Microalbuminuria as an early predictor of hypertensive complications in pregnant women at high risk

Microalbuminuria as an Early Predictor of Hypertensive Complications in Pregnant Women at High Risk Jacob Bar, MD, Moshe Hod, MD, Arie Erman, PhD, Samuel Friedman, MD, llana Gelerenter, MA, Boris Kaplan, MD, Geoffrey Boner, MD, and Jardena Ovadia, MD 0 The value of microalbuminuria in predicting hypertensive complications in pregnant patients at high risk was tested in a prospective trial. A secondary aim was to compare the urinary albumin excretion rate between highrisk hypertensive pregnant patients (study group) and pregnant patients at high risk of other complications, normal pregnant subjects, and nonpregnant subjects. Over the last 5 years, 276 patients were studied (142 in the study group v 134 controls). Albumin was measured in an 6-hour overnight urine collection throughout pregnancy using a radioimmunoassay technique. The pregnant women in both the study and control groups demonstrated a statistically significant increase in albumin excretion rate in the second and third trimesters compared with the first. Mean albumin excretion rate values were significantly higher in the study group dp = 0.0001). Using logistic and linear regression models, the presence of microalbuminuria in the early third trimester was proven to be predictive of hypertensive complications (odds ratio, 2.1; confidence intervals, 1.26 to 3.53) and birth weight (R* = 0.7, P < 0.05) in the study group. Intrauterine growth retardation and neonatal outcome were less predictable. With the introduction of radioimmunoassays and in light of these significant clinical results, we believe that highrisk patients in whom abnormal proteinuria develops usually have a microalbuminuric phase weeks earlier, and this test has some predictive value for severe disease. In addition, the accepted definition of gestational proteinuria should be reconsidered. 0 1996 by the National Kidney Foundation, Inc. INDEX

WORDS:

Microalbuminuria;

prediction;

pregnancy;

C

LINICAL proteinuria exceeding 300 mgl 24 hr is usually considered a late sign in the evolution of preeclampsia’ and is associated with a poor perinatal prognosis.’ Proteinuria is usually preceded by reduced uric acid clearance, and the use of increasing maternal plasma urate concentration as an early renal indicator of preeclampsia is under consideration.3-8 Other predictive factors, such as family history and parity, angiotensin II infusion test, rollover test, plasma fibronectin concentration, and urinary calcium/ creatinine ratio, have been suggested but lack clinical consensus.9V’o Microalbuminuria may be defined as a subclinical elevation of urinary albumin, below that of dipstick detection (ie, <170 mg/L) but above normal. Type I (insulin-dependent) diabetic patients with urinary albumin excretion rates

From the Department of Obstetrics and Gynecology and the Institute of Nephrology and Hypertension, Rabin Medical Center, Petah Tiqva, Israel; and the Statistical Laboratory, Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel. Received July 27, 1995; accepted in revised form March 19, 1996. Address reprint requests to Jacob Bar, MD, Department of Obstetrics and Gynecology, Rabin Medical Center, Beilinson Campus, Petah Tiqva 49 100, Israel. 0 1996 by the National Kidney Foundation, Inc. 0272-6386/96/2802-0009$3.00/O 220

American

Journal

hypertension.

(AERs) above normal are more likely to acquire diabetic nephropathy, eventually progressing to renal failure.“-14 Microalbuminuria is correlated with increased blood pressure not only in diabetic patients, but also in those with benign essential hypertension,‘5v16 and the use of antihypertensive treatment reduces urinary albumin excretion significantly. 16,” Higher AERs have been detected in normal women in the third trimester of pregnancy than in pregnant women in the second and first trimesters and in nonpregnant women.r8 Microalbuminuria was found in 30% of women who had a record of gestational diabetes mellitus.” Published results on pregnancy-induced hypertension negated the assumption that microalbuminuria is an early predictor of hypertensive complications.20~2’ In the present prospective study of a large group of pregnant women at risk of hypertensive complications, mainly severe preeclampsia, radioimmunoassay technique was used to detect elevated urinary AERs throughout pregnancy and to determine whether microalbuminuria could serve as an early predictor of hypertensive complications and perinatal outcome. PATIENTS

AND METHODS

Two hundred seventy-six patients were studied last 5 years. Of these, 142 were pregnant women of Kidney

