Onwude 1995.

Methods	RCT
Participants	232 women randomised (161 multigravida and 72 primigravida) Inclusion criteria: mean 24 weeks' gestation with high‐risk singleton pregnancy: history of 1 or more small babies (birthweight < 3rd percentile), history of pregnancy hypertension, history of unexplained stillbirth, or primigravida with abnormal uterine Doppler at 24 weeks' gestation Exclusion criteria: history of diabetes mellitus, chronic hypertension, asthma, use of anticoagulants, multiple pregnancy Setting: antenatal clinic, St James's University Hospital, Leeds, UK
Interventions	SUPPLEMENTATION: EPA + DHA versus placebo Group 1: 2.7 g of MaxEPA/day (n = 113): 9 capsules/day provided 1.62 g EPA and 1.08 g DHA Group 2: control: matching air‐filled capsules (n = 119) Timing of supplementation: from a mean of 24 weeks to 38 weeks' gestation EPA + DHA dose/day: high: 1080 mg DHA + 1620 mg EPA
Outcomes	Women/birth: length of gestation; preterm birth < 37 weeks; preterm birth < 32 weeks; duration and mode of onset of labour; mode of birth; PIH, PE, adverse events (75 women only) Babies/infants/children: stillbirth; neonatal mortality; birthweight; birthweight < 3rd percentile
Notes	Sample size estimate was given for proteinuric hypertension All women were asked to avoid NSAIDs Adherence: 50% of women in the fish oil group and 57% of women in the placebo group took < 70% of capsules. Protocol variations: 1 woman in the omega‐3 arm took aspirin and 1 women in the placebo arm purchased fish oil privately. Funding: Yorkshire Region Locally Organised Research; GLAXO (Leeds); Sevens Seas (Hull) Declarations of interest: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random numbers were generated by computer.
Allocation concealment (selection bias)	Low risk	Random numbers were kept in sealed, opaque, numbered envelopes in the hospital pharmacy and pharmacy staff allocated the trial treatments.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Placebo capsules were identical to treatment capsules (44% of a subgroup of women identified that they were in the fish oil group).
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Outcome assessments were blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	1/233 (0.4%) post‐randomisation exclusions (one woman with a multiple pregnancy was randomised in error) Adverse events were reported only for a small subsample; 76/233 (33%) women returned the questionnaires.
Selective reporting (reporting bias)	High risk	Limited range of outcomes reported; no SDs reported for continuous outcomes (length of gestation; birthweight).
Other bias	Low risk	Baseline characteristics were similar in supplement and placebo groups.