1. Objectives

To assess the effectiveness of interventions using diet, exercise, or both for the prevention excessive weight gain during pregnancy

2. How studies were identified

The following databases were searched in November 2014:

• Cochrane Pregnancy and Childbirth Group’s Trials Register
• CENTRAL (The Cochrane Library)
• MEDLINE
• Embase

Reference lists, relevant journals and conference proceedings were handsearched and the review authors directly contacted researchers

3. Criteria for including studies in the review

3.1 Study type

Randomized controlled trials. Quasi-randomized trials were not eligible

3.2 Study participants

Pregnant women of any body mass index (BMI) category

(Studies recruiting women with a normal BMI (generally >18 to <25 kg/m²) were considered to have a low risk of weight-related complications of pregnancy at baseline; those recruiting women with any BMI were considered mixed-risk; and studies recruiting overweight or obese women (BMI ≥25 to <30 kg/m² and ≥30 kg/m², respectively) were considered to be at high risk)

3.3 Interventions

Any intervention for preventing excessive weight gain in pregnancy using diet, exercise, or both, compared with standard or routine care for preventing excessive weight gain in pregnancy

3.4 Primary outcomes

• Excessive gestational weight gain (as defined by trialists)

(Gestational diabetes is the primary outcome of separate Cochrane reviews and was therefore not included in this review)

Secondary outcomes for the mother included mean weight gain, low weight gain (as defined by trialists), preterm birth, preterm prelabour rupture of membranes, preeclampsia or eclampsia, hypertension (not pre-specified), induction of labour, caesarean delivery, postpartum complications (postpartum haemorrhage, wound infection, endometritis, antibiotic treatment, perineal trauma, thromboembolic disease, maternal death), behaviour modification outcomes (diet, physical activity), and maternal weight retention postpartum. For the infants, secondary outcomes included birth weight (not pre-specified), macrosomia (birth weight >4000 g or greater than the 90^th centile for gestational age and sex), complications related to macrosomia (hypoglycaemia; hyperbilirubinaemia; infant birth trauma including palsy, fracture, and shoulder dystocia; respiratory distress syndrome), and childhood weight

4. Main results

4.1 Included studies

Sixty-five randomized controlled trials were included in this review, of which 49 studies enrolling 11,444 women contributed data to quantitative meta-analysis

• Two trials were cluster-randomized and the remaining trials were individually randomized
• Sample sizes ranged from 12 to more than 2000 women
• Forty-eight trials recruited women at <20 weeks’ gestation, 13 trials recruited women >20 weeks’ gestation, and in the remaining trials gestational age at recruitment was not specified
• Thirty-four trials recruited women from the general population, with eight of these trials reporting results separately for low/normal weight versus overweight/obese groups; 24 trials recruited women with overweight or obesity; seven trials recruited women with, or at high risk of, gestational diabetes; and one trial recruited women who had previously had a baby with macrosomia
• Eight studies used diet-only interventions; diet and exercise counselling interventions were employed in 25 trials; exercise interventions were used in 20 trials; diet counselling/other interventions were employed in seven trials; and diet and supervised exercise interventions were used in five trials
• Diet interventions were usually a low glycaemic load, diabetic, calorie-controlled, or low fat diet, while exercise interventions were most often of moderate intensity, and involved walking, dance, or aerobic classes

4.2 Study settings

• Australia (11 trials), Belgium (2 trials), Brazil (4 trials), Canada (5 trials), Colombia, Denmark (3 trials), Finland (3 trials), Germany, Ireland, Italy (2 trials), Kosovo, the Netherlands (2 trials), Norway (2 trials), Spain (3 trials), Sweden (2 trials), the United Kingdom of Great Britain and Northern Ireland, and the United States of America (20 trials), and Taiwan, China
• Most studies (59/65 trials) were conducted in high-income settings, and only two of these trials specifically recruited women of low socioeconomic position
• Most trials recruited women via clinics or hospitals providing antenatal care

4.3 Study settings

How the data were analysed
Six comparisons were made: i) all diet and/or exercise interventions versus routine care; ii) Low glycaemic load diet interventions versus routine care; iii) diet and exercise counselling versus routine care; iv) exercise versus routine care; v) diet and supervised exercise versus routine care; and vi) diet counselling versus routine care. Cluster-randomized data were combined with individually randomized data provided they had been adjusted for clustering. For studies including multiple arms, the control group was divided between the arms for comparison. Random effects meta-analysis was used to summarize the data. For dichotomous outcomes, risk ratios (RR) with corresponding 95% confidence intervals (CI) were calculated. For continuous outcomes, mean differences (MD) or standardized mean differences (SMD) and 95% CI were used, depending on whether measurements were made on different or the same scales, respectively. Sensitivity analyses were conducted excluding trials at risk of bias. Potential sources of heterogeneity were explored using subgroup analyses for the outcomes excessive gestational weight gain, mean gestational weight gain, low weight gain, preterm birth, caesarean section, preeclampsia, and macrosomia. Subgroup analysis was based on the underlying risk of the study population of suffering adverse effects due to poor weight control during pregnancy:

