Effect of aspirin in pregnant women is dependent on increase in bleeding time

OBJECTIVES: Randomized trials with low-dose aspirin to prevent preeclampsia and intrauterine growth restriction have yielded conflicting results. In p articular, 3 recent large trials were not conclusive. Study designs, howeve r, varied greatly regarding selection of patients, dose of aspirin, and tim ing of treatment, all of which can be determinants of the results. Retrospe ctively analyzing the conditions associated with failure or success of aspi rin may therefore help to draw up new hypotheses and prepare for more speci fic randomized trials. STUDY DESIGN: We studied a historical cohort of 187 pregnant women who were considered at high risk for preeclampsia, intrauterine growth restriction, or both and were therefore treated with low-dose aspirin between 1989 and 1994. Various epidemiologic, clinical, and laboratory data were extracted f rom the files. Univariate and multivariate analyses were performed to searc h for independent parameters associated with the outcome of pregnancy. RESULTS: Age, parity, weight, height, and race had no influence on the outc ome. The success rate was higher when treatment was given because of previo us poor pregnancy outcomes than when it was given for other indications, an d the patients with successful therapy had started aspirin earlier than had those with therapy failure (17.7 vs 20.0 weeks' gestation, P = .04). After multivariate analysis an increase in Ivy bleeding time after 10 days of tr eatment by >2 minutes was an independent predictor of a better outcome (odd s ratio 0.22, 95% confidence interval 0.09-0.51). Borderline statistical si gnificance was observed for aspirin initiation before 17 weeks' gestation ( odds ratio 0.44, 95% confidence interval 0.18-1.08). Abnormal uterine arter y Doppler velocimetric scan at 20-24 weeks' gestation (odds ratio 3.31, 95% confidence interval 1.41-7.7), abnormal umbilical artery Doppler velocimet ric scan after 26 weeks' gestation (odds ratio 37.6, 95% confidence interva l 3.96-357), and use of antihypertensive therapy (odds ratio 6.06, 95% conf idence interval 2.45-15) were independent predictors of poor outcome. CONCLUSIONS: Efficacy of aspirin seems optimal when bleeding time increases greater than or equal to 2 minutes with treatment, indicating a more power ful antiplatelet effect. This suggests that the dose of aspirin should be a djusted according to a biologic marker of the antiplatelet effect. A prospe ctive trial is warranted to test this hypothesis.