The role of glucagon-like peptide-1 receptor agonists in weight regain treatment or prevention after bariatric surgery: a systematic review and meta-analysis

Abstract

Background

Weight regain (WR) and insufficient weight loss (WL) occur in 20–25% of patients after bariatric surgery due to various factors. Glucagon-like peptide-1 receptor agonists (GLP-1RAs) have shown promise in promoting WL; however, evidence regarding their effectiveness in managing WL and preventing WR post-bariatric surgery remains limited.

Objective

This study aims to evaluate the role of GLP-1RAs in treating and preventing WR following bariatric surgery.

Methods

A systematic search was conducted across PubMed, Scopus, and Web of Science for studies assessing the impact of GLP-1RAs on WR after bariatric surgery.

Results

Our search identified 27 original studies, with 10 included in the meta-analysis involving 769 participants (392 treated with GLP-1RAs). The mean age was 44.05 years, with 30.47% male. The time interval from surgery to the initiation of GLP-1RAs treatment ranged from 1.5 to 86.7 months, with treatment durations between 4 and 12 months. The analysis showed significantly greater WL in the GLP-1RAs group compared to placebo (SMD = 0.82, 95% CI 0.23 to 1.42). Subgroup analysis for treatment durations ≤ 6 months indicated a higher WL in the GLP-1RAs group (SMD = 0.79, 95% CI 0.25 to 1.34). Adverse events were primarily gastrointestinal, with nausea significantly more frequently in the GLP-1RAs group (OR = 2.01, 95% CI 1.24 to 3.27).

Conclusion

GLP-1RAs effectively promote WL among participants experiencing WR after bariatric surgery. Initiating GLP-1RAs therapy shortly after surgery may help prevent WR. Further research is warranted to explore long-term outcomes and optimize treatment protocols for this patient population.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1007/s40519-025-01770-z.

Highlights

• GLP-1RAs promoted weight loss in post-bariatric surgery patients faced with weight regain or insufficient weight loss.

• Significant weight loss in the group treated with GLP-1RA was observed compared to the group got placebo in short-term (≤6 months) subgroup analysis.

• Nausea was the most common adverse event, significantly higher in the GLP-1RA group compared to placebo.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40519-025-01770-z.

Introduction

Weight regain (WR) and insufficient weight loss (IWL) are significant challenges faced by 20–25% of patients following bariatric surgery, often stemming from a combination of anatomical variations, genetic predispositions, hormonal imbalances, dietary noncompliance, and inadequate physical activity [1–3]. Despite the substantial benefits of bariatric metabolic surgery in reducing body mass index (BMI) and improving comorbid conditions such as diabetes mellitus (DM), hypertension (HTN), and hyperlipidemia (HLP), WR, IWL, and other complications can hinder long-term success [4–7]. The variability in WL outcomes post-surgery is influenced by factors such as baseline weight and surgical technique, with reductions typically ranging from 20 to 30% of pre-surgical weight [8]. Understanding the prevalence and causes of WR is crucial for developing effective management strategies to enhance sustained WL in this population.

Glucagon-like peptide-1 receptor agonists (GLP-1RAs) have emerged as a promising therapeutic option for promoting WL beyond lifestyle modifications, demonstrating efficacy in non-surgical populations [9, 10]. Additionally, GLP-1RAs have shown potential benefits in improving glycemic control and reducing major adverse cardiovascular events [11–14]. However, the evidence supporting their use specifically in patients’ post-bariatric surgery remains limited, with findings from individual studies not yet justifying widespread clinical adoption. Despite the potential benefits of GLP-1RAs, several challenges hinder their implementation in managing WR after bariatric surgery [15, 16]. These include uncertainties regarding optimal treatment duration, management of side effects, and the high cost associated with these medications. Furthermore, there is a lack of consensus on prioritizing access to GLP-1RAs for patients with diabetes versus those seeking WL alone. As interest grows in utilizing GLP-1RAs as an adjunct therapy for patients experiencing WR or as a preventive measure post-surgery, further research is essential to elucidate their role and establish evidence-based guidelines for effective use in this population. Therefore, we conducted a systematic review and meta-analysis to evaluate the current evidence regarding the outcomes associated with GLP-1RAs treatment in patients following bariatric surgery.