• Low-risk women: normal weight (BMI <25 kg/m²)
• Mixed-risk women: general population (any BMI category)
• High-risk women: overweight/obese (BMI ≥25 kg/m²), or women with, or at risk of, diabetes mellitus

Results
Diet and/or exercise interventions versus routine care in all pregnant women
Excessive gestational weight gain
The risk of excessive gestational weight gain was reduced by 20% among women randomized to interventions employing diet, exercise, or both in comparison with those randomized to routine care (RR 0.80, 95% CI [0.73 to 0.87], p<0.00001; 24 trials/7096 women). This finding was relatively consistent across most intervention types, with the greatest effect found for diet and supervised exercise interventions, and no significant effect found for diet counselling/other interventions. Sensitivity analysis excluding five trials considered to be at high risk of bias also produced a similar effect.

Maternal additional outcomes
Thirteen of 36 trials reporting on mean gestational weight gain found a statistically significant effect in favour of the intervention group (data not pooled). In meta-analysis of 11 trials involving 4422 women, a 14% increase in the risk of low weight gain was observed in the intervention group (RR 1.14, 95% CI [1.02 to 1.27], p=0.025). The risk of maternal hypertension was reduced by 30% with diet and/or exercise interventions (RR 0.70, 95% CI [0.51 to 0.96], p=0.027; 11 trials/5162 women). Overall, there was no effect of diet and/or exercise on mean postpartum weight retention in kilograms; however, the risk of postpartum weight retention (time frame defined by trialists) was reduced by 22% (RR 0.78, 95% CI [0.63 to 0.97], p=0.023; 5 trials/902 women). Mean energy intake (MD -570.8 kilojoules, 95% CI [-894.3 to -247.3], p=0.00054; 12 trials/4065 women), mean fibre intake (MD 1.53 g, 95% CI [0.94 to 2.12], p<0.00001; 8 trials/3466 women), and physical activity score at 26 to 29 weeks’ gestation (SMD 0.40, 95% CI [0.18 to 0.61], p=0.00029; 9 trials/2851 women) were all improved with diet and/or exercise interventions. The risks of preterm birth, preeclampsia, induction of labour, caesarean delivery and postpartum haemorrhage were not statistically significantly different between treatment and control groups.

Infant additional outcomes
The risk of neonatal respiratory distress syndrome was reduced by 53% in pooled analysis of two diet and exercise counselling trials involving 2256 high-risk women (RR 0.47, 95% CI [0.26 to 0.85], p=0.012). No statistically significant differences between intervention and control groups were found for the outcomes mean birth weight, birth weight >4000 g, birth weight >90th centile, low birth weight <2500 g or <10^th centile, shoulder dystocia, neonatal hypoglycaemia, neonatal birth trauma, or neonatal hyperbilirubinaemia.

Low glycaemic load diet versus routine care
Excessive gestational weight gain
Pooled analysis of two trials involving 833 women found a statistically significant effect of low glycaemic load diets on the risk of excessive gestational weight gain in comparison with routine care (RR 0.74, 95% CI [0.55 to 0.99], p=0.045). When restricted to high-risk women, the effect became non-statistically significant (RR 0.81, 95% CI [0.61 to 1.08], p=0.15; 2 trials/516 women).

Additional outcomes
For the outcomes low gestational weight gain, preterm birth, caesarean delivery, and birth weight >4000 g, results were not indicative of an effect. Data were not pooled for the outcome mean gestational weight gain due to heterogeneity, with two of five trials reporting on this outcome finding a reduced weight gain in the intervention group. No data were available for other outcomes.

Diet and exercise counselling versus routine care
Excessive gestational weight gain
While low-risk women had a reduced risk of excessive weight gain with diet and exercise counselling versus routine care (RR 0.72, 95% CI [0.55 to 0.95], p=0.022; 2 trials/247 women), the effect was not statistically significant in mixed-risk or high-risk populations (RR 0.98, 95% CI [0.83 to 1.15], 1 trial/219 women, and RR 0.85, 95% CI [0.71 to 1.02], 10 trials/2725 women, respectively).

Additional outcomes
Mean gestational weight gain was lower in the intervention group in mixed-risk and high-risk populations (MD -1.80 kg, 95% CI [-3.36 to -0.24], p=0.023; 3 trials/444 women, and MD -0.71 kg, 95% CI [-1.34 to -0.08], p=0.028; 12 trials/2741 women, respectively), but not in low-risk groups (MD -0.92 kg, 95% CI [-2.12 to 0.29], 2 trials/241 women). Overall, the risk of caesarean delivery was borderline statistically significantly reduced with diet and exercise counselling (RR 0.89, 95% CI [0.80 to 1.00], p=0.045; 10 trials/3406 women). Among high-risk women, infant birth weight >4000 g was also borderline statistically significantly reduced (RR 0.85, 95% CI [0.73 to 1.00], p=0.046; 10 trials/3252 women). For the outcomes low gestational weight gain, preterm birth, preeclampsia, results were not indicative of an effect. No other outcomes were reported on in the included trials.

