Author + information
- Rajiv Agarwal, MD∗ ()
- Indiana University School of Medicine and Richard L. Roudebush Veterans Administration Medical Center, Indianapolis, Indiana
- ↵∗Reprint requests and correspondence:
Dr. Rajiv Agarwal, Department of Medicine, Indiana University School of Medicine, Richard L. Roudebush Veterans Administration Medical Center, 1481 West 10th Street, Indianapolis, Indiana 46033.
The debate continues about the ideal target blood pressure (BP) goal in those with essential hypertension. However, it has unambiguously emerged that compared with BP measured in the clinic, BP measured at home has greater prognostic significance (1). Perhaps more importantly, an analysis of randomized trials comparing clinic with home BP suggests that small (2.63/1.68 mm Hg) but significant reductions in BP are possible with home BP monitoring (2). Physicians are more likely to increase antihypertensive drugs in response to home BP monitoring; thus, home BP has the power to reduce therapeutic inertia and therefore hypertension control (2). Home BP monitoring is now recommended by guidelines (3,4). Yet few data exist on the comparative evaluation of home or clinic BP with hard outcomes in randomized trials. Furthermore, few data exist with cohort studies with specific outcomes, such as strokes and myocardial infarctions.
The purpose of HONEST (Home blood pressure measurement with Olmesartan Naive patients to Establish Standard Target blood pressure) study of 21,591 Japanese patients, reported in this issue of the Journal, was to investigate the relationship between BP measured either at home (early morning and before bedtime) or in a physician's office and the subsequent incidence of cardiovascular events in olmesartan medoxomil-naive hypertensive patients taking this drug for a 2-year follow-up period (5). The overall results of this study were previously reported (6). This report addresses the comparison of home BP with clinic BP on the separate outcomes of stroke and coronary artery disease. Included were: 1) patients whose BP levels could be measured in a physician's office at least once within 28 days before starting the study drug; and 2) patients whose early morning BP levels could be measured by an arm cuff BP monitor at home at least twice on different days within 28 days before administration of the study drug. Excluded were patients with previous myocardial infarction, stroke, hospitalization due to heart failure within the previous 6 months, or those who had undergone coronary intervention surgery or procedures. Specifically, there was no BP goal targeted, and the management of the patient’s BP was left to the treating physician.
The investigators report the outcomes of stroke and coronary artery disease in 5 BP categories for morning home BP and clinic BP. The BP categories represent not the baseline BP (which was excluded from the analysis) but the average BP during antihypertensive drug treatment over ∼2 years of follow-up. The reference category of systolic home BP was <125 mm Hg and of systolic clinic BP was <130 mm Hg. Compared with the reference category, an average on-treatment systolic home BP in the morning of ≥155 mm Hg was associated with a hazard ratio (HR) of stroke of 6.01 and of coronary artery disease of 6.24. Compared with the reference category, the average on-treatment systolic clinic BP in the morning of ≥160 mm Hg was associated with an HR of stroke of 5.82 and an HR of coronary artery disease of 3.51. Although the ability to predict strokes was similar for morning systolic home BP and clinic BP, compared with clinic BP, the ability of morning systolic home in predicting coronary artery disease was superior.
The use of on-treatment average BP is difficult to translate into clinical practice. Most of the BP reduction in this study occurred in 16 weeks, after which BP on average was stable throughout the 2 years. Using baseline and the change from baseline to 16 weeks in the 2 BP types is easier to translate into clinical practice.
The association of morning surge in systolic BP using ambulatory BP monitoring and elderly hypertensive patients with stroke was previously reported (7). A clinician would certainly ask whether one needs morning home BP to draw the association between the level of BP and outcomes. The study in fact reported evening home BP and overall home BP (the average of morning and evening BPs), and the data show that using all home BP recordings instead of just the morning ones may be easier and at least as strongly associated if not more strongly associated with stroke and coronary artery disease. Thus, this study does not specifically demonstrate the value of morning home BP over the overall home BP.
