Spain 1997.

Methods	Computer‐generated random numbers used to prepare a table for the sequence of allocation. Tablets in identical blister packs. Allocated to 6 groups, according to treatment and timing of administration. 7 women excluded, because poor compliance or incomplete blood pressure assessments.
Participants	107 women aged 18‐40 years at < 16 weeks' gestation and at moderate risk of PE. For example, family or own history of PIH, PE, chronic HT, cardiovascular or endocrine problem, bleeding or endocrine disease. Excluded: multiple pregnancy.
Interventions	Exp: 100 mg aspirin. Control: placebo. Each treatment group could also be allocated to 3 different times of the day.
Outcomes	Women: GH; PE; caesarean section; abruption. Baby: death; preterm birth (< 37 weeks); IUGR.
Notes	Testing the hypothesis that aspirin effects are time‐dependent, being greater in the evening.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated random numbers.
Allocation concealment (selection bias)	Low risk	Coded sequentially number packs of tablets: Placebo and ASA quote: "were prepared in an identical manner and provided monthly to the volunteers in a box containing three blister packs, each with 10 tablets. The boxes, grouped in packs of seven (to cover medication for the duration of pregnancy) and labeled with the randomisation number, were assigned to each patient at the time of her recruitment."
Blinding (performance bias and detection bias) All outcomes	Low risk	Quote: "double‐blind" and placebo used.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Quote: "7 subjects who did not comply with all requirements set a priori for this clinical trial were eliminated from the study. Reasons for elimination included the use of additional medication during the trial, noncompliance with the assigned medication (missing more than six tablets during any given month), and the impossibility of providing all required BP profiles. Subjects providing fewer than five profiles of ambulatory BP monitoring (ABPM) were eliminated from the study." The distribution of exclusions/withdrawals between groups was not reported.
Selective reporting (reporting bias)	Low risk	Protocol not available, but all expected outcomes appear to be reported.
Other bias	Unclear risk	Quote: "Baseline characteristics related to age, weight, height, and 24‐hour mean BP values obtained from the first profile of ABPM (before treatment started) were similar for all six groups." 70/100 women were primipara, but distribution across groups not reported?