Efficacy of lifestyle modification combined with GLP-1 receptor agonists on body weight and cardiometabolic biomarkers in individuals with overweight or obesity: a systematic review and meta-analysis

Summary

Background

Glucagon-like peptide-1 receptor agonists (GLP-1RAs) are established treatments for obesity. However, it remains inconclusive whether the combination of lifestyle modifications and GLP-1RA interventions can lead to greater weight loss and better control of cardiovascular biomarkers. We aimed to evaluate the efficacy of this combination therapy on weight loss and cardiometabolic markers in adults with overweight or obesity.

Methods

We searched PubMed, Embase, and the Cochrane Library to identify randomized controlled trials published from inception until May 10, 2025 that assessed the effects of lifestyle modifications combined with GLP-1RAs in adults with overweight or obesity. The standard control group in these trials was set as lifestyle modifications combined with placebo. The efficacy outcomes were the changes of body weight, waist circumference, blood pressure, fasting blood glucose, glycated hemoglobin and lipids. Mean differences (MDs) were calculated using a random-effects model to assess the effects of lifestyle modifications combined with GLP-1RAs on weight loss and cardiometabolic markers. Risk of bias was assessed with the Cochrane risk-of-bias algorithm, with overall certainty of evidence evaluated by Grading of Recommendations Assessment, Development and Evaluation (GRADE) tool. This study is registered with PROSPERO (CRD42024600250).

Findings

A total of 33 randomized controlled trials involving 12,028 participants were included. Lifestyle modification combined with GLP-1RAs results in a significant mean weight loss of 7.13 kg compared with control groups (MD: −7.13 kg, 95% CI: −9.02, −5.24, P < 0.001). Significant improvements were also observed in other body composition parameters and cardiometabolic biomarkers, including waist circumference (MD: −5.74 cm, 95% CI: −7.17, −4.31, P < 0.001), lean mass (MD: −1.29 kg, 95% CI: −2.17, −0.41, P = 0.004), fat mass (MD: −2.93 kg, 95% CI: −4.70, −1.12, P = 0.001), systolic blood pressure (MD: −3.99 mmHg, 95% CI: −5.66, −2.33, P < 0.001), diastolic blood pressure (MD: −1.11 mmHg, 95% CI: −1.71, −0.42, P = 0.002), glycated hemoglobin (MD: −0.31%, 95% CI: −0.47, −0.15, P < 0.001), fasting blood glucose (MD: −6.51 mg/dL, 95% CI: −7.31, −4.71, P = 0.004), total cholesterol (MD: −5.85 mg/dL, 95% CI: −9.78, −1.91, P = 0.004), triglycerides (MD: −13.44 mg/dL, 95% CI: −20.38, −6.50, P < 0.001), low-density lipoprotein cholesterol (MD: −4.78 mg/dL, 95% CI: −7.35, −2.22, P = 0.003), with the exception of high-density lipoprotein cholesterol (MD: −0.14 mg/dL, 95% CI: −1.05, 0.76, P = 0.750). Longer treatment duration, use of semaglutide or tirzepatide, weekly dosing, and studies conducted in North America showed more pronounced weight loss effects. Risk of bias assessment indicated no high-risk studies among the included trials. The GRADE assessment indicated a range of certainty from low to high across outcomes, with high certainty for changes in high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, and body fat percentage, and moderate to low certainty for changes in body weight, blood pressure, glycemic outcomes, and other metabolic outcomes, largely influenced by heterogeneity and potential publication bias.

Interpretation

Lifestyle interventions combined with GLP-1RAs may help reduce body weight and improve cardiometabolic biomarkers in adults with overweight or obesity. In light of the varying certainty of evidence across outcomes, these results should be interpreted cautiously. Treatment duration, drug type, dosage, and geographic region were key influencing factors of intervention effectiveness.

Funding

Startup Fund for Young Faculty at Shanghai Jiao Tong University (Grant No. KJ3-0214-24-0011).

Research in context.

Evidence before this study

Obesity has been redefined as a chronic systemic disease and one of the most critical public health challenges. Lifestyle intervention, comprising caloric restriction and increased physical activity, remains the cornerstone of obesity management and is universally recommended as the first-line therapeutic approach. Glucagon-like peptide-1 receptor agonists (GLP-1RAs), due to their substantial effects on weight reduction and metabolic improvement, have become an important adjunctive treatment for obesity. However, to date, no systematic review has comprehensively examined the effects of lifestyle interventions combined with GLP-1RAs on weight loss and cardiometabolic outcomes. We searched PubMed, Embase, and Cochrane Library for studies published up to May 10, 2025, using the terms “GLP-1 Receptor Agonists” and “weight loss”. We identified 33 studies for the efficacy of combination of lifestyle modification and GLP-1RAs on weight loss. Most of the included studies were of high methodological quality.

Added value of this study

This study is the first meta-analysis to provide up-to-date evidence evaluating the effects of lifestyle modification combined with GLP-1RAs on weight reduction and cardiometabolic biomarkers. By pooling data from 33 studies, we demonstrated that lifestyle modification combined with GLP-1RAs result in a significant mean weight loss compared with control groups. Significant improvements were also observed in other body composition parameters and cardiometabolic biomarkers. Treatment duration, drug type, dosage, and geographic region were key influencing factors of intervention effectiveness.

Implications of all the available evidence

Our study confirms that combining lifestyle interventions with GLP-1RAs is a feasible and effective strategy for obesity treatment. However, the observed concurrent decrease in lean mass highlights a potential unintended consequence that may compromise overall health benefits. Future clinical approaches should consider incorporating higher-intensity physical activity or resistance training to optimize combined treatment protocols that maximize fat loss while minimizing lean mass reduction, ultimately improving cardiometabolic biomarkers and patient well-being.

Introduction

Overweight and obesity have become one of the most critical public health issues. Recently, obesity has been redefined as a chronic systemic disease characterized by functional impairment of tissues, organs, or the entire organism due to excessive fat accumulation.¹ This updated definition aims to inform clinical priorities and guide the development of more effective public health strategies to tackle this complex, multifactorial condition. According to the 2021 Global Burden of Disease (GBD) study, overweight and obesity affect approximately 2.11 billion people worldwide,² accounting for 3.71 million deaths and 129 million disability-adjusted life years (DALYs) lost.³ A recent global forecasting study suggested that the number of adults with overweight/obesity is expected to surge to 3.8 billion by 2050.² The global rise in obesity is expected to contribute substantially to the growing burden of chronic non-communicable diseases, including diabetes and cardiovascular diseases.^4,5 Evidence suggests that, under the continued trajectory of the obesity epidemic, the number of individuals with diabetes worldwide is projected to surpass 1.31 billion by 2050,⁴ while the prevalence of cardiovascular events could double in certain countries.⁵ These projections highlight the urgent need for effective, evidence-based interventions to curb the obesity epidemic and mitigate the foreseeable increase in associated disease burden.

The primary goal of obesity treatment is to reduce excess body fat accumulation, thereby lowering the risk of obesity-related complications. Lifestyle intervention remains the first-line therapy for obesity management, which includes caloric restriction, increased physical activity, and behavioral counseling.^6,7 Clinical guidelines recommend that lifestyle modifications should be maintained throughout the entire course of obesity treatment, supplemented with pharmacotherapy when necessary.^8,9 In recent years, glucagon-like peptide-1 receptor agonists (GLP-1RAs), initially developed as antidiabetic agents, have demonstrated potential for weight reduction. Their mechanisms include glucose-dependent stimulation of insulin secretion, suppression of glucagon release during hyperglycemia, delayed gastric emptying, prevention of postprandial glucose spikes, and reduced caloric intake, collectively contributing to weight loss.10, 11, 12, 13 Multiple meta-analyses have confirmed that GLP-1RAs effectively reduce body weight, waist circumference, and cardiometabolic markers among individuals with diabetes.^14,15 However, whether combining GLP-1RAs with lifestyle interventions can yield superior outcomes in obesity management and further improve cardiometabolic markers remains to be fully elucidated.

