Heterogeneous treatment effects of GLP-1RA on kidney outcomes in patients with type 2 diabetes.

INTRODUCTION

On average, glucagon-like peptide-1 receptor agonists (GLP-1RA) lower the risk of kidney outcomes in type 2 diabetes, but individual responses may differ. Here, we aimed to identify patient subgroups more likely to derive kidney benefit from GLP-1RA therapy.

METHODS

Using a target trial emulation framework, we conducted a new-user, active comparator cohort study that included over 28,000 individuals with type 2 diabetes initiating either GLP-1RA or dipeptidyl peptidase-4 inhibitors (DPP-4i) in Stockholm (2008-2021). Heterogeneous treatment effects on a composite kidney outcome were estimated using complementary approaches: risk modelling with the kidney failure risk equation (KFRE) and effect modelling with causal survival forests. For interpretability, Shapley Additive Explanations identified clinical features driving benefit.

RESULTS

We included 11,772 GLP-1RA and 16,775 DPP-4i initiators. Compared with DPP-4i, GLP-1RA initiation was associated with a lower five-year risk of kidney outcomes (risk difference -0.72% [95% confidence interval -1.53 to 0.13]; hazard ratio [HR] 0.84 [0.72-0.97]). Risk modelling revealed pronounced heterogeneity: patients in the highest KFRE-risk quarter experienced a substantial absolute risk reduction (-2.76% [-5.60 to -0.08]; HR 0.70 [0.56-0.86]), whereas those in the lowest quarter showed no benefit. Effect modelling similarly contrasted quarters 4 vs. 1 (HR 0.79 [0.65-0.94] vs. 1.06 [0.54-1.74]). Key predictors of treatment benefit included lower baseline estimated glomerular filtration rate (eGFR), faster preceding eGFR decline, higher albuminuria, higher glycated hemoglobin, and greater healthcare utilization prior to treatment start.

CONCLUSIONS

GLP-1RA therapy confers heterogeneous kidney benefits in type 2 diabetes, with certain high-risk subgroups experiencing the largest risk reduction. Such findings support personalized GLP-1RA prescribing guided by kidney function/damage, trajectories of eGFR decline, and KFRE risk stratification.