Exercise for Counteracting Weight Recurrence After Bariatric Surgery: A Systematic Review and Meta-Analysis of Randomized Controlled Trials

Abstract

Background—

Exercise is recommended to prevent post-surgical weight recurrence. Yet, whether exercise interventions are efficacious in this regard has not been systematically evaluated. Moreover, clinicians lack evidence-based information to advise patients on appropriate exercise Frequency, Intensity, Time and Type (FITT) for preventing weight recurrence.

Objectives—

Conduct a meta-analysis of randomized controlled trials (RCTs) involving exercise interventions specifying FITT and weight measurement ≥12-months post-surgery.

Setting—

Health-care system, USA

Methods—

We reviewed scientific databases up through February 2022 for RCTs comparing exercise interventions reporting FITT and a non-exercise control group on weight≥12-months post-surgery. Procedures following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) were registered at the International Prospective Register of Systematic Reviews (PROSPERO: CRD42022342337).

Results—

Of 1,368 studies reviewed, 5 met inclusion criteria (n=189; 47.8±4.2y, 36.1±3.8kg∙m⁻², 83.2±9.5% female; 61.7% underwent Roux-en-Y gastric bypass). Exercise interventions were largely supervised, lasted 12–26 weeks, and prescribed 80–210 minutes/week of moderate-to-vigorous intensity combined aerobic and resistance exercise over ≤5 days. Within-group effects showed non-statistically significant weight loss for exercise (d=−0.15, 95%CI: −1.96, 1.65; −1.4 kg, p=.87) and weight gain for control (d=0.11, 95%CI: −1.70,1.92; +1.0 kg, p=.90), with no difference between these groups (d=−0.26, 95%CI: −2.07, 1.55; −2.4 kg, p=.78).

Conclusions—

Exercise elicited an additional 2.4 kg weight loss versus control, although this effect was small and statistically non-significant. Ability to draw definitive conclusions regarding efficacy of exercise interventions for counteracting post-surgical weight recurrence was limited by the small number of trials and methodological issues. Findings highlight the need for more rigorous RCTs of exercise interventions specifically designed to reduce post-surgical weight recurrence.

INTRODUCTION

Weight recurrence is common after bariatric surgery. A large study of 1406 Roux-en-Y gastric bypass surgery patients found 91% experienced some amount of weight recurrence within 2 years after reaching their nadir weight which occurred at a median of 2 years post-surgery. Of these patients, more than half (57%) experienced clinically significant weight recurrence (≥20% of maximum weight loss).¹ Thus, maintaining weight loss after bariatric surgery can be challenging for many patients and adjunctive interventions to help patients change key lifestyle behaviors that counteract weight recurrence are needed.

Multiple national (e.g., American Society of Metabolic and Bariatric Surgery [ASMBS]) and international (e.g., The European Association for the Study of Obesity [EASO]) professional societies emphasize that is important for patients who have had bariatric surgery to increase their exercise and physical activity behaviors to attenuate post-surgery weight recurrence.^2,3 Observational research also supports the role of physical activity, specifically performed at moderate-to-vigorous intensities (i.e., MVPA), for enhancing long-term weight loss and limiting weight recurrence following bariatric surgery.^4–10 Importantly, this support is increasingly derived from research involving objective daily monitoring of patients’ MVPA which reduces bias inherent to self-reported MVPA.^{4,7,8,10–12} While a recent observational study found that total ambulatory physical activity, but not MVPA, was related to less weight recurrence, this finding was likely due to most patients engaging in low levels of MVPA.¹³ Indeed, while higher MVPA levels are associated with better long-term weight loss and less weight recurrence in most studies, patients who have had bariatric surgery have low levels of MVPA before surgery and make only modest changes in MVPA after surgery.^11,13–17 This is especially true for MVPA that is accumulated in sustained bouts >10 minutes which is used as a measurement proxy for exercise.^11,14–17 Thus, it appears many bariatric patients need additional intervention and support to increase exercise behaviors.

