Optimizing healthy gestational weight gain in women at high risk of gestational diabetes: A randomized controlled trial



This article is corrected by:

  1. Errata: Erratum: Optimizing healthy gestational weight gain in women at high risk of gestational diabetes: A randomized controlled trial Volume 24, Issue 1, 268, Article first published online: 22 December 2015

  • Clinical Trial Registration: Australian New Zealand Clinical Trial Registry Number: ACTRN12608000233325. Registered 7/5/2008. [www.anzctr.org.au].


    The authors have nothing to disclose. No financial disclosures were reported by the authors of this article.

  • Funding agencies: This project is supported by a BRIDGES grant from the International Diabetes Federation. BRIDGES, an International Diabetes Federation project, is supported by an educational grant from Lilly Diabetes (Project Number: LT07-121). The Jack Brockhoff Foundation also provided funding for this study. Helena Teede is an NHMRC research fellow.



Optimizing gestational weight gain (GWG) in early pregnancy is of clinical and public health importance, especially in higher risk pregnancies.

Design and Methods:

In a robustly designed, randomized controlled trial, 228 pregnant women at risk of developing gestational diabetes mellitus (GDM) were allocated to either control (written health information only) or intervention (four-session lifestyle program). All women received standard maternal care. Measures were completed at 12-15 and 26-28 weeks gestation. Measures included anthropometrics (weight and height), physical activity (pedometer and International Physical Activity Questionnaire), questionnaires (risk perception), and GDM screening.


The mean (SD) age [31.7 (4.5) and 32.4 (4.7) years] and body mass index [BMI; 30.3 (5.9) and 30.4 (5.6) kg/m2] were similar between control and intervention groups, respectively. By 28 weeks, GWG was significantly different between control and intervention groups [6.9 (3.3) vs. 6.0 (2.8) kg, P < 0.05]. When stratified according to baseline BMI, overweight women in the control group gained significantly more weight compared to overweight women in the intervention group [7.8 (3.4) vs. 6.0 (2.2) kg, P < 0.05], yet in obese women, GWG was similar in both groups. Physical activity levels declined by 28 weeks gestation overall (P < 0.01); however, the intervention group retained a 20% higher step count compared to controls [5,203 (3,368) vs. 4,140 (2,420) steps/day, P < 0.05]. Overall, GDM prevalence was 22%, with a trend toward less cases in the intervention group (P = 0.1).


Results indicate that a low-intensity lifestyle intervention, integrated with antenatal care, optimizes healthy GWG and attenuates physical activity decline in early pregnancy. Efficacy in limiting weight gain was greatest in overweight women and in high-risk ethnically diverse women.


Weight gain in younger women of reproductive age is increasing, contributing to a rise in the proportion of women entering pregnancy overweight or obese (1). Overweight and obesity in pregnancy are an established risk factor for maternal complications including, but not limited to, miscarriage, hypertension, gestational diabetes mellitus (GDM), and cesarean delivery (2). Gestational weight gain (GWG) exceeding international Institute of Medicine (IOM) recommendations exacerbates health risks further and is independently linked to maternal obesity progression long term (3), yet is common with ∼60% of overweight and obese women excessively gaining weight by term (4). Therefore, with increasing weight gain in women and associated health risks, prevention of excess GWG is a key public health priority.

Antenatal care provides an ideal opportunity for behavioral intervention, presenting a “teachable moment” (5) with a captive audience more likely motivated toward a healthy lifestyle to optimize pregnancy outcomes (6, 7). However, existing studies have been low quality (8) and there is only one previous study targeting women at high risk of complications including GDM (9). Furthermore, many previous studies fail to recruit women in early pregnancy, missing the opportunity to target the second trimester where maternal fat accretion is highest (10, 11). Maternal fat accretion is a key driver of maternal pregnancy complications, including GDM (12) and postpartum weight retention. Given the dramatic increase in GDM prevalence and the recent literature suggesting lifestyle intervention in pregnancy may prevent GDM, further research is warranted targeting women at high GDM risk (8).

