Prognostic significance of blood pressure measured in the office, at home and during ambulatory monitoring in older patients in general practice

(2005) 19, 801 807 & 2005 Nature Publishing Group All rights reserved 0950-9240/05 $30.00 www.nature.com/jhh ORIGINAL ARTICLE Prognostic significance of blood pressure measured in the office, at home and during ambulatory monitoring in older patients in general practice RH Fagard, C Van Den Broeke and P De Cort Department of Molecular and Cardiovascular Research, Hypertension and Cardiovascular Rehabilitation Unit, Faculty of Medicine, University of Leuven, UZ Gasthuisberg, Leuven, Belgium The purpose of the study was to assess the prognostic significance of out-of-the-office blood pressure (BP) measurement in older patients in general practice, and to compare the results for BP measured in the office, at home and during 24-h ambulatory monitoring. All registerd patients who were 60 years or older were eligible for the study, except when bedridden, demented or admitted in a home for sick elderly people, or when they had suffered a myocardial infarction or stroke. After baseline measurements in 1990 1993, incidence of major cardiovascular events (cardiovascular death, myocardial infarction and stroke) was ascertained in 2002 2003 and related to the BPs by use of multivariate Cox regression analysis. Age of the 391 patients averaged 7179 years; 40% were men. During median follow-up of 10.9 years, 86 patients (22%) suffered a cardiovascular event. The adjusted relative hazard rate, associated with a 1 s.d. increment in systolic BP was 1.13 for office BP (NS), and, respectively, 1.32, 1.33 and 1.42, for home, daytime and night time BP (Pp0.01 for all). Results were similar for diastolic BP. The prognostic significance of all out-of-the-office BPs was independent of office BP. The prognostic value of home BP was equal to (systolic) or even better (diastolic) than that of daytime BP. Night time BP predicted cardiovascular events independent of all other BPs. Prognosis of whitecoat hypertension was similar to that of true normotension, but better than in sustained hypertension. In conclusion, the prognostic value of home BP is better than that of office BP in older patients in primary care, and is at least equal to that of daytime ambulatory BP. The prognosis of patients with white-coat hypertension is similar to that of true normotensives. (2005) 19, 801 807. doi:10.1038/sj.jhh.1001903; published online 16 June 2005 Keywords: ambulatory blood pressure; cardiovascular events; home blood pressure; office blood pressure; prognosis; white-coat hypertension Introduction Most of the management of hypertension occurs in the primary care setting, in which decisions on antihypertensive treatment are mainly based on blood pressure (BP) measurements in the office. In the last decade, a number of studies have reported on the prognostic significance of out-of-the-office BP and showed that out-of-the-office BP has greater Correspondence: Professor RH Fagard, Department of Molecular and Cardiovascular Research, Hypertension and Cardiovascular Rehabilitation Unit, Faculty of Medicine, University of Leuven, UZ Gasthuisberg Hypertensie, Herestraat 49, KU Leuven, B-3000 Leuven, Belgium. E-mail: robert.fagard@uz.kuleuven.ac.be The study was presented at the 14th European Meeting on Hypertension, Paris, France, June 13 17, 2004. Received 16 February 2005; revised and accepted 17 May 2005; published online 16 June 2005 prognostic value than office or clinic BP. 1 16 These studies have been perfomed in the population, 3,5 in the elderly, 10 in hypertensive individuals in general, 1,2,4,7,11,13,15 in patients with isolated systolic hypertension, 8,9 in hypertensives under treatment 12,14 and in patients with refractory hypertension. 6 Out-of-the-office BP was assessed by ambulatory BP monitoring in all but two of these studies, in which home BP measurement was used. 5,14 In addition, only one study addressed this issue in general practice. 14 The study population was restricted to treated elderly hypertensives and out-of-the-office BP was assessed by home BP measurements. There are currently no data on ambulatory BP monitoring and outcome in primary care. Furthermore, the prognostic value of BP measured in the office, at home and during ambulatory monitoring has not been directly compared in the same study population. In the late