Compared with clinic BP, the reasons that associate home BP and coronary artery disease more strongly need an explanation. Here we can only speculate, but I will posit that the association is causal. The average baseline home BP was 151.2 and clinic BP was 153.6. Compared with clinic BP, there was greater precision in the measurement of home BP (SD: 16.3 vs. 19.0 mm Hg). Patients were treated with olmesartan without a specified target BP. For clinic BP, those with a BP <130 mm Hg were treated with 1.5 ± 0.9 antihypertensive medications and those with a BP ≥160 mm Hg with 1.6 ± 1.1 medications. In contrast, for home BP, those with <125 mm Hg were treated with 1.3 ± 0.8 medications and those with ≥155 mm Hg with 1.7 ± 1.2 medications. If the physicians titrated antihypertensive medications based on home BP and better controlled home BP, then it follows that the relationship between better home BP control and coronary artery disease would be stronger. Treatment of home BP-guided therapy may better control masked uncontrolled hypertension and ignore the relatively benign white-coat effect. In an analysis of home BP data from 5 countries, data show that among treated hypertensive patients (n = 1,451) compared with treated controlled patients, patients with masked hypertension (low office and high home BP) and uncontrolled hypertension (high office and home BP) had higher cardiovascular risk (adjusted HR: 1.76; 95% confidence interval: 1.23 to 2.53; p = 0.002 for masked uncontrolled hypertension and adjusted HR: 1.40; 95% confidence interval: 1.02 to 1.94; p = 0.04 for uncontrolled hypertension) (8). However, this reasoning still fails to explain why the HRs of stroke were similar whether BP was measured in the clinic or at home.
However, it is possible that the association is not causal, and we can speculate on several confounding variables. Before we consider confounding, we should reflect on validity of the endpoint. Coronary artery disease was defined as myocardial infarction and revascularization for angina. There were 45 myocardial infarctions but 77 revascularizations for angina in the study; sudden death was excluded. The results were driven by the softer endpoint of revascularization, and therefore a larger study with more hard endpoints is required to confirm the findings. The cause-and-effect relationship in this study is difficult to link because it was not a randomized trial. The association of BP and outcomes may be confounded by adherence to medications and response to treatment. Second, clinic BP can only be measured if the patients visit the clinic. If missing data on clinic BP or home BP are not at random, they may selectively affect the association of home or clinic BP with outcomes. Multiple comparisons and false discovery are other reasons to be cautious.
The strengths of the study are its large size, with >20,000 patients, and impressive follow-up, because only 9% patients were lost. With respect to strokes, it shows that home and clinic BPs are similar and, for coronary artery disease, that home BP is superior. In light of the SPRINT (Systolic Blood Pressure Intervention Trial) that showed better cardiovascular outcomes with a lower clinic systolic BP goal, the question that emerges is whether lower BP goals can be achieved using home BP monitoring (9). Several randomized trials suggest that this is feasible (10,11), and the large pragmatic HONEST study suggests that it is time to design a randomized trial similar to SPRINT with treatment targets dictated by—not clinic BP—but home BP assessments. If home BP can predict cardiovascular outcomes at least as well as clinic BP, it is time to test this strategy for cost-effectiveness, convenience, and, ultimately, its capability to relieve cardiovascular morbidity and mortality with the simple expedient of home BP monitoring.
↵∗ Editorials published in the Journal of the American College of Cardiology reflect the views of the authors and do not necessarily represent the views of JACC or the American College of Cardiology.
Dr. Agarwal is supported by grant 5I01CX000829-04 from the Veterans Affairs Research and Administration; and he has received consulting and personal fees from Bayer, Abbvie, Johnson & Johnson, Boehringer Ingelheim, Takeda, Daiichi-Sankyo, Amgen, Celgene, and Eli Lilly.
- American College of Cardiology Foundation
- Pickering T.G.,
- Miller N.H.,
- Ogedegbe G.,
- et al.
- Kario K.,
- Saito I.,
- Kushiro T.,
- et al.
- Kario K.,
- Pickering T.G.,
- Umeda Y.,
- et al.
- McManus R.J.,
- Mant J.,
- Roalfe A.,
- et al.