Examining the Effect of Fish Oil Supplementation in Chinese Pregnant Women on Gestation Duration and Risk of Preterm Delivery.

BACKGROUND: Intervention studies have shown that long-chain (LC) n-3 PUFA intake can prolong gestation but the dose-time-effect relations remain unresolved. OBJECTIVES: We examined the effect on gestation duration of 2 doses of supplemental LC n-3 PUFAs. METHODS: We undertook a 3-group parallel randomized controlled trial in areas of China with low (median: 2.1 g/d) and higher (14.3 g/d) fish intake. Unselected women (median: age, 26.2 y; BMI, 20. kg/m2) were randomly assigned at midgestation to take four 0.72-g identical gelatin capsules per day until the last day of the preterm period (<259 days of gestation), when they were asked to stop. Capsules contained fish oil [high fish oil (HFO) group (60% w/w LC n-3 PUFAs)], a 1:3 mixture of fish oil to olive oil [low fish oil (LFO) group (20%)], or olive oil [control (CON) group (0%)], providing 2.0, 0.5, and 0 g/d of LC n-3 PUFAs, respectively. Habitual fish intake was recorded at baseline. Hazard rate ratios (HRRs) for spontaneous delivery <259 days of gestation and <273 days of gestation across groups were estimated by Cox regression. RESULTS: Among 5531 women randomly assigned, 92.5% were included in analyses (1706/1825, 1695/1851, and 1717/1855, respectively). The groups were similar with respect to hazard rates <259 days of gestation [HRR: 1.04 (95% CI: 0.70, 1.53) for LFO compared with CON and 0.90 (95% CI: 0.60, 1.35) for HFO compared with CON], hazard rates <273 days of gestation [HRR: 1.00 (95% CI: 0.86, 1.18) and 0.91 (95% CI: 0.77, 1.07), respectively], and mean gestation durations [differences: 0.2 d (95% CI: -0.5, 0.8) and 0.6 d (95% CI: -0.06, 1.2), respectively]. Inspection of pregnancy survival curves suggested that LC n-3 PUFAs delayed delivery in low fish consumers by 5-10 d and that this effect ceased rapidly after stopping taking the capsules. CONCLUSION: This trial could not substantiate that fish oil prevents preterm birth in Chinese women, possibly because statistical power was too low. This trial was registered at clinicaltrials.gov as NCT02770456.