802 1980s, we designed a study in which BP was measured in the office, in the patient s usual living situation at home and by ambulatory BP monitoring in older patients in one general practice. Our primary aim was to compare the prognostic significance of the three types of BP measurement in this setting. In addition, the data allowed us to compare the outcome of white-coat hypertensives with true normotensives recruited from the same population, whereas most studies except one 14 recruited hypertensives and normotensive controls from different populations. 2,10,13,15 The importance of the issue is highlighted by the high prevalence of hypertension in older patients in the primary care setting, which amounted to 74% based on office BP measurement in 3761 patients in 253 general practices in Belgium. 17 Materials and methods The study protocol was approved by the institutional Ethics Committee and all participants gave informed consent. The study was performed in one primary care practice in Flanders, Belgium, in which 2044 patients were registered, of whom 462 were at least 60 years old. Patients were excluded from the current study when they were bedridden, demented, admitted in a home for sick elderly people or had suffered a myocardial infarction or a stroke. Baseline examinations were performed between November 1990 and May 1993. Patients were visited at home or contacted by telephone to explain the study protocol, which involved a visit to the office, a home visit and 24-h ambulatory BP monitoring. The office and home visits were performed within 2 weeks and in random order. The office visit consisted of a medical history, including medication and smoking habits, a standardized clinical examination and measurements of height, weight, heart rate and BP. After 5 min of rest, BP was measured three times in the sitting position by the auscultatory technique using a mercury sphygmomanometer and an appropriately sized cuff. BP was measured on the right arm, except when BP was lower by X10 mmhg on the right arm than on the left arm. Finally, a blood sample was taken for measurement of serum cholesterol. The same BP measurements were performed at the patient s home; the patient was informed on the date a few days before the visit. Office and home BPs were always measured by the same investigator, most often by the primary care physician, otherwise by one assistant physician. We used the average of the three sitting BPs for all analyses involving office and home BP. Finally, 24-h ambulatory BP monitoring was performed by use of the Spacelabs 90202 or 90207 device, with appropriate cuff size. BP was measured every 15 min from 0800 to 2200 h and every 30 min from 2200 to 0600 h. For the analysis, daytime BP was defined as the weighted average of all measurements between 1000 and 2000 h; night time was taken from 0000 to 0600 h. 18 We also divided the patients into four subgroups on the basis of office BP and daytime ambulatory BP. Office BP was considered normal when systolic BP was o140 mmhg and diastolic BP o90 mmhg; the cut off values for daytime ambulatory BP were 135 and 85 mmhg, respectively. 19 True normotension was defined as a normal office and ambulatory BP, and sustained hypertension corresponded to an elevated office and ambulatory BP. White-coat hypertension and masked hypertension were defined as isolated office and isolated ambulatory hypertension, respectively. Follow-up The vital and health status of the patients was ascertained between January 2002 and December 2003. The aggregate of stroke, myocardial infarction and cardiovascular death was the primary end point of the current study. All events were corroborated by the investigators according to established criteria 20 and without knowledge of the baseline data. Statistical analysis Database management and statistical analyses were performed using SAS software, version 6.12 (SAS Institute Inc.). Data are reported as mean7standard deviation (s.d.) or as percentages. BP measurements in different conditions were compared by use of paired t-tests and relationships between BPs by univariate regression analysis. Differences among groups were analysed by analysis of variance and Scheffe s multiple means test. The w 2 -test was used for categorical data. We used Cox regression analysis to assess the prognostic significance of baseline characteristics and of the various BPs. For BP, the relative hazard rates (RHR) represent the risk associated with a 1 s.d. increment in BP. In multivariate Cox regression analysis, we adjusted for age, gender, body mass index (BMI), use of BP-lowering drugs, smoking, serum total cholesterol and diabetes at baseline. To assess whether the effect was independent from other BP measurements, we performed additional consecutive adjustments for the other BPs. We also assessed the prognostic significance of white-coat hypertension, masked hypertension and sustained hypertension in comparison with true normotension. A two-tailed value of Pp0.05 was considered significant. Results Patients Of the 462 patients, 42 patients (9.1%) were not eligible because they were bedridden (n ¼ 3), demented (n ¼ 2), admitted in a home for sick elderly people (n ¼ 7), or had suffered a myocardial infarction (n ¼ 16), a stroke (n ¼ 12) or both (n ¼ 2) at entry.

