|
|
| |||||||||||||||||||||||||||||||||||||||
Background Although low-dose aspirin has been reported to reduce the incidence of preeclampsia among women at high risk for this complication, its efficacy and safety in healthy, nulliparous pregnant women are not known.
Methods We studied 3135 normotensive nulliparous women who were 13 to 26 weeks pregnant to determine whether treatment with aspirin reduced the incidence of preeclampsia. Of this group, 1570 women received 60 mg of aspirin per day and 1565 received placebo for the remainder of their pregnancies. We also evaluated the effect of aspirin on maternal and neonatal morbidity.
Results Of the original group of 3135 women, 2985 (95 percent) were followed throughout pregnancy and the immediate puerperium. The incidence of preeclampsia was lower in the aspirin group (69 of 1485 women [4.6 percent]) than in the placebo group (94 of 1500 women [6.3 percent]) (relative risk, 0.7; 95 percent confidence interval, 0.6 to 1.0; P = 0.05), whereas the incidence of gestational hypertension was 6.7 and 5.9 percent, respectively. There were no significant differences in the infants' birth weight or in the incidence of fetal growth retardation, postpartum hemorrhage, or neonatal bleeding problems between the two groups. Subgroup analysis showed that preeclampsia occurred primarily in women whose initial systolic blood pressure was 120 to 134 mm Hg (incidence among such women, 5.6 percent in the aspirin group vs. 11.9 percent in the placebo group; P = 0.01). The incidence of abruptio placentae was greater among the women who received aspirin (11 women, vs. 2 in the placebo group; P = 0.01).
Conclusions Low-dose aspirin decreases the incidence of preeclampsia among nulliparous women, primarily through its effect in those who have elevated systolic blood pressure initially. This treatment does not decrease perinatal morbidity but increases the risk of abruptio placentae.
Whether aspirin will benefit all pregnant women is not known. We therefore undertook a double-blind, placebo-controlled study to determine whether treatment with aspirin reduced the incidence of preeclampsia among nulliparous pregnant women, the group at highest risk for this disorder.
Methods
Subjects
We enrolled nulliparous women seeking prenatal care who were 13 to 25 weeks pregnant if their blood pressure was below 135/85 mm Hg and they had no proteinuria on testing with a dipstick (sensitivity, less than 30 mg per deciliter if the dipstick result is zero or trace). Women with chronic hypertension, renal disease, diabetes mellitus, and other medical illnesses were excluded.
Protocol
The protocol was approved by the National Institute of Child Health and Human Development and the institutional review board at each center, and each woman gave written informed consent. The same protocol was used at all participating centers, and an independent data-coordinating center (the George Washington University Biostatistics Center) was responsible for developing the treatment-allocation code and performing periodic analyses as needed.
A total of 4241 eligible women were assigned to a single-blind compliance test. Each woman was given 10 placebo tablets with instructions to take 1 tablet daily and return 10 days later. A woman was considered compliant if she ingested at least half the tablets. A total of 3135 women found to be compliant on this test were assigned randomly to receive aspirin (60 mg plus starch) or placebo (lactose plus starch). (The identical-appearing tablets were prepared by Sterling Winthrop, New York.) The women were instructed to take one tablet daily until the onset of labor, asked to avoid taking any products containing aspirin, and given a supply of acetaminophen for headache or other pain (if needed).
The women were followed according to the usual schedule at each center, typically every 4 weeks until 26 to 28 weeks of gestation, every 2 to 3 weeks until 36 weeks of gestation, and then weekly until delivery or the onset of preeclampsia. At each visit, the women's blood pressure, weight, and urinary protein excretion were measured. Clinic staff members were trained to measure blood pressure with use of a sphygmomanometer with the women seated and after a rest period of at least 10 minutes. Urinary protein was measured with a dipstick in a fresh, clean, midstream urine sample. A value of zero to trace was considered negative; a 1+ value was equivalent to 30 mg per deciliter, and a 2+ value equivalent to 100 mg per deciliter.
Compliance with treatment was assessed by asking the women about tablet intake and by counting tablets. The women were also contacted periodically by a research nurse to check on compliance.
Outcome Variables
Hypertension (defined as a systolic blood pressure of 
140 mm Hg or a diastolic blood pressure of 90 mm Hg) was designated as severe if two or more systolic values obtained four or more hours apart were 160 mm Hg or if two or more diastolic values were 110 mm Hg, or if there was one diastolic value 110 mm Hg and an antihypertensive drug was subsequently given. Gestational hypertension was defined as hypertension without proteinuria after week 20 of gestation or during the postpartum period. If hypertension occurred only during labor, a serum uric acid concentration 6 mg per deciliter (357 mmol per liter) was required for the diagnosis of gestational hypertension.