To address the above research gaps, we conducted this systematic review and meta-analysis to evaluate the efficacy of lifestyle modifications combined with GLP-1RAs in improving weight loss and cardiometabolic markers in individuals with overweight or obesity, utilizing evidence from the most recent trials to date.

Methods

This systematic review was conducted in accordance with the guidelines outlined in the Cochrane Handbook for Systematic Reviews of Interventions and was reported following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement.¹⁶ The protocol was registered at the International Prospective Register of Systematic Reviews (PROSPERO), under identification number CRD42024600250.

Search strategy

We searched MEDLINE (via PubMed), the Cochrane Central Register of Controlled Trials, and the Embase database from their inception until May 10, 2025 using the following terms: (“Glucagon-Like Peptide-1 Receptor Agonists”, “GLP-1 Receptor Agonists”, “dual receptor agonists”, “triple receptor agonists”, “semaglutide”, “liraglutide”) and (“Obesity”, “overweight”, “Weight Loss”, “Weight Reduction”). There were no language restrictions. The detailed search strategy is shown in Supplementary Table S1.

Inclusion and exclusion criteria

Inclusion criteria were as follows: (1) overweight or obese individuals, with or without weight-related comorbidities such as hypertension, dyslipidemia, type 2 diabetes, cardiovascular disease, or obstructive sleep apnea. Obesity was defined as body mass index (BMI) ≥30 kg/m² and overweight as BMI 25–29.9 kg/m²¹⁷; (2) the study was a randomized controlled trial; (3) the intervention group received lifestyle modification combined with GLP-1RAs, while controls received lifestyle modification with placebo; (4) the lifestyle intervention involved counseling on diet or physical activity; (5) body weight measurements were recorded for both treatment and control groups throughout the study period; (6) both dual and triple agonists were considered for inclusion in this study.

Exclusion criteria were as follows: (1) reviews, abstracts, or reports; (2) studies not designed as randomized controlled trials (RCTs); (3) animal-based research; (4) articles with incomplete data; (5) duplicate publications; (6) trials involving additional treatments such as dipeptidyl peptidase IV inhibitors, insulin, or other weight-influencing medications.

Outcomes

The primary outcome is the change in body weight (kg). The secondary outcomes include the change in percentage of body weight, waist circumference, lean mass, fat mass, percentage of body fat, systolic blood pressure (SBP), diastolic blood pressure (DBP), glycated hemoglobin (HbA_1c), fasting blood glucose (FBG), total cholesterol (TC), triglycerides (TG), high-density lipoprotein cholesterol (HDL-C) and low-density lipoprotein cholesterol (LDL-C).

Study selection and data extraction

Data for this meta-analysis were extracted independently by 2 investigators (JC and HZ) using a standard protocol. Any disagreements or unresolved differences were resolved by discussion with a third reviewer (ZZ), with reference to the original study protocol or published article to ensure accurate data extraction. Extracted information included the first author’s name, publication year, sample size for each intervention group, study location, follow-up duration, intervention dosage, lifestyle modification, concomitant diseases, baseline information (age, sex, baseline weight, BMI, SBP) and outcomes. We measured the changes from baseline in all outcomes. Medication doses were standardized to mg/w when necessary. For studies lacking mean value and standard deviations (SDs), we used sample size, median, interquartile range, standard error and 95% confidence interval based on Cochrane Handbook guidelines (https://training.cochrane.org/handbook/current/chapter-06) to calculate the missing values.¹⁸

Quality assessment

Risk of bias was independently assessed by two reviewers (JC and HZ) using the Cochrane risk-of-bias tool, which evaluates domains such as randomization, deviation, missing outcome, measurement, reporting, and overall biases. Risk-of-bias graphs and summaries were performed using the Cochrane Risk of Bias tool (version 2.0) for the included RCTs. The quality and strength of the evidence were evaluated with the Grading of Recommendations Assessment, Development and Evaluation (GRADE) tool. Evidence was classified as high quality, moderate quality, low quality, or very low quality.

Statistical analysis

Categorical variables were shown as frequencies (%), while continuous variables were reported as means with SDs. Mean difference (MD) and 95% confidence intervals (CIs) were calculated to estimate differences between the intervention and control groups. Subgroup analyses were conducted based on participant characteristics and intervention situations. Treatment duration was categorized as <52 weeks and ≥52 weeks; category of GLP 1-RAs was categorized as semaglutide, liraglutide, exenatide, tirzepatide and others, comorbidities were categorized as metabolic syndrome, diabetes, other comorbidities and healthy participants; pharmacological targets was categorized as GLP-1 receptor agonists alone and dual-/triple-receptor agonists; dosing frequency was categorized as daily and weekly; regions were categorized as Europe, North America, Asia and mixed; ethnicity was categorized based on the proportion of white participants (<80% vs. ≥80%). Meta-regression analysis was conducted using the similar variables to further explore potential influencing factors.

Data were pooled across all trials according to the random-effect model based on the DerSimonian-Laird method. The Higgins & Thompson I² statistic and prediction intervals were employed to assess heterogeneity. Sensitivity analyses were conducted by sequentially excluding each individual study (leave-one-out) to assess the robustness of the pooled MD. Additional analyses were performed by excluding studies with moderate to high risk of bias and those lacking clearly defined lifestyle interventions. Publication bias was examined using Begg’s funnel plot and Egger’s test, with a P-value of <0.05 and an asymmetrical funnel plot suggesting potential bias. We further performed a trim-and-fill analysis, which adjusted for potential publication bias by removing and imputing studies to achieve funnel plot symmetry, and recalculating the pooled effect size. All analyses were conducted using the “meta” package and “metafor” package in R software (version 4.4.1).

Ethics statement

Ethics committee approval was not required. All data analyzed in this study were extracted from previously published RCTs, and no new data involving human participants were collected. Each included RCT had received ethical approval from its respective institutional review board and obtained informed consent from all participants, as reported in the original publications.

Role of the funding source

This work was supported by the Startup Fund for Young Faculty at Shanghai Jiao Tong University (Grant No. KJ3-0214-24-0011), awarded to ZZ. As the corresponding author and fund recipient, ZZ was responsible for the study design, data interpretation, manuscript writing, and the decision to submit the manuscript for publication.

Results

Search results

A total of 2783 articles were identified in the initial search. After screening titles and abstracts, 2592 articles were excluded for the following reasons: not being RCTs, lacking body weight outcomes, or being protocols/registries/abstracts only. 158 articles were further excluded after the full-text review because of duplicate reporting, absence of lifestyle interventions and complete outcome data, concomitant use of GLP-1RAs with other hypoglycemic agents or medications in the treatment, being conference papers and retracted papers. The excluded studies are shown in the Supplementary Table S2. Finally, 33 eligible articles met our inclusion criteria and were included in the final analysis.19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30^,31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51

Study and participants characteristics

Table 1 summarizes the study designs of the 33 included RCTs, comprising a total of 12,028 participants with overweight or obesity. The intervention arms across all studies utilized GLP-1RAs, with the following specific agents and distribution: semaglutide (n = 7), liraglutide (n = 16), exenatide (n = 3), dulaglutide (n = 1), tirzepatide (n = 1), retatrutide (n = 1), lotiglipron (n = 1), survodutide (n = 1), orforglipron (n = 1), and noiglutide (n = 1). All participants across the included trials received lifestyle interventions of varying intensity, primarily comprising calorie restriction and regular physical activity, although 8 studies did not specify the intervention details. Among the 33 included studies, 23 were conducted in populations with metabolic syndrome, 3 in patients with diabetes, 1 in individuals with prediabetes, 2 in patients with obstructive sleep apnea, 1 in patients with craniopharyngioma, 1 in individuals with polycystic ovary syndrome, 1 in post-metabolic surgery patients, and 1 in a healthy population.

