Aspirin's effect on preterm birth did not depend on starting week or adherence in the ASPIRIN trial

Design

Post hoc analysis of a multicentre placebo-controlled RCT

Population

11,908 nulliparous women at 6 to 13 weeks in Africa, Asia and Latin America

Primary outcome

Preterm delivery, by week of initiation and adherence

Effect

Per-week change in treatment RR 0.97 (0.93 to 1.02); per 5% adherence 1.01 (0.99 to 1.04)

The ASPIRIN trial randomised nulliparous women in low-resource settings across Africa, Asia (including two Indian sites) and Latin America to low-dose aspirin or placebo from 6 to 13 weeks. This post hoc analysis of 11,908 women asked whether the effect on preterm delivery changed with the week aspirin began or with how reliably it was taken. Median start was 10.1 weeks; 85.5% of women took more than 90% of doses by pill count.

For each week later that aspirin started, the treatment risk ratio for preterm delivery did not shift (RR 0.97, 95% CI 0.93 to 1.02), nor for delivery before 34 weeks (0.98, 0.89 to 1.07) or perinatal death (1.04, 0.96 to 1.12). Each 5% increase in adherence made no difference either (preterm delivery 1.01, 0.99 to 1.04). There was no variation by region.

The practical reading is that any first-trimester start is reasonable, and a woman who books at 12 weeks has not missed the window. Because adherence was high across the trial, the analysis says little about women who take far fewer doses.

• Start low-dose aspirin at booking if it falls anywhere between 6 and 13 weeks — a later first-trimester start was not less effective
• Do not delay aspirin waiting for a dating scan if gestation is already clear
• Adherence was high in this trial (85.5% above 90%), so it cannot tell you whether poor adherence is harmless
• Pill counts at each visit are a simple adherence check that worked at scale in these settings

Why it matters

It removes the pressure to start at the earliest possible week, which late-booking women cannot meet anyway.

Don't overread it

This is a post hoc analysis within a trial with high adherence; it cannot show that missed doses do not matter.

The statistics, in plain English

An interaction ratio near 1.0 with a narrow interval (0.93 to 1.02 per week) means the aspirin effect looked much the same whenever it began. Post hoc analyses were not planned in advance, so they suggest rather than prove.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.