Different measures of blood pressure and incident atrial fibrillation

Hypertension is a strong predictor of cardiovascular diseases and its prevalence is increasing worldwide [1]. Atrial fibrillation, which is also associated with increased cardiovascular risk, is the most common cardiac arrhythmia [2]. Hypertension and atrial fibrillation often coexist and their incidence increases with ageing [1,2]. Epidemiological studies have shown that hypertension is an important risk factor for the development of atrial fibrillation [3,4]. Various mechanisms have been proposed to link hypertension and atrial fibrillation, including those involved in the pathophysiology of hypertension and those related to hypertension-mediated organ damage [5,6]. These potential mechanisms are summarized in Fig. 1. Diagnosis and management of hypertension is historically based on detection of clinic brachial blood pressure (BP). However, it has been showed that other types of BP measure, such as clinic central BP [7] and mostly ambulatory brachial BP [8] are more strongly associated with cardiovascular risk than clinic brachial BP. In such a context, some studies have also attempted to evaluate the impact of clinic central [9–11] or ambulatory brachial BP [12–15] (Table 1) on the occurrence of incident atrial fibrillation. Although some data suggest that clinic central [9] or ambulatory brachial BP [15] are superior to clinic brachial BP in predicting future atrial fibrillation, current evidence is not yet conclusive and no study compared the predictive value of these three different BP measures.FIGURE 1: Mechanisms linking hypertension with atrial fibrillation. LA, left atrial; LV, left ventricular; RAAS, renin-angiotensin-aldosterone system; SNS, sympathetic nervous system. TABLE 1 - Previous studies on ambulatory blood pressure and incident atrial fibrillation Ref. Patients/Events Age at entry (years) Mean follow-up (years) Daytime BP HR (95% CI) Night-time SBP HR (95% CI) 24-h BP HR (95% CI) Nondipping HR (95% CI) Ciaroni et al.[12] 597/28 >50 7 – – 1.16 (1.06–2.47) – Pierdomenico et al.[13] 1141/43 ≥40 6 – – 1.32 (1.05–1.67) 2.02 (1.08–3.79) Perkiömäki et al.[14] 903/91 40–59 16 1.05 (0.98–1.13) 1.07 (1.004–1.15) 1.09 (1.01–1.17) – Tikhonoff et al.[15] 2276/111 >18 14 1.22 (1.02–1.46) 1.20 (1.02–1.42) 1.27 (1.07–1.51) – Ref. [12] included individuals with clinic hypertension; hazard ratio (HR) is expressed per 10-mmHg increment of SBP and adjusted for age, sex and BMI. Ref. [13] included individuals with clinic and ambulatory hypertension; HR is expressed per 10-mmHg increment of SBP and adjusted for age, left atrial enlargement and nondipping. Ref. [14] included individuals with normotension and clinic hypertension (49%); HR is expressed per 5-mmHg increment of SBP and adjusted for age, sex, BMI, height, smoking, alanine aminotransferase, uric acid and glucose. Ref. [15] included individuals with normotension and clinic hypertension (about 30%); HR is expressed per 1 standard deviation (≈ 12 mmHg) increment of SBP and adjusted for sex, age, BMI, serum cholesterol, tobacco and alcohol use, history of cardiovascular disease and diabetes mellitus and antihypertensive drug treatment.BP, blood pressure; CI, confidence interval; HR, hazard ratio. Matsumoto et al.[16] in this issue of the Journal evaluated the impact of clinic brachial, clinic central and ambulatory brachial BP on incident atrial fibrillation. Among participants in the Cardiovascular Abnormalities and Brain Lesions study, 769 (mean age 70.5 years) with baseline sinus rhythm were followed for 9.5 ± 3.0 years and atrial fibrillation occurred in 83 (10.8%). No clinic brachial BP variable had significant association with incident atrial fibrillation. Clinic central SBP and pulse pressure were marginally associated with incident atrial fibrillation. Among ambulatory brachial BP parameters, per 10-mmHg increment, 24-h SBP (hazard ratio 1.24; 95% confidence interval, 1.07–1.44), daytime SBP (hazard ratio 1.21; 95% confidence interval, 1.04–1.40) and night-time SBP (hazard ratio 1.22; 95% confidence interval, 1.07–1.39) were significantly associated with incident atrial fibrillation. In patients with incident atrial fibrillation, both clinic brachial and clinic central SBP were correlated with 24-h, daytime and nigh-ttime BP, but clinic central SBP showed stronger correlation with all ambulatory brachial BP parameters. The likelihood ratio test showed that the