Unfortunately, to date, there have been few exercise-focused interventions specifically designed to counteract post-surgical weight recurrence. As a result, clinicians have minimal evidence-based guidance, let alone professional exercise prescription that specific Frequency, Intensity, Time and Type (FITT), to advise their patients on exercising to limit weight recurrence. To partially address these scientific and clinical gaps, we executed a systematic review of randomized controlled trials (RCTs) that compared exercise interventions to usual care conditions that provided minimal exercise and/or dietary advice and measured weight ≥12 months post-surgery to capture progressive weight recurrence. Additionally, we selected only articles with exercise interventions that specified the amount of exercise performed framed by the FITT principle of exercise prescription.^18,19 We aimed to: (1) determine the effect of exercise interventions on weight change ≥12 months following bariatric surgery, and (2) assemble and summarize information about the amount of exercise applied by the FITT to better inform the future development of exercise recommendations for reducing weight recurrence after bariatric surgery.

METHODS

The current study was conducted and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA).²⁰ The study protocol is registered at PROSPERO (ID: CRD42022342337).

Literature Search and Study Selection

With the assistance of a medical librarian, we searched for RCTs of exercise interventions in five databases (PubMed, Cochrane, Scopus, SPORTDiscus, and CINAHL) from inception until February 2022. See Online Supplemental Material A for the search terms.

Articles were included if they: 1) were peer-reviewed and published in English language journals; 2) employed an RCT design; 3) included both an exercise intervention that specified the FITT and a control group that included at least usual postoperative care; 4) measured weight change beyond 12 months post-surgery; and 5) involved male or female adults (≥18 years old) of any race/ethnicity, who had undergone Roux-en-Y gastric bypass, sleeve gastrectomy, gastric banding, or biliopancreatic diversion with duodenal switch surgery. Of note, studies were excluded if they only relied on patient self-reports of exercise participation. However, studies were included if they relied on an objective measure of physical activity when exercise intervention was not supervised. Identified reports were screened by title, abstract, and full text review based on the inclusion and exclusion criteria. Reference lists of all included studies, relevant reviews, and meta-analyses were manually searched and cross-referenced for additional reports. All study selection was completed by two trained investigators (K.M., Y.W., and/or M.L.) independently. All disagreements were resolved through discussion.

Data Extraction, Risk of Bias and Publication Bias

The full text of each article identified for inclusion was read and pertinent data (i.e., participants characteristics, study characteristics, and features of the exercise intervention [i.e., FITT elements and whether participants were supervised], primary outcome of weight change [mean and standard deviation of weight at pre- and post-intervention and control) were extracted using a standardized data extraction/coding form.^21,22 We evaluated risk of bias using the Cochrane Risk of Bias 2.0 tool for individually randomized, parallel group trials and assigned an overall judgement of low, high, or some concerns.²³ For each study, data extraction and evaluation of risk of bias were completed by two trained investigators (K.M., Y.W., and/or M.L.) independently. All disagreements were resolved through discussion. Last, we evaluated the potential for publication and other biases by examining the distribution and asymmetry of funnel plots and performing tests introduced by Begg and Mazumdar and Egger et al.^24–26

Effect Size Calculation

Standardized mean difference effect sizes (d) were calculated following random-effects models for: 1) the within group effects as the mean difference in weight between pre and post intervention/control divided by the pooled standard deviation correcting for small sample size; and 2) the between group effects as the mean difference in weight between the intervention and control groups after vs before intervention divided by the pooled standard deviation, correcting for small sample size bias and baseline differences between groups.^27,28 When the effects reached statistical significance, we interpreted ds of <.20 as insufficient, .20–.49 as small, 0.50–0.79 as moderate, and ≥.80 as large.²⁹ We assessed inconsistencies in d with the I² statistic transformed from the Q statistic with its confidence intervals (95% CIs).^30,31 To facilitate clinical interpretation of the ds, we back converted the standardized estimate (i.e., d) into kg of weight change by multiplying the ds by the SD of baseline weight calculated on the combined sample from all included studies.^31,32

