Current Evidence on Exercise for Gestational Diabetes Mellitus in Overweight and Obese Pregnant Women: A Systematic Review

The main purpose of this systematic review was to summarize the recent randomized controlled trials (RCTs) regarding the effects of exercise on gestational diabetes mellitus (GDM) in overweight and obese pregnant women. A literature search was performed on Web of Science, ScienceDirect, and Google Scholar. The keywords were “exercise”, “physical activity”, “overweight”, “obese”, “gestational diabetes mellitus” and “gestational diabetes”. Only RCTs related to this field that have been published in peer-reviewed journal were covered. Twelve related studies were included in this review, 5 trials employed comprehensive lifestyle programs, 4 trials employed comprehensive exercise training programs, and the remaining 3 trials employed single exercise training programs, respectively. Although there are some conflict results in the effects of exercise on GDM in overweight or obese pregnant women, it seems that single exercise training programs (i.e. moderate-intensity cycling intervention) may contribute to more benefit on reducing the incidence of GDM and improving glucose metabolism. However, further rigorously well-designed RCTs are much warranted for examining the appropriate setting of this intervention and for further verification.


Introduction
Gestational diabetes mellitus (GDM), similar to type-2 diabetes (T2D), is one of the most common complications of pregnancy, especially for these overweight or obese women [1][2][3]. It is featured by the disorder of glucose metabolism and carbohydrate risk of developing T2D and its complications in the future [10].
Therefore, it is of great clinical importance to explore some easy interventions that can be used during pregnancy to alleviate and even cure GDM.
In recent years, a number of studies have indicated that exercise has several benefits on women with GDM [11][12][13][14][15]. For example, Garnaes et al. [14] found that both the incidence of gestational diabetes mellitus and systolic blood pressure were significantly lower after exercise intervention than the control.
Several systematic review and meta-analysis on this topic add more evidence. Ming et al. [3] demonstrated that exercise can decrease the incidence of GDM without any adverse results on normalweight pregnant women.
Several factors are highly associated with the occurrence and development of GDM, with overweight or obese being one of the most common one [16,17]. However, there is still less evidence based on the systematic evaluation of the effects of exercise on GDM in overweight and obese pregnant women up to now. In addition, emerging studies have revealed that exercise may not benefit overweight pregnant women with GDM [18,19]. Therefore, the purpose of this study was to review and summarize the recent randomized controlled trials (RCTs) regarding the effects of exercise on gestational diabetes mellitus (GDM) in overweight and obese pregnant women.

Eligibility Criteria
In order to be included in this review, previous related studies must meet the following several eligibility criteria: 1) study types, only randomized controlled trials (RCTs) that have been published in peer-reviewed English journal and with the aim to investigate the effects of exercise on gestational diabetes mellitus in overweight and/or obese pregnant women were considered for inclusion, other types of research, such as cross-sectional studies, case reports, and thesis etc., were not covered; 2) participants, pregnant women who have been diagnosed as overweight and/or obese based on body mass index (BMI) were included, and there is no restriction on ethnicity and age; 3) intervention, participants in the intervention group should perform exercise or take part in some comprehensive exercise programs, while this is no restriction on control group; 4) outcomes, studies evaluated the effects of exercise on parameters related to gestational diabetes mellitus were included.

Searching Strategy
In order to examine the current evidence regarding this topic, a thorough literature search was applied to identify all related studies from 2010 until 13 October 2020 from the following three English electronic databases: Google Scholar, Web of Science, ScienceDirect.
The searching strategy was modified according to the requirement of each database. In addition, in order to ensure that the searching process is thorough and rigorous, all the retrieved records were independently assessed by two researchers and discussed with the third researcher if any disagreements happened. The citation snowballing method was also applied to check the reference lists of all included studies and locate any potential related studies [20,21].

Data Management
Two researchers independently extracted and summarized the following study characteristic from each trial, authors, country (where the trial was carried out), published year, participant characteristics (sample size, age, and BMI), interventions (exercise types, frequency, and duration), outcome parameters, and primary results. Any differences that happened during data management would be discussed and resolved with the third researchers. In addition, Mendeley Reference Management Software (Elsevier Ltd., Amsterdam, Netherlands) was used for further study management.

Risk of Bias Assessment
The Cochrane Risk of Bias Assessment Tool was used to evaluate the quality of all the included studies [22]. To be more specified, seven sub-domains including random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other biases, was assessed, and each sub-domain was graded as three levels, low risk of bias, unclear risk of bias, and high risk of bias.

