- Email: [email protected]
The value of elevated second-trimester β-human chorionic gonadotropin in predicting development of preeclampsia
The value of elevated second-trimester [3-human chorionic gonadotropin in predicting development of preeclampsia Adnan M.N. Ashour, MB, BS, Ellice S. Lieberman, MI), DrPH, Louise E. W'~ins Haug, MD, PhD,
and John T. Repke, MD Boston, Massachusetts OBJECTIVE: Our purpose was to investigate the association of an elevated second-trimester serum IS-human chorionic gonadotropin concentration with the subsequent development of hypertension in pregnancy and to evaluate its utility as a screening test for preeclampsia. STUDY DESIGN: We examined 6286 nondiabetic women with singleton pregnancies who, as part of triple-screen testing, had a serum 18 human chorionic gonadotropin level drawn between 15 and 22 weeks' gestation between November 1, 1991, and November 30, 1994. Medical records of women with hypertension (n = 675) were reviewed, patients with chronic hypertension were excluded, and the remainder were classified as having gestational hypertension (n = 333), mild preeciampsia (n = 110), or severe preeclampsia (n = 84). The !8-human chorionic gonadotropin level expressed as multiples of the median adjusted for maternal weight and gestational age was compared between normotensive and hypertensive complicated pregnancies. RESULTS: In the overall population IS-human chorionic gonadotropin levels ->2.0 multiples of the median during the second trimester were significantly associated with development of hypertension in pregnancy. The rate ratio for development of overall hypertension was 1.6 (95% confidence interval 1.3 to 2.0) and for preeclampsia 1.8 (95% confidence interval 1.3 to 2.6). When stratified by parity, a statistically significant association remained only among multiparous women, for overall hypertension (rate ratio 2.2, 95% confidence interval 1.6 to 3.2) and for preeclampsia (rate ratio 3.4, 95% confidence interval 2.1 to 5.6). Adjusting for confounding factors did not alter the results, in the overall population, with the use of 2.0 multiples of the median of IS-human chorionic gonadotropin as a cutoff value, the sensitivity of !3-human chorionic gonadotropin as a screen for development of hypertension was 15.6%, the specificity was 90.0%, and the positive predictive value was 12.8%. CONCLUSION: Overall, second-trimester serum IS-human chorionic gonadotropin levels were elevated among women who had hypertension during pregnancy. In our population this association was statistically significant only among multiparous women. The utility of an elevated second-trimester IS-human chorionic gonadotropin level as a screening test for preeclampsia is limited. (Am J Obstet Gynecol 1997;176:438-42.)
Key words: [3-Human chorionic gonadotropin, gestational hypertension, preeclampsia
Preeclampsia is a disorder u n i q u e to h u m a n pregnancy, affecting 5% to 10% of pregnancies. The etiology of preeclampsia remains unclear in spite of extensive clinical and basic research. In spite of improvements in antenatal and neonatal care, preeclampsia remains a major cause of maternal and neonatal morbidity and mortality. Although the exact nature of the primary event causing preeclampsia is n o t known, evidence accumulated over the past few years indicates From the Division of Maternal-Fetal Medicine and the Department of Obstetrics and Gynecology, Brigham and Women's Hospital and Harvard Medical School. Received for publication June 25, 1996; revised August 7, 1996; accepted September 19, 1996. Reprint requests:John T. Repke, MD, Brigham and Women's Hospital, Department of Obstetrics and Gynecology, Division of Maternal-Fetal Medicine, CWN584.4, 75Francis St., Boston, MA 02115. Copyright © 1997 by Mosby-Year Book, Inc. 0002-9378/97 ~5.00 + 0 6/1/78132 438
that abnormal placentation may be one of the initial events in the disease process. TM Several tests have been proposed to identify p r e g n a n t women at risk for development of preeclampsia, including the angiotensin sensitivity test, 5'6 the roll-over test, 7'8 secondtrimester Doppler studies, 9-12 and plasma fibronectin. ~3-15 None of these tests have been widely accepted by the obstetric community because of their low predictive value or, for some tests, their invasive and time-consuming nature. Several studies have demonstrated elevated third-trimester levels of h u m a n chorionic g o n a d o t r o p i n (hCG) and [3-hCG in patients with preeclampsia. ~6-a9 Others have shown patients with elevated second-trimester hCG or 13-hCG levels to be at higher risk for the development of pregnancyi n d u c e d hypertension 2°'21 and poor pregnancy outcome.22, 23 The aim of our study was to investigate the association
Volume 176, Number2 AmJ Obstet Gynecol
of elevated second-trimester [3-hCG levels with the subsequent development of hypertension in pregnancy and to evaluate its utility as a screening test for later development of preeclampsia.
Material and methods There were 7296 women with singleton pregnancies who had a triple-panel screen (maternal serum [3-hCG, unconjugated estriol, and c~-fetoprotein levels) for neural tubal defects and Down syndrome during the second trimester (15 to 22 weeks) between Nov. 1, 1991, and Nov. 30, 1994, at the antenatal diagnostic center at Brigham and Women's Hospital. Of those women, 298 had more than one pregnancy during that period and for these women only the first pregnancy on record was included in the current analysis (n = 6998). Pregnancies with known fetal chromosomal abnormalities diagnosed during or after pregnancy or major fetal anomalies detected by ultrasonography (n = 14), maternal diabetes (n = 99), or ending in abortion (n = 24) were excluded. Of the remaining 6861 pregnancies, 6286 (91.6%) were delivered at Brigham and Women's Hospital and therefore had adequate follow-up data. The remaining 575 women (8.4%) were delivered elsewhere and were excluded. Approval for medical records review was obtained from the institution's h u m a n research committee. By use of hospital discharge data, we determined that 675 of 6286 women (10.7%) had hypertension coded in the discharge diagnosis. Records on these women were reviewed and classified into chronic hypertension with or without superimposed pregnancy-induced hypertension (n = 148), gestational hypertension (n = 333), mild preeclampsia (n = 110); and severe preeclampsia or eclampsia (n = 84). Women with chronic hypertension defined as the presence of persistent hypertension before 20 weeks of gestation or persistent hypertension beyond 6 weeks post partum with or without superimposed pregnancy-induced hypertension were excluded from the analysis. The remaining 6138 women constituted our study population. Gestational hypertension was defined as a diastolic blood pressure ->90 mm Hg or a systolic pressure ->140 mm Hg on at least two occasions -->6 hours apart in the absence of significant proteinuria. Preeclampsia was defined as the development of hypertension with proteinuria (presence of ->300 mg of protein in a 24-hour urine collection or -> + 1 protein in at least two random urine specimens collected ---6 hours apart). Preeclampsia was regarded as mild unless one or more of the following signs or symptoms were present, in which case it was classified as severe: blood pressure readings ->160 mm Hg systolic or 110 mm Hg diastolic ->6 hours apart, proteinuria ->5 gm in a 24-hour collection or 3+ to 4+ on a dipstick random sample, oliguria (24-hour urinary output <400 ml), cerebral or visual disturbances, pulmo-
Ashour et al.
