Research brief · 2026-09-09
Why Do Two Meta-Analyses on the Same Question Disagree?
Systematic reviews rank high because they pool many trials. But pooling does not guarantee agreement. Two careful teams can build different reviews from overlapping trials and land on opposite conclusions. A pair of 2024 meta-analyses on GLP-1 receptor agonists and bone density shows how. These drugs, including semaglutide and liraglutide, are prescription medications, and decisions about them belong with a physician.
Li and colleagues, in the International Journal of Endocrinology, pooled 26 randomized trials and 2,268 people with type 2 diabetes and, against other antidiabetic drugs or placebo, found a significant increase in lumbar spine and femoral neck bone density. Kim and colleagues, in Diabetes/Metabolism Research and Reviews, kept only seven trials with a true placebo or untreated control and formal risk-of-bias screening, and found no significant change at the femoral neck, hip or spine. Both passed peer review. Neither corrected the other.
The first mechanic is what counted as a control. The 26-trial review included comparisons against other diabetes medications, some of which are independently linked to worse bone outcomes, so GLP-1 can look protective by contrast. The second is the statistic. Standardized mean difference is unit-less and can look large when trials are small or varied; raw mean difference in grams per square centimeter, used by the 7-trial review, sat at 0.01, minus 0.01 and zero. The third is time. Bone remodels slowly, and a 2025 network meta-analysis of 33 trials and 33,107 people in BMC Musculoskeletal Disorders warned that bone density findings under 52 weeks need cautious interpretation.
Counted by reviews, the bone-improves conclusion has more company, including a 2025 analysis in Acta Diabetologica. But counting reviews is not weighing them. The 7-trial review is the only one restricted to placebo-controlled comparisons with a current bias tool. A review can be the outlier and the most conservative at the same time. What none of them settle is whether any of this changes fracture rates. Sources are listed below.
Sources
- Li X, Li Y, Lei C, Effects of GLP-1 receptor agonists on bone metabolism in type 2 diabetes: systematic review and meta-analysis (26 trials, n=2,268)
- International Journal of Endocrinology, 2024
- Kim HJ et al., Effects of GLP-1 receptor agonist on bone mineral density and bone turnover markers: a meta-analysis (7 placebo-controlled trials)
- Diabetes/Metabolism Research and Reviews, 2024
- Alalwani YJ et al., Differential effects of GLP-1 receptor agonists on diabetic osteopathy: patient-stratified network meta-analysis (33 trials, n=33,107)
- BMC Musculoskeletal Disorders, 2025
- Tan Y, Liu S, Tang Q, Effect of GLP-1 receptor agonists on bone mineral density, bone metabolism markers, and fracture risk in type 2 diabetes (25 trials)
- Acta Diabetologica, 2025
Peptide Proof reports what studies found. We do not give medical advice, dosing guidance, or treatment recommendations.