The potential benefits of using fish oil in relation to preterm labor: the case for a randomized controlled trial?

Preterm delivery is a leading cause of perinatal mortality and morbidity. Preterm labor, which occurs in approximately 10% of pregnancies, is an underlying factor for 85% of the deaths of normally formed babies. In addition, the sequelae of premature delivery may include significant physical and neurological handicap. Survival of the very premature infant has improved with advances in neonatal intensive care, in which these babies often stay for weeks at a cost of approximately 1500 Euros per day. Tocolytics are drugs that inhibit uterine contractions (1). They are widely used in emergency situations to stop or delay premature labor. Drugs for tocolysis include betamimetics (e.g. ritodrine, terbutaline), oxytocin receptor antagonists (e.g. atosiban), magnesium sulfate, calcium blockers (e.g. nifedipine) and cyclooxygenase inhibitors (e.g. indomethacin). The purpose of using tocolytics is primarily to delay preterm delivery and gain time to prepare the mother and the fetus. This may include transfer of the mother to a hospital with neonatal intensive care facilities, time for glucocorticoid treatment for fetal lung maturation, and any causal treatment targeted specifically at the presumed underlying cause of the premature labor (e.g. antibiotics to fight infections). Although several tocolytics are able to prolong pregnancy (2), it has proved difficult to show that they do in fact improve neonatal health as indicated by hard outcomes (2, 3), or even that they reduce the occurrence of preterm delivery (4). Moreover, several tocolytics may give rise to adverse effects in the woman or the fetus (1, 3). Therefore, it is mandatory to develop new treatment regimens for tocolysis that are more effective than the existing ones, and that minimize any untoward effects. Long-chain n-3 fatty acids have been shown to delay the timing of spontaneous delivery (5, 6) and to reduce the recurrence of preterm delivery (6), and there are also indications that they may have tocolytic properties (7). In this paper we review the potential benefits of long-chain n-3 fatty acids in relation to premature labor. The long-chain n-3 fatty acids are primarily eicosapentaenoic acid (EPA, 20:5n3) and docosahexaenoic acid (DHA, 22:6n6). They are abundant in fat from animals of marine origin (8) and may thus be obtained by eating seafood or taking fish oil or fish liver oil. The evidence that these fatty acids can delay spontaneous delivery comes primarily from randomized controlled trials. One trial was carried out in Aarhus in Denmark (5). Pregnant women who received fish oil from the 30th week of gestation had longer mean duration of pregnancy of 4.0 days [95% confidence interval (CI) 1.5–6.4] and a corresponding increase in mean birthweight of 100 (1–214) g compared with women in a control group receiving olive oil. The effect on birthweight could be fully attributed to the longer gestation, and there seemed to be no effect on the fetal growth rate. Habitual intake of fish was recorded at the randomization. Among the 20% of women reporting the lowest intake of fish at entry, the mean gestation length was 7.4 (2.2–12.6) days longer in the fish oil group, while among the 20% with highest intake of fish no differences (− 8.4 to 5.1 days) could be detected; among the middle 40%, the mean gestation length was 4.8 (1.8–7.8) days longer in the fish oil group. FOTIP (fish oil trials in pregnancy), a multicenter randomized controlled trial, was conducted in 19 hospitals in seven European countries during the period 1990–96 (6). The primary aim was to test the possible preventive effect of fish oil supplementation on preterm delivery, intrauterine growth retardation and pregnancy-induced hypertension in high-risk pregnancies. Six different trials were undertaken simultaneously to address these questions. When all six trials were aggregated (n = 1619), women receiving fish oil had delayed onset of spontaneous delivery compared to women receiving olive oil, as estimated by Cox's regression model, which could account for elective delivery that occurred among 40% of the women (proportional hazards ratio 1.22; 95% CI 1.07–1.39, p = 0.002). In the trial with women who had had a preterm delivery in an earlier pregnancy (n = 228), the recurrence risk was reduced from 33% to 21%[odds ratio (OR) 0.54; 0.30–0.98]; in this trial, risk of early delivery (i.e. 6 weeks before term) was reduced from 13.3% to 4.6% (OR 0.32; 0.11–0.89), mean gestation increased by 8.5 (1.9–15.2) days and mean birthweight by 209 (27–390) g. No effect was detected on preterm risk in the other high-risk group, which was women pregnant with twins (n = 504), and no effects were detected on fetal growth or on preeclampsia risk. Of particular interest here is that two of the trials that recruited women with problems in the current pregnancy (either threatening preeclampsia or suspected intrauterine growth retardation, total n = 136) provided a higher dose of long-chain n-3 fatty acids, namely 6.1 g n-3 