Sensitivity Analysis

Of note, in one of the included studies (Mundbjerg et al., 2018),³³ the exercise intervention was conducted within 12-months post-bariatric surgery. However, this study provided weight data at both post-exercise intervention (i.e., at 12 months post-surgery) and at 24-month post-surgery follow up. Due to the small number of studies meeting with our inclusion criteria, we included the 12- and 24-months post-surgery data from this study to calculate the weight change beyond one-year post-surgery, while acknowledging this individual study effect size represents weight change after the exercise intervention was completed, while the other study effect sizes represent weight change concurrent with the exercise intervention being conducted. We performed sensitivity analysis and found that the overall ds were the same (p=0.78) with (d=−0.26, 95%CI: −2.07, 1.55) and without (d=−0.30; 95%CI: −1.93, 1.34) the Mundbjerg study³³ included. This result from the sensitivity analysis supported our decision to include this study.

Statistical Computing

Analyses were performed using Stata version 17.0 (Stata Crop, College Station, TX) with macros for meta-analysis, and incorporated random-effects assumptions.³⁴ Descriptive statistics are reported as mean±SD unless stated otherwise. Two-sided significance level was p<0.05.

RESULTS

The search for RCTs yielded 1,368 studies. After screening and selection processes, five RCTs met the inclusion criteria.^33,35–38 The study selection process is detailed in the PRISMA flow chart (Figure 1).

Figure 1.

PRISMA flow diagram of study selection process

Study Characteristics

The five qualifying RCTs were published in English language journals from 2011 to 2020 (2016±3 years) with a total of 189 (38±17 per study) participants. RCTs were conducted in the United States (k=2, 40%),^37,38 United Kingdom (k=1, 20%),³⁵ Spain (k=1, 20%),³⁶ and Denmark (k=1, 20%).³³ Using the Cochrane risk of bias tool, three of the five (60%) RCTs carried high risk of bias,^33,36,37 one (20%) carried some concerns of bias,³⁸ and one (20%) carried low risk of bias.³⁵ For the individual RCT risk of bias scores, see Online Supplemental Content B. The domain in which most of the RCTs exhibited high level of risk of bias was Deviation from Intended Interventions (k=3, 60%).^33,36,37

Participant Characteristics

As shown in Table 1, participants in the exercise intervention group were generally between 40 and 55 (47.8±4.2) years of age. Most participants identified as female (83.2±9.5%). When reported (k=2, 40%),^37,38 participants identified as White (59.0%), Black (22.1%), and others (18.9%). More than half (61.7%) of the participants had Roux-en-Y gastric bypass surgery, followed by sleeve gastrectomy (24.5%) and gastric banding (13.8%). The mean time from surgery to study baseline was 18.9±11.4 months, ranging from 12 to 37 months. At study baseline, participants had a mean body mass index (BMI) of 36.1±3.8 kg.m⁻² and a mean weight of 96.5±9.3 kg.

Table 1.

Summaries of Participant and Intervention Characteristics of Individual Studies (N=5)