Materials and methods

This study was designed following the framework delineated by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [17] with rigorous attention given to registering the study protocol in the International Prospective Register for Systematic Reviews (PROSPERO) (registration code: CRD42023466256).

Search strategy and study selection

We conducted a comprehensive search through PubMed, Scopus, and Web of Science databases to identify pertinent studies from their inception until March 2024. Language constraints were consciously omitted. To ensure comprehensive coverage, similar articles and references of eligible studies and pertinent reviews were done. Detailed search strategies are available in Supplementary Table 1. Two authors (AM and SS) independently screened study titles and abstracts to ascertain eligibility for full-text assessment, with any discrepancies resolved through consultation with a senior author (HR). We included studies that initiate GLP-1RAs therapy in patients following bariatric surgery. Our exclusion criteria comprised non-original designs, studies investigated on internal GLP-1, studies lacking weight change outcomes, studies without all population undergoing bariatric surgery, absence of baseline patient characteristics.

Risk-of-bias assessment

The methodological quality of included RCTs was assessed utilizing the Cochrane Risk of Bias 2 (RoB2) tool [18], while observational studies were evaluated with the Risk Of Bias In Non-randomized Studies—of Interventions (ROBINS-I) [19]. RoB2 examines 5 domains to evaluate the risk of bias in each RCT, whereas ROBINS-I evaluates observational studies in 7 domains to determine the risk of bias as low, some, or high concern. In a collaborative effort, two designated authors (AM and SS) independently undertook the assessment, with any incongruences resolved through consultation with a senior authority (HR). Detailed findings are demonstrated in Supplementary Table 2.

Data extraction

Two designated authors (AM and SS) meticulously extracted pertinent data from the included studies, including authorship details, publication timeline, samples of the study, mean demographic parameters (e.g., age, sex, baseline weight, and baseline BMI), type of bariatric surgery, prevalent background diseases, type of the GLP-1RAs therapy, the time of the initiation of the medical treatment, laboratory indices, amount of weight-change, and a comprehensive array of outcome measures, such as nausea, vomiting, diarrhea, and serious adverse events.

Statistical analysis

In conducting the meta-analysis, we employed the R programming language (Version 4.1.3, R for Windows, Vienna, Austria) along with RStudio (Version 1.1.463, Posit PBC, Boston, MA, USA), utilizing the 'tidyverse' and 'meta' statistical packages. We investigated how GLP-1RAs impact weight changes in patients who underwent bariatric surgery compared to a placebo arm. The researchers determined the impact of GLP-1RAs on weight fluctuations after bariatric surgery by analyzing the percentage of absolute differences in body weight (kg) or percentage of body changes. Mean absolute weight changes (kg), standardized mean difference (SMD) and 95% confidence intervals (CIs) were computed for these continuous variables (WL and WR). In our overall analysis of studies reporting outcomes at different time frames, we included data with the longest follow-up periods. To mitigate treatment duration bias, we conducted subgroup analyses based on the duration of GLP-1RAs therapy (≤ 6 months and > 6 months). Secondary analyses were performed to assess complication rates as a combined odds ratio (OR) compared to the placebo group in the meta-analyses. ORs and 95% CIs were computed for these binary variables. To evaluate inter-study statistical heterogeneity, we calculated I2 parameters, with values exceeding 50% indicative of noteworthy heterogeneity. In instances where significant heterogeneity was anticipated, we opted for a random-effect model to estimate the effect size of pooled data; conversely, a common-effect model was applied without significant heterogeneity. To assess the robustness of our meta-analytic results and identify potential outlier studies, we conducted a leave-one-out sensitivity analysis. This analysis was performed using the InfluenceAnalysis() function from the dmetar package in R. The function systematically omits each study one at a time from the meta-analysis and recalculates the overall effect estimate, thereby evaluating the influence of each individual study on the pooled result. Additionally, funnel plots were generated for outcomes with a sufficient number of studies included in them.