fatty acids per day from around week 33 of gestation. In these women the delaying effect of fish oil appeared to be greater, as expressed by the proportional hazards ratio (2.00; 1.16–3.45, p = 0.01), than in the other four trials that had provided 2.8 g per day to the women; by inspection of the survival curves, delivery seemed to be delayed by 5–15 days by fish oil in these two trials. Two smaller trials with fish oil, carried out in Leeds (9) and Groningen (10) (n = 62 and n = 263), were not able to detect an effect on duration of pregnancy. However, the studies focused on women at high risk of preeclampsia and a large proportion of the deliveries were likely to be elective, which limits the possibility of detecting any effect on the timing of spontaneous delivery with conventional statistical methods. In 1938–39 a trial with alternate allocation was conducted in London with more than 5000 pregnant women 11-14). Half of the women were supplemented with a cocktail of minerals and vitamins, whereas the other half did not receive any. Vitamins A and D were provided as halibut-liver oil, which also contains n-3 fatty acids, albeit in small amounts relative to other fish liver oils. It is noteworthy that fewer women delivered early in the intervention group compared to the control group (505/2510 vs. 603/2511, p = 0.0008), and that later trials with the other substances provided in the study (e.g. vitamin A) have not been able to show convincing effects on duration of pregnancy (11). However, if this were an effect of n-3 fatty acids, this would mean that even very low doses of n-3 fatty acids (0.1 g/day in the London trial) can delay delivery in women with a low intake of n-3 fatty acids. There is also observational evidence in support of the notion that n-3 fatty acids may delay spontaneous delivery. A positive association has been shown between biomarkers for intake of these fatty acids and length of gestation (15), and in a study with more than 8000 Danish women, low fish consumption was shown to be a strong risk factor for preterm delivery (16). Furthermore, some ecological comparisons agree with this possibility. Gestations are longer and birthweight higher in the fish-eating community of the Faroe Islands compared to Denmark (this comparison gave rise to the hypothesis) (17, 18), and a similar pattern has been identified in the Orkney Islands compared to the Scottish mainland (19). A group of Indians living in London, presumably consuming no fish and relative high amounts of linoleic acid (20), have also been shown to have substantially shorter gestations (21) than a comparable group of Caucasians in London (15). Finally, some rat experiments have shown that a high intake of long-chain n-3 fatty acids, relative to n-6 fatty acids, throughout pregnancy is associated with delayed timing of delivery (22). For tocolysis in a clinical setting, the effects of these fatty acids on the delivery mechanism need to be fast, or immediate. There are indications that this might well be the case. As mentioned, in the randomized trials fish oil was shown to be able to delay timing of delivery substantially when it was given from week 30 (15), and even from week 33 of gestation (6). An observational study used fatty acids measured in erythrocyte phopholipids as a biomarker for intake (23); this study also supports the notion of a relatively fast effect. The association seen with length of gestation was stronger when fatty acids were measured in a phospholipid fraction (i.e. phophatidylcholine) with fast turnover of fatty acids – and therefore probably reflecting intake during a relatively narrower time window prior to the blood sampling that took place in week 37 of gestation – than if the fatty acids were measured in another phospholipid fraction (phosphatidyl ethanolamine) (23) that has been shown to have slower turnover (24). Linoleic acid, an n-6 fatty acid, is the parent fatty acid of the prostaglandins, PGF2alpha and PGE2, which are essential to the delivery process. Leat & Northrop performed and interesting experiment (25). They showed that rats fed a diet throughout pregnancy that was low in linoleic acid but high in linolenic acid (the parent n-3 fatty acid) have impaired parturition; this was expected form earlier studies. Interestingly in the present context was that, when the investigators from day 19 of pregnancy (which is only two days before expected date of delivery in the rat) exchanged linolenic acid with linoleic acid, parturition was normalised. Thus, there was clearly a fast effect of a dietary fatty acid on the parturition mechanism in the rat. Recently, an intriguing experiment with sheep was published (7). In this animal model premature labor can be induced by glucocorticoids, allowing various tocolytics to be tested. Six sheep had intravenous infusion with n-3 fatty acids after the glucocorticoid treatment, whereas six control sheep had an infusion with neutral lipids. In the fish oil group the onset of both labor and delivery was delayed. In two animals receiving n-3 fatty acids, premature labor stopped completely, a