439
Table I. Characteristics of study populations Characteristic
Parity* Nulliparous Multiparous Unknown Ethnicity(%) *,t White Black Other Maternal age (yr)*,+,+ Birth weight (gin)*,%+ Gestational age (wk)*,t,++
No hypertension (n = 5611)
Hypertension (n = 527)
2392 (42.6%) 2429 (43.3%) 790 (14.1%)
345 (65.5%) 182 (34.5%) 0
2661 (47.8%) 232 (45.6%) 1103 (19.8%) 125 (24.6%) 1804 (32.4%) 152 (29.9%) 28.2 (5.3) 27.6 (5.7) 3358.6(564.3) 3083.0 (713.2) 39.5 (2.4) 38.6 (2.6)
*p < 0.05. tMissing data (ethnicity 61, birth weight 64, gestational age 29, and maternal age 8). +Value in parentheses is SD.
nary edema, epigastric or right-upper-quadrant pain, impaired liver function of unclear cause, or thrombocytopenia. Eclampsia was defined as the occurrence of convulsions, not caused by any coincidental neurologic disease, in a woman whose condition also met the criteria for preeclampsia. The maternal serum [3-hCG level was quantified by assay (Abbott IMX) and expressed as multiples of the median corrected for maternal weight and gestational age. Gestational age was calculated from the first day of the last menstrual period unless ultrasonography before 16 weeks demonstrated a discrepancy of -->10 days, in which case ultrasonographic dating of the pregnancy was used for calculations. With use of various [3-hCG cutoff values, risk ratios were calculated for the development of overall hypertension, for preeclampsia, and for severe preeclampsia in the overall population as well as separately for nulliparous and multiparous patients. We examined maternal and neonatal characteristics by hypertension status. Logistic regression was used to adjust for maternal age and black race. Power calculations were based on hospital statistics showing a 12.2% rate of pregnancy-related hypertension in the overall population, 15% among nulliparous women, 8% among multiparous women, and a 10% rate of [3-hCG multiples of the median ~2.0. The projected n u m b e r of women needed to detect a 50% increase in the rate of hypertension in pregnancies With 13-hCG multiples of the median -->2.0 in the overall population was 3200 and to detect a 75% difference was 1570 women. The projected n u m b e r of nulliparous women needed to detect a 50% and a 75% difference were 2480 and 1230, respectively. The projected n u m b e r of multiparous women needed was 5150 to detect a 50% difference and 2510 to detect a 75% difference. All power calculations assumed c~ = 0,05 (two-sided), [3 = 0.20, and power = 0.80.
440
Ashour et al.
Februaly1997 AmJ Obstet Gynecol
Table II. Association between maternal serum [3-hCG levels and hypertension status in overall population All hypertension* (n = 527) ~-hCG level (multiples of median)
<2.0 >--2.0 2.0-2.99 >-3.0
Gestational hypertension (n = 333)
Preeclampsia (n = 194)
No.
% and No.
RRand 95%CI
% and No.
RR and 95%CI
% and No.
RRand 95%CI
5496 642 503 139
8.1 (445) 12.8 (82) 12.9 (65) 12.2 (17)
1.0 1.6 (1.3-2.0) 1.6 (1.3-2.1) 1.5 (1.0-2.4)
5.2 (285) 7.5 (48) 8.0 (40) 5.8 (8)
1.0 1.5 (1.1-2.0) 1.6 (1.1-2.2) 1.1 (0.6-2.2)
2.9 (160) 5.3 (34) 5.0 (25) 6.5 (9)
1.0 1.8 (1.3-2.6) 1.7 (1.1-2.6) 2.2 (1.2-4.3)
RR, Risk ratio; CI, confidence interval. *Gestational hypertension and preeclampsia.
Table III. Association between maternal serum [~-hCG levels and hypertension status in multiparous women All hypertension* (n = 182)
Gestational hypertension (n = 107)
Preeclampsia (n = 75)
~-hCG ~vel (mu~@~sofme~aN
No.
% and No.
R.Rand 95%CI
% and No.
RRand 95 % CI
% and No.
I
RR and 95%CI
<2.0 >-2.0 2-2.99 ~3.0
2359 252 209 43
6.2 (147) 13.9 (35) 14.4 (30) 11.6 (5)
1.0 2.2 (1.6-3.2) 2.3 (1.6-3.3) 1.9 (0.84.3)
3.9 (92) 5.6 (15) 6.2 (13) 4.7 (2)
1.0 1.6 (1.0-2.8) 2.3 (1.6-3.3) 1.9 (0.8-4.3)
2.3 (55) 7.9 (20) 8.1 (17) 7.0 (3)
1.0 3.4 (2.1-5.6) 3.5 (2.1-5.9) 3.0 (1.0-9.2)
RR, Risk ratio; CI, confidence interval. *Gestational hypertension and preeclampsia.