Result

Baseline characteristics

Our comprehensive search initially identified a total of 3,955 studies. After removing duplicates and screening titles and abstracts, 285 studies remained. Ultimately, 27 studies met our inclusion criteria for the systematic review. Among these, 10 studies with two arms comparing GLP-1RAs to placebo were included in the meta-analysis. Of the included studies, five were RCTs [20–24] and five were cohort studies [25–29]. The inter-reader agreement, as measured by the Kappa coefficient, was 90%. The PRISMA flowchart outlining the study selection process is presented in Fig. 1.

Fig. 1.

PRISMA flowchart. Summary of study selection

The total number of participants across all included studies was 769. Among these individuals, 392 (50.97%) were prescribed GLP-1RAs, while the remaining participants constituted the control group receiving a placebo. The mean age of participants was 44.05 years (ranging from 32.5 to 58.6 years), with approximately 30.47% being male (34.95% in the GLP-1RAs group and 36.87% in the placebo group). The types of bariatric surgery represented in the studies included laparoscopic sleeve gastrectomy (LSG), Roux-en-Y gastric bypass (RYGB), laparoscopic gastric banding, and one anastomosis gastric bypass. The mean time interval between surgery and the initiation of GLP-1RAs ranged from 1.5 to 87.6 months. The GLP-1RAs used in these studies included shorter-acting liraglutide and once-weekly semaglutide, with maximum doses of 3 mg and 2.4 mg, respectively. Dosages varied among studies, with most participants starting at a minimum dose that was gradually increased weekly. The mean duration of GLP-1RAs treatment ranged from 4 and 12 months. Further details regarding the characteristics of the included studies can be found in Tables 1–2.

Table 1.

Baseline characteristics of studies with case and control

Study characteristics			Case group						Control group
Author/year	Country	Design	Participants number	Mean age	Male%	DM%	Baseline weight (kg)	Baseline BMI	Participants number	Mean age	Male%	DM%	Baseline weight (kg)	Baseline BMI
Mok/2023 [20]	United Kingdom	RCT	35	46.7 ± 10.8	26	14	N/A	N/A	35	48.4 ± 10.6	26	11	N/A	N/A
Miras/2019 [21]	United Kingdom	RCT	53	55.3 ± 8.3	38	N/A	127.8 ± 25.7	N/A	27	57.6 ± 9.3	48	N/A	130.8 ± 29.7	N/A
Horber/2021 [28]	Switzerland	Cohort	34	56 ± 10	8.82	N/A	120 ± 19	45 ± 8	30	55 ± 10	20	N/A	124 ± 20	44 ± 9
Hany/2023 [22]	Egypt	RCT	38	38.2 ± 9	34.2	7.9	N/A	46.6 ± 8	31	37.1 ± 12.8	22.6	6.5	N/A	43.3 ± 6.2
Gazda/2021 [27]	USA	Cohort	84	47.6 ± 10.8	20.2	50	138.4 ± 27.4	49.6 ± 9	86	42.4 ± 11.3	16.3	11.6	145.2 ± 43.4	50.8 ± 12.9
Badurdeen (1)/2020 [26]	Brazil	Cohort	26	40.6 ± 8.6	34.6	N/A	101 ± 10.6	35.8 ± 2.3	26	41.1 ± 10.6	38.5	N/A	101.8 ± 10.7	35.5 ± 1.6
Badurdeen (2)/2020 [25]	Brazil	Cohort	53	38.5 ± 12.1	69.8	N/A	100.3 ± 11	N/A	53	38.5 ± 11	69.8	N/A	100 ± 11.8	N/A
Coelho/2023 [23]	United Kingdom	RCT	13	53.4 ± 8.3	23.1	N/A	107.9 ± 28.4	40 ± 9	14	51.2 ± 8.5	21.4	N/A	102 ± 23.8	36.5 ± 8.8
Thakur/2020 [24]	India	RCT	12	40.2 ± 11.8	50	N/A	117.6 ± 20.8	42.6 ± 6.3	11	44.7 ± 11.7	45.45	N/A	103.1 ± 15.1	41.6 ± 5.1
Mosli/2017 [29]	Saudi Arabia	Cohort	44	32.5 ± 8.4	45	N/A	103.8 ± 19.1	37 ± 5.9	64	34.9 ± 9.8	55	N/A	99.3 ± 19.9	37 ± 5.9

