Hypertension and Atrial Fibrillation: Balancing Stroke and Bleeding Risks

Hypertension (HTN) is the commonest cardiovascular comorbidity related to atrial fibrillation (AF), with 49–90% of AF patients included in clinical trials having a previous history of HTN. Elevated blood pressure (BP) has been one of the strongest predictors of the development of AF and HTN per se increases the risk of AF by 2-fold.1 The intimate association between HTN and AF has led to the proposal that AF is yet another manifestation of hypertensive target organ damage.1 The relationship of HTN to stroke and ischemic events is well known, in patients with and without AF.2 In a large “real world” Swedish AF cohort study,3 there was significant independent association between HTN and ischemic stroke [hazard ratio (HR) 1.19, 95% confidence interval (CI) (1.12–1.25)]. In the PROGRESS trial,4 active treatment with perindopril reduced the relative risk of major vascular events by 38% in patients with AF and previous stoke. Given that HTN plays an essential role in stroke and thromboembolism risk in AF patients, stroke risk stratification scores for AF, such as the CHADS2 and CHA2DS2-VASc scores, assign 1 point to the presence of HTN or uncontrolled BP; thus, the combination of antihypertensive treatment with stroke prevention using oral anticoagulation is an important consideration for the management of AF patients.5 On the other hand, AF patients also have high risk of intracranial and other major bleeding events, not simply related to oral anticoagulation, but also by a higher intrinsic risk of bleeding.6 In the large Swedish AF cohort study,3 the presence of HTN was also an independent risk factor for intracranial bleeding [HR 1.32, 95% CI (1.15–1.52)] and major bleeding [HR 1.25, 95% CI (1.16–1.33)]. Indeed, uncontrolled HTN is an important risk factor for bleeding, both in anticoagulated and non-anticoagulated AF patients.7 Thus, bleeding risk should be evaluated in all AF patients before starting oral anticoagulation with the principle aim to “flag up” those patients at risk for careful follow-up and more regular review, with the latter focused on reversible bleeding risk factors.6 Based on common bleeding risk factors, risk stratification scores (e.g., HEMORR2HAGES, HAS-BLED, ATRIA, ORBIT and more recently, ABC-Bleeding) have been developed to assess bleeding risk in AF patients, with different degrees of complexity ranging from simple clinical or demographic variables to complex genetic or biomarkers factors.8–10 The HEMORR2HAGES score includes 1 point for uncontrolled HTN, while the HAS-BLED score gives 1 point for uncontrolled BP (systolic BP (SBP) >160 mm Hg). ATRIA score also assigns 1 point if previous diagnosis of HTN is identified, but not for uncontrolled BP despite the latter aspect having a stronger relationship to bleeding risk. More recent bleeding risk scores (e.g., ABC-Bleeding) do not include HTN as a risk factor for bleeding. Indeed, the INTERSTROKE study11 showed that HTN is the most important risk factor for spontaneous intracranial bleeding. Senoo et al.9 compared the HAS-BLED, ATRIA, and ORBIT scores and, although all 3 bleeding risk scores demonstrated modest predictive value for clinically relevant bleeding, the HAS-BLED score performed best, and was the only bleeding score that has shown to be predictive of intracranial hemorrhage,6 and so the control of BP in all patients, especially in AF, is imperative. A frequently debated issue is to determine whether a history of HTN has the same weight as elevated SBP per se, for stroke and bleeding risk. Previous studies have found an elevated risk of stroke or bleeding events in AF patients, regardless of BP control, whereas some studies have found a high(er) risk of adverse events in AF patients without previous history of HTN but with uncontrolled SBP.12 The SPAF trial13 showed that a history of HTN (relative risk = 2.0, P < 0.001) and SBP >160 mm Hg (relative risk = 2.3, P < 0.001) were independently associated with increased stroke risk but did not analyze the relationship with bleeding events. The SPORTIF trials14 showed in anticoagulated AF patients, event rates stroke/systemic embolism increased substantially at mean SBP levels ≥140 mm Hg during the follow-up. For the top quartile of SBP (140.8–191.7 mm Hg) compared with the lowest quartile (84.0–122.6 mm Hg), the HR for stroke and systemic embolic events was 1.83 [95% CI (1.22–2.74), P = 0.0037], but there was no relationship between bleeding and quartiles of BP. Indeed, in the ACTIVE-W trial,15 a history of HTN (but not ‘baseline’ SBP) did not predict stroke or bleeding, but lower SBP during the