STUDIES	PARTICIPANTS CHARACTERISTICS		INTERVENTION CHARACTERISTICS
	Exercise Intervention Group	Control Group	Exercise Intervention Group	Control Group
Shah et al., 2011	N=21 Age=53.9 yr Female%=92.0% BMI=41 kg.m−2 Weight=110.3 kg Ethnicity: 75% White; 25% Black	N=12 Age=47.3 yr Female%=90.0% BMI=42.4 kg.m−2 Weight=101.4 kg Ethnicity: 43% White; 48% Black; 9% Hispanic	Prescribed FITT: 5d/wk × 60–70%VO2max × 42min/d# × 12wk of aerobic exercise; goal was 2000 Kcal/wk Progression: Based on exercise intensity, individualized Supervision and logging: Supervised, logged, tracked with HR monitor Adherence: NA Additional intervention: Diet same as control Average time from surgery to baseline: 17.7 months	Prescribed activity: Diet: 1200–1500 Kcal/wk + multivitamin and mineral supplements
Coleman et al., 2017	N=26 Age=52.0 yr Female%=84.6% BMI=32.7 kg.m−2 Weight=90.8 kg Ethnicity: 53.8% White; 11.5% Black; 19.2% Hispanic	N=23 Age=46.6 yr Female%=84.0% BMI=33.1 kg.m−2 Weight=87.0 kg Ethnicity: 64% White; 4% Black; 32% Hispanic	Prescribed FITT: 2d/wk × moderate-vigorous × 60min/d × 26wk of aerobic and resistance exercise; supplemented by flexibility and neuromotor exercise; goal was 150min/wk Progression: Based on exercise intensity, individualized; focused on reducing barriers to exercise at the beginning Supervision and logging: Supervised and logged in the health center; logged outside the health center; tracked with pedometer Adherence: On average, participants performed 60 min/wk of supervised + 139.4 min/wk of unsupervised exercise Additional intervention: Weekly counselling over the phone Average time from surgery to baseline: 13.7 months	Prescribed activity: Usual care that included lab tests, weight measures, counselling on physical activity and diet
Herring et al., 2017	N=12 Age=44.3 yr Female%*=91.7% BMI=38.2 kg.m−2 Weight=106.5 kg Ethnicity: NA	N=12 Age=52.4 yr Female%*=91.7% BMI=39.4 kg.m−2 Weight=106.0 kg Ethnicity: NA	Prescribed FITT: 3d/wk × 64–77%HRmax if aerobic and 60% 1-RM if resistance exercise × 60min/d × 12wk of aerobic and resistance exercise; goal was 180min/wk Progression: Based on exercise intensity and time, individualized Supervision and logging: Supervised Adherence: On average, participants performed 171 min/wk of supervised exercise; 95% attendance rate Additional intervention: None Average time from surgery to baseline: 19.3 months	Prescribed activity: Usual care
Mundbjerg et al., 2018	N=32 Age=42.3 yr Female%=74.4% BMI=33.3 kg.m−2 Weight=91.6 kg Ethnicity: NA	N=28 Age=42.3 yr Female%=80.0% BMI=34.1 kg.m−2 Weight=91.9 kg Ethnicity: NA	Prescribed FITT: 2d/wk × 50–70%VO2max × 40min/d × 26wk of aerobic exercise; supplemented by upper-body resistance exercise; goal was 80min/wk Progression: Based on exercise intensity, individualized Supervision and logging: Supervised Adherence: 19(59.4%) subjects attended >50% of all sessions; 8(25%) attended < 10% Additional intervention: Standard diet recommendations Average time from surgery to baseline: 12.0 months	Prescribed activity: Standard diet and physical activity recommendations
Hernandez et al., 2020	N=11 Age=50.6 yr Female%=70.0% BMI=34.4 kg.m−2 Weight=92.5 kg Ethnicity: NA	N=10 Age=46.4 yr Female%=87.5% BMI=32.8 kg.m−2 Weight=84.6 kg Ethnicity: NA	Prescribed FITT: 2–4d/wk × 60–80%HRmax if continuous aerobic + 60–95% HRmax if HIIT + 75%1-RM if resistance exercise × 50–80min/d × 20wk of aerobic and resistance exercise; supplemented by flexibility exercise; goal was ~160min/wk Progression: Based on exercise frequency, intensity, time and type, individualized Supervision and logging: Supervised, logged Adherence: Participants attended >85% of all sessions Additional intervention: Diet same as control Average time from surgery to baseline: 37.0 months	Prescribed activity: Usual care

Note.

^*=data was reported for participants in the exercise intervention and control groups combined;

^#=For Shah et al., 2011, we estimated the 42min/d based on time spent in moderate intensity physical activity data reported for baseline, 6wks, and 12 wks.

Abbreviations: Female%=percent of female participants; BMI=Body Mass Index; VO_2max=maximal oxygen uptake; FITT=frequency, intensity, time and type; HR=heart rate; NA=not available; 1-RM=one repetition maximum; HIIT=high intensity interval training.