Results Table I shows selected demographic and reproductive characteristics of women by hypertension status. Nulliparous women had higher rate of hypertension compared with multiparous women. Birth weights were lower in pregnancies with hypertension and these women were delivered at earlier gestational ages. A significant association between elevated secondtrimester [3-hCG levels and subsequent development of hypertension in pregnancy was observed in the overall population (Table II). When stratified by parity, only multiparous women retained a statistically significant association between elevated second-trimester [3-hCG multiples of the median and development of hypertension (Table III). In this group the association was of marginal statistical significance in patients who had gestational hypertension where with use of 2.0 multiples of the median of [3-hCG as a cutoff, the rate ratio was 1.6 (95% confidence interval 1.0 to 2.8). This association was more clearly demonstrated in patients who had preeclampsia and strongest in patients who had severe preeclampsia; the rate ratio with use of 2.0 multiples of the median of [3-hCG as the cutoff value was 5.2 (95% confidence interval 2.5 to 10.6). Although the risk of pregnancy-related hypertension among nulliparous women was approximately 30%
higher in the presence of an elevated second-trimester [3-hCG level (Table IV), this association was not statistically significant. A logistic regression model to correct for increasing prevalence of the disease with increased maternal age and black race did not alter the association in either the multiparous or the nulliparous group. Table V shows the sensitivity, specificity, positive predictive value, and negative predictive value for 2.0 and 3.0 multiples of the median of [3-hCG in predicting overall hypertension, preeclampsia, and severe preeclampsia. At a [3-hCG multiple of the median level of 2.0, the sensitivity for predicting hypertension in the overall population was 15.6% with a positive predictive value of 12.8%. For predicting preeclampsia, the sensitivity was 17.5% with a positive predictive value of 5.3%. The sensitivity for predicting severe preeclampsia was 21.4% with a positive predictive value of 2.8%. The sensitivity for predicting preeclampsia at a cutoff value of 2.0 multiples of the median of [3-hCG was better for muhiparous women and best for predicting severe preeclampsia in this group (sensitivity 35.5%). In nulliparons women, at a cutoffvalue of 2.0 multiples of the median of [3-hCG, the sensitivity for predicting severe preeclampsia was only 13.2%. The use of other [3-hCG multiple of the median levels
Volume 176, Number 2 AmJ Obstet Gynecol
Ashour et al,
441
Table IV. Association between maternal serum [3-hCG levels and hypertension status in nulliparous women All hypertension* (n = 345) ~-hCG level (multiples of median)
No.
% and No.
RR and 95 % CI
<2.0 >-2.0 2.0-2.99 >-3.0
2433 304 232 72
12.2(298) 15.5(47) 15.1(35) 16.7(12)
1.0 1.3(0.9-1.7) 1.2(0.9-1.7) 1.4(0.8-2.3)
Gestational hypertension (n = 226)
Preedampsia (n = 119)
RR and 95% C1
% and No.
7.9(193) 10.9(33) 11.6(27) 8.3(6)
1.0 1.4(1.0-1.9) 1.5(1.0-2.1) 1.1(0.5-2.3)
% and No.
RR and 95%CI
4.3(105) 4.6(14) 3.4(8) 8.3(6)
1.0 1.1(0.6-1.8) 0.8(0.41.6) 1.9(0.9-4.2)
RR, Risk ratio; C/, confidence interval. *Gestational hypertension and preeclampsia.
Table V. Value of second-trimester serum [3-hCG values for prediction of hypertension by [3-hCG multiples of median >-2.0 ~-hCG multiples of median Parity
Overall population Sensitivity Specificity PPV NPV Nulliparous Sensitivity Specificity PPV NPV Multiparous Sensitivity Specificity PPV NPV
>-3.0 ~-hCG multiples of median
Overall hypertension
Gestational hypertension
PET
Severe PET
Overall hypertension
Gestational hypertension
PET
SeverePET
15.6 90.0 12.8 91.9
14.4 90.0 7.9 94.7
17.5 89.8 5.3 97.1
21.4 89.7 2.8 98.8
3.2 97.8 12.2 91.5
2.4 97.8 6.2 94.4
4.6 97.8 6.5 96.9
6.0 97.8 3.6 98.7
13.6 89.3 "15.5 87.8
14.6 89.3 11.4 91.7
11.8 88.9 4.6 95.7
13.2 88.9 2.3 98.1
3.5 97.5 16.7 87.5
2.7 97.5 9.1 91.4
5.0 97.5 8.3 95.8
5.7 97.4 4.2 98.1
19.2 91.1 13.9 93.8
14.0 91.1 6.5 96.0
26.7 90.9 7.9 97.7
35.5 90.7 4.4 99.2
2.7 98.4 11.6 93.1
1.9 98.4 5.0 95.8
4.0 98.4 7.0 97.2
6.5 98.4 4.7 98.9
PET, Preeclampsia; PPV, positive predictive value; NPV, negative predictive value.
did not improve the sensitivity and predictive value for predicting pregnancy-induced hypertension or preeclampsia in any part of the population.
Comment Our data confirm previous reports that demonstrated an association between elevated second-trimester maternal serum [3-hCG levels and the subsequent development of pregnancy-induced hypertension and preeclampsia.20, 21 Our study differs from other reports because we evaluated this association among multiparous and nulliparous women separately. Stratification by parity demonstrated this association to be statistically significant only among multiparous women. In this group the association of elevated second-trimester serum [3-hCG levels and tile later development of hypertension in pregnancy was stronger in patients with preeclampsia and was strongest among women who had severe preeclampsia. Although the risk of development of hypertension was approximately 30% higher in the presence of elevated
second-trimester [3-hCG levels among nulliparous women, this association was not statistically significant. Our power was adequate to detect a 50% increase in the risk of hypertension in pregnancies with multiple of the median [3-hCG levels ->2.0; however, we would have needed 7018 nulliparous women to have adequate power (80%) to demonstrate a 30% increase in the risk for the development of hypertension in this group. Fisher et al. 24 demonstrated that when nulliparous patients with hypertension during pregnancy underwent kidney biopsy, 75% of them showed glomerular endotheliosis, 16.3% had chronic renal lesions, and 6.7% had both preeclampsia and chronic renal disease. When the same study was carried out in multiparous patients with hypertension in pregnancy, the characteristic histologic lesion of preeclampsia was found in only 23%, chronic renal disease was present in 51%, chronic renal lesions plus preeclampsia were found in 13%, and normal histologic features were found in 11%. A possible explanation for the stronger association between elevated [3-hCG levels and later development of hypertension in
442
Ashour et al.