Since the main aim of the study was to compare the prognostic value of different BP measurements, only patients in whom at least two types of BP measurement were available were retained for the analyses. On this basis we excluded 20 of the 420 patients (4.8%) because only one type of BP was available, mainly due to lack of collaboration or to failed BP measurements. Finally, nine of the 400 remaining patients (2.2%) were lost to follow-up. Patient characteristics at baseline Table 1 summarizes the general characteristics of the 391 remaining patients. The oldest patient was 99 years old. Women were on average 3.0 years older than men (Po0.001). All systolic BPs were significantly different from one another (Po0.001), except office and home BP (P ¼ 0.54). The paired differences for diastolic BP were significant (Po0.01), except for the comparison of office and daytime BP (P ¼ 0.20). The regression equations and coefficients of determination of the relationships between office, home, daytime and night time BP are given in Table 2. All systolic BPs increased significantly with age ( þ 0.12prp þ 0.25; Pp0.05). Office, home and daytime diastolic BP decreased with age ( 0.13prp 0.24; Pp0.01), but this was not the case for night time BP (P ¼ 0.65). The night day ratio of systolic (r ¼þ0.31; Po0.001) and of diastolic BP (r ¼þ0.21; Po0.001) was higher in older than in younger subjects. 803 Table 1 Baseline characteristics of all patients and of untreated and treated patients, and univariate relative hazard rates for cardiovascular events for all patients All Untreated Treated RHR (95% CI) Number (%) 391 (100) 265 (68) 126 (32) Age (year) 7179 7079 7279* 3.03 (2.44 3.76)*** Gender, men (n (%)) 157 (40) 126 (48) 31 (25)*** 1.21 (0.79 1.85) NS Body mass index (kg/m 2 ) 27.574.7 26.774.6 29.174.6*** 0.92 (0.73 1.15) NS Current smokers (n (%)) 74 (19) 62 (23) 12 (9.5)*** 1.19 (0.71 2.00) NS Serum cholesterol concentration (mmol/l) 6.3971.27 6.3171.31 6.5571.17 NS 0.79 (0.64 0.98)* Diabetes mellitus (n (%)) 33 (8.4) 19 (7.2) 14 (11) NS 2.70 (1.52 4.80)*** Blood pressure (mmhg) a Systolic Office 142.8722.1 136.7720.6 155.6719.6*** 1.44 (1.16 1.80)*** Home 143.6722.9 137.8720.6 156.2722.7*** 1.63 (1.33 1.99)*** Daytime 130.7713.9 128.0712.6 136.2714.9*** 1.46 (1.19 1.79)*** Night time 117.5717.3 114.5715.8 124.2718.5*** 1.89 (1.56 2.30)*** Night/day ratio 0.9070.10 0.9070.10 0.9270.10 NS 1.63 (1.32 2.01)*** Diastolic Office 77.5711.2 76.2710.5 80.0712.1** 0.78 (0.62 0.98)* Home 74.9710.1 74.3710.0 76.2710.3 NS 1.17 (0.94 1.45) NS Daytime 76.779.1 76.678.1 76.8711.0 NS 0.93 (0.74 1.17) NS Night time 65.1710.2 64.979.8 65.4711.2 NS 1.28 (1.03 1.58)* Night/day ratio 0.8570.11 0.8570.11 0.8670.11 NS 1.49 (1.19 1.85)*** Values are means7s.d. or number (percentages). Relative hazard rates (RHR) (95% CI) reflect the risk associated with a 1 s.d. increment for a continuous variable and with presence vs absence for a dichotomous variable. n: number of patients. *Pp0.05; **Pp0.01; ***Pp0.001; NS P40.05, for respectively, the comparison of treated and untreated patients, and the relative hazard rates. a Number of BP measurements: office: 376; home: 373; daytime: 374; night time: 364. Table 2 Relationships between different methods of BP measurement BP measurement n Systolic BP Diastolic BP Intercept Slope R 2 * Intercept Slope R 2 * Home vs office 358 44.9 0.61 0.44 38.8 0.47 0.25 Daytime vs office 359 77.6 0.37 0.35 45.6 0.40 0.25 Night time vs office 352 57.4 0.42 0.28 38.0 0.35 0.14 Daytime vs home 356 78.5 0.36 0.36 51.0 0.34 0.14 Night time vs home 347 60.2 0.40 0.28 46.9 0.24 0.06 Daytime vs night time 364 68.5 0.53 0.45 41.7 0.54 0.36 Abbreviations: BP: blood pressure; n: number or BP measurements in each comparison; R 2 : coefficient of determination. *Po0.001 for all coefficients of determination.