Table 1.

Summary of study characteristics in each of the randomized controlled trials included in this meta-analysis.

Trial name	Countries or Regions	Participants and comorbidity	Types and Dosages of GLP-1 receptor agonists	Control group	Duration (weeks)	Lifestyle modification	Risk of bias
Amin 202519	Canada, Japan and the USA	Metabolic Syndrome	Intervention Group1: lotiglipron 80 mg/w; Intervention Group 2: lotiglipron 140 mg/w; Intervention Group 3: lotiglipron 200 mg/w (five-step titration); Intervention Group4: lotiglipron 200 mg/w (four-step titration); Intervention Group 5: lotiglipron 260 mg/w.	Placebo	20	All participants received general dietary and exercise guidance; Detailed information about the intervention was not provided in the article or protocol.	Medium risk
Malhotra 202420	60 sites across nine countries	Obstructive sleep apnea	Tirzepatide 15 mg/w	Placebo	52	500 kcal/day energy deficit; At least 150 min per week of moderate intensity physical activity	Low risk
Le Roux 202421	43 sites in 12 countries (USA, Australia, Belgium, Canada, China, Germany, South Korea, Netherlands, New Zealand, Poland, Sweden, and UK)	Metabolic Syndrome	Intervention Group 1: survodutide 0.6 mg/w; Intervention Group 2: survodutide 2.4 mg/w; Intervention Group 3: survodutide 3.6 mg/w; Intervention Group 4: survodutide 4.8 mg/w.	Placebo	46	500 kcal/day energy deficit; 150–300 min per week of physical activity	Low risk
McGowan 202422	30 trial sites in Canada, Denmark, Finland, Spain, and the UK	Prediabetes	Intervention Group 5: semaglutide 2.4 mg/w	Placebo	52	All participants received general dietary and exercise guidance; Detailed information about the intervention was not provided in the article or protocol.	Low risk
Li 202423	China	Metabolic Syndrome	Intervention Group1: noiiglutide 0.12 mg/d; Intervention Group 2: noiiglutide 0.24 mg/d; Intervention Group 3: noiiglutide 0.36 mg/d.	Placebo	24	500 kcal/day energy deficit, setting a floor at 1200 kcal/day; Moderate-intensity physical activity	Medium risk
Gatta 202424	France	Craniopharyngioma-related obesity	Exenatide 20 μg/d	Placebo	26	600 kcal/day energy deficit; 30 min per day of moderate intensity physical activity	Low risk
McElroy 202425	USA	Type 2 diabetes	Liraglutide 3 mg/d	Placebo	40	Adherence to U.S. government guidelines for healthy diet and physical activity. The protocol did not prescribe a standardized energy deficit; At least 150–300 min of moderate-intensity aerobic activity	Medium risk
Jastreboff 202326	USA	Metabolic Syndrome	Intervention Group 1: retatrutide 1 mg/w; Intervention Group 2: retatrutide 4 mg/w [initial dose, 2 mg]; Intervention Group 3: retatrutide 4 mg/w [initial dose, 4 mg]; Intervention Group 4: retatrutide 8 mg/w [initial dose, 2 mg]; Intervention Group 5: retatrutide 8 mg/w [initial dose, 4 mg]; Intervention Group 6: retatrutide 12 mg/w [initial dose, 2 mg].	Placebo	48	Adherence to U.S. government guidelines for healthy diet and physical activity. The protocol did not prescribe a standardized energy deficit; At least 150 min per week of moderate intensity physical activity or 75–150 min per week of vigorous-intensity aerobic physical activity	Medium risk
Zhang 202327	China	Polycystic ovary syndrome	Dulaglutide 1.5 mg/w	Placebo	24	30 min of moderate-intensity aerobic exercise per day, 5–7 days a week, and 10–20 min of resistance exercise, 3 times a week	Low risk
Coppin 202328	Switzerland	Metabolic Syndrome	Liraglutide 3 mg/d	Placebo	16	All participants received general dietary and exercise guidance; Detailed information about the intervention was not provided in the article or protocol.	Low risk
Wharton 202329	Canada, USA, and Hungary	Metabolic Syndrome	Intervention Group 1: orforglipron 12 mg/d; Intervention Group 2: orforglipron 24 mg/d; Intervention Group 3: orforglipron 36 mg/d; Intervention Group 4: orforglipron 45 mg/d.	Placebo	36	Education regarding healthy eating and exercise was provided by trial personnel to all participants. Detailed information about the intervention was not provided in the article or protocol.	Medium risk
Mok 202330	UK	Post-Metabolic Surgery Patients	Liraglutide 3 mg/d	Placebo	24	500 kcal/day energy deficit; 150 min per week of moderate to vigorous exercise	Low risk
Knop 202331	East Asia, Europe, and North America	Metabolic Syndrome	Semaglutide 50 mg/d	Placebo	68	500 kcal/day energy deficit; 150 min per week of physical activity	Low risk
Rubino 202232	USA	Metabolic Syndrome	Intervention Group 1: semaglutide 2.4 mg/w; Intervention Group 2: liraglutide 3.0 mg/d.	Placebo	68	500 kcal/day energy deficit; At least 150 min per week of physical activity	Low risk
Garvey 202233	USA, Europe, the United Kingdom, and Canada	Metabolic Syndrome	Semaglutide 2.4 mg/w	Placebo	104	500 kcal/day energy deficit; At least 150 min per week of physical activity	Low risk
Fox 202234	USA	Metabolic Syndrome	Exenatide 2 mg/w	Placebo	52	1200–1500 kcal/day energy; 200 min per week of moderate-vigorous intensity activity	Low risk
Rosenstock 202135	52 medical research centres and hospitals in India, Japan, Mexico, and the USA	Type 2 diabetes	Intervention Group 1: tirzepatide 5 mg/w; Intervention Group 2: tirzepatide 10 mg/w; Intervention Group 3: tirzepatide 15 mg/w.	Placebo	40	All participants received general dietary and exercise guidance; Detailed information about the intervention was not provided in the article or protocol.	Low risk
Ludgren 202136	Denmark	Metabolic Syndrome	Intervention Group: liraglutide 3 mg/d	Placebo	8	800 kcal/day energy deficit; 150 min per week of moderate-intensity physical activity	Low risk
Wilding 202137	Asia, Europe, North America, and South America	Metabolic Syndrome	Semaglutide 2.4 mg/w	Placebo	68	500 kcal/day energy deficit; 150 min per week of physical activity	Low risk
Rubino 202138	10 countries	Metabolic Syndrome	Semaglutide 2.4 mg/w	Placebo	68	500 kcal/day energy deficit; 150 min per week of physical activity	Low risk
Neeland 202139	USA	Metabolic Syndrome	Intervention Group 1: liraglutide 0.6 mg/d; Intervention Group 2: liraglutide 1.2 mg/d; Intervention Group 3: liraglutide 1.8 mg/d; Intervention Group 4: liraglutide 2.4 mg/d; Intervention Group 5: liraglutide 3.0 mg/d.	Placebo	40	500 kcal/day energy deficit; At least 150 min per week of moderate-intensity physical activity	Low risk
Lau 202140	10 countries (Canada, Denmark, Finland, Ireland, Japan, Poland, Serbia, South Africa, the UK, and the USA)	Metabolic Syndrome	Liraglutide 3 mg/d	Placebo	26	500 kcal/day energy deficit; 150 min per week of physical activity	Low risk
Grannel 202141	Ireland	Metabolic Syndrome	Liraglutide 3 mg/d	Placebo	16	High-protein and low-glycemic-index food intake (males 1500 kcal, females 1200 kcal; 1 g protein per kilogram of body weight); 150 min per week of physical activity	Medium risk
Wadden 202142	41 sites in the USA	Metabolic Syndrome	Semaglutide 2.4 mg/w	Placebo	68	Hypocaloric diet (1200–1800 kcal/d) of conventional food; 100 min of physical activity per week, which increased by 25 min every 4 weeks.	Low risk
Kelly 202043	Belgium, Mexico, Russia, Sweden, and the USA	Type 2 diabetes	Liraglutide 3 mg/d	Placebo	56	All participants received general dietary and exercise guidance; Detailed information about the intervention was not provided in the article or protocol.	Low risk
Chao 201944	USA	Metabolic Syndrome	Liraglutide 3 mg/d	Placebo	52	All participants received 21 sessions of Intensive Behavioral Therapy. Detailed information about the intervention was not provided in the article or protocol.	Medium risk
Peradze 201945	USA	Metabolic Syndrome	Liraglutide 3 mg/d	Placebo	5	500 kcal/day energy deficit; 150 min per week of physical activity	Medium risk
O’Neil 201846	8 countries (Australia, Belgium, Canada, Germany, Israel, Russia, the UK, and the USA)	Metabolic Syndrome	Intervention Group 1: semaglutide 0.05 mg/d; Intervention Group 2: semaglutide 0.1 mg/d; Intervention Group 3: semaglutide 0.2 mg/d; Intervention Group 4: semaglutide 0.3 mg/d; Intervention Group 5: semaglutide 0.4 mg/d; Intervention Group 6: liraglutide 3.0 mg/d.	Placebo	52	500 kcal/day energy deficit; At least 150 min per week of physical activity	Low risk
Le Roux 201747	191 clinical research sites in 27 countries in Europe, North America, South America, Asia, Africa, and Australia	Metabolic Syndrome	Liraglutide 3 mg/d	Placebo	160	500 kcal/day energy deficit; At least 150 min per week of physical activity	Medium risk
Blackman 201648	40 sites in the USA and Canada	Obstructive sleep apnea	Liraglutide 3 mg/d	Placebo	32	500 kcal/day energy deficit; At least 150 min per week of physical activity	Medium risk
Lepsen201549	Denmark	Healthy participants	Liraglutide 1.2 mg/d	Placebo	8	600 kcal/day energy deficit	Medium risk
Rosenstock 201050	The USA and Puerto Rico	Metabolic Syndrome	Exenatide 10 μg/d	Placebo	24	All participants received general dietary and exercise guidance; Detailed information about the intervention was not provided in the article or protocol.	Medium risk
Astrup 201051	19 clinical research sites in eight European countries	Metabolic Syndrome	Intervention Group 1: liraglutide 1.2 mg/d; Intervention Group 2: liraglutide 1.8 mg/d; Intervention Group 3: liraglutide 2.4 mg/d; Intervention Group 4: liraglutide 3.0 mg/d.	Placebo	20	500 kcal/day energy deficit; Encouraged to maintain or increase physical activity.	Low risk