addition of 24-h or daytime or night-time BP as continuous variables to the model including clinic central BP and other covariates significantly improved prediction of incident atrial fibrillation. Moreover, the optimal cutoffs for predicting incident atrial fibrillation were 127, 131, 130 and 123 mmHg for clinic central, 24-h, daytime and night-time SBP, respectively, and BP values higher than these cutoffs were significantly associated with outcome. However, the likelihood ratio test showed that the addition of 24-h or night-time BP as categorical variables (above cutoffs) to the model including clinic central BP as a categorical variable (above cutoff) and other covariates significantly improved prediction of outcome. This study not only has various strengths but also some limitations. The studied population included mainly elderly treated hypertensive patients of Hispanic ethnicity and results cannot be completely extrapolated to other types of patients. Clinic central SBP calibrated by clinic brachial SBP and DBP has been suggested to be less close to the invasive aortic SBP than that calibrated by clinic brachial mean and DBP [17] potentially suggesting that the method used in the present study might have contributed to explain the marginal association between clinic central SBP and incident atrial fibrillation. Target organ damage, such as left ventricular and atrial structural and functional alterations, which could be intermediate steps towards atrial fibrillation were not included in the multivariate analyses. In any case, data on the association between ambulatory brachial BP and incident atrial fibrillation in this study are in line with those obtained in other types of populations [12–15]; in a previous study calibrating clinic central SBP by clinic brachial mean and DBP, the superiority of ambulatory brachial over clinic central BP was confirmed in predicting cardiovascular outcome [18]; and studies including organ damage in multivariate analyses confirmed the independent prognostic value of BP in predicting incident atrial fibrillation [4,13]. Thus, it could be argued that the limitations of this study do not weaken their relevant findings. Globally, the present study suggests that the hierarchical order in predicting incident atrial fibrillation is brachial ambulatory (mainly 24-h and night-time), clinic central and clinic brachial BP. The superiority of brachial ambulatory BP over the other BP measures may be related to its capability to better integrate the pathophysiological background of hypertension, which may also directly influence the occurrence of atrial fibrillation (upper part of Fig. 1) and its consequences that are directly involved in the predisposition to atrial fibrillation (lower part of Fig. 1). The advantage of clinic central over clinic brachial and ambulatory brachial BP may be that it is more representative of the BP to which the heart, kidneys and cerebral-afferent arteries are exposed [7]. However, like clinic brachial BP, for known reasons, it may not be entirely representative of the real BP of the 24-h period. The advantage of ambulatory brachial over clinic brachial and clinic central BP lies in its detection of the real diurnal and nocturnal BP [8], though it is a peripheral BP. A potential improvement in the future might be the recording of 24-h ambulatory central BP and the evaluation of its prognostic impact in comparison with ambulatory brachial BP [19–21]. In the meanwhile, brachial ambulatory BP remains the best measure to evaluate hypertension characteristics and its complications. Various studies [22–28] evaluated the impact of BP control on future occurrence of atrial fibrillation but results are controversial. This aspect may partly be explained by the partial capability of clinic brachial BP to detect real BP control [29]. Given the aforementioned aspect and the results of the present study, future research evaluating the impact of ambulatory BP reduction on the occurrence of incident atrial fibrillation might be helpful to plan prevention strategies. Another aspect that might deserve future investigation is to assess whether invasive treatment of atrial fibrillation, as recently reported [30], might influence pathophysiological background of hypertension, ambulatory BP and circadian BP variations potentially predisposing to a reduced recurrence of atrial fibrillation. ACKNOWLEDGEMENTS Conflicts of interest There are no conflicts of interest.