We have previously demonstrated the efficacy of a simple low-intensity behavior change lifestyle intervention (HeLP-her) to prevent weight gain in nonpregnant women over a 1-year period (13). In the current study rather than in this study, we aimed to build on this health promotion work in a different setting (antenatal care) and population (pregnant women). We recruited overweight and obese women in early pregnancy (12-15 weeks gestation) at high risk of developing GDM, identified using our previously developed and validated screening tool (14). In this group, we aimed to optimize GWG and improve adherence to IOM recommendations. Given the contribution of maternal fat accretion to pregnancy complications, we focused on prevention of maternal fat accretion, which occurs during the second trimester and peaks at 28-30 weeks gestation. Hence, weight gain was measured at 28 weeks gestation following the active intervention phase and at 6 weeks postpartum following the maintenance phase. Here we report 28 week results following the intervention phase of this randomized controlled study.

Methods and Procedures

Participants and setting

Recruitment took place at three large metropolitan tertiary teaching hospitals in Victoria, Australia, combining over 8,900 births per year (15). Eligibility criteria included women 12-15 weeks gestation, overweight (body mass index; BMI ≥ 25 or ≥23 kg/m2 if high-risk ethnicity [Polynesian, Asian, and African populations] (16)) or obese (BMI ≥ 30 kg/m2), and at increased risk for developing GDM identified by a validated risk prediction tool (14). The validated risk tool, developed by the authors, was based on first trimester data in women attending the same large tertiary hospital as this intervention study. With a derivation group (n = 2,880 pregnancies) and a validation group (n = 1,396 pregnancies), multivariate analysis was used to identify predictive factors for GDM and to generate a simple scoring tool (14). Additional criteria included participant agreement to complete an oral glucose tolerance test at 28 weeks gestation rather than a standard glucose challenge test at GDM screening. Exclusion criteria included multiple pregnancies, diagnosed type 1 or 2 diabetes, a BMI ≥ 45 kg/m2, pre-existing chronic medical conditions, and non-English-speaking women.


Participating women were randomly assigned to intervention or control through computer-generated randomized sequencing. Allocation concealment was achieved by using sealed opaque envelopes. Care providers, investigators, and outcome data analyzers were blinded to group allocation. The primary outcome was GWG with secondary outcomes including GDM diagnosis, physical activity, and risk perception. Women completed measures at baseline, between 12 and 16 weeks gestation, and at ∼28 weeks gestation in line with GDM screening. All women received standard maternal care with the study integrated with routine maternity visits. The trial was designed to meet criteria for a high-quality study with a low risk of bias. CONSORT criteria are available in Figure 1.

Figure 1.

CONSORT criteria for included participants.

Control group

Women allocated to the control group received a brief, single education session based on the widely available generic Australian Dietary and Physical Activity Guidelines. Written pamphlet versions of these guidelines were also provided (17, 18). GWG was not discussed and participants received no further study support.

Intervention group

Women allocated to the intervention received an individual four-session behavior change lifestyle intervention based on the Social Cognitive Theory, adapted from our previously successful lifestyle intervention program (HeLP-her) (13). Sessions were provided in the antenatal clinic setting and scheduled around routine antenatal visits at 14-16, 20, 24, and 28 weeks gestation. They were provided by a health coach (CH; exercise physiologist); however, the intervention was designed to be delivered by generic health care providers. The aim was to support and empower pregnant women to optimize their lifestyle and GWG in an interactive, individualized environment. The sessions provided pregnancy-specific dietary advice in addition to simple healthy eating and physical activity messages. In addition, simple behavioral change strategies were progressively practised to identify short-term goals, and increase self-efficacy and self-monitoring. Goals were determined individually by the participants, informed by the lifestyle messages, and included goals such as reducing high fat or convenience foods, increase fruit and vegetable intake, and increase frequency of physical activity. Self-monitoring strategies included pedometers and the use of weight gain charts based on IOM recommendations for weight gain throughout pregnancy (11). Intervention participants received the same written information as controls in addition to resources promoting optimal health, GWG, and lifestyle.

The Southern Health Research Advisory and Ethics Committee approved the study and all participants gave written consent.



Anthropometric assessment included weight on an electronic scale to the nearest 0.1 kg (Tanita model BWB-800 Digital Scale, Wedderburn Scales, Melbourne, Australia) and height measured by a registered nurse unaware of participant allocation.


