnancy, 29, 318-29.

doi:10.3109/10641950902968676

[17] Alper, A.B., Yi, Y., Webber, L.S., Pridjian, G., Mumeney,

A.A., Saade, G., Morgan, J., Nuwayhid, B., Belfort, M.

and Puschett, J. (2007) Estimation of glomerular filtra-

tion rate in preeclamptic patients. American Journal of

Perinatology, 24, 569-574. doi:10.1055/s-2007-986697

[18] Geddes, C.C., Woo, M.Y. and Brady, S. (2008) Glom eru-

lar filtration rate – What is the rationale and justification

of normalizing GFR for body surface area? Nephrol Dial

Transplant, 23, 4-6. doi:10.1093/ndt/gfm662

[19] Stevens, L.A., Coresh, J., Deysher, A.E. and Levey, A.S.

(2007) Evaluation of the MDRD study equation in a

large diverse population. Journal of the American Society

Nephrology, 18, 2749-2757.

doi:10.1681/ASN.2007020199

[20] Froissart, M., Rossert, J., Jacquot, C., et al. (2005) Pre-

dictive performance of the MDRD and Cockcroft-

Gault equations for estimating renal function. Journal of

the American Society Nephrology, 16, 763-773.

doi:10.1681/ASN.2004070549

[21] Smith, M., Moran, P., Ward, M.K., Davison, J.M. (2008)

Assessment of glomerular filtration rate during pregnan-

cy using the MDRD formula. BJOG, 115, 109-112.

[22] Ahmed, S.B., Bentley-Lewis, R., Hollenberg, N.K.,

Graves, S.W., Seely, E.W. (2009) A comparison of pre-

diction equations for estimating glomerular filtration rate

in pregnancy. Hypertens Pregnancy, 28, 243-55.

doi:10.1080/10641950801986720