Current issues of ACP Journal Club are published in Annals of Internal Medicine


Aspirin to prevent pregnancy-induced hypertension: a meta-analysis

ACP J Club. 1991 Nov-Dec;115:80. doi:10.7326/ACPJC-1991-115-3-080

Source Citation

Imperiale TF, Petrulis AS. A meta-analysis of low-dose aspirin for the prevention of pregnancy-induced hypertensive disease. JAMA. 1991 Jul 10;266:260-4.



To evaluate the effectiveness of low-dose aspirin for the prevention of pregnancy-induced hypertension through a meta-analysis of data from controlled trials.

Data sources

MEDLINE, January 1980 to June 1990, was searched for English-language articles, using the key words pregnancy, hypertension, preeclampsia, aspirin, clinical trials, and humans. A manual search included references from retrieved reports, review articles, and textbooks.

Study selection

Inclusion criteria were clinical trials of aspirin in doses of less than 325 mg/d, with assessment of ≥ 1 of the following outcomes: pregnancy-induced hypertension (with or without proteinuria), cesarean section, severely low-birth-weight (SLBW) infants (< tenth percentile), or fetal or neonatal deaths. 6 of 15 identified studies met the inclusion criteria. All women in the selected trails were considered to be at high risk for pregnancy-induced hypertension.

Data extraction

The 2 authors independently assessed the methodologic quality of the trials (kappa, 0.84) and abstracted quantitative data from each trial.

Main results

The mean occurrence of pregnancy-induced hypertension in the control groups of the trials was 33% (range, 17% to 52%). The use of low-dose aspirin produced a relative risk (RR) of pregnancy-induced hypertension of 0.35 95% CI, 0.22 to 0.55), with a number-needed-to-treat (NNT) to prevent 1 case per 4 patients (CI, 3 to 7). Delivery by cesarean section occurred in 36% of the women in the control group; the incidence was reduced in the experimental group (RR, 0.34; CI, 0.25 to 0.48; NNT, 8; CI, 5 to 32); however, the specific indications for cesarean section were not described in the original studies. SLBW infants were born to 28% of the control group; low-dose aspirin led to a relative risk of 0.56 (CI, 0.36 to 0.88); NNT, 6 (CI, 4 to 12). Aspirin did not significantly decrease the rate for fetal and neonatal death, which occurred in only 6% of the control group (RR, 0.88; CI, 0.32 to 2.46). No adverse effects were noted in the 194 women who received low-dose aspirin or in their babies.


The use of low-dose aspirin in women at high risk for pregnancy-induced hypertension reduces the risk for hypertension and severely low-birth-weight infants and may reduce the risk for cesarean section.

Source of funding: Not stated

Address for article reprint: Dr. T.F. Imperiale, MetroHealth Medical Center, 3395 Scranton Road, Cleveland, OH 44109.


Antiplatelet agents are being used with increasing frequency for the prevention of pregnancy-related hypertension and its sequelae. The efficacy of the intervention has not been clearly demonstrated in the literature, mostly because of low event rates of the outcomes of interests and consequent low study power. This issue therefore lends itself well to data aggregation techniques.

The study is a review of existing published data on the topic. Like most meta-analyses, it is limited to data in the public domain and does not include unpublished data or studies missed by the search strategies. Despite these limitations, the data presented are convincing. They concur with the findings in the Oxford Database of Perinatal Trials (1). For women taking aspirin during pregnancy, it was reported that reductions of odds ratios for hypertension were 0.57 (95% CI, 0.36 to 0.92); for severe intrauterine growth retardation, 0.61 (CI, 0.33 to 1.15); for cesarean section, 0.44 (CI, 0.25 to 0.76); and for neonatal death, 0.52 (CI, 0.21 to 1.24)

The meta-analysis did not assess the effects of differences among the studies in criteria for initiation of treatment. All women studied were identified as high risk, based on known risk factors for pregnancy-related hypertensive disorders. These included a history of pre-eclampsia, intrauterine growth retardation, multiple gestation, abnormal Doppler uteroplacental flow-velocity wave-forms, and a positive rollover test. Risk factors and prognosis for this condition have recently been reviewed (2).

The identification of those individuals who will benefit most from the therapy remains to be determined. Similarly, a better assessment of the risks associated with treatment should be forthcoming from larger trials in progress.

Donna M. Federkow, MD
McMaster University Hamilton, Ontario