Diseases,

Vol 28, No 2 (August),

1996:

over the attending

pp 220-225

MICROALBUMINURIA

IN HYPERTENSIVE

PREGNANCY

our high-risk antenatal clinic (group 1). All were at high risk for hypertensive complications, mainly severe preeclampsia: women with a history of preeclampsia in a previous pregnancy, chronic hypertension, or renal disease or other vascular disease such as systemic lupus erythematosus, and those with signs or symptoms of preeclampsia in the current pregnancy (diastolic blood pressure [DBP] >90 mm Hg). The criteria for diagnosis were in place before we started to prospectively collect urine samples for determination of AERs. Women in whom microalbuminuria or clinical proteinuria was present at admission were excluded from the trial. The patients were studied monthly from admission until 28 weeks’ gestation, then every 2 weeks until 32 weeks’ gestation, and then weekly until term. At each visit an g-hour overnight urine collection was obtained, and weight and blood pressure were recorded in addition to the careful medical history and physical examination. The first trimester was defined as 1 to 13 weeks, the second as 14 to 28 weeks, and the third as 29 weeks to term. Simultaneous analysis of plasma concentrations of blood urea nitrogen, creatinine, and uric acid and platelet count were performed using the automated hydroxylamine method (SMA 12/60; Technicon, Tarrytown, PA) and thrombocounter (H.l System; Technicon). The conventional definition of hypertension is usually based on blood pressure values alone (DBP >90 mm Hg). This is insufficient to determine the severity of hypertension in an ongoing pregnancy. Consequently, both in the determination of the status of patients at entry and in the assessment of outcome, more specific definitions were used? 1. Chronic hypetiension: DBP 290 mm Hg either in the first half of pregnancy or 6 weeks after delivery, or the patient began pregnancy while already receiving antihypertensive treatment. 2. Nonproteinuric pregnancy-induced hypertension: initial DBP 590 mm Hg with an increase of at least 25 mm Hg to DBP >90 mm Hg, or initial DBP >90 mm Hg with an increase of at least 15 mm Hg (irrespective of the use of antihypertensive agents). 3. Proteinuric pregnancy-induced hypertension: blood pressure changes as above, with the additional finding of more than trace protein (equivalent to more than 300 mg/24 hr on urine collection). 4. Renal disease: renal structural or histologic abnormalities or abnormal function demonstrated radiologically or by biopsy before or at any time after pregnancy. Those patients in whom proteinuric pregnancy-induced hypertension developed (definition 3) or in whom anticonvulsive therapy (magnesium sulfate) was initiated because of severe disease and/or impending eclampsia, were allocated to group la (hypertensive complications, n = 24). The remaining patients in the study group were considered to have no hypertensive complications (group lb, n = 118). The remaining 134 patients served as controls. These were further divided into three groups: group 2, those at high risk for reasons other than preeclampsia or intrauterine growth retardation (IUGR) (patients with imminent preterm labor, pregnancy with history of infertility, and other medical complications, such as thyroid, cardiac, or pulmonary disease); group 3, 78 healthy pregnant women; and group 4, 28 nonpregnant volunteers who understood the purpose of the study. All pregnant controls underwent I-hour overnight urine col-

221 lection at intervals identical to those used for the study group. Nonpregnant control results were obtained after three consecutive I-hour overnight collections. Determination of urinary AERs was performed by radioimmunoassay, as described by Christensen and Orskov,23 with a modification used in our laboratory.24 Antihuman albumin was purchased from Sigma (Holon, Israel). Specificity of antibodies, tested by immunoelectrophoresis technique versus human serum and human albumin, was proven to be monospecific for human albumin. We used thymol as a presetvative; thymol had no effect on the radioimmunoassay. Results were expressed in milligrams per 24 hours by multiplying the results of S-hour overnight collection by 3 for simplification. Linear regression analysis indicated that ABRs expressed as milligrams per 24 hours were highly correlated with values expressed as micrograms per milliliters in the same urine samples (n = 276; r = 0.903, P = 0.001). The confidence interval in our laboratory for nonpregnant women is 0,13.4 mg/24 hr; for healthy pregnant women after 29 weeks’ gestation (third trimester), it is 11.8, 16.1 mg/24 hr. The intra-assay and interassay coefficients of variation are 6% and 5.6%, respectively, with a sensitivity of 1.5 mg/ L or 2 mg124 hr obtained for urine samples with albumin concentrations of 20 and 100 mg/L. A midstream specimen of urine sediment and culture was examined to exclude urinary infection or bleeding as a cause of microalbuminuria. Blood pressure was measured by a specially trained nurse using a standard sphygmomanometer. The recorded blood pressure was the mean of three measurements taken on the right arm, after the subject was supine for 5 minutes. The fifth Korotkoff sound was used as an indicator of DBP.