RCT randomized controlled trial, DM diabetes mellitus, BMI body mass index

Table 2.

Baseline characteristics of studies with case and control

Study characteristics	Case group						Control group
Author/year	Surgery type	The time between surgery to drug initiation (months)	Weight before GLP-1RA (kg)	BMI before GLP-1RA	GLP-1RA type and maximum dose (mg)	GLP-1RA treatment duration (months)	Surgery type	The time between surgery to drug initiation (months)	Weight before placebo (kg)	BMI before placebo	Treatment	Treatment duration (months)
Mok/2023 [20]	RYGB or SG	55.1 ± 33.3	116.1 ± 23.6	41.6 ± 6.9	Liraglutide 3.0	6	RYGB or SG	49.1 ± 33.7	123.5 ± 24.8	44.6 ± 8.3	Placebo	6
Miras/2019 [21]	RYGB or SG	45.6 ± 24	100.7 ± 20.7	36.1 ± 7.8	Liraglutide 1.8	4	RYGB or SG	45.6 ± 24	103.5 ± 27	37 ± 7.7	Placebo	4
Horber/2021 [28]	RYGB	At least 72	84 ± 13	31.2 ± 4	Liraglutide 3.0	6	RYGB	At least 72	75 ± 15	27.1 ± 5	N/A	6
Hany/2023 [22]	RYGB	1.5	84 ± 13	31.2 ± 4	Liraglutide 3.0	6	RYGB	1.5	75 ± 15	27.1 ± 5	Placebo	6
Gazda/2021 [27]	Band ligation, gastric reduction, and bypass	87.6 ± 67.2	117.9 ± 24.5	42.3 ± 7.9	GLP-1RA	12	Band ligation, gastric reduction, and bypass	112.8 ± 91.2	110.8 ± 30.4	39.1 ± 9.6	Non-pharmacologic treatment	12
Badurdeen (1)/2020 [26]	ESG	5	87.13 ± 8.93	N/A	Liraglutide 3.0	7	ESG	5	87.59 ± 7.73	N/A	Without any treatment	7
Badurdeen (2)/2020 [25]	IGB insertion	12	N/A	N/A	Liraglutide 3.0	9	N/A	12	N/A	N/A	Without any treatment	9
Coelho/2023 [23]	LAGB	1.5	107.93 ± 28.4	N/A	Liraglutide 1.8	12	N/A	1.5	103.06 ± 23.8	N/A	Placebo	12
Thakur/2020 [24]	LSG	1.5	118.6 ± 24.6	42.6 ± 6.3	Liraglutide 3.0	6	LSG	1.5	103.1 ± 16.4	41.6 ± 5.1	Placebo	6
Mosli/2017 [29]	IGB	*					IGB				Without any treatment

GLP-1RA glucagon-like peptide 1 receptor agonists, RYGB Roux-en-Y gastric bypass, SG sleeve gastrectomy, ESG endoscopic sleeve gastroplasty, IGB intragastric balloon, LAGB laparoscopic adjustable gastric banding, LSG laparoscopic sleeve gastrectomy

Quality assessment

The quality of the included studies was assessed using RoB2 for RCTs and ROBINS-I for observational studies. Among the RCTs, one study had some concerns due to the randomization process [20], while another exhibited some concerns in the intervention domain [30]. For the observational studies, two were identified as having some concerns regarding participant selection [26, 31]. Additionally, one study had high risk of bias in the confounding domain and some concerns in the selection of reported results [32]. Lastly, one article was assessed as having some concerns in the missing data domain [29]. Beside the aforementioned studies, all other domains across the included studies were demonstrated low risk of bias according to the respective assessment tools. A summary of the quality assessment for the included studies can be found in Supplementary Table 2.