multiparous w o m e n is an increased i n c i d e n c e of underlying renal disease resulting in a decreased renal excretion rather than increased secretion loy the trophoblast. Sorensen et al. 21 d e m o n s t r a t e d an increased risk of preeclampsia in w o m e n with increased second-trimester hCG levels and suggested that a second-trimester screening p r o g r a m m i g h t be d e v e l o p e d to assess those at risk for pregnancy-induced hypertension. Second-trimester maternal serum screening for fetal aneuploidy and neural tubal defects usually includes h C G or [~-hCG is currently a widely p e r f o r m e d test. Thus second-trimester serum [~-hCG m e a s u r e m e n t would be a potentially costeffective tool for identifying w o m e n at risk for the d e v e l o p m e n t of preeclampsia. Such a screening p r o g r a m would facilitate closer surveillance and possible intervention with low-dose aspirin or calcium supplementation. A l t h o u g h our data demonstrate that elevated secondtrimester [3-hCG levels is associated with an increased risk for the d e v e l o p m e n t of hypertension in pregnancy, especially a m o n g multiparous women, the p o o r sensitivity and predictive value of the test would suggest that its use as a screening test for later d e v e l o p m e n t of gestational hypertension and preeclampsia is limited. REFERENCES
1. Robertson WB, Khong TY, Brosens I, De Wolf F, Sheppard BL, Bonnar J. The placental bed biopsy': review from three European centers. Am J Obstet Gynecol 1986;155:401-12. 2. Roberts JM, Robert NT, Musci TJ, Rodgers GM, Hubel CA, McLaughin MFL Preeclampsia: an endothelial cell disorder: Am J Obstet Gynecol 1989;161:1200-4. 3. Kaar K, Jouppila P, Kuikka J, Luotola H, Toivanen J, Rekonen A. Intervillous blood flow in normal and complicated late pregnancy measured by means of an intravenous 133 XE method. Acta Obstet Gynecol Scand 1980;59:7-10. 4. Frusca T, Morassi L, Pecorelli S, Grigolato P, Gastaldi A. Histological features of uteroplacental vessels in normal and hypertensive patients in relation to birth weight. BrJ Obstet Gypaecol 1989;96:835-9. 5. MorrisJA, O'GradyJP, Hamilton CJ, Davidson EC. Vascular reactivity to angiotensin II infusion during gestation. Am J Obstet Gynecol 1978;130:379-84. 6. Oney T, Kaulhausen H. The value of the angiotensin sensitivity test in the early diagnosis of hypertensive disorders in pregnancy. A m J Obstet Gynecol 1982;142:17-20. 7, GudsonJPJr, Andreson SG, May ~¢~. A clinical evaluation of the roll-over test for pregnancy-induced hypertension. AmJ Obstet Gynecol 1977;127:1-3. 8. Kuntz WD. Supine pressor (roll-over) test: an evaluation. AmJ Obstet Gynecol 1980;137:764-8. 9. Bower S, Bewley S, Campbell S. Improved prediction of
February 1997 Am J Obstet Gynecol
10. 11. 12.
13.
14.
15. 16. 17. 18.
19.
20.
21.
22.
23.
24.
preeclampsia by two stage screening of uterine arteries using the early diastolic notch and color Doppler imaging. Obstet Gynecol 1993;82:78-83. McParand P, Pearce JM, Chamberlain GVP. Doppler ultrasound and aspirin in recognition and prevention of pregnancy-induced hypertension. Lancet 1990;335:1552-5. Steel SA, PearceJM, McParland P, Chamberlain GVP. Early Doppler ultrasound screening in prediction of hypertensive disorders of pregnancy. Lancet 1990;335:1548-51. Jacobson SL, Imhof R, Manning N, Mannion V, Little D, Rey E, et al. The value of Doppler assessment of the uteroplacental circulation in predicting preeclampsia or intrauterine growth retardation. Am J Obstet Gynecol 1990;162: 110-4. Lazarachick J, Stubbs TM, Romein L, Van Dorsten JP, Loadholt CB. Predictive value of fibronectin levels in normotensive gravid women destined to become preeclamptic. A m J Obstet Gynecol 1986;154:1050-2. Stubbs TM, Lazarchick J, Horger EO. Plasma fibronectin levels in preeclampsia: a possible biochemical marker for vascular endothelial damage. Am J Obstet Gynecol 1984; 150:885-7. Lockwood CJ, Peters JH. Increased plasma levels of EDI+ cellular fibronectin precede the clinical signs of preeclampsia. Am J Obstet Gynecol 1990;162:358-62. Smith GVS, Smith OW. The anterior pituitary-like hormone in late pregnancy toxemia. Am J Obstet Gynecol 1939;38:618-24. Smith GVS, Smith OW. Excessive gonad-stimulating hormone and subnormal amounts of oestrin in the toxemias of late pregnancy. Am J Physiol 1934;107:128-45. Said ME, Campbell DM, Azzam ME, MacGillivray I. Betahuman chorionic gonadotrophin levels before and after the development of preeclampsia. Br J Obstet Gynaecol 1984; 91:772-5. Hsu CD, Chan DW, Iriye B, Johnson TRB, Hong SF, Repke JT. Elevated serum human chorionic gonadotropin as evidence of secretory response in severe preeclampsia. Am J Obstet Gynecol 1994;170:1135-8. Gonen R, Perez R, David M, Dar H, Merksamer R, Sharf M. The association between unexplained second-trimester maternal serum hCG elevations and pregnancy induced complications. Obstet Gynecol 1992;80:83-6. Sorensen TK, Williams MA, Zingheim RW, Clement SJ, Hickok DE. Elevated second-trimester human chorionic gonadotropin and subsequent pregnancy-induced hypertension. Am J Obstet Gynecol 1993;169:834-8. Gravett CP, Buckmaster JG, Watson PT, Gravett MG. Elevated second trimester maternal serum [3-HCG concentrations and subsequent adverse pregnancy outcome. Am J Med Genet 1992;44:485-6. Wenstrom ILl2), Owen J, Boots LR, DuBard MB. Elevated second-trimester human chorionic gonadotropin levels in association with poor pregnancy outcome. Am J Obstet Gynecol 1994;171:1038-41. Fisher KA, Luger A, Spargo BH, Lindheimer MD. Hypertension in pregnancy: clinical-pathological correlations and remote prognosis. Medicine (Baltimore) 1981;60:267-76.