The demographic and clinical characteristics of 33 RCTs are also detailed in Table 2. The median sample size was 206 and ranged from 40 to 2210. The range of intervention durations was 5–160 weeks, with 15 trials lasting over 52 weeks. 13 studies were from multi-centers on different continents, 7 studies were from Europe and 11 studies were from North America, only 2 studies were from China (Fig. 1).

Table 2.

Baseline demographic and clinical characteristics of study participants in each of the randomized controlled trials included in this meta-analysis.

Trial name	Intervention	Number of Participants	Female (%)	White participants (%)	Age (years)	BMI (kg/m2)	SBP (mmHg)	Weight (kg)
Amin 2025	Intervention Group 1: lotiglipron 80 mg/w	66	38 (57.6)	56 (84.8)	49.3 (13.0)	37.1 (5.7)	–	105.7 (22.3)
	Intervention Group 2: lotiglipron 140 mg/w	64	45 (70.3)	44 (68.8)	48.1 (12.4)	38.4 (7.1)	–	106.7 (25.3)
	Intervention Group 3: lotiglipron 200 mg/w (five-step titration)	65	36 (55.4)	56 (86.2)	47.6 (13.5)	37.9 (5.7)	–	108.2 (21.7)
	Intervention Group 4: lotiglipron 200 mg/w (four-step titration)	66	42 (63.6)	55 (83.3)	50.0 (10.7)	37.0 (4.5)	–	104.1 (15.9)
	Intervention Group 5: lotiglipron 260 mg/w	64	40 (62.5)	52 (81.3)	48.5 (11.7)	38.3 (7.3)	–	109.3 (24.9)
	Control Group	64	36 (56.3)	55 (85.9)	50.9 (13.3)	33.7 (6.9)	–	109.7 (22.1)
Malhotra 2024	Intervention Group: tirzepatide 15 mg/w	234	69 (29.5)	159 (67.9)	47.9 (11.5)	39.1 (7.0)	129.4 (11.5)	114.7 (23.7)
	Control Group	235	73 (31.1)	166 (70.9)	51.7 (11.0)	38.7 (6.0)	130.5 (13.5)	115.5 (22.0)
Le Roux 2024	Intervention Group 1: survodutide 0.6 mg/w	77	51 (66.0)	59 (77.0)	48.6 (12.6)	37.8 (6.3)	125.0 (13.4)	107.8 (18.7)
	Intervention Group 2: survodutide 2.4 mg/w	78	54 (69.0)	60 (77.0)	49.0 (13.1)	37.6 (6.3)	125.4 (13.3)	106.6 (23.0)
	Intervention Group 3: survodutide 3.6 mg/w	76	51 (63.0)	63 (83.0)	50.3 (11.8)	37.0 (5.7)	127.4 (13.2)	104.7 (19.6)
	Intervention Group 4: survodutide 4.8 mg/w	76	53 (70.0)	59 (78.0)	47.6 (13.5)	37.6 (6.0)	122.6 (12.3)	105.9 (17.4)
	Control Group	77	53 (69.0)	60 (78.0)	50.0 (13.5)	35.8 (5.0)	127.5 (14.2)	104.3 (23.0)
McGowan 2024	Intervention Group: semaglutide 2.4 mg/w	138	100 (72.5)	124 (90.0)	53.0 (11.0)	39.9 (6.6)	131.0 (15.0)	111.9 (21.5)
	Control Group	69	47 (68.1)	59 (86.0)	53.0 (11.0)	40.4 (7.6)	129.0 (15.0)	111.0 (23.5)
Li 2024	Intervention Group 1: noiiglutide 0.12 mg/d	64	35 (54.7)	0 (0.0)	35.3 (8.4)	31.7 (2.7)	120.8 (11.4)	87.3 (11.6)
	Intervention Group 2: noiiglutide 0.24 mg/d	65	35 (53.8)	0 (0.0)	34.4 (8.7)	32.3 (3.2)	121.1 (10.6)	88.7 (12.2)
	Intervention Group 3: noiiglutide 0.36 mg/d	63	35 (55.6)	0 (0.0)	36.3 (9.6)	32.5 (3.6)	122.7 (11.3)	90.3 (15.1)
	Control Group	62	33 (53.2)	0 (0.0)	35.7 (7.7)	32.4 (3.4)	120.4 (11.0)	90.5 (14.9)
Gatta 2024	Intervention Group: exenatide 20 μg/d	20	8 (40.0)	–	40.4 (14.8)	37.3 (7.5)	132.0 (12.0)	109.1 (19.9)
	Control Group	20	9 (45.0)	–	40.4 (12.4)	37.4 (7.3)	132.0 (16.0)	106.1 (17.4)
McElroy 2024	Intervention Group: liraglutide 3 mg/d	29	35 (50.0)	27 (93.0)	41.4 (12.3)	39.0 (7.4)	119.7 (9.9)	110.6 (24.0)
	Control Group	31	35 (49.3)	28 (90.0)	43.9 (11.0)	37.2 (4.5)	120.8 (11.6)	101.9 (14.8)
Jastreboff 2023	Intervention Group 1: retatrutide 1 mg/w	69	33 (48.2)	61 (88.0)	50.6 (13.3)	37.5 (5.9)	–	106.4 (19.8)
	Intervention Group 2: retatrutide 4 mg/w [initial dose, 2 mg]	33	16 (48.2)	29 (88.0)	50.8 (11.9)	37.3 (5.9)	–	108.0 (26.3)
	Intervention Group 3: retatrutide 4 mg/w [initial dose, 4 mg]	34	16 (47.2)	30 (88.0)	46.8 (14.1)	37.4 (4.7)	–	107.0 (21.3)
	Intervention Group 4: retatrutide 8 mg/w [initial dose, 2 mg]	35	17 (49.2)	30 (86.0)	46.1 (13.5)	37.4 (6.0)	–	106.5 (21.6)
	Intervention Group 5: retatrutide 8 mg/w [initial dose, 4 mg]	35	17 (49.2)	33 (94.0)	48.7 (11.1)	37.0 (5.5)	–	108.6 (20.9)
	Intervention Group 6: retatrutide 12 mg/w [initial dose, 2 mg]	62	30 (48.2)	56 (90.0)	45.8 (12.2)	37.4 (6.0)	–	108.0 (21.7)
	Control Group	70	34 (49.2)	59 (84.0)	48.0 (12.5)	37.3 (5.9)	–	109.2 (20.9)
Zhang 2023	Intervention Group: dulaglutide 1.5 mg/w	35	27 (77.1)	0 (0.0)	30.3 (5.1)	29.7 (3.6)	124.1 (12.7)	77.6 (10.0)
	Control Group	33	23 (69.7)	0 (0.0)	28.6 (4.2)	29.7 (3.7)	127.7 (18.3)	78.8 (13.0)
Coppin 2023	Intervention Group: liraglutide 3 mg/d	20	6 (30.0)	–	37.4 (11.2)	35.9 (3.0)	–	102.3 (17.0)
	Control Group	23	8 (34.5)	–	40.0 (14.1)	34.9 (2.9)	–	101.7 (9.8)
Wharton 2023	Intervention Group 1: orforglipron 12 mg/d	50	31 (62.0)	47 (94.0)	49.8 (10.5)	37.7 (7.7)	129.4 (12.1)	107.5 (25.3)
	Intervention Group 2: orforglipron 24 mg/d	53	30 (56.6)	46 (87.0)	57.0 (9.1)	38.1 (7.7)	129.7 (10.8)	112.1 (30.2)
	Intervention Group 3: orforglipron 36 mg/d	58	36 (62.1)	50 (86.0)	55.9 (11.3)	38.0 (6.3)	131.4 (11.9)	108.3 (25.5)
	Intervention Group 4: orforglipron 45 mg/d	61	35 (57.4)	59 (96.0)	53.7 (11.9)	37.7 (6.6)	127.7 (11.3)	108.0 (24.5)
	Control Group	50	29 (58.0)	45 (90.0)	54.0 (8.8)	37.8 (6.5)	128.5 (9.5)	107.6 (25.2)
Mok 2023	Intervention Group: liraglutide 3 mg/d	35	26 (74.3)	22 (63.0)	46.7 (10.8)	41.6 (6.9)	131.3 (15.0)	116.1 (23.6)
	Control Group	35	26 (74.3)	22 (63.0)	48.4 (10.6)	44.6 (8.3)	131.3 (14.5)	123.5 (24.8)
Knop 2023	Intervention Group: oral semaglutide 50 mg/d	334	247 (74.0)	246 (74.0)	49.0 (13.0)	37.3 (6.3)	129.0 (16.0)	104.5 (22.0)
