Harper 2010.

Methods	RCT: NCT00135902
Participants	852 women randomised Inclusion criteria: women with at least 1 prior spontaneous preterm birth; singleton pregnancies; GA at randomisation between 16 and 22 (21 6/7) weeks. An US examination was required between 14 weeks GA and enrolment to screen for major anomalies. Exclusion criteria: major fetal anomaly or demise; regular intake of fish oil supplements (> 500 mg per week at any time during the preceding month); daily use of NSAIDs; allergy to fish or fish products; gluten intolerant; heparin use or known thrombophilia; haemophilia; planned termination; current hypertension or current use of antihypertensive medications; uncontrolled thyroid disease; type D, F or R diabetes; maternal medical complications; current or planned cerclage; illicit drug or alcohol abuse during current pregnancy; plan to, or give birth at a non‐network hospital; participation in another pregnancy intervention study; participation in this trial in a previous pregnancy; seizure disorder. Characteristics: 30% women never consumed fish or consumed fish < once per month; 9% consumed fish > 3 times a week. Setting: antenatal clinics in 13 network centres, USA: recruitment between January 2005 and October 2006
Interventions	SUPPLEMENTATION: DHA + EPA + PG versus placebo + PG Group 1: 1200 mg EPA; 800 mg DHA for a total of 2000 mg of omega‐3 long‐chain polyunsaturated acids, divided into 4 capsules per day: total number randomised: n = 434 (Source of omega‐3 LCPUFA was deep ocean fish; each capsule contained 10 IU of vitamin E as a preservative.) Group 2: matching placebo (4 capsules containing a minute amount of inert mineral oil per day). Total number randomised: n = 418 Timing of supplementation: 16‐22 weeks GA (mean 19.6 weeks) to 36 weeks GA All women: received 17α‐hydroxyprogesterone caproate (weekly intramuscular injections: 250 mg); participants received no dietary advice as part of the study and otherwise received usual clinical care. DHA + EPA dose/day: high: 800 mg DHA + 1200 mg EPA
Outcomes	Women/birth: fatty acid status; diet (fish intake); DNA; PE; PPH; adverse events; gestational diabetes; preterm birth < 37 weeks; spontaneous preterm birth < 37 weeks; preterm birth < 34 weeks; birth > 40 weeks; low birthweight < 2500 g; SGA; LGA; GA reported as IQR Babies/infants/children: birthweight (median); NEC; RDS (clinical diagnosis of RDS and oxygen therapy (FiO2 > 0.40); neonatal sepsis; BPD; ROP, IVH, perinatal mortality (pregnancy loss and neonatal death); NICU/intermediate care nursery admission; neonatal morbidity composite (ROP, grade III or IV IVH, Patent ductus arteriosus, NEC, culture‐proven sepsis, respiratory morbidity, and perinatal death)
Notes	Compliance run‐in: consenting women received an injection of 250 mg 17α‐hydroxyprogesterone caproate (17P) and a 7‐day supply of placebo capsules. Those who did not return after 5 days and before 21 6/7 weeks GA or had taken < half of the placebo capsules were not allowed to participate. Women passing the compliance run‐in were randomly assigned to EPA/DHA or placebo. Funding: The Eunice Kennedy Shriver National Institute of Child Health and Human Development Grants (HD27860, HD27917, HD40560, HD34208, HD40485, HD21410, HD27915, HD40500, HD40512, HD40544; MO1‐RR‐000080; HD34136; HD27869; HD40545; HD36801 and HD19897). Declarations of interest: "Dr Esplin serves on the scientific advisory board and holds stock in Sera Prognostics, a private company that was established to create a commercial test to predict preterm birth and other obstetric complications. Dr Manuck is also on the scientific advisory board for Sera Prognostics". None of the other authors reported any conflicts of interest.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: “randomly assigned”; “simple urn method of randomization with stratification according to clinical center to create a randomization sequence for each center”
Allocation concealment (selection bias)	Low risk	Quote: “randomly assigned”; “simple urn method of randomization with stratification according to clinical center to create a randomization sequence for each center”
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: “double‐masked”; "Study group assignment was not known by study participants, their health care providers or the research personnel”; placebo control
Blinding of outcome assessment (detection bias) All outcomes	Low risk	After giving birth, “records of the participants and their newborns were reviewed by study personnel, unaware of treatment assignments, who abstracted delivery date, birth weight, occurrence of maternal or neonatal complications and interventions”
Incomplete outcome data (attrition bias) All outcomes	Low risk	No losses to follow‐up (for primary outcome) Denominators for liveborns = 427 and 410 (2 neonates in the omega‐3 group and 7 in the placebo group died before admission to NICU and were not included in liveborn neonate outcomes).
Selective reporting (reporting bias)	Low risk	Large number of relevant outcomes reported (some as medians).
Other bias	Low risk	Baseline demographics, risk factors for recurrent preterm birth and dietary fish intake were similar between omega‐3 and placebo groups.