and John T. Repke, MD Boston, Massachusetts OBJECTIVE: Our purpose was to investigate the association of an elevated second-trimester serum IS-human chorionic gonadotropin concentration with the subsequent development of hypertension in pregnancy and to evaluate its utility as a screening test for preeclampsia. STUDY DESIGN: We examined 6286 nondiabetic women with singleton pregnancies who, as part of triple-screen testing, had a serum 18 human chorionic gonadotropin level drawn between 15 and 22 weeks' gestation between November 1, 1991, and November 30, 1994. Medical records of women with hypertension (n = 675) were reviewed, patients with chronic hypertension were excluded, and the remainder were classified as having gestational hypertension (n = 333), mild preeciampsia (n = 110), or severe preeclampsia (n = 84). The !8-human chorionic gonadotropin level expressed as multiples of the median adjusted for maternal weight and gestational age was compared between normotensive and hypertensive complicated pregnancies. RESULTS: In the overall population IS-human chorionic gonadotropin levels ->2.0 multiples of the median during the second trimester were significantly associated with development of hypertension in pregnancy. The rate ratio for development of overall hypertension was 1.6 (95% confidence interval 1.3 to 2.0) and for preeclampsia 1.8 (95% confidence interval 1.3 to 2.6). When stratified by parity, a statistically significant association remained only among multiparous women, for overall hypertension (rate ratio 2.2, 95% confidence interval 1.6 to 3.2) and for preeclampsia (rate ratio 3.4, 95% confidence interval 2.1 to 5.6). Adjusting for confounding factors did not alter the results, in the overall population, with the use of 2.0 multiples of the median of IS-human chorionic gonadotropin as a cutoff value, the sensitivity of !3-human chorionic gonadotropin as a screen for development of hypertension was 15.6%, the specificity was 90.0%, and the positive predictive value was 12.8%. CONCLUSION: Overall, second-trimester serum IS-human chorionic gonadotropin levels were elevated among women who had hypertension during pregnancy. In our population this association was statistically significant only among multiparous women. The utility of an elevated second-trimester IS-human chorionic gonadotropin level as a screening test for preeclampsia is limited. (Am J Obstet Gynecol 1997;176:438-42.)
Key words: [3-Human chorionic gonadotropin, gestational hypertension, preeclampsia
Preeclampsia is a disorder u n i q u e to h u m a n pregnancy, affecting 5% to 10% of pregnancies. The etiology of preeclampsia remains unclear in spite of extensive clinical and basic research. In spite of improvements in antenatal and neonatal care, preeclampsia remains a major cause of maternal and neonatal morbidity and mortality. Although the exact nature of the primary event causing preeclampsia is n o t known, evidence accumulated over the past few years indicates From the Division of Maternal-Fetal Medicine and the Department of Obstetrics and Gynecology, Brigham and Women's Hospital and Harvard Medical School. Received for publication June 25, 1996; revised August 7, 1996; accepted September 19, 1996. Reprint requests:John T. Repke, MD, Brigham and Women's Hospital, Department of Obstetrics and Gynecology, Division of Maternal-Fetal Medicine, CWN584.4, 75Francis St., Boston, MA 02115. Copyright © 1997 by Mosby-Year Book, Inc. 0002-9378/97 ~5.00 + 0 6/1/78132 438
that abnormal placentation may be one of the initial events in the disease process. TM Several tests have been proposed to identify p r e g n a n t women at risk for development of preeclampsia, including the angiotensin sensitivity test, 5'6 the roll-over test, 7'8 secondtrimester Doppler studies, 9-12 and plasma fibronectin. ~3-15 None of these tests have been widely accepted by the obstetric community because of their low predictive value or, for some tests, their invasive and time-consuming nature. Several studies have demonstrated elevated third-trimester levels of h u m a n chorionic g o n a d o t r o p i n (hCG) and [3-hCG in patients with preeclampsia. ~6-a9 Others have shown patients with elevated second-trimester hCG or 13-hCG levels to be at higher risk for the development of pregnancyi n d u c e d hypertension 2°'21 and poor pregnancy outcome.22, 23 The aim of our study was to investigate the association
Volume 176, Number2 AmJ Obstet Gynecol
of elevated second-trimester [3-hCG levels with the subsequent development of hypertension in pregnancy and to evaluate its utility as a screening test for later development of preeclampsia.
Material and methods There were 7296 women with singleton pregnancies who had a triple-panel screen (maternal serum [3-hCG, unconjugated estriol, and c~-fetoprotein levels) for neural tubal defects and Down syndrome during the second trimester (15 to 22 weeks) between Nov. 1, 1991, and Nov. 30, 1994, at the antenatal diagnostic center at Brigham and Women's Hospital. Of those women, 298 had more than one pregnancy during that period and for these women only the first pregnancy on record was included in the current analysis (n = 6998). Pregnancies with known fetal chromosomal abnormalities diagnosed during or after pregnancy or major fetal anomalies detected by ultrasonography (n = 14), maternal diabetes (n = 99), or ending in abortion (n = 24) were excluded. Of the remaining 6861 pregnancies, 6286 (91.6%) were delivered at Brigham and Women's Hospital and therefore had adequate follow-up data. The remaining 575 women (8.4%) were delivered elsewhere and were excluded. Approval for medical records review was obtained from the institution's h u m a n research committee. By use of hospital discharge data, we determined that 675 of 6286 women (10.7%) had hypertension coded in the discharge diagnosis. Records on these women were reviewed and classified into chronic hypertension with or without superimposed pregnancy-induced hypertension (n = 148), gestational hypertension (n = 333), mild preeclampsia (n = 110); and severe preeclampsia or eclampsia (n = 84). Women with chronic hypertension defined as the presence of persistent hypertension before 20 weeks of gestation or persistent hypertension beyond 6 weeks post partum with or without superimposed pregnancy-induced hypertension were excluded from the analysis. The remaining 6138 women constituted our study population. Gestational hypertension was defined as a diastolic blood pressure ->90 mm Hg or a systolic pressure ->140 mm Hg on at least two occasions -->6 hours apart in the absence of significant proteinuria. Preeclampsia was defined as the development of hypertension with proteinuria (presence of ->300 mg of protein in a 24-hour urine collection or -> + 1 protein in at least two random urine specimens collected ---6 hours apart). Preeclampsia was regarded as mild unless one or more of the following signs or symptoms were present, in which case it was classified as severe: blood pressure readings ->160 mm Hg systolic or 110 mm Hg diastolic ->6 hours apart, proteinuria ->5 gm in a 24-hour collection or 3+ to 4+ on a dipstick random sample, oliguria (24-hour urinary output <400 ml), cerebral or visual disturbances, pulmo-
Ashour et al.