	Control Group	333	238 (71.5)	248 (74.0)	50.0 (12.0)	37.7 (6.8)	130.0 (15.0)	106.2 (22.3)
Rubino 2022	Intervention Group 1: semaglutide 2.4 mg/w	126	102 (81.0)	94 (74.6)	48.0 (14.0)	37.0 (7.4)	125.0 (14.0)	102.5 (25.3)
	Intervention Group 2: liraglutide 3.0 mg/d	127	97 (76.4)	95 (74.8)	49.0 (13.0)	37.2 (6.4)	126.0 (16.0)	103.7 (22.5)
	Control Group	85	66 (77.6)	60 (70.6)	51.0 (12.0)	38.8 (6.5)	123.0 (14.0)	108.8 (23.1)
Garvey 2022	Intervention Group: semaglutide 2.4 mg/w	152	123 (80.9)	141 (92.8)	47.3 (11.7)	38.6 (6.7)	126.0 (14.0)	105.6 (20.8)
	Control Group	152	113 (74.3)	142 (93.4)	47.4 (10.3)	38.5 (7.2)	125.0 (15.0)	106.5 (23.1)
Fox 2022	Intervention Group: exenatide 2 mg/w	33	18 (54.5)	26 (79.0)	15.9 (1.6)	36.5 (4.3)	117.0 (10.0)	105.6 (17.7)
	Control Group	33	13 (39.4)	28 (85.0)	16.1 (1.5)	37.3 (4.6)	117.0 (8.0)	111.4 (17.2)
Rosenstock 2021	Intervention Group 1: tirzepatide 5 mg/w	121	65 (54.0)	38 (31.0)	54.1 (11.9)	32.2 (7.0)	128.2 (15.7)	87.0 (21.2)
	Intervention Group 2: tirzepatide 10 mg/w	121	49 (40.0)	43 (36.0)	55.8 (10.4)	32.2 (7.6)	127.8 (12.6)	86.2 (19.5)
	Intervention Group 3: tirzepatide 15 mg/w	121	58 (48.0)	43 (36.0)	52.9 (12.3)	31.5 (5.5)	126.8 (13.8)	85.4 (18.5)
	Control Group	115	59 (51.0)	46 (40.0)	53.6 (12.8)	31.7 (6.1)	127.8 (14.1)	84.8 (20.0)
Ludgren 2021	Intervention Group: liraglutide 3 mg/d plus exercise	49	31 (65.0)	–	42.0 (12.0)	32.8 (2.4)	122.0 (13.0)	98.3 (11.5)
	Control Group	48	31 (63.0)	–	43.0 (12.0)	32.7 (3.0)	122.0 (14.0)	96.8 (13.2)
Wilding 2021	Intervention Group: semaglutide 2.4 mg/w	1306	955 (73.1)	973 (74.5)	46.0 (13.0)	37.8 (6.7)	126.0 (14.0)	105.4 (22.1)
	Control Group	655	498 (76.0)	499 (76.2)	47.0 (12.0)	38.0 (6.5)	127.0 (14.0)	105.2 (21.5)
Rubino 2021	Intervention Group: semaglutide 2.4 mg/w	535	429 (80.2)	446 (83.4)	47.0 (12.0)	34.5 (6.9)	121.0 (13.0)	96.5 (22.5)
	Control Group	268	205 (76.5)	226 (84.3)	46.0 (12.0)	34.1 (6.7)	121.0 (13.0)	95.4 (22.7)
Neeland 2021	Intervention Group: liraglutide 3.0 mg/d	73	67 (92.0)	43 (59.0)	49.6 (9.8)	37.2 (6.0)	130.3 (14.9)	101.0 (17.9)
	Control Group	53	51 (93.0)	35 (64.0)	50.9 (8.8)	38.1 (6.1)	125.8 (13.9)	102.3 (17.9)
Lau 2021	Intervention Group: liraglutide 3.0 mg/d	99	65 (65.7)	82 (82·8)	51.5 (9.3)	38.4 (7.4)	131.8 (13.7)	107.8 (24.1)
	Control Group	101	59 (58.4)	71 (70·3)	51.4 (11.9)	37.4 (5.7)	131.9 (13.2)	106.2 (21.6)
Grannel 2021	Intervention Group: liraglutide 3.0 mg/d	59	29 (49.2)	–	53.7 (9.2)	43.0 (6.2)	–	123.2 (23.3)
	Control Group	19	9 (47.4)	–	56.7 (12.7)	43.0 (5.7)	–	119.5 (24.1)
Wadden 2021	Intervention Group: semaglutide 2.4 mg/w	407	315 (77.4)	307 (75.4)	46.0 (13.0)	38.1 (6.7)	124.0 (15.0)	106.9 (22.8)
	Control Group	204	180 (88.2)	158 (77.5)	46.0 (13.0)	37.8 (6.9)	124.0 (15.0)	103.8 (22.9)
Kelly 2020	Intervention Group: liraglutide 3.0 mg/d	125	71 (56.8)	105 (84.0)	14.6 (1.6)	35.3 (5.1)	116.0 (10.0)	99.3 (19.7)
	Control Group	126	78 (61.9)	115 (91.3)	14.5 (1.6)	35.8 (5.7)	117.0 (12.0)	102.2 (21.6)
Chao 2019	Intervention Group: liraglutide 3.0 mg/d	50	39 (78.0)	27 (54.0)	49.5 (11.0)	38.0 (4.3)	–	105.8 (14.7)
	Control Group	50	42 (84.0)	27 (54.0)	45.2 (12.3)	38.5 (5.4)	–	107.8 (17.9)
Peradze 2019	Intervention Group: liraglutide 3.0 mg/d	20	–	–	–	35.1 (1.3)	133.8 (2.6)	104.5 (4.3)
	Control Group	20	–	–	–	36.4 (1.6)	133.3 (2.8)	101.3 (3.6)
O’Neil 2018	Intervention Group 1: semaglutide 0.05 mg/d	103	36 (35.0)	88 (85.0)	47.0 (13.0)	39.1 (6.5)	–	111.3 (23.2)
	Intervention Group 2: semaglutide 0.1 mg/d	102	36 (35.3)	76 (755.0)	45.0 (13.0)	39.6 (7.4)	–	111.3 (21.5)
	Intervention Group 3: semaglutide 0.2 mg/d	103	37 (35.9)	72 (70.0)	44.0 (11.0)	40.1 (6.9)	–	114.5 (24.5)
	Intervention Group 4: semaglutide 0.3 mg/d	103	37 (35.9)	74 (72.0)	47.0 (12.0)	39.6 (7.1)	–	111.5 (23.0)
	Intervention Group 5: semaglutide 0.4 mg/d	102	36 (35.3)	71 (70.0)	48.0 (13.0)	39.9 (8.8)	–	113.2 (26.4)
	Intervention Group 6: liraglutide 3.0 mg/d	103	36 (35.0)	78 (76.0)	49.0 (11.0)	38.6 (6.6)	–	108.7 (21.9)
	Control Group	136	48 (35.3)	97 (71.0)	46.0 (13.0)	40.1 (7.2)	–	114.2 (25.4)
Le Roux 2017	Intervention Group: liraglutide 3 mg/d	1505	1141 (75.8)	1256 (83·5)	47.5 (11.7)	38.8 (6.4)	124.7 (12.9)	107.5 (21.6)
	Control Group	749	573 (76.5)	628 (83·8)	47.3 (11.8)	39.0 (6.3)	125.0 (12.8)	107.9 (21.8)
Blackman 2016	Intervention Group: liraglutide 3 mg/d	180	51 (28.3)	130 (72.2)	48.6 (9.9)	38.9 (6.4)	–	–
	Control Group	179	50 (27.9)	135 (75.4)	48.4 (9.5)	39.4 (7.4)	–	–
Lepsen2015	Intervention Group: liraglutide 1.2 mg/d	27	22 (81.5)	27 (100.0)	46.0 (2.0)	35.0 (0.6)	124.0 (3.0)	98.3 (1.9)
	Control Group	25	22 (88.0)	25 (100.0)	45.0 (2.0)	34.0 (0.5)	124.0 (4.0)	96.9 (1.9)
Rosenstock 2010	Intervention Group: exenatide 10 μg/d	73	–	–	–	–	–	109.5 (2.7)
	Control Group	79	–	–	–	–	–	107.6 (2.6)
Astrup 2010	Intervention Group 1: liraglutide 1.2 mg/d	95	73 (76.8)	–	47.2 (9.7)	34.8 (2.6)	127.0 (13.1)	96.2 (13.5)
	Intervention Group 2: liraglutide 1.8 mg/d	90	68 (75.6)	–	45.5 (10.9)	35.0 (2.6)	123.0 (13.0)	98.0 (12.5)
	Intervention Group 3: liraglutide 2.4 mg/d	93	71 (76.3)	–	45.0 (11.1)	35.0 (2.8)	126.0 (13.9)	98.4 (13.0)
	Intervention Group 4: liraglutide 3.0 mg/d	93	70 (75.3)	–	45.9 (10.7)	34.8 (2.8)	124.0 (11.3)	97.6 (13.7)
	Control Group	98	75 (76.5)	–	45.9 (10.3)	34.9 (2.8)	124.0 (11.1)	97.3 (13.2)