The Yamax Digiwalker SW-700 Pedometer (Yamax Corporation, Tokyo, Japan) was used to assess the number of free-living steps per day as a tool with demonstrated accuracy in pregnancy, as previously described (19, 20). Pedometers were sealed to blind participants to their step count (21). Readings were processed to provide daily step count.

International Physical Activity Questionnaire long version

The International Physical Activity Questionnaire (IPAQ) was used as a validated subjective tool to assess physical activity across four domains including leisure time, domestic, work, and transport-related physical activity (22). Weekly and daily average metabolic equivalent (MET) scores were calculated for total physical activity, as well as for walking, moderate and vigorous activity performed per week as continuous variables (23).

Psychosocial measures

Risk perception for development of GDM and excess GWG was assessed using a four-point Likert scale adapted from the theory of health Stage of Change (24). Data from the Likert scale were processed to reflect two categories of either no perceived risk or perceived risk.

GDM diagnosis

GDM was diagnosed based on the Australasian Diabetes in Pregnancy Society (ADIPS) criteria including the presence of either a fasting venous plasma glucose level of ≥99 mg/dl (≥5.5 mmol/l) and/or a 2-h level of ≥144 mg/dl (≥8.0 mmol/l) (25). Recently released International Association of Diabetes and Pregnancy Study Groups (IADPSG) diagnostic criteria were also evaluated. IADPSG criteria include a fasting venous plasma glucose of ≥91.8 mg/dl (≥5.1mmol/l), a 1-h plasma glucose of ≥180 mg/dl (≥10.0 mmol/l), or a 2-h glucose level of ≥153 mg/dl (≥8.5 mmol/l) (26).

Statistical analysis

All data are presented as mean (SD) with 95% confidence intervals (CIs) unless otherwise stated. Frequency data are presented as % (n). Two-tailed statistical analysis was performed using SPSS for Windows 17.0 software (SPSS, Chicago, IL) with statistical significance set at α level of P < 0.05. At baseline, data were assessed using independent samples T-tests and chi-square tests for continuous and categorical variables, respectively (control vs. intervention). Analysis of covariance was used to assess outcomes as well as interaction effects between group and the categorical variables country of birth (Australian born vs. non-Australian born immigrant) or baseline BMI (overweight vs. obese) on outcomes of weight gain and physical activity. Logistic regression was used to assess categorical outcomes including risk perception (no perceived risk vs. perceived increased risk). Regression analysis included group (intervention vs. control) and the categorical variables country of birth or baseline BMI in addition to the interaction between them as covariates. Weekly weight gain was estimated by dividing weight gained over the intervention period by 14 weeks. Rate of weight gain was compared against IOM, BMI-specific recommendations for the second and third trimester (11). Age and parity were used as a covariate in all analyses. Change in variable was defined as the percentage change over time (baseline to 28 weeks gestation).

The sample size for this intervention was calculated based on a difference in BMI of 0.8 kg/m2 between groups at 6 weeks postpartum (equivalent to 2-3 kg postpartum weight retention in women of average BMI in Australia based on results from the Australian Longitudinal Women's Health Study (27)). With a 10% dropout rate (informed by previous research (13)), 222 women in total were required. The sample size calculation was based on a power of 0.80, a significance level of 5% (two-sided), and a standard deviation of 2.0 kg/m2, in line with previous studies (28).



A total of 1,331 women were invited to participate through an invitation flyer and follow-up phone call at hospital booking. Of the 1,331 women, 329 expressed interest (25% response rate) and 228 were recruited and randomized with 121 women allocated to intervention and 107 to control (Figure 1). The mean age of women allocated to control and intervention at baseline was 31.7 (4.5) and 32.4 (4.7) years, respectively, with no significant differences between groups. Overall, mean gestation at baseline was 14 (0.8) weeks. Demographic characteristics are presented in Table 1.