Preeclampsia was defined as hypertension plus proteinuria (either 300 mg per 24 hours or 2+ or more by dipstick on two or more occasions 4 hours apart) in the absence of a urinary tract infection. Severe preeclampsia was defined as severe hypertension and proteinuria; urinary protein excretion 5 g per day with any degree of hypertension; hypertension complicated by pulmonary edema or a low platelet count (<100,000 per cubic milliliter); or hemolysis, an elevated serum aspartate aminotransferase concentration (>70 U per liter), and a low platelet count (the HELLP syndrome). Women were considered to have eclampsia if they met the criteria for gestational hypertension and had convulsions. Intrauterine fetal growth retardation was defined as a birth weight below the 10th percentile according to the growth tables of Brenner et al.14. Abruptio placentae was diagnosed according to clinical findings (vaginal bleeding and uterine tenderness) or placental examination.
Statistical Analysis
A group sequential-analysis plan (the Lan and DeMets method15) was chosen before the study started. For this study, a generalization of the O'Brien and Fleming boundary was chosen16. Two interim analyses that corresponded to 34 percent and 63 percent of the total planned results were performed; the critical P value for the final analysis was therefore 0.047 rather than 0.05 (two-tailed). All the women were included in the treatment group to which they were assigned, regardless of compliance with treatment. Comparisons between the aspirin and placebo groups were performed with the chi-square test, Fisher's exact test, the Wilcoxon rank-sum test, or the Mantel-Haenszel test.
Results
Of the 3135 women found to be compliant in the initial test, 1570 women were assigned to the aspirin group and 1565 to the placebo group. The demographic and clinical characteristics of the women in the two groups were similar except that more women in the placebo group had a systolic blood pressure of 120 to 134 mm Hg (297 [19.0 percent], vs. 245 women [15.6 percent] in the aspirin group; P = 0.01), although the mean systolic blood pressures in the groups were similar (Table 1). During the study 150 women (4.8 percent) were lost to follow-up (85 in the aspirin group and 65 in the placebo group). Thus, the results reported here are for 1485 women who received aspirin and 1500 women who received placebo. The characteristics of these 2985 women were similar to those of the 3135 women.
|
The frequency of all hypertensive complications is shown in Table 2. The incidence of preeclampsia was 4.6 percent in the aspirin group and 6.3 percent in the placebo group (relative risk, 0.7; 95 percent confidence interval, 0.6 to 1.0; P = 0.05). There were no significant differences between the groups in the incidence of gestational hypertension (6.7 percent in the aspirin group vs. 5.9 percent in the placebo group) or preeclampsia and gestational hypertension combined (11.4 percent vs. 12.2 percent). One hundred women (57 in the aspirin group and 43 in the placebo group) who had gestational hypertension also had serum uric acid concentrations 6 mg per deciliter (357 mmol per liter). If these 100 women were considered to have preeclampsia, the incidence of preeclampsia was 8.5 percent in the aspirin group and 9.1 percent in the placebo group. Eight women had eclampsia (three in the aspirin group and five in the placebo group) and six had the HELLP syndrome (one in the aspirin group and five in the placebo group). Aspirin appeared to have a greater prophylactic effect in the subgroup of 519 women with higher systolic blood pressure (120 to 134 mm Hg) at the beginning of the study. Among these 519 women, 13 of the 234 in the aspirin group (5.6 percent) had preeclampsia, as compared with 34 of 285 (11.9 percent) in the placebo group (P = 0.01). Aspirin had little effect when the base-line systolic blood pressure was below 120 mm Hg (Figure 1). Controlling for this imbalance in base-line blood pressure by the Mantel-Haenszel test yielded a relative risk of 0.8 (95 percent confidence interval, 0.6 to 1.0).
|
|
The incidence of oligohydramnios, premature rupture of the membranes, and preterm and post-term delivery was similar in both groups, as were the rate of cesarean delivery, average gestational age, birth weight, and length at delivery (Table 3 and Table 4). The incidence of low birth weight (below the 10th percentile) and high birth weight (more than 4 kg) was similar, as was the number of infants admitted to the neonatal intensive care unit.
|
|
In the aspirin group, 30 pregnancies (2.0 percent) were lost; there were 8 spontaneous abortions, 17 fetal deaths at 20 weeks' gestation or later, and 5 neonatal deaths. In the placebo group, 21 pregnancies (1.4 percent) were lost; there were 7 spontaneous abortions, 7 fetal deaths, and 7 neonatal deaths. Two fetal deaths (at 26 and 27 weeks' gestation) in the aspirin group and one (at 28 weeks) in the placebo group were related to severe preeclampsia or eclampsia.
