Glucagon-like peptide-1 receptor agonists may be associated with increased gastrointestinal adverse events and could show associations with lower risks of serious infections and respiratory disease; however, most noncardiometabolic findings were supported by low- or moderate-certainty evidence.
Glucagon-like peptide (GLP)-1 receptor agonists (RA) were evaluated across 60 meta-analyses, including 1,751 randomized clinical trials and 3.6 million participants, most of whom were patients with type 2 diabetes or obesity, with follow-up ranging from 3 months to more than 5 years. Investigators conducted a systematic umbrella review, reanalyzing pooled trial-level data using random-effects models and grading evidence certainty with the Grading of Recommendations Assessment, Development, and Evaluation framework while assessing methodological quality using AMSTAR 2.
The most consistent associations were observed for gastrointestinal adverse events. GLP-1 RAs were linked to 2.47 higher odds of nausea, 2.78 higher odds of vomiting , and 1.94 higher odds of diarrhea, with evidence graded as high to moderate quality. However, variability across the studies and prediction interval analyses indicated ongoing uncertainty in these estimates.
Conversely, GLP-1 RAs were associated with 0.89 times the odds of serious infections, supported by high-quality evidence. A similar pattern was observed for incident respiratory disease, which had 0.85 times the odds), although this association was less certain and didn't consistently meet the prespecified credibility thresholds.
Across fracture, respiratory, neurologic, psychiatric, and endocrine outcomes, most associations didn't meet the prespecified high-credibility thresholds, although several of them showed possible safety signals. These included a lower likelihood of fracture (0.67 times the odds) and all-cause dementia (0.55 times the odds), as well as a greater likelihood of thyroid disease (1.27 times the odds). Gastrointestinal disease outcomes, including gastroesophageal reflux disease (2.19 times the odds) and gallbladder or biliary disease (1.34 times the odds), didn't meet stringent credibility criteria and were considered exploratory.
Cancer outcomes also didn't meet the prespecified credibility thresholds. Nominal associations were observed for colorectal cancer (1.24 times the odds) and pancreatic cancer (0.51 times the odds), but these findings remained exploratory and weren't sustained in sensitivity analyses.
Subgroup analyses examining GLP-1 RA class, dosing, and treatment duration were based on a limited number of studies and were therefore interpreted as exploratory. Sensitivity analyses suggested that several associations weren't robust, with the direction of effect flipping in 9 of 39 outcomes and statistical significance no longer observed in 6 of 39 outcomes after exclusion of a single trial.
Methodological concerns were frequently identified, including an absence of prespecified protocols and incomplete reporting of funding sources in many meta-analyses. In addition, numerous outcomes were based on adverse event reporting rather than predefined endpoints, which the investigators noted may introduce inconsistent definitions and increase the risk of bias.
“[A]lthough GLP-1 RAs have been widely studied across a spectrum of health outcomes, the current evidence base is insufficient for definitive conclusions of high certainty,” wrote lead study author Kaijie Yang, MD, of the Department of Endocrinology and Metabolism and the Institute of Endocrinology at the National Health Commission Key Laboratory of Diagnosis and Treatment of Thyroid Diseases at the First Hospital of China Medical University, and colleagues.
The study authors reported no conflicts of interest. Funding was provided by the National Natural Science Foundation of China.
Source: JAMA Network Open