439
Table I. Characteristics of study populations Characteristic
Parity* Nulliparous Multiparous Unknown Ethnicity(%) *,t White Black Other Maternal age (yr)*,+,+ Birth weight (gin)*,%+ Gestational age (wk)*,t,++
No hypertension (n = 5611)
Hypertension (n = 527)
2392 (42.6%) 2429 (43.3%) 790 (14.1%)
345 (65.5%) 182 (34.5%) 0
2661 (47.8%) 232 (45.6%) 1103 (19.8%) 125 (24.6%) 1804 (32.4%) 152 (29.9%) 28.2 (5.3) 27.6 (5.7) 3358.6(564.3) 3083.0 (713.2) 39.5 (2.4) 38.6 (2.6)
*p < 0.05. tMissing data (ethnicity 61, birth weight 64, gestational age 29, and maternal age 8). +Value in parentheses is SD.
nary edema, epigastric or right-upper-quadrant pain, impaired liver function of unclear cause, or thrombocytopenia. Eclampsia was defined as the occurrence of convulsions, not caused by any coincidental neurologic disease, in a woman whose condition also met the criteria for preeclampsia. The maternal serum [3-hCG level was quantified by assay (Abbott IMX) and expressed as multiples of the median corrected for maternal weight and gestational age. Gestational age was calculated from the first day of the last menstrual period unless ultrasonography before 16 weeks demonstrated a discrepancy of -->10 days, in which case ultrasonographic dating of the pregnancy was used for calculations. With use of various [3-hCG cutoff values, risk ratios were calculated for the development of overall hypertension, for preeclampsia, and for severe preeclampsia in the overall population as well as separately for nulliparous and multiparous patients. We examined maternal and neonatal characteristics by hypertension status. Logistic regression was used to adjust for maternal age and black race. Power calculations were based on hospital statistics showing a 12.2% rate of pregnancy-related hypertension in the overall population, 15% among nulliparous women, 8% among multiparous women, and a 10% rate of [3-hCG multiples of the median ~2.0. The projected n u m b e r of women needed to detect a 50% increase in the rate of hypertension in pregnancies With 13-hCG multiples of the median -->2.0 in the overall population was 3200 and to detect a 75% difference was 1570 women. The projected n u m b e r of nulliparous women needed to detect a 50% and a 75% difference were 2480 and 1230, respectively. The projected n u m b e r of multiparous women needed was 5150 to detect a 50% difference and 2510 to detect a 75% difference. All power calculations assumed c~ = 0,05 (two-sided), [3 = 0.20, and power = 0.80.
440
Ashour et al.
Februaly1997 AmJ Obstet Gynecol
Table II. Association between maternal serum [3-hCG levels and hypertension status in overall population All hypertension* (n = 527) ~-hCG level (multiples of median)
<2.0 >--2.0 2.0-2.99 >-3.0
Gestational hypertension (n = 333)
Preeclampsia (n = 194)
No.
% and No.
RRand 95%CI
% and No.
RR and 95%CI
% and No.
RRand 95%CI
5496 642 503 139
8.1 (445) 12.8 (82) 12.9 (65) 12.2 (17)
1.0 1.6 (1.3-2.0) 1.6 (1.3-2.1) 1.5 (1.0-2.4)
5.2 (285) 7.5 (48) 8.0 (40) 5.8 (8)
1.0 1.5 (1.1-2.0) 1.6 (1.1-2.2) 1.1 (0.6-2.2)
2.9 (160) 5.3 (34) 5.0 (25) 6.5 (9)
1.0 1.8 (1.3-2.6) 1.7 (1.1-2.6) 2.2 (1.2-4.3)
RR, Risk ratio; CI, confidence interval. *Gestational hypertension and preeclampsia.
Table III. Association between maternal serum [~-hCG levels and hypertension status in multiparous women All hypertension* (n = 182)
Gestational hypertension (n = 107)
Preeclampsia (n = 75)
~-hCG ~vel (mu~@~sofme~aN
No.
% and No.
R.Rand 95%CI
% and No.
RRand 95 % CI
% and No.
I
RR and 95%CI
<2.0 >-2.0 2-2.99 ~3.0
2359 252 209 43
6.2 (147) 13.9 (35) 14.4 (30) 11.6 (5)
1.0 2.2 (1.6-3.2) 2.3 (1.6-3.3) 1.9 (0.84.3)
3.9 (92) 5.6 (15) 6.2 (13) 4.7 (2)
1.0 1.6 (1.0-2.8) 2.3 (1.6-3.3) 1.9 (0.8-4.3)
2.3 (55) 7.9 (20) 8.1 (17) 7.0 (3)
1.0 3.4 (2.1-5.6) 3.5 (2.1-5.9) 3.0 (1.0-9.2)
RR, Risk ratio; CI, confidence interval. *Gestational hypertension and preeclampsia.