Fig. 1.

Flow diagram of the selection process of studies.

Efficacy of lifestyle modification combined with GLP-1RAs

Random-effects models were employed to synthesize the effect of GLP-1RAs combined with lifestyle modification on weight loss among the 33 eligible trials. Compared with the control group, lifestyle modification combined with GLP-1RAs resulted in a 7.13 kg weight reduction (MD: −7.13, 95%CI: −9.02, −5.24, P < 0.001, Fig. 2).

Fig. 2.

Forest plot of the effect of lifestyle modifications combined with GLP-1 receptor agonists and lifestyle modifications combined with placebo on weight loss.

Table 3 presented the results of secondary endpoints. Compared with the control group, the combination of lifestyle intervention and GLP-1RAs showed significant reductions across body composition parameters and cardiometabolic parameters, includingpercentage of body weight (MD: −7.02, 95% CI: −8.85, −5.18, P < 0.001), waist circumference (MD: −5.74 cm, 95% CI: −7.17, −4.31, P < 0.001), lean mass (MD: −1.29 kg, 95% CI: −2.17, −0.41, P = 0.004), fat mass (MD: −2.93 kg, 95% CI: −4.70, −1.12, P = 0.001), systolic blood pressure (MD: −3.99 mmHg, 95% CI: −5.66, −2.33, P < 0.001), diastolic blood pressure (MD: −1.11 mmHg, 95% CI: −1.71, −0.42, P = 0.002), glycated hemoglobin (MD: −0.31%, 95% CI: −0.47, −0.15, P < 0.001), fasting blood glucose (MD: −6.51 mg/dL, 95% CI: −7.31, −4.71, P = 0.004), total cholesterol (MD: −5.85 mg/dL, 95% CI: −9.78, −1.91, P = 0.004), triglycerides (MD: −13.44 mg/dL, 95% CI: −20.38, −6.50, P < 0.001), low-density lipoprotein cholesterol (MD: −4.78 mg/dL, 95% CI: −7.35, −2.22, P = 0.003), with the exception of high-density lipoprotein cholesterol (MD: −0.14 mg/dL, 95% CI: −1.05, 0.76, P = 0.750).

Table 3.

Effects of lifestyle modifications combined with GLP-1 receptor agonists on all outcomes.