Statistical Analysis Urinary ABR measurements in the study and control groups were not distributed normally, and therefore the oneway nonparametric Km&al-Wallis test was used. Albumin excretion rates were compared between the study group (group 1) and the controls (groups 2, 3, and 4) and between trimesters of pregnancy within groups. To compare antenatal and postnatal data of the study subgroups (la and lb), Student’s t and chi-square tests were used. To predict hypertensive complications and IUGR in the study group by serial measurements of AERs, stepwise logistic regression models were fitted. To predict birth weight and gestational week of delivery by AERs, linear regression models were fitted to the data. P 5 0.05 was considered statistically significant. RESULTS

Clinical details of the patients and controls are given in Table 1. Table 2 presents the characteristics and antenatal and postnatal data of the hypertensive subjects. The patients who had hypertensive complications (group la) were significantly older than those who did not (group lb) and had a significantly worse maternal and neonatal outcome, as determined by the variables in Table 2. Table 3 demonstrates the significant differ-

BAR

222

Table 1. Characteristics No. of Patients Group 1: high complications

Group Group Group

risk of hypertensive

2: high risk for other 3: normal pregnant 4: not pregnant

Abbreviation;

reasons

PIH, pregnancy-induced

of the Study and Control Groups Age, yr (mean 2 SD)

142

31 t

28 78 28

5

23%

Subgroups

(%)

Group la Proteinuric PIH Group 1b Chronic hypertension Chronic renal disease Nonproteinuric PIH

17% 47% 19% 18%

27% 24%

hypertension.

and Antenatal

ematical formula used to calculate the chances of a high-risk patient having hypertensive complications is given in Table 4, which indicates that the risk of hypertensive complications is predictable by early third-trimester AERs (ALB31) and systolic blood pressure @BP) values. A natural logarithm (In) of AERs was used to fit the results to the regression model. Any increase in a unit of In (ALB31) resulted in an odds ratio of 2.1 to develop hypertensive complications. For example, for a patient with a SBP of 140 mm Hg, an increase in AERs from 2.7 to 54.6 mg/ 24 hr resulted in a significant increase in the risk of hypertensive complications from 26% to 6 1%. Albumin excretion rates that were measured at

and Postnatal

Patients in Whom Hypertensive Complications Developed (Group la; n = 24) Mean age (yr) Parity Nulliparity No. of days hospitalization before delivery Antihypertensive treatment Maximum systolic blood pressure (mm W Maximum diastolic blood pressure (mm Hg) Maximum uric acid (mg/dL) Cesarean section for severe hypertension or fetal distress Preterm deliveries (~37 wk) IUGR Hospitalization in a neonatal unit

Percentage of Nulliparity

31 t5 29 t 6 30 t 8

ences in AERs between the study and control groups in the second (P = 0.0087) and third (P = 0.0001) trimesters of pregnancy; no significant differences were noted in the first trimester (P = 0.91). Normal pregnant women showed a significant increase in AERs in the third trimester compared with the second (P = 0.016). This was also observed in the hypertensive group (P = 0.007). The mean AER of the nonpregnant group was 6.1 + 3.3 mg/24 hr (3.7 + 1.9 pg/mL) in our laboratory. Using a logistic regression model for AERs in the early third trimester, we predicted hypertensive complications in the study group (odds ratio, 2.1; confidence interval, 1.26 to 3.53). The math-

Table 2. Characteristics

ET AL

36.2 + 4.4 2.9 2 2.6 21%

Data of the Study Subgroups Patients Wiih No Hypertensive Complications (Group lb; n = 118)