Outcomes

In our analysis of nine studies comparing GLP-1RAs to placebo for WL following bariatric surgery, we found no significant differences between the two groups (SMD = 0.56, 95% CI -0.19 to 1.30) (Fig. 2A). However, sensitivity analyses, which involved excluding individual studies, revealed that omitting one outlier study [22], characterized by substantial variability compared to the others, resulted in a significant increase in WL for the GLP-1RAs group (SMD = 0.82; 95% CI 0.23 to 1.42) compared to the placebo group (Fig. 3A).

Fig. 2.

Comparison of weight loss after therapy between groups treated with GLP-1RAs and placebo based on SMD. A Overall analysis, B subgroup analysis of studies with treatment duration of ≤ 6 months, C subgroup analysis of studies with treatment duration of > 6 months

Fig. 3.

Sensitivity analysis was conducted by omitting each study for comparison of weight loss after therapy between GLP-1RAs and placebo-treated groups. A Overall analysis, B subgroup analysis of studies with treatment duration of ≤ 6 months, C subgroup analysis of studies with treatment duration of > 6 months

Given the high heterogeneity among studies, we conducted subgroup analyses based on treatment duration, categorizing studies into those including participants receiving treatment for ≤ 6 months and those including participants treated for > 6 months. Both subgroup analyses demonstrated non-significant differences in WL between the GLP-1RAs and placebo groups with (SMD = 0.44, 95% CI -0.37 to 1.26) and (SMD = 0.74, 95% CI -0.32 to 1.79), respectively (Fig. 2B-C). Notably, sensitivity analysis by omitting one of the studies [22] for the subgroup with treatment durations of ≤ 6 months indicated a significantly greater WL in the GLP-1RAs group compared to placebo (SMD = 0.79; 95% CI 0.25 to 1.34) (Fig. 3B-C).

Moreover, regarding absolute weight change, we observed that there was a meaningful increase in WL for the GLP-1RAs group (4.17; 95% CI 1.41 to 6.93) comparing placebo group. In subgroup analyses, this difference remained significant for participants receiving ≤ 6 months treatment (5.07; 95% CI 1.70 to 8.45). However, regarding patients treated for > 6 months, this difference was non-significant (3.28; 95% CI − 0.62 to 7.17) (Fig. 4).

Fig. 4.

Comparison of weight loss after therapy between groups treated with GLP-1Ras and placebo based on absolute weight change (kg). A Overall analysis, B subgroup analysis of studies with treatment duration of ≤ 6 months, C subgroup analysis of studies with treatment duration of > 6 months

When assessing WL post-bariatric surgery and WR thereafter before initiation of therapy, our analysis found no significant differences between the GLP-1RAs and placebo groups, with (SMD = 0.23, 95% CI − 0.36 to 0.81) for WL and SMD = 0.33, 95% CI − 0.11 to 0.77) for WR, respectively (Supplementary Figs. 1 and 2). Sensitivity analyses for WL outcome also remained non-significant when each study was omitted individually (Supplementary Fig. 3). However, excluding the outlier study [27] revealed a significant increase in WR in the GLP-1RAs treatment group compared to placebo (SMD = 0.49, 95% CI 0.12 to 0.86) (Supplementary Fig. 4). A summary of the evaluated outcomes is presented in Table 3. Moreover, funnel plots of outcomes were symmetrical and we observed no significant publication bias. These figures are available in supplementary materials (Supplementary Figs. 5–7).

Table 3.