Maternal and Fetal Complications of Therapy
Eleven women (0.7 percent) in the aspirin group had abruptio placentae, as compared with two women (0.1 percent) in the placebo group (P = 0.01) (Table 5). There were cases of abruptio placentae in the aspirin group at five of the seven centers, whereas the two cases of abruptio placentae in the placebo group occurred at a single center. Among the 11 women in the aspirin group who had abruptio placentae, there were two fetal deaths, two preterm deliveries (at 24 and 34 weeks), and two deliveries by cesarean section for fetal distress; the other 5 women had uncomplicated deliveries. There were no differences between the groups in the incidence of excessive blood loss (especially at delivery), postpartum hemorrhage, change in the hematocrit, or need for transfusions (Table 5). In addition, there were no bleeding complications associated with the use of epidural anesthesia among the women who received aspirin. Among the neonates, there was no difference in the incidence of cephalohematoma, cerebral hemorrhage, petechiae, purpura, excessive bleeding at the time of circumcision, any bleeding disorder, or the need for transfusion.
|
The average number of visits and the mean number of tablets ingested were similar in the two groups. The estimated percentage of the prescribed tablets taken averaged 83 percent in the aspirin group and 84 percent in the placebo group. Seventy-two percent of the women in the aspirin group and 74 percent of the women in the placebo group took 80 percent or more of the tablets prescribed between the start of the study and delivery.
A total of 256 women (8.6 percent) discontinued the assigned drug (140 [9.4 percent] in the aspirin group and 116 [7.7 percent] in the placebo group) because of perceived side effects, at the request of their physicians, or for other reasons. Unlike the women lost to follow-up, however, these women were followed through delivery and are included in the analysis. About twice as many women in the aspirin group as in the placebo group discontinued the assigned drug because of perceived side effects (gastrointestinal symptoms, headaches, and rash) (23 vs. 11, P = 0.04).
Discussion
The results of seven prospective studies and two meta-analyses suggest that the prophylactic ingestion of low doses of aspirin reduces the incidence of preeclampsia5,6,7,8,9,10,11,17,18. These findings have been widely publicized by the media. Analysis of the literature, however, reveals that in all but one of the reports,17 the women treated with aspirin were identified by an adverse obstetrical history or by certain predictive tests7,8,9. We designed this trial to determine the efficacy and safety of aspirin in healthy, nulliparous pregnant women. We observed a 26 percent reduction in the incidence of preeclampsia, which was confined mainly to women with an initial systolic blood pressure of 120 to 134 mm Hg, but no reduction in the overall rate of hypertension. Furthermore, despite the reduction in the incidence of preeclampsia among the aspirin-treated women, there was no beneficial effect on overall perinatal morbidity, and aspirin was associated with an increased incidence of abruptio placentae.
In an earlier study of women at high risk for the disorder,19 the incidence of preeclampsia among those treated with aspirin averaged 2.2 percent (range, 0 to 5), a rate substantially lower than that in women given placebo (average, 15 percent; range, 9 to 35). However, in two studies -- one focusing on women at low risk17 and the other on women at moderate risk18 -- the results were equivocal. In the former study,17 there was a 70 percent reduction in the incidence of preeclampsia among women who received aspirin, whereas in the latter, aspirin had no benefit18. In our study (which involved more women than the other studies), not only was there a small reduction in the incidence of preeclampsia among the women who received aspirin, but also the women in this group in whom preeclampsia did develop were more likely to deliver their infants earlier.
Other benefits that have been ascribed to aspirin include a lower incidence of intrauterine growth retardation and of preterm delivery and higher birth weight -- all due in part to the prevention of early preeclampsia19,20. In still other reports, women treated with aspirin in whom preeclampsia did not develop had longer pregnancies, and their infants had higher birth weights6,21,22. No such benefits were found in our trial. The reasons for the discrepancies are not apparent, but they may be related to the large size of our study and the more heterogeneous groups of women studied by others.