Exercise Intervention Characteristics

The FITT characteristics of each exercise intervention/study are presented in Table 1. On average, exercise interventions were initiated at 18.9±11.5 months (including the Mundjberg study³³) and 22.1±10.3 months (not including the Mundjberg study³³) after bariatric surgeries and lasted an average of 19.2±7.0 weeks. Exercise interventions primarily prescribed a combination of aerobic and resistance training with warm-up and cool-down periods, using a variety of exercise options, namely walking on the treadmill, stair climbing, rowing, and bicycling for aerobic exercise; and workouts that target different muscle groups such as leg press and extension for resistance exercise. Participants completed the exercise interventions either only at fitness facilities while supervised by exercise professionals,^33,35–37 or at both a fitness facility while supervised and on their own.³⁸ Flexibility training was also included in two of the five studies.^36,38 Participants were instructed to perform 80–210 weekly minutes of MVPA over a maximum frequency of five days/week. All studies progressed the exercise volume with strategies such as increasing exercise frequency, increasing the length of each exercise session, and maintaining or increasing exercise intensity as the cardiorespiratory and muscular fitness of the participants improved over time. Two studies also progressed participants based on the type of exercise. Specifically, Hernandez et al. started participants with resistance, flexibility, and continuous aerobic exercise, and added high intensity interval training from the third week on.³⁶ Meanwhile, Coleman et al. emphasized the importance of using functional training that mimics activities of daily living for improving overall mobility at the beginning of the program.³⁸

Weight Change Within and Between the Exercise Intervention and Control Groups.

Analysis of within-group effects showed non-statistically significant (p=0.87) weight loss among the exercise groups (d=−0.15, 95%CI: −1.96, 1.65; −1.4 kg) and non-statistically significant (p=0.90) weight gain among the control groups (d=0.11, 95%CI: −1.70, 1.92; 1.0 kg). Analysis of between-group effects showed a non-statistically significant (p=0.78) weight loss among the exercise compared to control groups (d=−0.26, 95%CI:− 2.07, 1.55; −2.4 kg). We did not observe heterogeneity (I² =0.0%, 95%CIs: 0.0, 15.8).

Publication Bias

Based on the tests introduced by Begg and Mazumdar (p=0.81) and Egger et al. (p=0.52), we did not observe any publication or small sample bias, whereas the funnel plots suggested publication or other reporting bias (see Online Supplemental Material C).

DISCUSSION

This study aimed to conduct a systematic review and meta-analysis of RCTs on the potential benefit of exercise interventions for counteracting weight recurrence≥ 12-months after bariatric surgery. We focused only on those interventions that specified the amount of exercise performed by the FITT principle of exercise prescription with the intention of obtaining information on the exercise intervention characteristics that could better inform future exercise recommendations for reducing post-surgical weight recurrence. This systematic review yielded five RCTs comparing exercise interventions framed by the FITT principle to control groups that provided at least usual care. We found participants who were assigned to the exercise interventions achieved modest weight loss (−1.4 kg), whereas those assigned to usual care control conditions experienced modest weight gain (+1.0 kg), although these within-group effects were not statistically significant. Similarly, the magnitude of difference in weight change between the exercise and control groups represented a small effect (d=−0.26) and was not statistically significant.

Our findings run contrary to those from multiple observational studies indicating higher levels of physical activity, including MVPA, associate with lower weight recurrence after bariatric surgery.^4–10,13 This discrepancy likely owes to multiple factors, as detailed below.

First, none of the exercise intervention FITT prescriptions aligned with the 250–300 weekly minutes of moderate-intensity physical activity recommended by the American College of Sports Medicine (ACSM) and the 2018 Physical Activity Guidelines for management of obesity and prevention of weight recurrence.^18,39 Indeed, exercise intervention prescriptions/goals varied widely with respect to intensity (light-to-vigorous) and time (80–210 MVPA minutes/week). Similarly, prescribed exercise frequency ranged from 2 to 5 days, fewer than the ≥5 days/week recommended by the ACSM, especially for adults with overweight and obesity.¹⁸

