Risk of Bone Mineral Density Loss Associated with GLP-1 Receptor Agonists in Overweight COPD Populations [Response to Letter]

Dear editor

We thank Hong and Liu for their thoughtful comments on our article examining glucagon-like peptide-1 receptor agonists (GLP-1RAs) and clinical outcomes in patients with chronic obstructive pulmonary disease (COPD) and obesity.1 Their letter raises two points: the lack of an osteoporosis- and fracture-related outcome in our original analysis, and limitations in follow-up duration and the characterization of all-cause mortality.2 Below we address both points and report additional analyses conducted in response.

Osteoporosis and fragility fracture are clinically relevant outcomes in COPD, given the established contributions of chronic systemic inflammation, corticosteroid exposure, physical inactivity, and malnutrition to bone loss in this population. We therefore performed an additional propensity score-matched cohort analysis within the same TriNetX database and study population (COPD with obesity), comparing incident osteoporosis and pathological fracture between new users of GLP-1RAs and the comparator weight-loss medication group. After 1:1 propensity score matching on the same baseline covariates as the original analysis, 10,487 patients were included in each group, consistent with our primary matched cohort. Over a median follow-up of 530 days in the GLP-1RA group and 467 days in the comparator group, incident osteoporosis occurred at 20.6 per 1000 person-years in the GLP-1RA group versus 21.9 per 1000 person-years in the comparator group (hazard ratio [HR], 0.90; 95% confidence interval [CI], 0.79–1.03). Fragility fracture risk was lower with GLP-1RA use (HR, 0.76; 95% CI, 0.66–0.88), driven largely by reductions in hip fracture (HR, 0.68; 95% CI, 0.53–0.88) and vertebral compression fracture (HR, 0.74; 95% CI, 0.61–0.89); pathological fracture showed a similar downward trend that did not reach statistical significance (HR, 0.71; 95% CI, 0.49–1.02). Within the follow-up period captured in our database, GLP-1RA use was not associated with a measurably higher risk of osteoporosis or fracture in this COPD-with-obesity phenotype, and was in fact associated with a lower risk of fragility fracture. As the letter notes, the relationship between GLP-1RA use and bone health remains unsettled: some studies report increased fragility fracture risk among older adults with type 2 diabetes and accelerated bone loss in non-diabetic GLP-1RA users.3,4 The discrepancy between our findings and these prior reports may reflect differences in study population and comparator group. However, we agree that bone mineral density and fracture incidence merit routine consideration in future studies of GLP-1RA use in COPD populations.

We also performed additional subgroup analysis stratifying all-cause mortality by baseline osteoporosis status and fracture history. Among patients with pre-existing osteoporosis or fragility fracture, the association between GLP-1RA use and all-cause mortality was attenuated relative to the overall cohort (HR, 0.62; 95% CI, 0.53–0.73; p-interaction = 0.006), while patients without such a history showed a stronger association (HR, 0.38; 95% CI, 0.28–0.52). This indicates that baseline skeletal fragility may modify the magnitude of mortality benefit associated with GLP-1RA use, although the benefit remained substantial and statistically significant even among patients with pre-existing osteoporosis or fracture.

Lastly, we agree that the median follow-up in our original analysis (530 days in the GLP-1RA group and 467 days in the comparator group) may not fully capture the longer-term skeletal effects of GLP-1RA use, particularly given that current GOLD guidance recommends bone density reassessment in COPD patients every 3–5 years. To address this, we extended the follow-up window in the TriNetX database to 5 years and repeated the primary and osteoporosis/fracture analyses using a landmark approach starting 30 days after index. Over this extended follow-up, the association between GLP-1RA use and all-cause mortality remained directionally consistent with our primary analysis (HR, 0.45; 95% CI, 0.37–0.50). The risk of osteoporosis remained non-significantly different between groups (HR, 0.96; 95% CI, 0.85–1.08), while fragility fracture risk remained significantly lower in the GLP-1RA group (HR, 0.76; 95% CI, 0.66–0.87), and the reduction in pathological fracture reached statistical significance with longer follow-up (HR, 0.70; 95% CI, 0.50–0.99). These extended-follow-up results are now presented alongside our original short-term finding.1

In conclusion, the additional analyses performed in response to Hong and Liu’s comments do not indicate that GLP-1RA use increases the risk of osteoporosis or fracture in patients with COPD and obesity. The mortality benefit associated with GLP-1RA use was attenuated but remained significant among patients with pre-existing skeletal fragility. We thank Hong and Liu for raising these points, which have allowed us to strengthen our original findings with longer follow-up and a more complete assessment of skeletal outcomes. We agree that bone health should be incorporated as a standard outcome in future pharmacoepidemiological studies of GLP-1RAs in COPD population.

Disclosure

The authors report no conflicts of interest in this communication.