Searching Results
The searching and study selecting process is showed in Figure   1. A total of 21,602 records were retrieved from three literature databases, and 452 studies remained after all the irrelevant or duplicate items were removed based on titles. According to the eligibility criteria, 437 papers were then further excluded from this review because of the following several reasons, 1) not randomized controlled trials; 2) studies not related to overweight and/or obese pregnant women; 3) study not involved in exercise intervention; 4) study protocols without any outcomes. After that, 5 duplicates between databases were excluded while 2 additional articles were included after reference searching of the included studies and retrieved reviews. Finally, 12 related studies met the criteria were included in this review.

Study Quality
The study quality was evaluated in terms of the risk of bias, and the detailed information was presented in Figure 2 & Table 1. More

American Journal of Biomedical Science & Research
Copy@ Rongrong Xuan than half of the included studies described the random sequence generation process in detail (n=9, 75.00%) and concealed the allocation from both participants and researchers (n=8, 66.67%).
However, there is still some studies that did not presented enough detailed information regarding these two domains. Regarding the following two domains, all studies presented high risk of bias because it is usually not possible to blind both participants and researchers when performing interventions (n=12, 100.00%).
Nevertheless, all the studies reported low risk of bias in incomplete outcome data and selective reporting (n=12, 100.00%). It is unclear if there is any other bias that existed in these studies (n=12, 100.00%).

Basic Study characteristics
The basic study characteristics were presented in Table 2 The primary outcomes of these included studies were presented in Table 3. Summary of these resulted were displayed in 3 parts, including 1) effects of comprehensive lifestyle program on GDM; 2) effects of comprehensive exercise training program on GDM; 3) effects of single exercise training program on GDM.

Effects of Comprehensive Lifestyle Program on GDM
Most studies (n=5) in this review investigated the effects of comprehensive lifestyle program on GDM.
Two studies related to this topic were conducted in 2015.
Hawkins and his colleagues [19] compared the effects of pregnancy lifestyle intervention, which is consisted of personalized diet and moderate-intensity physical activity, with standard prenatal care on pregnant women. Subjects were asked to perform this intervention at least 30 minutes per day for 6 months in total.
Gestational weight gain, infant birth weight, and parameters related to insulin resistance were measured before and after intervention, and they found that only the ability for vigorous physical activity was significantly improved when compared to the control group.
Although there are some differences in study design, Simmons et al. [28] reported some positive results. Subjects were divided into three groups, healthy eating group, physical activity group, and healthy eating + physical activity group, and were required to perform corresponding intervention from 20 weeks to 35 weeks of gestation. They found that the fasting glucose and gestational weight gain were significantly reduced after healthy eating intervention, but not for the other two groups. However, using the same study design, Simmons et al. [23] presented some different outcomes. They demonstrated that although the gestational weight was significantly reduced after healthy eating + physical activity intervention, none of these three interventions can significantly reduce the incidence of gestational diabetes mellitus.
Finally, two recent studies in this category also reported some controversial results. Bruno et al. [12] compared the effects of pregnancy lifestyle intervention (30 minutes per time for 3 times per week from enrollment to 36-week gestation) with standard prenatal care, and they found that this comprehensive lifestyle program significantly reduced the incidence of GDM. Vinter et al.
[25] conducted a similar trial on pregnant women. Subjects were asked to perform comprehensive intervention 30 to 60 minutes per day for 6 months in total, and they reported that lifestyle intervention was not effective in improving metabolic and obstetric parameters. weeks after delivery, and the results shown that no significant effect was observed after exercise training programme when compared to control group. The subsequent two studies also put some conflict results. In Garnaes et al. [14] study, subjects were randomly assigned to a 60-min moderate-intensity endurance and strength training intervention at 3 times per week or standard prenatal care from 12-18 weeks of gestation to delivery, and they found that although the gestational weight gain was not significantly reduced, both the incidence of gestational diabetes mellitus and systolic blood pressure seemed to be lower after moderate-intensity endurance and strength training intervention than control group. However, using the similar study design, Daly et al. [18] demonstrated that the comprehensive exercise training program can significantly reduce the gestational weight gain but have no effect on maternal glycemia.

Effects of Single Exercise Training Program on GDM
There are also 3 studies focused on investigating the effects of single exercise training program on GDM.  [14]. However, there is a relatively different trend for the single exercise training program, with 2 positive results [11,13] and 1 neutral result [24]. To sum up, from the recent evidence, it seems that moderate-intensity cycling intervention may have

Conclusions
In conclusion, although there are some conflict results in the effects of exercise on GDM in overweight or obese pregnant women, it seems that moderate intensity cycling intervention may contribute to more benefit on reducing the incidence of GDM and improving glucose metabolism. However, more rigorously welldesigned RCTs are warranted to examine the appropriate setting and for further verification.