Second, the precise adherence rates to the exercise intervention prescriptions/goals was reported by only two of the five studies,^35,38 making it difficult to assess intervention feasibility and determine the precise amount of exercise performed. Of the five included RCTs, Herring and colleagues demonstrated the best adherence with a 95% session attendance rate and participants performing an average of 171 minutes/week of supervised aerobic and resistance exercise at moderate intensity.³⁵ As a previous review indicated,⁴⁰ it is imperative that future studies report key indicators of exercise adherence (i.e., attendance rate, dropout rate, average FITT of exercise performed at multiple assessment points, and percentage of participants who achieved exercise prescription/goals) to obtain the necessary information to inform the development of both exercise guidelines and the optimal FITT of exercise interventions to counteract post-surgical weight recurrence. Additionally, investigators should measure participant barriers to exercise adherence (e.g., lack of time, motivation, pain levels), and effect moderators of the exercise intervention weight outcome response (e.g., sex, age, initial fitness level, race/ethnicity, FITT) to better optimize and tailor interventions for patients following bariatric surgery. Related to intervention tailoring, it is important to recognize that using general exercise guidelines for weight management to specify FITT of exercise interventions might not be appropriate or feasible for some bariatric surgery patients (e.g., those who are initially inactive versus insufficiently active or who have musculoskeletal impairments and related pain that limit ability to engage in exercise at higher intensities for longer durations). For these patients, the initial focus might be less on higher exercise volume to attenuate weight recurrence and more on finding opportunities throughout the day to move more regardless of intensity (and sit less) and performing functional exercises, similar to the intervention conducted by Coleman et al. (see Table 1),³⁸ that help with everyday activities and can prepare the body for eventual progression to higher-intensity aerobic and resistance exercises. Alternatively, the FITT may be modified to focus on performing low-impact weight-bearing exercises in multiple shorter bouts versus a single longer bout.

Finally, only two of the studies were designed to reduce weight recurrence,^33,36 and none of the studies provided information on whether participants were weight stable or regaining weight. Thus, our analysis was limited to the few studies that measured weight at post-intervention and at post-surgery intervals (18–37 months post-surgery) when weight recurrence typically occurs. As a result, we were only able to investigate whether exercise contributes to positive weight outcomes during periods of likely weight recurrence but not whether exercise prevents and/or reverses weight recurrence per se. Additionally, it is important to note that there was modest non-significant weight gain even in the control groups (d=0.11, 1.0 kg; p=0.9). This might be due to that all but one of the interventions were initiated before 19 months post-surgery which may not have afforded sufficient time for patients to experience significant weight recurrence.¹ It is also possible that some study participants were already exercising and/or were very compliant with other aspects of their care and thus may not be representative of the general bariatric surgery population.

Our findings underscore the need for exercise intervention trials with clear focus, participant selection, specification of the exercise amount by the FITT principle, and outcomes related to weight recurrence. Such trials should specify whether the goal is to prevent or reverse clinically significant weight recurrence and match participant selection accordingly—i.e., trials focused on prevention of weight recurrence should include participants who are relatively weight stable, whereas trials focused on reversal of weight recurrence should include participants who are experiencing clinically significant weight recurrence. Careful selection and matching of participants with the weight recurrent focus/outcome will enhance clarity of data interpretation regarding the specific role(s) of exercise in weight recurrence following bariatric surgery.

A unique feature of this literature review was the framing of the review and analysis around the FITT principle of exercise prescription.^18,19 While we aimed to compile and leverage the FITT characteristics from each exercise intervention to both guide clinicians in advising patients on exercise for counteracting weight recurrence and better inform future exercise recommendations, the small number and size of RCTs and wide variability in FITT applied undermined this effort. We also note that the participants were primarily middle-aged, White, and female. It is important future trials recruit larger and more diverse samples to better understand the roles of age, sex, and race/ethnicity in exercise intervention feasibility, acceptability, and efficacy following bariatric surgery. Finally, while sleeve gastrectomy is currently the most commonly performed bariatric surgery procedure,⁴¹ most participants in this study underwent Roux-en-Y gastric bypass. It remains unclear if the role of exercise in weight recurrence is moderated by surgery type. Notably, while study inclusion criteria were broad to ensure there was sufficient study heterogeneity to investigate important moderators (e.g., exercise intervention duration, initial participant exercise level, exercise measurement method) of intervention effects on weight recurrence, the number of studies was insufficient to perform such analyses. This limitation further highlights the need for additional RCTs of exercise interventions designed to prevent or reduce post-surgical weight recurrence.