Table 1. Demographic Characteristics of Participants (n = 228)
VariableControl (n = 7)Intervention (n = 121)P
Age (years)31.7 (4.5)32.4 (4.6)0.31
Weight (kg)77.9 (17.6)78.4 (17.8)0.83
BMI (kg/m2)30.3 (5.9)30.4 (5.6)0.83
Body mass index (kg/m2) % (n)
 ≤29.9960 (64)57 (69)0.67
 ≥30.0040 (43)43 (52) 
Country of birth % (n)
 Australia38 (41)36 (44)1.00
 Southeast Asia12 (13)14 (16) 
 Southern/Central Asia36 (38)36 (43) 
 Other14 (15)14 (18) 
Education % (n)
 Year 11 or below11 (12)16 (20)0.32
 Year 127 (7)11 (13) 
 Certificate/diploma28 (30)21 (25) 
 University degree or higher46 (49)47 (57) 
 Declined/no answer8 (9)5 (6) 
Work % (n)
 Full time26 (28)23 (28)0.69
 Part time28 (30)28 (34) 
 No paid work37 (39)43 (52) 
 Declined/no answer9 (10)6 (7) 
Income ($) % (n)
 <40,00030 (32)26 (32) 
 40–60,00015 (16)20 (24) 
 60–80,00012 (13)15 (18) 
 >80,00019 (20)18 (22) 
 Unsure/declined24 (26)21 (25) 
Childbearing history % (n)
 First pregnancy43 (46)42 (51)0.51
 Second pregnancy37 (40)36 (43) 
 Third pregnancy or higher20 (21)22 (27) 
Smoker % (n)

Weight gain and IOM recommendations

At baseline, weight and BMI were similar with no significant differences between groups (Table 1). At 26-28 weeks, weight gain within groups significantly increased in line with advancing gestation (P < 0.001); however, the intervention group gained 0.9 kg or 14% less than the control group overall [6.0 (2.8) vs. 6.9 (3.3) kg (95% CI: −1.7 to −0.1), P < 0.05], which persisted after adjustment for baseline BMI. Women in the intervention group had a significantly lower estimated rate of weekly GWG compared to control women [0.43 (0.22) vs. 0.51 (0.22) kg/week, respectively, P < 0.05].

There was a significant interaction between group and baseline BMI, with overweight women (<30 kg/m2) allocated to the control group gaining significantly more (26%) than women allocated to the intervention group [7.8 (3.4) vs. 6.0 (2.2) kg (95% CI: −2.8 to −0.6), P < 0.05, Figure 2A]. Overweight women in the intervention group had a significantly lower estimated weekly rate of GWG compared to overweight control women [0.43 (0.18) vs. 0.59 (0.28) kg/week, P < 0.01], with GWG in the intervention group more comparable to IOM trimester specific recommendations (<0.32 kg/week). Obese women (>30 kg/m2) gained less weight with no significant difference between groups [5.2 (2.6) vs. 5.9 (3.5) kg (95% CI: −0.7 to 2.1), P = 0.32, Figure 2A].

Figure 2.

(A) Mean weight gain between intervention and control groups, according to baseline body mass index (BMI). (B) Mean weight gain between intervention and control groups, according to country of birth.

The interaction between country of birth (Australian born vs. immigrants) and group allocation was also significant. Immigrant women in the intervention group gained significantly less than those in the control group [5.8 (2.7) vs. 7.6 (3.5) kg (95% CI: −2.8 to −0.6), P < 0.05, Figure 2B] by 28-week gestation. Women born in Australia had a similar GWG with no significant between group differences [5.7 (2.8) vs. 6.0 (2.8) kg (95% CI: −1.0 to 1.6), P = 0.60, Figure 2B].Additionally, estimated rate of GWG was significantly reduced in the intervention group compared to controls [0.43 (0.23) vs. 0.57 (0.29) kg/week, P < 0.01].

Gestational diabetes outcome

Across the whole group, prevalence of GDM using the ADIPS criteria was 21.9% (n = 50) and using the revised IADPSG (26) was 27.2% (n = 62). The study was not powered for GDM outcomes; however, when applying the IADPSG criteria the control group had eight extra cases of GDM diagnosis (35 vs. 27 cases, P = 0.1).

Risk perception

As a whole group, 65.6% of overweight women at baseline did not believe they were at any risk of developing GDM during pregnancy compared with 34.8% of obese women (P < 0.01). Australian born women were more likely to perceive themselves at increased risk of GDM at baseline (63.0%) when compared with 37.7% of immigrant women (P < 0.01).