Exercise interventions versus routine care
Excessive gestational weight gain
Mixed-risk populations (RR 0.77, 95% CI [0.66 to 0.88], p=0.00022; 3 trials/1592 women) and high-risk populations (RR 0.84, 95% CI [0.73 to 0.95], p=0.0077; 5 trials/690 women) were at significantly lower risk of excessive gestational weight gain with exercise interventions in comparison to routine care, while low-risk populations were not (RR 0.69, 95% CI [0.47 to 1.02], 2 trials/953 women).

Additional outcomes
Mean gestational weight gain was lower in the intervention group in low-risk populations (MD -1.50 kg, 95% CI [-2.08 to -0.92], p<0.00001; 1 trial/687 women), but not in mixed- or high-risk populations (MD -1.00 kg, 95% CI [-2.01 to 0.01], 4 trials/1196 women and MD -0.34 kg, 95% CI [-1.15 to 0.47], 5 trials/548 women, respectively). The risk of low gestational weight gain was statistically significantly increased among low-risk women in the exercise intervention group (RR 1.29, 95% CI [1.06 to 1.58], p=0.011; 1 trial/687 women), but not among mixed- or high-risk women (RR 1.20, 95% CI [1.00 to 1.43], p=0.055; 2 trials/1336 women, and RR 1.03, 95% CI [0.66 to 1.60], 3 trials/504 women, respectively). Although no statistically significant differences were found between the intervention and control groups for preterm birth, in all risk categories the trend favoured the control group. For the outcomes caesarean delivery, preeclampsia, and infant birth weight >4000 g, results were not indicative of an effect. No other outcomes were reported on.

Diet and supervised exercise interventions versus routine care
Excessive gestational weight gain
Overall, diet and supervised exercise interventions in comparison to routine care reduced the risk of excessive gestational weight gain by 25% (RR 0.75, 95% CI [0.61 to 0.92], p=0.0059; 5 trials/689 women). The effect remained significant in subgroup analysis for mixed-risk women (RR 0.64, 95% CI [0.47 to 0.88], p=0.0053; 2 trials/235 women), but not for low-risk (RR 0.53, 95% CI [0.16 to 1.71], 2 trials/106 women) or high-risk women (RR 0.83, 95% CI [0.66 to 1.06], 2 trials/348 women).

Additional outcomes
Overall, mean gestational weight gain was not different between treatment groups, but was statistically significantly lower in the intervention group in low-risk women (MD -3.33 kg, 95% CI [-5.45 to -1.21], p=0.0021; 1 trial/57 women). No statistically significant difference between treatment groups was found for the outcomes low weight gain, caesarean delivery, preeclampsia, and infant birth weight >4000 g, and no other outcomes were reported on.

Diet counselling/other versus routine care
Excessive gestational weight gain
The risk of excessive gestational weight gain was not statistically significantly reduced with diet counselling/other interventions among mixed-risk or high-risk populations (RR 0.47, 95% CI [0.08 to 2.85], 3 trials/443 women, and RR 0.92, 95% CI [0.53 to 1.62], 1 trial/84 women, respectively). No data were available on low-risk women.

Additional outcomes
Mean gestational weight gain was statistically significantly reduced in four of the seven trials reporting on this outcome (data not pooled). The risk of low gestational weight gain was statistically significantly increased among low-risk women in the diet counselling/other group (RR 2.97, 95% CI [1.75 to 5.01], 1 trial/171 women); no data on other risk groups were available. For the outcomes caesarean delivery, preterm birth, preeclampsia, and infant birth weight >4000 g, results were not indicative of an effect, and no other outcomes were reported on.

5. Additional author observations*

The overall methodological quality of the included trials was fair, with more than half of all studies reporting adequate methods of allocation concealment. For the main comparison, all diet and/or exercise interventions versus routine care, the GRADE quality of evidence assessment was high for the outcomes excessive gestational weight gain, preeclampsia, caesarean delivery, and infant birth weight >4000 g, and moderate for the outcomes low weight gain and preterm birth. Few trials were conducted in low- and middle-income settings, limiting the generalizability of the findings to these settings. Only four trials included women with gestational diabetes, and thus it is unclear whether the findings are applicable to women with this condition.

The evidence reviewed here demonstrates a reduced risk of excessive gestational weight gain with interventions employing diet, exercise, or both in comparison to routine care. These interventions may also reduce the risk of maternal hypertension and neonatal respiratory distress syndrome, and increase the risk of low gestational weight gain. Caesarean delivery and macrosomia may also be reduced, particularly among women with a high risk of weight-related complications of pregnancy. No long-term data on childhood weight were available. Exercise appears to be an important factor in controlling weight gain in pregnancy, although it may increase the risk of preterm birth.

Further trials in resource-constrained settings are needed, as are implementation studies. Research into the effect of exercise interventions on preterm birth is also warranted.