Association of blood glucose-6-phosphate dehydrogenase activity with atrial fibrillation and post-ablation recurrence: a single-center retrospective study

Abstract Background Oxidative stress contributes to atrial fibrillation (AF), while glucose-6-phosphate dehydrogenase (G6PD) supports cellular redox homeostasis by maintaining NADPH availability. We investigated the associations of blood G6PD activity with prevalent AF and 1-year recurrence after catheter ablation and evaluated its discriminatory performance. Methods This single-center retrospective study included 175 consecutive cardiology inpatients: 92 patients with AF and 83 contemporaneous hospitalized controls. Multivariable logistic regression assessed the association between G6PD activity and AF after adjustment for age, sex, body mass index, hypertension, diabetes mellitus, coronary artery disease, estimated glomerular filtration rate, and hemoglobin. Hemoglobin-normalized G6PD activity was examined in a post hoc sensitivity analysis. Receiver operating characteristic analyses with bootstrap 95% confidence intervals were performed. Patients with AF were followed for 12 months after catheter ablation. Results Blood G6PD activity was lower in patients with AF than in controls [2280.0 (2077.8–2468.0) vs. 2548.0 (2342.5–2713.0) U/L, P < 0.001], and the difference remained significant after hemoglobin normalization [6.786 (6.299–7.440) vs. 7.503 (6.936–8.077) U/g Hb, P < 0.001]. Each 100-U/L increase in G6PD activity was associated with lower odds of AF (OR 0.731, 95% CI 0.637–0.837; P < 0.001). G6PD activity yielded an AUC of 0.738 (bootstrap 95% CI 0.664–0.808), and its addition to the clinical reference model increased the apparent AUC from 0.658 to 0.778 (ΔAUC 0.120, bootstrap 95% CI 0.040–0.197). During follow-up, 21 patients experienced recurrence. Preprocedural G6PD activity was lower in patients with recurrence [2116.0 (2068.0–2175.0) vs. 2326.0 (2145.0–2578.0) U/L, P < 0.001] and yielded an AUC of 0.749 (bootstrap 95% CI 0.646–0.845). Adding G6PD activity to a limited clinical reference model increased the apparent AUC from 0.678 to 0.768, although the confidence interval for the increment included zero (ΔAUC 0.091, bootstrap 95% CI − 0.020 to 0.202). Conclusions Lower blood G6PD activity was independently associated with prevalent AF and provided incremental in-sample discrimination beyond conventional clinical variables. Lower activity was also associated with post-ablation recurrence, although its incremental prognostic value remains uncertain.