Although there was considerable variability in the FITT applied, all exercise interventions involved supervised aerobic exercise alone or combined with resistance training for ≤ 26 weeks in length. Yet, none of the exercise interventions reported advising patients how to sustain exercise adherence after supervision ended. This lack of disclosure of behavioral support contradicts the current understanding of obesity as a chronic and relapsing disease,⁴² with the potential for weight regain to develop into a serious long-term complication,⁴³ and data showing lower PA associates with higher risk for weight regain after both surgical and non-surgical weight loss treatments.^{4–10,13,44–47} Interventions should thus both teach behavioral strategies (e.g., exercise self-monitoring, goal-setting, and planning/scheduling) to sustain exercise over the long-term after formal exercise program completion.^48–50

We acknowledge that there are several limitations of the current meta-analysis. First, although we systematically searched five electronic databases with the aid of a medical librarian, we did not include the grey literature (e.g., conference proceedings, dissertations) or articles published in languages other than English. In addition, we advise the reader to place a moderate level of confidence in our statistical findings due to the above described methodological limitations in the included studies (e.g., small number of studies that had sufficient follow up to measure weight change after 12 months post-surgery; small sample sizes of trials; wide variability in exercise FITT; and insufficient data regarding the amount of exercise performed and weight change trajectory of participants prior to intervention), as well as the observed high level of risk of bias and publication bias that both indicated the observed effect of intervention may not accurately represent the true benefits of exercise on preventing weight recurrence.⁵¹ Lastly, due to the small number of studies that met with our inclusion criteria, we did not have adequate statistical power to examine if important variables, such as the baseline level of physical activity, the amount of free-living physical activity accumulated during the intervention period, and the length of exercise intervention, impacted intervention effects on weight recurrence. Nonetheless, this was the first meta-analysis that aimed to address the gaps in the literature by reviewing and meta-analyzing RCTs that compared the effects of exercise interventions, specifying FITT to control groups on weight outcomes ≥ 12-months following bariatric surgery. In addition, the methodological limitations we have outlined may guide future researchers to conduct well-designed, rigorous exercise interventions to prevent or reverse weight recurrence in larger, more diverse patient samples.

Finally, this systematic review and meta-analysis focused on the role of exercise in counteracting weight recurrence given that it is highly prevalent and contributes to reemergence of comorbidities, impaired quality of life and reduced surgery satisfaction.^1,43 However, it is important to recognize that the studies included in this review and others have shown that physical activity and exercise can confer a host of other physical (e.g., improved body composition, physical function, , and cardiometabolic risk profile [e.g., glucose levels and insulin sensitivity])^{33,35–38,52} and mental (e.g., improved self-efficacy, emotional and social well-being, and depression symptoms)^35,37,53 health benefits beyond weight loss in bariatric surgery patients.

CONCLUSIONS

Weight recurrence after bariatric surgery is common and has potential to develop into a serious long-term complication. While exercise is recommended to prevent post-surgical weight recurrence, the effectiveness of exercise interventions in doing so is unclear, providing clinicians with limited evidence-based guidance on prescribing the optimal FITT exercise prescription for counteracting weight regain. Our results showed modest nonsignificant weight loss among the exercise participants and modest nonsignificant weight gain among control participants, although the between-group effect was of small magnitude and nonsignificant. While the results of this systematic review and meta-analysis do not definitively support that exercise interventions effectively counteract post-surgical weight recurrence, the small number of studies and several important methodological limitations lower confidence that the current evidence reflects the true effect. Thus, it is premature to conclude that exercise is not an effective strategy for reducing weight recurrence. Highlighted limitations of the studies reviewed may guide future researchers to conduct well-designed, rigorous exercise interventions to prevent or reverse weight recurrence in larger, more diverse patient samples. Finally, while not examined in this review, there are likely numerous additional health-related benefits of physical activity and exercise that can be realized by patients following bariatric surgery that are independent of weight status.

HIGHLIGHTS.

• Exercise has potential to prevent weight recurrence after bariatric surgery.

• Exercise interventions had a small effect on weight beyond ≥1-year post-surgery.

• Methodological limitations undermine confidence that the observed effect is accurate.

• Rigorous testing of exercise effects on post-surgical weight recurrence is needed.