Keenan 2014.

Methods	RCT: NCT01158976: Nutrition and Pregnancy Study (NAPS)
Participants	64 women randomised (2:1 ratio) Inclusion criteria: pregnant women living in urban low‐income environments, enrolled at 16‐21 weeks' gestation and 'demographically eligible' (Medicaid insured or eligible, African American, and aged 20‐30 years) Exclusion criteria: 2 or more servings of sea fish per week, known medical complications (gestational diabetes, PE, subchorionic haematoma), regular use of steroid medications, regular alcohol use, cigarettes or use of illegal substances (by maternal report), use of blood thinners or anticoagulants, use of psychotropic medications, BMI > 40, allergy to iodine or soy Setting: Pittsburgh, USA: University of Pittsburgh Medical Center obstetric clinics (from 2010‐2012). The clinics are located in urban areas, and provide health services to women living in low‐income households.
Interventions	SUPPLEMENTATION: DHA + EPA + DPA + ETA versus placebo Group 1: omega‐3 LCPUFA (2 gel capsules providing 450 mg DHA, 40 mg DPA and ETA, 90 mg EPA and 10 mg vitamin E): n = 43 Group 2: placebo capsules ‐ matched in size, colour and smell to the study drug (900 mg soybean oil, 16.5 mg vitamin E, 10 mg EPA/DHA): n = 21 Timing of supplementation: 16‐21 weeks GA to birth To support adherence with the intervention, research assistants contacted participants by phone 3 times per week to ask the time of day that the supplement was taken, and gathered data on perception of taste and possible gastrointestinal side effects. DHA + EPA dose/day: mid: 450 mg DHA + 90 mg EPA
Outcomes	Women/birth: Perceived Stress Scale (self report) at 24 and 30 weeks' gestation; Trier Social Stress Test (cortisol response ‐ saliva samples before and after test completion; baseline, 24 and 30 weeks' gestation); symptoms of depression (Edinburgh Postnatal Depression Scale (EPNS)) at 24 and 30 weeks; Difficult Life Circumstances Scale; length of gestation Babies/infants/children: birthweight; 1‐minute Apgar score; cortisol concentrations; BSID‐III at 4 months; Face‐to‐Face Still Face at 4 months
Notes	Funding: capsules provided by Nordic Naturals; supported by NIH grant. “This study was supported by NIH grants R21 HD058269 and RO1HD084586 to Dr Keenan, with additional support from the University of Chicago Institute for Translational Medicine (UL1TR000430). DHA supplement and placebo were supplied by Nordic Naturals”. Declarations of interest: none reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated random assignment Quote: “The pharmacist at the University of Pittsburgh used computer‐generated random assignment of identification numbers to active supplement or placebo in blocks of nine"
Allocation concealment (selection bias)	Low risk	Pharmacist (third party)
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Placebo controlled
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Placebo capsules matched in size, colour and smell to the study drug; probably done
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	13/64 (20%) women were lost to follow‐up: 21% in the omega‐3 group and 19% in the placebo group. 2 participants in the placebo group withdrew due to miscarriage and mood changes; 2 also withdrew in the omega‐3 group ‐ 1 with headaches and 1 with an upset stomach. Data for 15/64 (23%) infants were not available at the 4 months postpartum assessment. End of pregnancy/birth mother and infant outcomes, including length of gestation and birthweight: at the 36‐week gestation data collection point (last before birth), data were collected from 36/43 (84%) and 17/21 (81%) of mothers randomised to the intervention and control groups respectively. Number of mothers and infants from whom data were collected at the end of pregnancy/birth time point was not reported. 3‐month postpartum outcomes, including infant neurological/neurosensory and developmental outcomes and maternal stress: data were collected from 34/43 (79%) and 15/21 (71%) of infants born to mothers randomised to the intervention and control groups respectively.
Selective reporting (reporting bias)	Unclear risk	Not all expected outcomes were reported (e.g. preterm birth).
Other bias	Unclear risk	Only baseline comparisons for maternal mental health characteristics were reported (baseline groups wrong way round in Table 1 of 2014 paper).