Outcomes analyses

Outcomes	Number of studies	SMD (95% CI)	I2	P-value
WL after surgery and before treatment	6	0.23 (− 0.36; 0.81)	92%	< 0.01
WR after surgery and before treatment	5	0.33 (− 0.11; 0.77)	84%	< 0.01
WL after treatment  <  = 6 months  > 6 months	9 7 4	0.56 (− 0.19; 1.30) 0.44 (− 0.37; 1.26) 0.74 (− 0.32; 1.79)	94% 94% 93%	< 0.01  < 0.01  < 0.01

SMD standardized mean difference, CI confidence interval, GLP-1RA glucagon-like peptide 1 receptor agonists, WL weight loss, WR weight regain

Adverse events

The analysis of adverse events between the GLP-1RAs and placebo groups revealed that the most frequently reported adverse effects were nausea, vomiting, and diarrhea. Analysis of six studies indicated a significantly higher incidence of nausea in GLP-1RAs group compared to the placebo group, with an odds ratio (OR) of 2.01 (95% CI 1.24 to 3.27) (Fig. 5a). In contrast, the occurrence of vomiting and diarrhea did not show significant differences between the two groups; the OR for vomiting was 1.00 (95% CI 0.54 to 1.85) based on six studies (Fig. 5b), while the OD for diarrhea was 0.70 (95% CI 0.22 to 2.23) from four studies (Fig. 5c) (Table 4).

Fig. 5.

Comparison of adverse events between GLP-1RAs and placebo-treated groups: A nausea, B vomiting, C diarrhea

Table 4.

Adverse events following GLP-1RA treatment in studies with case and control

Adverse event	Number of studies	OR (95% CI)	I2	P-value
Nausea	6	2.01 (1.24; 3.27)	45%	0.10
Diarrhea	4	0.70 (0.22; 2.23)	0%	0.68
Vomiting	6	1.00 (0.54; 1.85)	0%	0.84

OR odds ratio, CI confidence interval

Meta-regression

The meta-regression analysis results reveal several insights regarding WL and WR after surgery and also WL after therapy. For participants who received GLP-1RAs therapy for more than six months, the intercept estimates were positive for age and gender and negative for baseline weight, indicating a potential increase in WL; however, the coefficients for all variables (both case and control) were not statistically significant, suggesting that these demographic factors do not have a meaningful impact on WL in this context. In contrast, for those on GLP-1RAs therapy for 6 months or less, the intercepts were negative for age and positive for gender, but again, the estimates for age and sex did not yield significant results, indicating a lack of influence from these variables on weight changes. Additionally, in the analyses of WL after surgery and before treatment, the intercepts were also positive for age with p-value = 0.07 and p-value = 0.09 for GLP-1RAs and placebo treatment groups, respectively, suggesting that older age may be associated with slightly reduced WL. The results for WR after surgery demonstrated similarly non-significant findings across all demographic variables. Overall, these analyses indicate that while there may be trends in the data, demographic factors such as age and sex do not appear to significantly influence WL or WR outcomes following GLP-1RAs therapy or bariatric surgery. Further research may be needed to explore other potential predictors of weight change in these populations. Supplementary Table 3 indicates the meta-regressions.

Observational studies

In our systematic review, we included 17 observational studies involving 1592 participants who received GLP-1RAs after bariatric surgery. Among these studies, three were classified as case series [33–35], while the remaining were observational studies [4, 36–48]. The mean age of participants was 43.94 years, with 20.98% being male. The time interval between bariatric surgery and the initiation of GLP-1RAs treatment ranged from 12 to 100.8 months, and the duration of GLP-1RA therapy varied from 3 to 12 months. Supplementary Tables 4 and 5 summarize the baseline characteristics of these observational studies. The assessment of these studies revealed that WL following bariatric surgery varied significantly, with reductions ranging from 3.6% to 66.7%. Additionally, WR after surgery was observed in a range of 8% to 80%. After the initiation of GLP-1RA treatment, participants experienced WL between 3.6% and 17.6%. These findings underscore the variability in weight management outcomes among patients post-bariatric surgery and highlight the potential role of GLP-1RAs in addressing WR in this population.