Probability Value

32 ? 4.9 2.1 + 1.8 15%

0.0002 NS NS

1.9 2 3.2 10%

0.0001

153 2 36

124 2 32

0.0003

104 5 10 6.1 +- 1.5

84 2 12 5.1 2 1.5

0.0000

20% 62% 55% 53%

4% 22% 18% 9%

0.008 0.0002 0.0001 0.0008

8.5

2 10.7 78%

0.0000

0.006

MICROALBUMINURIA

IN HYPERTENSIVE

PREGNANCY

mean gestational week 30.9 ? 3.2 in the study group definitely anteceded the appearance of clinical proteinuria (>300 mg/24 hr), detected at mean gestational week 34 + 3.3, and mean week of delivery (37.6 + 3.2). Week of delivery was 35.3 2 3.5 in group la and 38.2 + 2.5 in group lb. Albumin excretion rates in the third trimester of pregnancy were nine times higher in pregnancies complicated by IUGR than in pregnancies with normal fetal growth: 411 rt 876 mg/24 hr v 48 + 84 mg/24 hr, respectively (P = 0.088). Fitting a model to predict IUGR by third-trimester AERs resulted in a statistically nonsignificant prediction (P = 0.25) with odds ratio = 1.002. Using a multiple linear regression stepwise model, early third-trimester AERs, gestational week, and DBP were able to significantly predict birth weight (R’ = 0.7, P < 0.05) using the following formula: Birth weight = -2,969.18

+ 185.63 X (gestational week)

X 10.61 (DBP early third trimester)

(AERs of early third trimester) However, third-trimester AERs could not significantly predict gestational week of delivery, nor could second- or third-trimester AERs pre-

Table 3. Albumin Excretion Rate Throughout Pregnancy in Patient Group

Study group High-risk pregnancy NOWlld pregnancy

Second Trimester

Third Trimester

21.0 * 37.9’^’ (13.4 2 23.3)

23.7 2 42.3" (14.2 2 24.2)

125.7 + 423St (71.6 _f 241.8)

13.6 + 6.5”’ (6.5 + 4.0)

10.6 -c 3.4" (6.4 + 2.0)

15.2 + 7.7’tt (6.7 2 5.2)

9.0 2 4.1" (5.4 k 2.5)

14.0 c 7.3’ttt (6.0 2 4.2)

12.4 2 6.3"' (7.6 -t 3.9)

Table 4. Stepwise Logistic Regression Analysis to Calculate Patient Risk of Hypertensive Complications

ALB31 (mg/24 hr)

In (ALB31) (mg/24 hr)

2.72 7.39 20.09 54.60 148.41 403.43

1 2 3 4 5 6

14 26 43 61 77 88

66 81 90 95 98 99

Abbreviations: P, calculated patient risk of hypertensive complications; P*, calculated risk for SBP of 140 mm Hg; Pt, calculated risk for a SBP of 160 mm Hg; SBP, systolic blood pressure (mm Hg); ALB31, albumin excretion rate measured in early third trimester; In (ALB31), used to fit results to the regression model by natural logarithm (In).

diet neonatal outcome (Apgar score at 5 minutes, fetal distress in labor, and hospitalization in a neonatal special care unit). DISCUSSION

- 125.81 In (ALB31)

First Trimester

223

NOTE. Data are expressed as mean values 2 SD in mg/24 hr (&ml). Probability values (calculated using the Kurskal-Wallis ANOVA): *differences in third trimester between study. high-risk, and normal pregnant patients (P = 0.0001); “differences in second trimeater between study, highrisk, and normal pregnant patients (P = 0.0067); “‘differences in first trimester between study, high-risk, and normal pregnant patients (P = NS): tdifferences in AER between third and second trimesters in the study group (P = 0.007): ttdifferences in AER between third and second trimesters in the high-risk for other reasons group (P = 0.02); tttdifferences in AER between third and second trimesters in the normal pregnant group (P = 0.016).