Recently, O'Brien et al.23. found that ingestion of acetaminophen in doses of 15 mg per kilogram of body weight was associated with a significant inhibition of prostacyclin production during the third trimester. Since acetaminophen was recommended for use if needed in this trial, it may be argued that the use of acetaminophen could mitigate the effect of low-dose aspirin. Unfortunately, we have no data on the use of acetaminophen by these women, but we think it unlikely that its use differed in the two groups.
Aspirin has been implicated in a number of adverse actions affecting the mother, fetus, or neonate19,20. The maternal risks include increased antepartum and postpartum hemorrhage; the fetal risks include oligohydramnios; and the neonatal risks are persistent pulmonary hypertension and a variety of bleeding problems19. We found no increase in the frequency of adverse effects among the women in the aspirin group, as was reported in previous studies5,6,7,8,9.
The women taking low-dose aspirin had abruptio placentae more often than those in the placebo group, although only 6 of the 11 women with abruptio placentae in the aspirin group had adverse outcomes of pregnancy. The incidence of abruptio placentae in other trials involving both women at high risk5,6,7,8,9,18 and those at low risk17 was similar in the aspirin and the placebo groups, and there is a single report in which women taking 150 mg of aspirin per day had a lower incidence of abruptio placentae and fetal loss (2 percent) than those in a placebo-treated group (12 percent)6. In addition, the incidence of abruptio placentae among the aspirin-treated women was similar to that found in the general population,24 whereas that in the placebo group was considerably lower. Alternatively, the 0.1 percent incidence of abruptio placentae in the placebo group may reflect the true incidence of this complication among healthy, nulliparous pregnant women25.
In conclusion, a number of randomized trials have suggested that low-dose aspirin therapy reduces the incidence of preeclampsia among certain pregnant women at risk for this complication. Our results, combined with those of Hauth et al.,17 suggest that the incidence of preeclampsia may be reduced by aspirin use among unselected healthy nulliparous pregnant women as well. The effect, however, was small. Moreover, because the women who received aspirin in this trial had no improvement in perinatal morbidity and were at increased risk for abruptio placentae, we do not recommend the routine use of low-dose aspirin in healthy nulliparous women. This conclusion may not apply to women with various risk factors for preeclampsia (previous preeclampsia, chronic hypertension, multiple gestation, and diabetes mellitus). Indeed, in this study, aspirin had a greater benefit among women with higher initial systolic blood pressure.
Supported by grants (HD 21410, HD 21434, HD 21366, HD 21380, HD 19897, HD 21414, HD 21386, and HD 21363) from the National Institute of Child Health and Human Development.
We are indebted to Sterling Health, New York, for the generous donation of the study drugs and to McNeil Laboratories, Lakewood, New Jersey, for the donation of acetaminophen tablets.
Source Information
Dr. Mortimer Rosen is deceased.The members of the network are listed in the Appendix.
Address reprint requests to Dr. Sibai at the Department of Obstetrics and Gynecology, 853 Jefferson Suite E102, Memphis, TN 38103.
References
The Network of Maternal-Fetal Medicine Units was established by the National Institute of Child Health and Human Development in 1986. The participating institutions and investigators are as follows: Columbia University -- Susan Godfrey, R.N., Mortimer Rosen, M.D. (deceased), and Rex Chao, M.D.; George Washington University Biostatistics Center -- Elizabeth Thom, Ph.D., Tavia Gordon, and John D. Greene, B.A.; Johns Hopkins University -- Frank Witter, M.D., and Laura Rocco, R.N., M.S.; Thomas Jefferson University -- Richard Depp, M.D., and Susan Tannenbaum, R.N.; University of North Carolina, Chapel Hill -- Robert C. Cefalo, M.D. (chairman); University of Pittsburgh and Magee-Women's Hospital -- Steve N. Caritis, M.D., and Margaret Cotroneo, R.N.; University of Southern California at Los Angeles -- Richard H. Paul, M.D., and Catherine Walla, R.N.; University of Tennessee, Memphis -- Baha M. Sibai, M.D., Brian M. Mercer, M.D., and Eileen Bray, R.N.; Yale University -- Roberto Romero, M.D., John C. Hobbins, M.D., and Ginny Sabo, R.N.; and the National Institute of Child Health and Human Development -- Donald McNellis, M.D. (program officer), Sumner Yaffe, M.D., Mark Klebanoff, M.D., and Charlotte Catz, M.D.
| |||||||||||||||||||||||||||||||||||||||
Related Letters:
Prevention of Preeclampsia with Low-Dose Aspirin
Moya F. R., Germain A., Sibai B. M.
Extract |
Full Text
N Engl J Med 1994;
330:794-795, Mar 17, 1994.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