At baseline obese women were more likely to report perceived increased risk of gaining excess weight in comparison to overweight women (83.1 vs. 67.8%, respectively, P < 0.05). Similar results persisted for subgroup analysis for country of birth, with Australian born women more likely to perceive increased risk of excess GWG compared to immigrant women (85.2 vs. 67.4%, P < 0.05).

Physical activity

Participants wore the blinded pedometer for an average of 4.5 (1.8) and 4.8 (1.8) days at baseline and 28 weeks, respectively. At baseline, there was no significant difference between intervention and control groups for mean daily pedometer steps [6,006 (3,034) vs. 5,574 (3,163) steps/day, P = 0.34] or for total MET min−1 estimated by the IPAQ [3,617 (5,941) vs. 3,810 (7,203) MET min−1/day, P = 0.82]. This persisted after stratification into light, moderate, and vigorous MET min−1. By 28 weeks gestation, average daily steps declined across the whole group (P < 0.001). On univariate analysis the intervention group had a 20% higher daily step in comparison to the control group [5,203 (3,368) vs. 4,140 (2,420) steps/day (95% CI: 91–2,035) P < 0.05]. There was no significant change over time in MET min−1 estimated by the IPAQ between or within groups.

When stratified according to baseline BMI and country of birth, both obese women (≥30 kg/m2) and immigrant women in the intervention group maintained greater activity with a 50% higher step count compared to control women within the same category, although neither reached significance (P = 0.1 and P = 0.4, respectively). Women in the lower BMI category (<30 kg/m2) and Australian born women showed a similar, nonsignificant reduction in daily step count between intervention and control groups.


We recruited overweight and obese women at risk of developing GDM and delivered a low–moderate intensity antenatal lifestyle intervention in early to mid-pregnancy. Here, we report that GWG was significantly optimized by 28 weeks gestation. Of importance is the efficacy of the intervention in limiting weight gain in overweight women and in those immigrant, ethnically diverse women, a high risk group for GDM. We also report a trend toward reduction in the incidence of GDM in the intervention group.

A recent systematic review highlighted that previous lifestyle interventions in pregnancy were of low to medium quality, most had not targeted early to mid-pregnancy when maternal fat accretion is greatest, and only one had focused on women at high risk of GDM (8). In this study, we targeted women in early to mid-pregnancy as maternal fat accretion is the key component of second trimester GWG, drives maternal complications including GDM, and peaks at 28-30 weeks gestation (11, 29). Given the high risk of GDM in our population (diagnosed at 28 weeks) and prior literature suggesting that lifestyle intervention in pregnancy reduces the risk of GDM, second trimester intervention was vital in this population (8). Although we focused our lifestyle intervention sessions in the second trimester, women were encouraged to continue monitoring weight gain and improving lifestyle behaviors throughout the third trimester and into the postpartum period. Finally, in this study, early intervention with end point data collection by 28 weeks was important to maintain internal study validity, as more than one in five women developed GDM and received additional intensive lifestyle intervention from 28 weeks onward. Here, we have demonstrated in a high-quality study that lifestyle intervention in early to mid-pregnancy, in high risk women, limits excess GWG by 28 weeks and may reduce GDM.

With a linear relationship between GWG and adverse pregnancy outcomes (30), the current prevention of ∼1 kg GWG is of public health significance. Indeed, for each extra kilogram gained during pregnancy above IOM recommendations, there is an estimated 10% increase in likelihood of adverse pregnancy-related outcomes (30). Risks also extend long term with excess GWG in early pregnancy shown to independently predict long-term obesity and adverse health outcomes including central adiposity and increased blood pressure (31). Therefore, from a public health perspective, prevention of excess GWG is likely to have both short- and long-term health implications. A recent retrospective cohort study of 8,000 women supports prevention of excess GWG in early pregnancy, reporting that a weekly rate of GWG above IOM recommendations in the second trimester predicts excessive weight gain at term (32). Results here demonstrate women receiving the intervention had an estimated mean rate of weight gain more comparable to IOM recommendations compared to control women. Recent longitudinal data indicate women who adhere to IOM recommendations are less likely to develop obesity in the years following pregnancy in comparison to women with a high GWG exceeding IOM recommendations who have a threefold increased risk of obesity development and central adiposity (33).