In all pregnant patients studied, there was a significant increase in urinary albumin excretion in the third trimester compared with the second, and in the second trimester compared with the first. A higher urinary AER seems to be a physiologic effect of pregnancy.9*20Y2’*24,25 This can be explained by the increase in glomerular filtration rate as well as the increase in prostacyclin production and angiotensin II production, and even renal tubular effects.26V28However, this study is the first demonstration of the ability of AERs to predict hypertensive complications in high-risk pregnant women. A major increase in urinary AERs was observed in the high-risk hypertensive group in the second and third trimesters of pregnancy. The difference in the mean AERs between the study and control groups was more pronounced in the third trimester, but still significant in the second trimester. In the pregnant patients at high risk for reasons other than hypertension, as well as in the normal pregnant women, the increase was less significant. Our results are in contrast to the finding of Lopez Espinoza et a12’ and Konstantin-Hansen et aL2’ who failed to demonstrate an early increase in AERs in chronic

BAR

224

hypertensive and preeclamptic pregnant women. Specifically, using both logistic and linear regression models, microalbuminuria in the early third trimester indicated worse pregnancy outcome. Results were significant even with our very broad inclusion criteria. Based on our results, it seems that in highrisk patients a microalbuminuric phase probably precedes the clinical proteinuric phase, even as early as the beginning of the third trimester. This contrasts with other earlier assumptions that renal involvement in preeclampsia can develop with a variable time course.3 We believe this difference in findings was due to the larger high-risk and control groups used here (n = 142 v 134, respectively), the frequent urinary albumin measurements, the use of 8hour overnight urine collection, and the adjustment of appropriate statistical models. When SBP was 140 mm Hg, any increase in AERs resulted in a significantly higher risk of hypertensive complications. When SBP was higher, a higher risk was detected even at low AERs (Table 4). The correlation between AERs and SBP was poor (r = 0.098). However, when analyzed separately, SBP had a slightly higher significance than AERs in predicting severe disease (P = 0.0001 v P = 0.0004,respectively). This may be explained by the assumption that SBP and AER are independent variables in predicting severe disease in the early third trimester of pregnancy. The prediction of IUGR and other neonatal parameters by AERs was less significant. This may have been due to the presence of other intrapartum or obscure antepartum parameters. Nevertheless, mean AERs in the third trimester were ninefold higher in the IUGR neonates than in the appropriate-for-gestational-age neonates (411 + 876 mgl24 hr v 48 2 84 mg/24 hr). However, prediction was not significant, probably because of the high standard deviation and small percentage of IUGR cases (14%). With the introduction of radioimmunoassays for microalbumin determination and the clinical significance of such measurements in hypertensive pregnant patients, perhaps we should reconsider the accepted definition of clinical gestational proteinuria (ie, 300 mg/24 hr), which is based on less sensitive methods and usually is a late sign of severe disease. From a practical point of view, early detection

ET AL

of high-risk patients by an efficient and simple predictive test will enhance surveillance, such as by weekly or biweekly visits to the hospital. Twenty-four-hour blood pressure monitoring to optimize use of antihypertensive drugs may be offered in complicated cases, although it has never been decisively proven to improve outcome. Fetal growth and well-being need to be more frequently assessed by daily to weekly electronic monitoring and biweekly ultrasound for biophysical profile and every 10 to 14 days for growth. This study has confirmed that patients in whom abnormal proteinuria develops usually have a microalbuminuric phase weeks earlier, and this test has some predictive value for severe disease. Further study is needed to establish whether special care in such patients may improve perinatal outcome. REFERENCES 1. Chasley LC: Hypertensive Disorders of Pregnancy. Norwalk, CT, Appleton-Century-Crofts, 1978, pp 157-162 2. Butler NR, Bonham DG: Perinatal Mortality: First Report of the British Perinatal Mortality Survey. Edinburgh, UK, Livingstone, 1963, pp 86-100 3. Redman GWG, Beilin LJ, Bonnar J: Renal function in preeclampsia. J Clin Path01 29:91-94, 1976 (suppl 10) 4. Pollad VE, Nettles JB: The kidney in toxemia of pregnancy. A clinical and pathologic study based on renal biopsies. Medicine 39:469-526, 1960 5. McEwan HP: Investigation of proteimuia in pregnancy by immunoelectrophoresis. Br J Obstet Gynaecol 75:283294, 1968 6. Redman

