It’s Time to Take This Method Seriously

The common phenomena of white-coat and masked hypertension have established the need for assessing blood pressure (BP) out of the office, particularly using 24-hour ambulatory monitoring (ABPM). In the last 2 decades, evidence on the usefulness of the alternative method for out-of-office BP assessment, namely home BP monitoring (HBPM), has accumulated, and guidelines on using this method have recently been published in the United States and Europe.1,2 As with ABPM, HBPM allows the detection of white-coat and masked hypertension and has additional advantages, such as wide availability, low cost, and excellent acceptability by hypertensive patients for repeated use.1,2 In this issue of Hypertension, Niiranen et al present the results of Finn-Home, an outcome study of HBPM in the general population in Finland.3 Strengths of this study are the large data set ( 14,000 subject-years, with 162 documented cardiovascular events) and the use of optimal methodology for office BP measurement (nurses using mercury sphygmomanometers) and HBPM (validated electronic device and guidelinesrecommended monitoring schedule with 7-day duplicate morning and evening measurements). This study contributes to the HBPM and outcome database that now comprises 8 large long-term prospective studies3–10 including 17,688 subjects and almost 100,000 person-years of follow-up (Table). There are important differences among these studies regarding the study population, the BP measurements, the definition of primary outcomes, and the methods of statistical adjustment. Studies have been performed in Europe and Japan. Five studies included general population samples,3–5,7,9 1 included patients registered in a single primary care practice after excluding those with previous cardiovascular disease,8 and 2 focused on treated hypertensives.6,10 Office BP was measured using conventional mercury sphygmomanometers, apart from the Ohasama study,4 which used electronic devices; and that of Okumiya et al,5 which did not report on office measurements. In 3 studies, office BP was measured in a single visit (2 to 3 readings),3,4,8 and in another 3 studies, it was measured in 2 visits (6 readings)6,7,9 (unclear in 1 study10). HBPM was self-performed by participants using electronic devices, apart from the Flanders study,8 where HBPM was performed in a single occasion by a physician or assistant physician using a mercury device. The HBPM schedule also varied considerably among studies, with 5 of them having morning and evening measurements3,5–7,9 and 3 having only morning measurements.4,8,10 The Ohasama, SHEAF, FinnHome, and Okumiya et al studies obtained 20 to 28 home readings,3–6 Didima obtained 12 readings,9 Flanders obtained 3 readings8 and PAMELA obtained only 2 readings7 (readings not specified in J-HEALTH10). Variations in the definitions of outcomes and adjusting factors in the multivariate analyses are shown in the Table. Despite the considerable methodological and clinical heterogeneity of these studies, a comparison of the prognostic ability of home versus office BP measurements is possible, because in most of the studies both methods have been applied in the same subjects.3,4,6–9 All these studies consistently showed home BP to be a significant predictor of cardiovascular events (Table). More importantly, in most of the studies, the cardiovascular risk was better predicted by HBPM than by office measurements (Table).3,4,6–8 Even when the same number of home versus office measurements were compared, the prognostic ability of HBPM was again superior, suggesting that its advantage is not attributed only to the larger number of readings.4,7 A random-effects summary of the adjusted hazard ratios yields values of 1.015 (95% CI 1.010 to 1.020) and 1.024 (1.017 to 1.032) per mm Hg for systolic and diastolic home BP versus 1.007 (1.004 to 1.011) and 1.015 (1.003 to 1.026), respectively, for office BP. This difference is beyond chance for systolic BP. A more informative analysis would be to consider both HBPM and office BP measurements in the same adjusted model. This has been done in different analytic approaches in 5 studies. The Finn-Home,3 SHEAF,6 Ohasama,4 and Flanders8 studies found that HBPM, but not office BP, was significantly predictive when both were considered, whereas the PAMELA study7 found that both office and HBPM conferred independent information. One would like to see also formal reclassification analyses that examine whether the addition of HBPM changes the classification of participants into categories where different treatment would be indicated11 and thus outcomes could be improved. Ideally, reclassification evaluations could be done in an individuallevel data meta-analysis to allow standardization of the The opinions expressed in this editorial are not necessarily those of the editors or of the American Heart Association. From the Hypertension Center (G.S.S.), Third University Department of Medicine, Sotiria Hospital, Athens, Greece; Department of Hygiene and Epidemiology (K.C.M.S., J.P.A.I.), University of Ioannina School of Medicine, Ioannina, Greece. Correspondence to George S. Stergiou, Third University Department of Medicine, Sotiria Hospital, 152 Mesogion Avenue, Athens 11527, Greece. E-mail [email protected] (Hypertension. 2010;55:1301-1303.) © 2010 American Heart Association, Inc.