Outcomes	Number of studies	Number of participants	Mean difference (95%CI)	P	Prediction interval	I2 (%)
Body Weight, kg	33	12,028	−7.13 (−9.02, −5.24)	<0.001	(−18.64, 4.38)	99
Secondary outcomes
Percentage of Body Weight, %	21	8754	−7.02 (−8.85, −5.18)	<0.001	(−15.90, 1.87)	99
Waist Circumference, cm	20	7195	−5.74 (−7.17, −4.31)	<0.001	(−12.14, 0.66)	95
Lean Mass, kg	6	503	−1.29 (−2.17, −0.41)	0.004	(−3.94, 1.36)	75
Fat Mass, kg	5	425	−2.93 (−4.70, −1.12)	0.001	(−8.28, 2.42)	82
Percentage of Body Fat, %	4	271	−1.18 (−2.00, −0.35)	0.005	(−2.52, 0.16)	0
Systolic Blood Pressure, mmHg	24	8167	−3.99 (−5.66, −2.33)	<0.001	(−11.82, 3.84)	97
Diastolic Blood Pressure, mmHg	23	7960	−1.11 (−1.71, −0.42)	0.002	(−3.71, 1.49)	80
Glycated Hemoglobin, %	21	7634	−0.31 (−0.47, −0.15)	<0.001	(−0.42, −0.12)	100
Fasting Blood Glucose, mg/dL	21	7373	−6.51 (−7.31, −4.71)	<0.001	(−14.75, 1.73)	99
Total Cholesterol, mg/dL	14	2969	−5.85 (−9.78, −1.91)	0.004	(−19.67, 4.61)	67
Triglycerides, mg/dL	16	3061	−13.44 (−20.38, −6.50)	<0.001	(−37.58, 10.70)	74
High-Density Lipoprotein Cholesterol, mg/dL	13	2903	−0.14 (−1.05, 0.76)	0.750	(−2.49, 2.19)	40
Low-Density Lipoprotein Cholesterol, mg/dL	13	2903	−4.78 (−7.35, −2.22)	0.003	(−8.88, −0.86)	11

We also calculated the standardized mean differences (SMDs) for each outcome. Compared with the control group, significant reductions were observed in body weight, waist circumference, lean mass, fat mass, blood pressure, and blood lipids (including total cholesterol, triglycerides and low-density lipoprotein cholesterol) to varying degrees (Supplementary Table S3).

Subgroup analysis

We further conducted subgroup analyses to investigate key factors influencing treatment efficacy across all outcomes (Table 4). Greater weight loss was observed in subgroups with a treatment duration of more than 52 weeks (MD: −9.40, 95% CI: −11.41, −7.39, P for interaction = 0.002), those receiving semaglutide (MD: −11.99, 95% CI: −12.93, −11.05) or tirzepatide (MD: −11.56, 95% CI: −21.68, −1.43, P for interaction < 0.001), with weekly dosing frequency (MD: −10.04, 95% CI: −13.42, −6.65, P for interaction = 0.012) and studies conducted in North America (MD: −7.44, 95% CI: −9.92, −4.96, P for interaction = 0.031). No significant subgroup effects were found for comorbidity (P = 0.057), pharmacological targets (P = 0.121), and ethnicity (P = 0.229).

Table 4.

Subgroup analysis of the effect of lifestyle modifications combined with GLP-1 receptor agonists on weight loss.

	Number of studies	Number of participants	Mean difference (95%CI)	P-interaction	Prediction intervals	I2 (%)
Treatment duration				0.002
<52 weeks	19	3326	−5.37 (−7.00, −3.74)		(−12.93, 2.19)	96
≥52 weeks	14	8702	−9.40 (−11.41, −7.39)		(−17.90, −0.90)	98
Category of GLP 1-RAs				<0.001
Semaglutide	7	5292	−11.94 (−13.32, −10.55)		(−16.45, −7.43)	79
Liraglutide	15	4138	−4.76 (−5.62, −3.90)		(−7.70, −1.82)	76
Exenatide	3	257	−3.36 (−4.54, −2.17)		(−5.96, −0.76)	5
Tirzepatide	2	944	−11.56 (−21.68, −1.43)		(−124.94, 101.82)	99
Other agents	6	1397	−7.88 (−12.75, −3.00)		(−19.63, 6.03)	99
Comorbidity				0.057
Metabolic Syndrome	23	10,014	−7.49 (−9.18, −5.81)		(−16.18, 1.20)	97
Diabetes	3	753	−5.22 (−7.00, −3.44)		(−11.60, 1.16)	54
Other comorbidities	6	1209	−7.55 (−13.79, −1.30)		(−29.11, 14.02)	99
Healthy participants	1	52	−1.79 (−6.15, 2.35)		–	–
Pharmacological targets				0.121
GLP-1 receptor agonists alone	30	10,476	−6.69 (−8.72, −4.65)		(−18.55, 5.18)	99
Dual-/triple-receptor agonists	3	1382	−11.71 (−17.74, −5.69)		(−38.01, 14.59)	99
Dosing frequency				0.012
Daily	21	6029	−5.43 (−6.65, −4.22)		(−11.08, 0.21)	91
Weekly	12	5999	−10.04 (−13.42, −6.65)		(−23.91, 3.83)	99
Regions				0.031
Europe	7	568	−4.05 (−5.84, −2.27)		(−9.58, 1.47)	83
North America	12	2993	−7.44 (−9.92, −4.96)		(−52.79, 48.30)	95
Asia	2	274	−2.25 (−6.79, 2.30)		(−13.87, 9.32)	97
Ethnicity				0.229
<80% white participants	17	6640	−8.03 (−10.08, −5.98)		(−14.92, 2.30)	99
≥80% white participants	16	5388	−6.31 (−8.22, −4.39)		(−17.10, 1.04)	99

To further explore potential sources of heterogeneity, we performed a meta-regression analysis, which identified the category of GLP-1RAs as a significant moderator. Specifically, studies using semaglutide showed a significantly greater effect on weight loss (MD: −7.37, 95% CI: −11.85, −2.88, P = 0.001; Supplementary Table S4).

Subgroup analysis results for secondary outcomes are presented in Supplementary Tables S5–S14. Overall, longer treatment duration and weekly dosing regimens showed a consistent trend of benefit across secondary outcomes.

Heterogeneity and sensitivity analysis

Most of the outcomes in this meta-analysis exhibited substantial heterogeneity, with I² values exceeding 50% for all outcomes except percentage of body fat, HDL-C, and LDL-C. Prediction intervals were also calculated, and only the interval for glycated hemoglobin and LDL-C did not cross the null value. More than half of the variation in true effect sizes for body weight change was explained by the included covariates (treatment duration, category of GLP-1RAs, comorbidity, pharmacological targets, dosing frequency, regions, ethnicity) in the meta regression model (R² = 66.77%).

Sensitivity analyses showed that the results did not change significantly after omitting each trial in succession (Supplementary Fig. S1). We also conducted sensitivity analyses by excluding studies with undefined lifestyle interventions or at moderate and high risk of bias, and the results remained consistent (Supplementary Tables S15and S16).

Assessment of bias and evidence certainty

Of the 33 trials included in this analysis, 3 trials did not report sufficient details on the randomization process, 6 trials had missing outcome data, 1 trial was at risk of bias in outcome measurement, and 2 trials presented potential bias due to selective reporting of results (Supplementary Figs. S2 and S3). The GRADE assessment demonstrated varying levels of certainty across the outcomes (Supplementary Table S17). High-certainty evidence was observed for changes in high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, and body fat percentage. Moderate-certainty evidence supported changes in systolic blood pressure, fasting blood glucose, total cholesterol, lean mass, fat mass, and waist circumference. In contrast, low-certainty evidence was found for body weight (in kg and %), diastolic blood pressure, glycated hemoglobin, triglycerides, and several other metabolic outcomes. The main reasons for downgrading the certainty of evidence were heterogeneity across studies and suspected publication bias.