Discussion

The use of GLP-1RAs in the context of WL following bariatric surgery has been a topic of growing interest. GLP1-RAs effects on WL are well-established in the general population and in patients with diabetes [49–53]. However, fewer studies have specifically explored the efficacy of GLP-1RAs in preventing WR and inducing WL after bariatric surgery [54–57]. We conducted a meta-analysis, considering a total of 10 original studies. Consistent with previous studies, our findings suggest that GLP1-RAs use leads to significant WL in patients who have undergone bariatric surgery [54, 56]. Especially, liraglutide as one the main GLP1-Ras by reducing mean BMI and body weight percentage affects improving WL outcomes. [58] GLP-1 agonists were demonstrated to be an important part of standard of care after bariatric surgery, particularly in those with insufficient WL [59]. As observed in our results, nausea could be the most common complication of GLP1-Ras therapeutic plans, after constipation, and abdominal pain [60].

GLP-1RAs have been investigated for long-term use to achieve sustained WL effects [49, 61–63]. While the most pronounced weight loss with GLP-1RA therapy tends to occur within the first 6 months of treatment, continued use is generally associated with maintenance or further incremental weight loss. However, the rate of weight reduction might slow over time [52, 64, 65]. This pattern could be influenced by factors such as medication adherence and physiological adaptation. These studies indicate that the amount of WL may vary in real-world data compared to clinical trials [66].

Gastrointestinal adverse effects such as nausea, vomiting, ileus, and gastroparesis may be responsible for non-adherence to treatment in the long term. The maximum dose tolerated was 3 mg for liraglutide and 2.4 mg for semaglutide in the majority of cases [67]. Nevertheless, our results suggest that nausea is the only significant adverse effect in patients on medication, reported twice as much in cases undergoing treatment. It is important to note that these side effects are not very common and can often be managed effectively [61, 68–71]. In general, the use of GLP-1RAs was linked to tolerable rates of adverse events with favorable safety profiles [71, 72]. Thus, the adverse events mentioned earlier should not discourage healthcare providers from considering treatment with GLP-1RAs. Healthcare providers can address these concerns by discussing them with patients and weighing the potential benefits of GLP-1RAs therapy against the risks.

Considering the timing of GLP-1RAs initiation after bariatric surgery, recent studies indicate that the proper healing and metabolic adaptations that occur following bariatric procedures, such as changes in gut hormones and insulin sensitivity, may play a role in the effectiveness of GLP-1RAs [73]. This timing may be influenced by the enhanced endogenous secretion of GLP-1RAs following surgery, which has been shown to contribute to appetite management, insulin secretion, and WL [74–76]. Therefore, initiating GLP-1RAs early in the postoperative period may not yield optimal results compared to delayed initiation. Nevertheless, following the first year after surgery, there is an increased likelihood of experiencing more pronounced WR [77]. Alternatively, early GLP1-RAs initiation may help prevent WR altogether. The optimal timeframe involves balancing the enhanced response from later initiation with the need to avoid WR, considering the complex interplay between the timing of GLP1-RAs initiation and its impact on WL and WR post-surgery. While the available data on the effects of treatment termination are insufficient, minimizing WR post-treatment is crucial to achieving sustained management of obesity in patients who have undergone bariatric surgery [65].