CWG, Beilin VJ, Wilkinson RH: Plasma urate measurements in predicting fetal death in hypertensive pregnancy. Lancet 1:1370-1371, 1976 7. Sayen N, Haram K, Nilsen ST: Serum urate as a predictor of fetal outcome in severe preeclampsia. Acta Obstet Gynecol Stand 63:71-75, 1984 8. Hill LM: Metabolism of uric acid in normal and toxemic pregnancy. Mayo Clin Proc 53:743-751, 1978 9. O’Brien WF: Predicting preeclampsia. Obstet Gynecol 75:445-452,

1990

10. Dekker GA, Sibai BM: Early detection of preeclampsia. Am J Obstet Gynecol 165:160-172, 1991 11. Viberti GC, Hill BD, Jarrett RJ, Argyropoulos A, Mahmud U, Keen H: Microalbuminuria as a predictor of clinical nephropathy in insulin dependent diabetes mellitus. Lancet 2:1430-1432, 1982 12. Parving HH, Oxenbull B, Svendsen PA, Christiansen JS, Andersen AR: Early detection of patients at risk of developing diabetic nephropathy. A longitudinal study of urinary albumin excretion. Acta Endocrinol (Cope@ 100:530-535, 1982 13. Mogensen CE, Christisen CK: Predicting diabetic ne-

MICROALBUMINURIA

IN HYPERTENSIVE

PREGNANCY

phropathy in insulin dependent patients. N Engl J Med 311:89-93, 1984 14. Mathiensen ER, Oxenbull B, Johansen K, Svensen PA, Deckert T: Incipient nephropathy in type I (insulin-dependent) diabetes. Diabetologia 26:406-410, 1984 15. Parving HI-I, Jensen H, Mogensen CE, Ervin PE: Increased urinary albumin excretion rate in benign essential hypertension. Lancet 1:1190-l 192, 1974 16. Gosling P, Beevers PG: Urinary albumin excretion and blood pressure in the general population. Clin Sci 76:39-42, 1989 17. Pedersen EB, Mogensen CE: Effects of antihypertensive treatment on urinary albumin excretion, glomerular filtration rate. Stand J Clin Lab Invest 36:231-237, 1976 18. Bar J, Hod M, Erman A, Friedman S, Ovadia J: Microalbuminuria: Prognostic and therapeutic implications in diabetic and hypertensive pregnancy. Diabet Med 12:649656, 1995 19. Friedman S, Rabinetson D, Bar J, Erman A, Hod M, Kaplan B, Boner G, Gvadia J: Microalbuminuria following gestational diabetes. Acta Obstet Gynecol Stand 74356360.1995 20. Lopez Espinoza I, Dhar H, Humphreys S, Redman CWG: Urinary albumin excretion in pregnancy. Br J Obstet Gynaecol 93:176-181, 1986 21. Konstantin-Hansen KF, Hesseldahl H, Podersen SM: Microalbuminuria as a predictor of preeclampsia. Acta Obstet Gynecol Stand 71:343-346, 1992

225 22. CLASP Collaborative Group: CLASP-A randomised trial of low-dose aspirin for the prevention and treatment of preeclampsia among 9364 pregnant women. Lancet 343:619-629, 1994 23. Christensen C, Orskov C: Rapid screening PEG radioimmunoassay for qualification of pathological microalbuminuria. Dial Nephrol 3:92-94, 1984 24. Erman A, Neri A, Sharoni R: Enhanced urinary albumin excretion after 35 weeks of gestation and during labour in normal pregnancy. Stand J Clin Lab Invest 52:409-413, 1992 25. Wright A, Steele P, Bennett JR, Watts G, Polak A: The urinary excretion of albumin in normal pregnancy. Br J Obstet Gynaecol 94:408-412, 1987 26. Mogensen CE: Kidney function and glomerular permeability to macromolecules in early juvenile diabetes mellitus. Stand J Clin Lab Invest 28:79-90, 1971 27. Hakim RM, Goldszer RC, Brenner BM: Hypertension and proteinuria: Long-term sequelae of uninephrectomy in humans. Kidney Int 25:930-936, 1984 28. Brenner BM, Meyer TW, Hostetter TH: Dietary protein intake and the progressive nature of kidney disease: The role of a haemodynamically-mediated glomerular injury in the pathogenesis of progressive glomerular sclerosis in aging, renal ablation and intrinsic renal disease. N Engl J Med 307:652-659, 1982