Insulin resistance and the atrial fibrillation substrate across treatment strategies in obese non-diabetic patients: A post hoc analysis of the PRAGUE-25 randomized trial

Background: Insulin resistance (IR) is prevalent in obese atrial fibrillation (AF) patients, yet whether IR modifies outcomes differently across treatment strategies remains unknown. Aims: The aim of this study was to determine whether baseline Homeostatic Model Assessment of Insulin Resistance (HOMA-IR) predicts AF outcomes differently within catheter ablation (CA) vs. lifestyle modification plus antiarrhythmic drug (LFM + AAD) arms of the PRAGUE-25 trial. Methods: Post hoc analysis of 124 non-diabetic patients (body mass index 30–40 kg/m2; 63 CA, 61 LFM + AAD) with available HOMA-IR. Restricted cubic splines modeled the dose-response overall and within each arm separately. In the next step, 32 diabetic patients were included. Primary outcome was AF recurrence at 12 months. Results: IR was near-universal (91.1% with HOMA-IR ≥2.5; median 5.6). A nonlinear, U-shaped HOMA-IR–recurrence association (nonlinearity P = 0.006) was apparent in the LFM + AAD arm (P = 0.005; recurrence 47.4%/23.5%/47.8% across tertiles) but absent in the CA arm (P = 0.58; 17%–29%). The treatment × HOMA-IR interaction was non-significant (P = 0.87). The nonlinearity was reproducible across insulin-based indices and on bootstrapping but attenuated with extreme-value trimming. Adding 32 diabetic patients, a trend toward a CA benefit was observed in diabetes (hazard ratio, 6.72; 95% confidence interval, 0.88–51.48; P = 0.07). Conclusions: In this exploratory post hoc analysis, CA provided consistent rhythm control regardless of IR, whereas lifestyle-based therapy appeared less effective at higher HOMA-IR. The U-shaped association did not reach formal interaction significance and is hypothesis-generating, warranting prospective validation.