Begg’s funnel plot was shown in Supplementary Fig. S4. The plot was asymmetric. We further performed Egger's regression test to assess potential publication bias (P = 0.022). The trim-and-fill analysis showed slightly increased effect size (MD: −11.32; 95% CI: −13.30, −9.34; P < 0.001; Supplementary Fig. S5).

Discussion

This meta-analysis of RCTs investigates the effectiveness of lifestyle modification combined with GLP-1RAs for weight loss and cardiometabolic biomarkers in individuals with overweight or obesity, including 33 studies with a total of 12,028 participants. The addition of GLP-1RAs to lifestyle modification significantly reduced body weight, and improved blood pressure, glycemic control, and lipid profiles. Greater benefits were observed with longer treatment duration, use of semaglutide or tirzepatide, weekly dosing, and studies conducted in North America.

Lifestyle intervention, comprising caloric restriction and increased physical activity, remains the cornerstone of obesity management and is universally recommended as the first-line therapeutic approach.⁵² However, its effectiveness is often compromised by suboptimal adherence, difficulty in sustaining behavioral changes, and weight regain. GLP-1RAs have emerged as a promising adjunctive therapy that may overcome several limitations of lifestyle intervention.⁵³ Through mechanisms including appetite suppression, delayed gastric emptying, enhanced satiety, and improved glycemic regulation, GLP-1RAs contribute to more sustained and clinically significant weight loss.54, 55, 56 Previous randomized controlled trials and meta-analyses have consistently demonstrated the efficacy of GLP-1RAs in reducing body weight, body mass index, and waist circumference across diverse populations.57, 58, 59 However, these studies have not systematically evaluated the effectiveness of combining lifestyle modifications with GLP-1RAs therapy for weight loss. By synthesizing data from 33 randomized controlled trials, our meta-analysis demonstrates that the addition of GLP-1RAs to lifestyle modification results in a significant reduction in body weight and waist circumference. These findings are consistent with existing literature and provide robust evidence to support a comprehensive treatment strategy for individuals with overweight or obesity. Although asymmetry was observed in the funnel plot, the results of the trim-and-fill analysis, which accounts for potential unpublished studies, yielded a larger effect size (MD: −11.32 vs. −7.13). These findings suggested that the true treatment effect may have been slightly underestimated due to potential publication bias, and further support the robustness of our overall conclusions.

Our study also demonstrated that the addition of GLP-1RAs to lifestyle interventions significantly improved cardiometabolic biomarkers, including blood pressure, fasting blood glucose, and lipid profiles. These findings are consistent with previous studies. The Semaglutide Treatment Effect in People With Obesity (STEP) 4 trial showed that following a 20-week run-in period with subcutaneous semaglutide 2.4 mg/week, continued treatment with semaglutide, compared to placebo, resulted in significant reductions in systolic blood pressure, fasting plasma glucose, HbA_1c, total cholesterol, triglycerides, and LDL-C levels.³⁸ Similar results have been reported in other meta-analyses.^57,60 GLP-1RAs were initially developed for glycemic control in type 2 diabetes, exerting their effects primarily through stimulating insulin secretion and suppressing glucagon release.⁶¹ Increasing evidence also indicates their potential cardiovascular benefits, possibly through mechanisms involving nitric oxide upregulation and suppression of endothelial inflammation, which may lead to improvements in blood pressure and lipid metabolisms. Moreover, recent meta-analyses have demonstrated that GLP-1RA treatment significantly reduces the risk of major adverse cardiovascular events in patients with diabetes.^62,63 These findings further suggest that combining lifestyle interventions with GLP-1RAs may offer additional long-term cardiovascular benefits.

Subgroup analyses indicated that longer treatment duration, the use of semaglutide or tirzepatide, weekly dosing regimens, and studies conducted in North America were associated with greater weight reduction. These factors may reflect a combination of drug efficacy, patient adherence, and regional variations in baseline characteristics and healthcare systems. Longer treatment duration and the convenience of weekly dosing regimens may enhance patient adherence and help sustain therapeutic efficacy. Consistently, a meta-analysis by Wong et al. also demonstrated that longer GLP-1RA treatment duration is associated with greater weight loss.⁶⁴ Since the weight-reducing and metabolic benefits of GLP-1RAs typically emerge gradually over several weeks, continued administration is essential to maintain peak efficacy and prevent weight regain. In addition, prolonged treatment provides patients with sufficient time to develop and consolidate healthy dietary and physical activity habits. Notably, tirzepatide is a dual GIP/GLP-1 receptor agonist. Previous studies have demonstrated its superior efficacy in weight loss and improvements in blood glucose and HbA1c compared to semaglutide.⁵⁸ However, our findings indicate that both agents exhibit comparably favorable weight reduction effects, making them both viable options as GLP-1 receptor agonists.

To our knowledge, this is the first meta-analysis to provide up-to-date evidence on the effects of combining lifestyle modification with GLP-1RAs on weight reduction and cardiometabolic biomarkers. These novel findings underscore the added value of integrating pharmacologic and behavioral interventions, supporting a more comprehensive and individualized approach to obesity management. The results may help inform future clinical guidelines and promote the adoption of combination strategies in routine care. Based on the GRADE criteria, all outcomes in our study were supported by low to high certainty of evidence. However, several limitations should be noted. First, Egger’s test indicated significant publication bias and the funnel plot showed asymmetry. Trim-and-fill analysis suggested that the true treatment effect might have been slightly underestimated due to potential publication bias. Second, although lifestyle interventions in the included studies involved caloric restriction and physical activity, we were not able to assess the impact of different types or intensities of lifestyle components in detail. While we attempted sensitivity analyses, more granular data are needed to evaluate specific intervention strategies. Third, previous meta-analyses have shown that GLP-1RAs may significantly reduce lean mass in addition to body weight.⁶⁵ Our study attempted to collect data on changes in body composition, including lean mass, fat mass, and percentage of body fat. Despite the limited number of studies, a consistent trend was observed. Future research may need to consider incorporating higher-intensity physical activity or resistance training to offset the loss of lean mass. Fourth, our subgroup analysis found that studies conducted in North America reported greater efficacy. However, due to limited data, we were only able to perform limited stratification by ethnicity. Only two of the included trials were conducted in Asian populations; the others included participants from diverse racial. More research in diverse populations and regions is necessary to provide generalizable and culturally relevant insights. Fifth, our study included one trial conducted in individuals with craniopharyngioma-related obesity,²⁴ a rare and severe form of hypothalamic obesity resulting from damage to the hypothalamic-pituitary axis due to craniopharyngioma or its treatment. Although a leave-one-out sensitivity analysis (Supplementary Fig. S1) showed that excluding this study did not alter the overall results, our findings should be interpreted with caution when applied to this specific population.

In conclusion, lifestyle interventions combined with GLP-1RAs may help reduce body weight and improve cardiometabolic biomarkers in adults with overweight or obesity. In light of the varying certainty of evidence across outcomes, these results should be interpreted cautiously. Treatment duration, drug type, dosage, and geographic region were key influencing factors of intervention effectiveness.

Contributors

ZZ and HW planned and designed the study, ZZ and JC developed search strategies, JC, HZ and ZZ screened potential studies and extracted data from the included studies, JC and ZZ accessed and verified the data, performed the statistical analysis. ZZ and YW provided methodological support and helped to interpret findings. JC, YH, TZ and ZZ wrote the first draft, ZZ, YW and JC revised the draft. All authors approved the final version of the manuscript.

Data sharing statement

All data in this analysis are based on published studies. Data described in the manuscript, code book, and analytic code will be made available upon request pending.

Declaration of interests

All authors declare no competing interests.