Results Table I shows selected demographic and reproductive characteristics of women by hypertension status. Nulliparous women had higher rate of hypertension compared with multiparous women. Birth weights were lower in pregnancies with hypertension and these women were delivered at earlier gestational ages. A significant association between elevated secondtrimester [3-hCG levels and subsequent development of hypertension in pregnancy was observed in the overall population (Table II). When stratified by parity, only multiparous women retained a statistically significant association between elevated second-trimester [3-hCG multiples of the median and development of hypertension (Table III). In this group the association was of marginal statistical significance in patients who had gestational hypertension where with use of 2.0 multiples of the median of [3-hCG as a cutoff, the rate ratio was 1.6 (95% confidence interval 1.0 to 2.8). This association was more clearly demonstrated in patients who had preeclampsia and strongest in patients who had severe preeclampsia; the rate ratio with use of 2.0 multiples of the median of [3-hCG as the cutoff value was 5.2 (95% confidence interval 2.5 to 10.6). Although the risk of pregnancy-related hypertension among nulliparous women was approximately 30%
higher in the presence of an elevated second-trimester [3-hCG level (Table IV), this association was not statistically significant. A logistic regression model to correct for increasing prevalence of the disease with increased maternal age and black race did not alter the association in either the multiparous or the nulliparous group. Table V shows the sensitivity, specificity, positive predictive value, and negative predictive value for 2.0 and 3.0 multiples of the median of [3-hCG in predicting overall hypertension, preeclampsia, and severe preeclampsia. At a [3-hCG multiple of the median level of 2.0, the sensitivity for predicting hypertension in the overall population was 15.6% with a positive predictive value of 12.8%. For predicting preeclampsia, the sensitivity was 17.5% with a positive predictive value of 5.3%. The sensitivity for predicting severe preeclampsia was 21.4% with a positive predictive value of 2.8%. The sensitivity for predicting preeclampsia at a cutoff value of 2.0 multiples of the median of [3-hCG was better for muhiparous women and best for predicting severe preeclampsia in this group (sensitivity 35.5%). In nulliparons women, at a cutoffvalue of 2.0 multiples of the median of [3-hCG, the sensitivity for predicting severe preeclampsia was only 13.2%. The use of other [3-hCG multiple of the median levels
Volume 176, Number 2 AmJ Obstet Gynecol
Ashour et al,
441
Table IV. Association between maternal serum [3-hCG levels and hypertension status in nulliparous women All hypertension* (n = 345) ~-hCG level (multiples of median)
No.
% and No.
RR and 95 % CI
<2.0 >-2.0 2.0-2.99 >-3.0
2433 304 232 72
12.2(298) 15.5(47) 15.1(35) 16.7(12)
1.0 1.3(0.9-1.7) 1.2(0.9-1.7) 1.4(0.8-2.3)
Gestational hypertension (n = 226)
Preedampsia (n = 119)
RR and 95% C1
% and No.
7.9(193) 10.9(33) 11.6(27) 8.3(6)
1.0 1.4(1.0-1.9) 1.5(1.0-2.1) 1.1(0.5-2.3)
% and No.
RR and 95%CI
4.3(105) 4.6(14) 3.4(8) 8.3(6)
1.0 1.1(0.6-1.8) 0.8(0.41.6) 1.9(0.9-4.2)
RR, Risk ratio; C/, confidence interval. *Gestational hypertension and preeclampsia.
Table V. Value of second-trimester serum [3-hCG values for prediction of hypertension by [3-hCG multiples of median >-2.0 ~-hCG multiples of median Parity
Overall population Sensitivity Specificity PPV NPV Nulliparous Sensitivity Specificity PPV NPV Multiparous Sensitivity Specificity PPV NPV
>-3.0 ~-hCG multiples of median
Overall hypertension
Gestational hypertension
PET
Severe PET
Overall hypertension
Gestational hypertension
PET
SeverePET
15.6 90.0 12.8 91.9
14.4 90.0 7.9 94.7
17.5 89.8 5.3 97.1
21.4 89.7 2.8 98.8
3.2 97.8 12.2 91.5
2.4 97.8 6.2 94.4
4.6 97.8 6.5 96.9
6.0 97.8 3.6 98.7
13.6 89.3 "15.5 87.8
14.6 89.3 11.4 91.7
11.8 88.9 4.6 95.7
13.2 88.9 2.3 98.1
3.5 97.5 16.7 87.5
2.7 97.5 9.1 91.4
5.0 97.5 8.3 95.8
5.7 97.4 4.2 98.1
19.2 91.1 13.9 93.8
14.0 91.1 6.5 96.0
26.7 90.9 7.9 97.7
35.5 90.7 4.4 99.2
2.7 98.4 11.6 93.1
1.9 98.4 5.0 95.8
4.0 98.4 7.0 97.2
6.5 98.4 4.7 98.9
PET, Preeclampsia; PPV, positive predictive value; NPV, negative predictive value.
did not improve the sensitivity and predictive value for predicting pregnancy-induced hypertension or preeclampsia in any part of the population.
Comment Our data confirm previous reports that demonstrated an association between elevated second-trimester maternal serum [3-hCG levels and the subsequent development of pregnancy-induced hypertension and preeclampsia.20, 21 Our study differs from other reports because we evaluated this association among multiparous and nulliparous women separately. Stratification by parity demonstrated this association to be statistically significant only among multiparous women. In this group the association of elevated second-trimester serum [3-hCG levels and tile later development of hypertension in pregnancy was stronger in patients with preeclampsia and was strongest among women who had severe preeclampsia. Although the risk of development of hypertension was approximately 30% higher in the presence of elevated
second-trimester [3-hCG levels among nulliparous women, this association was not statistically significant. Our power was adequate to detect a 50% increase in the risk of hypertension in pregnancies with multiple of the median [3-hCG levels ->2.0; however, we would have needed 7018 nulliparous women to have adequate power (80%) to demonstrate a 30% increase in the risk for the development of hypertension in this group. Fisher et al. 24 demonstrated that when nulliparous patients with hypertension during pregnancy underwent kidney biopsy, 75% of them showed glomerular endotheliosis, 16.3% had chronic renal lesions, and 6.7% had both preeclampsia and chronic renal disease. When the same study was carried out in multiparous patients with hypertension in pregnancy, the characteristic histologic lesion of preeclampsia was found in only 23%, chronic renal disease was present in 51%, chronic renal lesions plus preeclampsia were found in 13%, and normal histologic features were found in 11%. A possible explanation for the stronger association between elevated [3-hCG levels and later development of hypertension in
442
Ashour et al.