phenomenon allegedly never seen with any other tocolytic agent in this animal model by the investigators, who have extensive experience in this field of research. These findings suggest that, in the sheep, long-chain n-3 fatty acids not only have acute effects on the timing of delivery but also may have tocolytic properties. The delaying effect on timing of delivery has been thought to be mediated through modulation of the formation of prostaglandins in the body. Thus, the long-chain n-3 fatty acids may reduce the activity of eicosanoid promoters of the parturition process, particularly prostaglandins F and E, and increase the activity of eicosanoids with myometrial-relaxant properties, particularly prostacyclins (18, 26). Some biochemical studies have indeed confirmed that EPA + DHA may influence the formation of prostaglandins in the fetal membranes (27, 28). However, the investigators who carried out the sheep experiment also showed that infusion with long-chain n-3 fatty acids was associated with relatively lower formation of myometrial prostaglandin H synthase 2 (PGHS2, which is the main rate-limiting enzyme for producing prostaglandins in the myometrium) mRNA compared to the control group, whereas they saw no difference between the groups in levels of myometrial PGHS1 mRNA and oxytocin receptor mRNA; they suggested that this might reflect part of the mechanisms (29). Long-chain n-3 fatty acids may have an effect on the electrical activity of the heart, and this may be the underlying mechanism for fish oil's antiarrhythmic properties (30), and the preventive effect against sudden cardiac death following a myocardial infarction seen in randomized controlled trials (31). As the myometrium has constant electrical and contractile activity (29), even in the prelabor period, one possibility is that the long-chain n-3 fatty acids could have a similar “antiarrhythmic” effect on the myometrium, which could then explain the delaying effect on initiation of labor as well as their possible tococlytic properties. Prolonging pregnancy may not always be advantageous, even for the premature fetus. The clinician in charge of each case will decide whether tocolytic therapy should be given or not, usually after careful evaluation of the clinical parameters and tests for fetal well-being. In the European multicenter trial with 1617 women (6), cases of postterm delivery was substantially increased in the fish oil arm. Therefore, any treatment with fish oil supplementation should be stopped whenever pregnancy has gone beyond the preterm period (6). Fish oil has been shown to increase bleeding time, a laboratory measure. However, the clinical relevance of this is uncertain, and in the trials mentioned (5, 6), where a total of 2150 women were randomized, no significant increases were detected in bleeding complications in the fish oil group. However, in any future trial of fish oil to pregnant women, women and infants should be carefully monitored for any possible bleeding complications, including vaginal bleeding and intracerebral hemorrhage. Indomethacin is known to give rise to in utero closure of ductus arteriosus (32), and in some animals cyclooxygenase 2 (COX2) inhibitors seem to have similar effects (33). Prenatal closure of ductus arteriosus may result in right ventricular overload and even in cardiac arrest in the fetus; in the neonate, it may give rise to pulmonary hypertension. As long-chain n-3 fatty acids share some of the actions of the prostaglandin synthase inhibitors, it has been suggested that these fatty acids may also have the potential to give rise to premature closure of ductus arteriosus (18). This too needs to be carefully monitored in any future trials. However, it should also be kept in mind that long-chain n-3 fatty acids are a natural food constituent, and that giving fish oil is a gentler approach for tocolysis than providing pharmaceutical agents with potent physiological actions. Randomised controlled trials indicate that long-chain n-3 fatty acids supplementation can influence the process of parturition. Onset of labor is delayed and recurrence risk of preterm delivery reduced, and in animal studies these fatty acids, appear to have a tocolytic effect. It is possible that fish oil may be effective in delaying reinitiation of labor after successful tocolytic treatment with other agents. However, the tocolytic properties demonstrated in the sheep experiment suggests that fish oil also has the potential to become a first-line tocolytic agent, either on its own or in conjunction with other tocolytic treatments. We think the evidence is sufficiently compelling that these possibilities should be tested within the framework of a randomized controlled trial. A rigorous testing would undoubtedly have to be carried out on a multicenter and multinational basis. We will hereby encourage comments and criticisms of the idea; these may be submitted either to Acta Obstetricia et Gynecologica Scandinavica or directly to us. The first author is supported by the March of Dimes Birth Defects Foundation and the Danish National Research Foundation.