Healthy eating in pregnancy

BMJ 2014; 348 doi: http://dx.doi.org/10.1136/bmj.g1739 (Published 04 March 2014) Cite this as: BMJ 2014;348:g1739
  1. Lucilla Poston, Tommy’s professor of maternal and fetal health; head of division of women’s health
  1. 1King’s College London, London, UK
  1. lucilla.poston{at}kcl.ac.uk

Always a good idea, now with more supporting evidence

Most pregnant woman want to know the best foods to eat and what to avoid. The possible dangers of eating liver, the need to avoid unprocessed cheeses and too much tuna, and, above all, the importance of folate supplements are widely appreciated. Two linked papers from Australia (Dodd and colleagues; doi:10.1136/bmj.g1285) and Norway (Englund-Ogge and colleagues; doi:10.1136/bmj.g1446) have explored whether eating “healthily” can improve clinically important pregnancy outcomes.1 2

Dodd and colleagues looked at the global prevalence of overweight and obesity among women in antenatal care and the associated complications, including risks of gestational diabetes and delivery of overlarge infants. Women in their LIMIT trial were advised to maintain a balance of carbohydrates, fat, and protein in their diet; to reduce consumption of foods rich in refined carbohydrates and saturated fat; and to increase intake of fruit, fibre, and vegetables; a regimen that should reduce insulin resistance, central to the complications of obesity in pregnancy. They were also encouraged to take physical exercise. This randomised controlled trial of a lifestyle intervention in overweight and obese women was adequately powered to address clinical outcomes. Previously small and often poorly designed studies have been undertaken with a surrogate clinical endpoint, notably gestational weight gain. Meta-analysis of these has shown no evidence for improved clinical outcomes.3

This well conducted trial shows how difficult it can be to improve health outcomes in overweight or obese individuals through lifestyle recommendations as they found no differences in the primary outcome, the delivery of infants large for gestational age (>90th centile) and no reduction in gestational diabetes or any of the common maternal complications. The 18% relative risk reduction in the incidence of macrosomia (birth weight above 4000 g), a pre-stated secondary outcome, will nonetheless attract interest. As highlighted by the authors, macrosomic infants are at increased risk of obesity. In a new report of 7738 adolescents in the United States, just 12% of these young people had macrosomia at birth, but they represented a remarkable 36% of those obese at the age of 144; the implication being that reducing macrosomia should lower the risk of childhood obesity. Ongoing trials such as the UK UPBEAT study5 will confirm or refute this reduction in macrosomia by dietary intervention, and, importantly, the LIMIT childhood follow-up will show whether the reduction in macrosomia affects childhood obesity.

LIMIT, an exemplary study in design and execution, was ineffective in improving maternal outcomes, and the challenge remains to find an effective prevention strategy. What are the alternatives? Could the LIMIT intervention have been too light a touch? The UK UPBEAT trial, due to complete in 2014, offers a more intensive intervention led by health trainers.5 Metformin treatment is also being evaluated in ongoing trials, but runs the risk of “medicalising” pregnancy. The LIMIT authors comment that a more intense lifestyle intervention, requiring more time from women and staff, would be impractical. Even more impractical is the notion that healthcare resources could support an intervention for all overweight and obese pregnant women, who make up about half of the antenatal population in Australia and the UK. Also, 70-80% of overweight and obese women do not develop complications. Should any of the ongoing dietary/lifestyle or pharmacological intervention trials prove effective, about three in four overweight/obese women would then be treated unnecessarily in clinical practice. While the focus should remain on finding effective interventions, the ability to identify early in pregnancy those obese women most at risk would be a bonus. Successful treatment strategies could then be targeted to those who would benefit. More importantly, we should look to a worldwide public health campaign recommending that women try to achieve a normal BMI before attempting to conceive and, particularly, after pregnancy as weight retention postpartum is a major health risk for the next pregnancy.6

Englund-Ogge’s study in pregnant Norwegian women adopted a different tactic to explore diet and pregnancy outcomes. The Norwegian Mother and Child Cohort Study provided an unparalleled depth and breadth of dietary information for this investigation. Using data from validated food frequency questionnaires from 66 000 women, the authors describe how the diets of the pregnant women could be grouped according to three dietary patterns; “prudent,” “western,” and “traditional”; this was achieved by complex analysis of all foods eaten, looking for those that grouped together in an individual’s diet. Dietary foodstuffs often fall into these categories as individuals tend to consume related foods that together describe healthy or unhealthy choices or ethnically associated patterns. Dietary “patterns” are also a readily understandable concept and translatable to dietary advice. Women who ate “prudently” had a lower incidence of premature delivery after adjustment for confounders including age, history of preterm delivery, maternal education, and parity. The prudent diet comprised vegetables, fruit, water as a beverage, fibre rich bread, whole grain cereals, and other wholesome foods. As preterm delivery arises from spontaneous labour as well as induced labour or caesarean section (iatrogenic labour), several subgroups of women were explored. A prudent diet was most strongly associated with a lower incidence of spontaneous labour and with late preterm delivery. As the authors acknowledge, causal links between healthy diets and preterm birth cannot be claimed because of the risk of residual confounding, and here one wonders if socioeconomic status was adequately accounted for. Nonetheless, how might a prudent diet protect against premature delivery? The consumption of probiotic milk products, and anti-inflammatory properties of fruit and vegetables are mentioned, somewhat speculative but perhaps more plausible than the suggested dampening of the hypothalamic pituitary axis, the role of which in the onset of labour in humans is not established.

The authors build on several studies that have proposed the benefit of a diet rich in fruit and/or vegetables in prevention of premature birth, including a recent report in this journal showing that fruit intake before pregnancy is a factor that relates to healthy outcome in nulliparous pregnant women.7 Health professionals would therefore be well advised to reinforce the message that pregnant women eat a healthy diet.


Cite this as: BMJ 2014;348:g1739


  • Research doi:, doi:10.1136/bmj.g1446 10.1136/bmj.g1285
  • Competing interests: I have read and understood the BMJ Group policy on declaration of interests and declare the following interests: I am chief investigator on the UPBEAT trial, and my institution received money to fund the SCOPE study, a prospective international cohort of 5690 nulliparous pregnant women. I have received payment from the food and pharmaceutical industry for lecturing, for one piece of research unrelated to the subject of this editorial, and for another related to the subject of this editorial.

  • Provenance and peer review: Commissioned, not externally peer reviewed.