multiparous w o m e n is an increased i n c i d e n c e of underlying renal disease resulting in a decreased renal excretion rather than increased secretion loy the trophoblast. Sorensen et al. 21 d e m o n s t r a t e d an increased risk of preeclampsia in w o m e n with increased second-trimester hCG levels and suggested that a second-trimester screening p r o g r a m m i g h t be d e v e l o p e d to assess those at risk for pregnancy-induced hypertension. Second-trimester maternal serum screening for fetal aneuploidy and neural tubal defects usually includes h C G or [~-hCG is currently a widely p e r f o r m e d test. Thus second-trimester serum [~-hCG m e a s u r e m e n t would be a potentially costeffective tool for identifying w o m e n at risk for the d e v e l o p m e n t of preeclampsia. Such a screening p r o g r a m would facilitate closer surveillance and possible intervention with low-dose aspirin or calcium supplementation. A l t h o u g h our data demonstrate that elevated secondtrimester [3-hCG levels is associated with an increased risk for the d e v e l o p m e n t of hypertension in pregnancy, especially a m o n g multiparous women, the p o o r sensitivity and predictive value of the test would suggest that its use as a screening test for later d e v e l o p m e n t of gestational hypertension and preeclampsia is limited. REFERENCES
1. Robertson WB, Khong TY, Brosens I, De Wolf F, Sheppard BL, Bonnar J. The placental bed biopsy': review from three European centers. Am J Obstet Gynecol 1986;155:401-12. 2. Roberts JM, Robert NT, Musci TJ, Rodgers GM, Hubel CA, McLaughin MFL Preeclampsia: an endothelial cell disorder: Am J Obstet Gynecol 1989;161:1200-4. 3. Kaar K, Jouppila P, Kuikka J, Luotola H, Toivanen J, Rekonen A. Intervillous blood flow in normal and complicated late pregnancy measured by means of an intravenous 133 XE method. Acta Obstet Gynecol Scand 1980;59:7-10. 4. Frusca T, Morassi L, Pecorelli S, Grigolato P, Gastaldi A. Histological features of uteroplacental vessels in normal and hypertensive patients in relation to birth weight. BrJ Obstet Gypaecol 1989;96:835-9. 5. MorrisJA, O'GradyJP, Hamilton CJ, Davidson EC. Vascular reactivity to angiotensin II infusion during gestation. Am J Obstet Gynecol 1978;130:379-84. 6. Oney T, Kaulhausen H. The value of the angiotensin sensitivity test in the early diagnosis of hypertensive disorders in pregnancy. A m J Obstet Gynecol 1982;142:17-20. 7, GudsonJPJr, Andreson SG, May ~¢~. A clinical evaluation of the roll-over test for pregnancy-induced hypertension. AmJ Obstet Gynecol 1977;127:1-3. 8. Kuntz WD. Supine pressor (roll-over) test: an evaluation. AmJ Obstet Gynecol 1980;137:764-8. 9. Bower S, Bewley S, Campbell S. Improved prediction of
February 1997 Am J Obstet Gynecol
10. 11. 12.
13.
14.
15. 16. 17. 18.
19.
20.
21.
22.
23.
24.
preeclampsia by two stage screening of uterine arteries using the early diastolic notch and color Doppler imaging. Obstet Gynecol 1993;82:78-83. McParand P, Pearce JM, Chamberlain GVP. Doppler ultrasound and aspirin in recognition and prevention of pregnancy-induced hypertension. Lancet 1990;335:1552-5. Steel SA, PearceJM, McParland P, Chamberlain GVP. Early Doppler ultrasound screening in prediction of hypertensive disorders of pregnancy. Lancet 1990;335:1548-51. Jacobson SL, Imhof R, Manning N, Mannion V, Little D, Rey E, et al. The value of Doppler assessment of the uteroplacental circulation in predicting preeclampsia or intrauterine growth retardation. Am J Obstet Gynecol 1990;162: 110-4. Lazarachick J, Stubbs TM, Romein L, Van Dorsten JP, Loadholt CB. Predictive value of fibronectin levels in normotensive gravid women destined to become preeclamptic. A m J Obstet Gynecol 1986;154:1050-2. Stubbs TM, Lazarchick J, Horger EO. Plasma fibronectin levels in preeclampsia: a possible biochemical marker for vascular endothelial damage. Am J Obstet Gynecol 1984; 150:885-7. Lockwood CJ, Peters JH. Increased plasma levels of EDI+ cellular fibronectin precede the clinical signs of preeclampsia. Am J Obstet Gynecol 1990;162:358-62. Smith GVS, Smith OW. The anterior pituitary-like hormone in late pregnancy toxemia. Am J Obstet Gynecol 1939;38:618-24. Smith GVS, Smith OW. Excessive gonad-stimulating hormone and subnormal amounts of oestrin in the toxemias of late pregnancy. Am J Physiol 1934;107:128-45. Said ME, Campbell DM, Azzam ME, MacGillivray I. Betahuman chorionic gonadotrophin levels before and after the development of preeclampsia. Br J Obstet Gynaecol 1984; 91:772-5. Hsu CD, Chan DW, Iriye B, Johnson TRB, Hong SF, Repke JT. Elevated serum human chorionic gonadotropin as evidence of secretory response in severe preeclampsia. Am J Obstet Gynecol 1994;170:1135-8. Gonen R, Perez R, David M, Dar H, Merksamer R, Sharf M. The association between unexplained second-trimester maternal serum hCG elevations and pregnancy induced complications. Obstet Gynecol 1992;80:83-6. Sorensen TK, Williams MA, Zingheim RW, Clement SJ, Hickok DE. Elevated second-trimester human chorionic gonadotropin and subsequent pregnancy-induced hypertension. Am J Obstet Gynecol 1993;169:834-8. Gravett CP, Buckmaster JG, Watson PT, Gravett MG. Elevated second trimester maternal serum [3-HCG concentrations and subsequent adverse pregnancy outcome. Am J Med Genet 1992;44:485-6. Wenstrom ILl2), Owen J, Boots LR, DuBard MB. Elevated second-trimester human chorionic gonadotropin levels in association with poor pregnancy outcome. Am J Obstet Gynecol 1994;171:1038-41. Fisher KA, Luger A, Spargo BH, Lindheimer MD. Hypertension in pregnancy: clinical-pathological correlations and remote prognosis. Medicine (Baltimore) 1981;60:267-76.