Effects of interventions in pregnancy on maternal weight and obstetric outcomes: meta-analysis of randomised evidence

This review claims that diet intervention leads to a reduction in weight gain of 4 kg. However, the 2 largest studies constituting 73% of the included patients in the diet subgroup, only reduced weight gain by 1.7 and 2.2 kg. So how can the meta-analysis come to this result? In the case of extreme heterogeneity, as in this case (I^2=92%) care in the interpretation and analysis of the data needs to be taken. Using a random effects model will mean that the two studies with a 1000 patients will receive the same weight as the studies that include 20 patients. With 8 smaller studies the 2 large studies will relatively not get any weight. This inappropriate use of random effects model has led the authors to the exaggerated estimate. A fixed effect model would give a result of around 2 kg, but these studies might be too different to combine.

The two large studies also allowed insulin treatment primarily in the "diet arm" which raises concerns of whether the effects were caused by diet or insulin.

We refer to the comprehensive systematic review and meta-analysis of randomized controlled interventions to reduce gestational weight gain (GWG) in pregnancy by Thangaratinam et al [1]. The mean reduction in GWG across all studies was relatively small at 1.42kg (0.95-1.89kg). Clearly, reducing GWG in pregnant women is challenging, even though small differences do contribute to a better outcome.

Anecdotally, women diagnosed with gestational diabetes (GDM) gain less weight than those without GDM after this diagnosis is made, presumably secondary to changes in their diet and activity levels. However, there is little formal documentation of this in the literature.

We prospectively recorded GWG in 108 women from diagnosis of GDM at 26-28 weeks, to 36 weeks gestation, compared with 339 women without GDM. In this brief 9 week period, GDM women gained a mean of 2.1 ± 2.51kg compared to 5.16 ± 2.42 kg in women without GDM (p<0.001), resulting in less total GWG (12.2 ± 5.17 kg versus 15.0 ± 5.03 kg, p<0.001) (figure 1). GWG prior to the diagnosis of GDM was similar in both groups using self-reported prepregnancy weight.

GDM women had a greater mean estimated prepregnancy body mass index (BMI) of 24.5 ± 4.9kg/m2 versus 22.8 ± 3.9 kg/m2 (p<0.001) and were more likely to be non Caucasian (57.4% versus 26.5%, p<0.001). Third trimester weight gain remained significantly decreased in GDM women after accounting for BMI category and ethnicity. GWG was not associated with blood glucose levels at either the fasting or 2 hour time point on the 75g diagnostic glucose tolerance test, suggesting that even women with mild GDM are capable of losing weight with interventions for GDM. As we used fasting and 2 hour cut-offs of 5.5mmol/L and 8.0mmol/L respectively, some of these women would no longer fit the criteria for GDM based on the recent IADPSG diagnostic criteria.

The primary management goal in GDM is achieving target blood glucose levels on home blood glucose monitoring, not weight reduction. Women at our centre were reviewed once by a diabetes educator (with an additional visit if they required insulin) for education on the importance of blood glucose control, home blood glucose monitoring and dietary advice, advocating regular and well-spaced low glycemic index carbohydrates and a low fat diet. They were advised to undertake 30 minutes of exercise per day such as walking. They were not routinely assessed by a dietician nor by an exercise physiologist. They were reviewed every 1-3 weeks by the diabetes clinic medical team that instituted pharmacotherapy if their blood glucose levels were above target range.

Perhaps a key difference between women with GDM and those who participate in weight loss interventional programs with the sole aim of reducing weight gain is the availability of daily performance feedback indicators such as home blood glucose monitoring, as well as a strong motivation to reduce the perinatal complications of GDM and to avoid insulin treatment. Yet most interventional trials to reduce GWG focus on adherence to a diet or exercise program as a primary goal, rather than the psychology or goal setting strategies behind creating behavioural change in pregnant women [2].

Further research should aim to evaluate the key psychological motivators and performance feedback indicators that facilitate lifestyle and dietary behavioral change in pregnancy for the benefit of mother and baby.

1. Thangaratinam, S., et al., Interventions to reduce or prevent obesity in pregnant women: a systematic review. Health technology assessment, 2012. 16(31): p. iii-iv, 1-191.
2. Brown, M.J., et al., A systematic review investigating healthy lifestyle interventions incorporating goal setting strategies for preventing excess gestational weight gain. PloS one, 2012. 7(7): p. e39503.

How have you calculated the mean values for all interevntions combined because I cannot work out how if you average 3.84, 0.72 and 1.06 you get 1.42kg instead of 1.87kg; the averages for the sample that exclude diabetics also don't add up to the expected average