Predictive ability of visit-to-visit blood pressure indices for adverse events in patients with non-valvular atrial fibrillation: Subanalysis of the J-RHYTHM Registry

BACKGROUND: We previously reported that standard deviation (SD) of systolic blood pressure (SBP), an index of BP variability, and SBP-time in target range (TTR), an index of BP consistency, were significantly associated with adverse events in patients with non-valvular atrial fibrillation (NVAF). Thus, this study aimed to compare predictive ability for adverse events among visit-to-visit BP variability/consistency indices using data from the J-RHYTHM Registry. METHODS: Of 7406 outpatients with NVAF, 7226 (age, 69.7 ± 9.9 years; men, 70.7%), in whom BP was measured 4 times or more (14.6 ± 5.0 times) during the 2-year follow-up period or until occurrence of an event, were included. As BP consistency for target SBP between 110 and 130 mmHg, SBP-TTR by the Rosendaal method and SBP-frequency in range (FIR) were calculated. Predictive ability was expressed by the area under receiver-operating-characteristic curve (AUC). AUCs of SBP-TTR and SBP-FIR for adverse events were compared with those of SBP-SD by the DeLong’s test. RESULTS: SBP-SD, SBP-TTR, and SBP-FIR were 11.0 ± 4.2 mmHg, 49.5 ± 28.3%, and 52.3 ± 23.0%, respectively. AUCs of these indices for thromboembolism, major hemorrhage, and all-cause death were 0.62, 0.64, and 0.63 for SBP-SD; 0.56, 0.55, and 0.56 for SBP-TTR; and 0.55, 0.56, and 0.58 for SBP-FIR; respectively. AUCs of SBP-SD were significantly larger than those of SBP-TTR for major hemorrhage (P = 0.010) and all-cause death (P = 0.014), and SBP-FIR for major hemorrhage (P = 0.016). CONCLUSION: Among visit-to-visit BP variability/consistency indices, predictive ability of SBP-SD for major hemorrhage and all-cause death was superior to that of SBP-TTR and SBP-FIR in patients with NVAF.