As stated, the WL effects of GLP-1RAs are notable, and potential additional benefits should also be considered, especially in managing comorbidities such as diabetes and cardiovascular risk factors. Our results suggest that implementing GLP-1RAs therapy could help prevent WR after bariatric surgery. As earlier demonstrated, most of the patients among included studies were women, as it is known that more women than men seek treatment for obesity [78]. Moreover, based on some studies which reported that women appear to respond better to GLP1-RA treatment [79], our findings may not be fully generalizable to male patients and might vary by sex. These sex-specific differences in treatment response require further investigation. The lack of long-term follow-up on effectiveness and adverse events is one of the main limitations in the present studies. Furthermore, little is known about the WL sustainability after GLP-1RAs cessation and the optimal time frame for initiating the agent post-bariatric surgery. Real-world data on effectiveness and safety may differ from findings in controlled studies, highlighting the need for additional larger-scale studies and real-world data to guide clinical decision-making. There are a few available analyses comparing the efficacy of liraglutide and semaglutide for patients after bariatric surgery with regain [56]. However, the need for further research to directly compare these two agents is evident. Future studies should focus on conducting trials to provide more robust evidence on the comparative effectiveness of liraglutide and semaglutide in managing WR post-bariatric surgery. Additionally, exploring the efficacy of interclass drug switches could be a valuable area for future investigation to optimize weight management strategies for patients who experience challenges with one medication. To provide a more comprehensive understanding of the impact of GLP-1RAs on post-bariatric surgery outcomes, future research should concentrate on resolving these limitations.

Our study has several potential limitations that should be acknowledged. Firstly, like other systematic reviews, our analysis heavily relies on published literature. It is possible that studies with negative or non-significant results were less likely to be published, which could lead to an overestimation of treatment effects. Moreover, not including trial registries and gray literature in order to maintain the robustness and reliability of our manuscript could also emphasize the publication bias in our study. Although we made significant efforts to conduct a comprehensive search and minimize publication biases, it is essential to acknowledge that some biases may still exist. Another limitation of our study is the significant variations in GLP-1RAs classifications and dosages across the included studies. These variations may have contributed to the observed heterogeneities in the data. Different types of surgery observed among studies could also influence the reported analysis. Unfortunately, due to the limited number of studies available, we could not provide a detailed subgroup analysis based on the specific GLP-1RAs classifications, procedure type, baseline patient characteristics, and the timing of therapy initiation. Especially, most of the observational studies reported treatment initiation after more than 50 months. Therefore, our analysis cannot demonstrate results regarding a shorter treatment initiation. Additionally, the lack of exact mean dosage reporting in the studies prevented us from conducting a thorough subgroup analysis based on dosage. Finally, higher percentage of included patients diagnosed with DM in some of our studies might have influence both surgical and pharmacological outcomes. Despite these limitations, our study provides valuable insights. In comparison to previous systematic reviews and meta-analyses, we were able to perform subgroup analyses based on the duration of treatment with GLP-1RAs. These variables can introduce biases in the results, and we have discussed their potential impact in the results and discussion section. It is essential to recognize these limitations when interpreting the findings of our study. Future research should address these limitations by including a more comprehensive range of studies, providing detailed GLP-1RAs classification information, and reporting precise dosages used in the treatment protocols.

Conclusion

In conclusion, our study findings suggest that the use of GLP-1RAs may offer significant benefits for patients who have undergone bariatric surgery and are facing WR or IWL Additionally, it is important to highlight that the safety profile of these medications was favorable, with no significant adverse events reported. These results support the potential of GLP-1RAs as a valuable therapeutic option in managing weight outcomes in this patient population. While utilizing GLP-1Ras in patients with diabetes or underlying cardiovascular disease were demonstrated to have diverse benefits, such as reductions in HbA1c levels, blood pressure control, improving endothelial function, and enhancing the lipid profile [80–82], our meta-analysis focused exclusively on weight-related outcomes. However, these findings make GLP-1RAs a valuable therapeutic option for the management of WR in post-bariatric surgery patients, offering additional benefits apart from WL.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization: SS, AM; methodology: HR, SS, AM; investigation: SS, AA; validation: HR, formal analysis: AA, RAB; supervision: HR; visualization: SA, project administration: SS, AM, writing—original draft: SS, AM, SM, SY, SM; writing—review and editing: AA, RAB, HR.

Funding

Not applicable.

Data availability

All data generated or analyzed during this meta-analysis are derived from peer-reviewed publications, as detailed in the references.

Declarations

Competing interests

The authors declare no competing interests.