follow-up was associated with lower risk stroke. In a recent subanalysis of the Japanese J-RHYTHM AF registry,16 the presence of HTN was an independent risk factor for major bleeding but not for thromboembolism; however, patients in the highest quartile of SBP (≥136 mm Hg) at the time closest to an event had higher risk of thromboembolism [HR 2.88, 95% CI (1.75–4.74); P < 0.001] and major bleeding [HR 1.61, 95% CI (1.02–2.53); P = 0.041]. Thus, BP control could be more important than history of HTN or baseline BP levels. In the current issue of American Journal of Hypertension, Ishii et al.17 investigated whether a history of HTN or SBP level is more associated with ischemic stroke or bleeding risk in the Fushimi AF Registry, a large community-based prospective survey of AF patients in 80 institutions from Fushimi-ku in Japan. Of 3,713 patients analyzed, 2,304 (62%) patients had previous HTN and the mean SBP at baseline was 128.2 ± 19.1 mm Hg. Patients with history of HTN had more comorbidities and higher CHADS2, CHA2DS2-VASc, and HAS-BLED score. The incidence of ischemic events (stroke and systemic embolism) was numerically higher in HTN (2.42%) than in the non-HTN group (1.53%), but nonsignificant [HR 1.19, 95% CI (0.89–1.60)]. For major bleeding, a similar trend was observed with numerically higher bleeding events in the HTN group (1.91%) than in non-HTN group (1.61%), but again, nonsignificant [HR 1.18, 95% CI (0.86–1.65)]. When the authors divided the HTN group between the presence of SBP ≥150 mm Hg at baseline, compared to SBP<150 mm Hg, patients with elevated SBP at baseline had significantly higher risk of ischemic events [HR 1.74, 95% CI (1.08–2.72)] and bleeding events [HR 2.01, 95% CI (1.21–3.23)], thus highlighting the importance of elevated SBP in patients with AF. Pulse pressure was also analyzed with similar results, and no differences were observed regarding the type of stroke. Nonetheless, this important analysis is limited by its dependence on baseline BP, with events recorded over follow-up, which may be years since the initial patient was recruited. This is the same approach by various biomarker studies that measure a biomarker (often nonspecific) and record events at 1 or 5 or 10 years later, when the risk profile (especially age) may have substantially changed. These results support the previous evidence in SPORTIF trials that BP at baseline or during the follow-up would be more important in predict clinically significant events, rather than simply history of HTN. More recently, a subanalysis of the ARISTOTLE trial,12 found that elevated BP at any point during the clinical trial was significantly associated with a 50% increase in risk of ischemic events [HR 1.53; 95% CI (1.25–1.86)] and hemorrhagic stroke [HR 1.85; 95% CI (1.26–2.72)]. The ROCKET-AF18 trial also showed that for every 10 mm Hg increase in screening SBP, the risk of stroke and systemic embolism increased but not bleeding risk. The study by Ischii et al. discusses the critical importance of BP control since stroke risk of the group with HTN and controlled BP is comparable to patients without history of HTN. These findings emphasize that ischemic and bleeding risks of AF patients are clearly dynamic and modifiable. Simple and dynamic assessment (using validated scores) are helpful in daily clinical practice to assess both ischemic and bleeding risks in anticoagulated AF patients (Figure 1). Association of hypertension with stroke and bleeding risk factors in atrial fibrillation. Nonetheless, many risk factors for bleeding are also risks factors for thromboembolism and many of them are reversible. The appropriate and responsible use of risk scores should be to “flag up” high-risk patients for more careful review and follow-up.19 Indeed, a high HAS-BLED score (≥3) is not an excuse to avoid oral anticoagulation in AF patients and periodic assessment and clinical correction of modifiable risk factors should be prioritized.6,20 Risk stratification is also not a “one off” assessment but a dynamic process and focus on potentially reversible risk factors should be emphasized. G.Y.H.L.: consultant for Bayer/Janssen, BMS/Pfizer, Biotronik, Medtronic, Boehringer Ingelheim, Microlife, and Daiichi-Sankyo. Speaker for Bayer, BMS/Pfizer, Medtronic, Boehringer Ingelheim, Microlife, Roche, and Daiichi-Sankyo. No fees are received personally. Other authors declared no conflict of interest. José Miguel Rivera-Caravaca is supported by Instituto Murciano de Investigación Biosanitaria (IMIB16/AP/01/06) and has received a grant from Sociedad Española de Trombosis y Hemostasia (SETH; grant for short international training stays 2016).