Interestingly, the intervention was most effective in overweight women (<30 kg/m2) representing the majority of women at high GDM risk. This group was most likely receptive to modification of lifestyle-related behaviors and intervention impact in the setting of higher background GWG (as observed in the control group) and lower baseline risk perception, both of which were amenable to intervention. Reciprocally, the intervention was not effective in women with established obesity. Potential contributing factors include a lower rate of weight gain seen here in the obese control group and in previous studies (34), and a higher baseline awareness of GDM risk. The lower background GWG in obese pregnant women limits the power of the current and of any future intervention studies to demonstrate effective weight management in obese women during pregnancy. IOM total GWG recommendations for obese women are relatively low at between 5 and 9 kg and increasing adherence is likely to involve intensive health professional contact. Supporting the dose-response effect in eliciting weight management in obese populations (35), previous intensive studies report better limitation of GWG in obese women than results seen here (36, 37).

The study population was ethnically diverse with ∼60% immigrant women (80% of these from Southern Asia/Southeast Asia) enabling evaluation of the intervention according to ethnicity. Immigrant women in the intervention gained ∼2 kg less by 28 weeks gestation compared to those in the control group (P < 0.05). With increased morbidities in pregnancy demonstrated at a lower BMI (1), targeting women in early pregnancy from high-risk ethnic backgrounds is important. However, implementation may be challenging with barriers including language, cultural beliefs and practices, and education/health literacy level. Here, encouragingly we demonstrate individualized goals and action plans supported by simple lifestyle messages are effective in those with low health literacy and diverse cultural beliefs with potential implications for broader lifestyle change and prevention settings. Potentially, the intervention may have been less effective in Australian born women because of higher baseline health literacy and healthy lifestyle knowledge.

Self-monitoring of GWG against IOM recommendations was integral to the current intervention, supporting the literature on the efficacy of self-monitoring as a broadly effective strategy in weight management (38). Regular weight monitoring at a clinical level may be effective; however, prior studies have demonstrated that weighing alone in pregnancy does not impact on GWG (34) and there is currently poor antenatal uptake of GWG monitoring in Australia. This study and previous literature in the area suggest that combined lifestyle intervention and weight monitoring is needed during pregnancy to effectively limit excess GWG. Perceptions around healthy weight gain in pregnancy also appear to be distorted in women and health professionals (39), and education on IOM recommendations would likely assist adherence with GWG guidelines. We also report that despite reduced physical activity overall, encouragingly the intervention attenuated the decline in physical activity by 20%, although the authors acknowledge greater engagement in activity would be preferable.

The strengths of this study include the study size, high-quality study design, implementation of a lifestyle intervention in early to mid-pregnancy, use of robust measures, including a validated screening tool to identify women at increased risk of GDM (1), and use of objective measures for assessing physical activity. The latter is particularly important given previous studies rely heavily on self-recalled measures. Additional strengths include novel recruitment of a high-risk group and a high retention rate. Limitations include inadequate power to show an impact on GDM outcome and a lack of weight data prepregnancy and at delivery. Also in recruiting a specific population of high-risk pregnant women, the results found may not be generalized to all pregnant women.

Effective low-intensity interventions aimed at optimizing GWG and adherence to international guidelines are urgently needed. Here, in a high-risk group, we present a robustly designed study demonstrating efficacy of a low- to moderate-intensity lifestyle intervention in limiting excess GWG during the second trimester, where maternal fat accretion is greatest. We also note a trend toward a reduction in GDM in the intervention group. Greatest efficacy was noted in women with a BMI < 30 kg/m2 and in ethnically diverse women. Overall, these results support the use of an effective early pregnancy lifestyle intervention with outcomes of public health and clinical relevance. Moving forward, we propose a risk stratification approach in pregnancy with high-intensity lifestyle intervention for women with established obesity, low- to moderate-intensity intervention for women at high risk of complications such as GDM, and the integration of simple low-intensity behavior change strategies into routine antenatal care for all pregnant women to optimize clinical outcomes with scarce resources.


The authors acknowledge Sanjeeva Ranasinha and Eldho Paul for statistical assistance, Amanda Hulley, Lauren Snell, and Melanie Gibson-Helm for recruitment and data collation, Deborah Thompson and Nicole Ng for data entry and Carolyn Allan for initial input into study design.