804 As shown in Table 1, patients under antihypertensive treatment were older, comprised more women than men, and had higher BMI and systolic BP than the untreated patients. Office hypertension, defined as sitting systolic BPX140 mmhg or diastolic BPX90 mmhg or taking antihypertensive drugs, was present in 62% of the patients, of whom 54% were on antihypertensive therapy. Finally, Table 3 summarizes baseline characteristics in the four subgroups according to office BP and daytime ambulatory BP, irrespective of antihypertensive treatment. Follow-up Median follow-up after the baseline examinations was 10.9 (range: 0.04 13.0) years and total follow-up amounted to 3378 patient-years. Overall, 165 patients died, 69 from a cardiovascular cause and 92 from a noncardiovascular cause; the cause of death could not be identified in four cases. Noncardiovascular causes of death included cancer (n ¼ 33), internal diseases (n ¼ 52) and peri-operative death (n ¼ 7). The primary end point comprised 86 first cardiovascular events. Among the 31 first nonfatal events, there were 13 strokes and 18 myocardial infarctions. The first event was cardiovascular death in 55 patients (stroke: 11; myocardial infarction: five; sudden death: 18; congestive heart failure: 17; peripheral vascular disease: two; cardiac surgery: one; and pulmonary embolism: one). Table 1 gives the crude RHRs for the primary endpoint for patient characteristics and BP. Age, diabetes mellitus and all systolic BPs were highly significant predictors in univariate analysis. Table 4 gives the adjusted RHRs for the various systolic and diastolic BPs. All BPs independently predicted cardiovascular events, except office BP. Among the covariates, only age, gender and diabetes at baseline were significant in the multivariate Cox models. The risk of a cardiovascular event increased by B15% per 1 year increment in age (Po0.001), was B3 times higher in men than in women (Pp0.01), and B2 times greater in diabetics than in nondiabetics (0.01oPo0.10). Table 5 summarizes the RHRs after additional consecutive adjustment for other BP measurements. Home BP, and daytime and night time ambulatory BP predicted cardiovascular events, independent Table 4 Adjusted relative hazard rates for cardiovascular events with different methods of BP measurement BP measurement a Relative hazard rate (95% CI) P-value Systolic Office 1.13 (0.88 1.45) 0.34 Home 1.32 (1.06 1.64) 0.01 Daytime 1.33 (1.07 1.64) o0.01 Night time 1.42 (1.16 1.74) o0.001 Diastolic Office 1.04 (0.82 1.34) 0.73 Home 1.40 (1.14 1.72) 0.001 Daytime 1.26 (1.00 1.59) 0.05 Night time 1.40 (1.12 1.75) o0.01 Relative hazard rates are for each 1 s.d. increment in BP, with adjustment for age, gender, BMI, use of BP-lowering drugs, smoking, serum total cholesterol and diabetes at baseline. a Number of BP measurements: office: 376, home: 373, daytime: 374, night time: 364. Table 3 Baseline characteristics in four subgroups, classified according to office and daytime ambulatory BP, irrespective of antihypertensive treatment NT WCHT MHT SHT P-value* Number (% of all patients) 136 (38) 87 (24) 31 (8.6) 105 (29) Age (year) 69.277.9 70.078.2 68.077.5 71.979.2 0.04 Gender, men (n (%)) 65 (48) 24 (28) a 17 (55) 37 (35) 0.005 Body mass index (kg/m 2 ) 27.275.0 27.173.7 28.374.4 28.075.2 0.39 Current smokers (n (%)) 35 (26) 7 (8.0) a 9 (29) 13 (12) o0.001 Serum cholesterol concentration (mmol/l) 6.4471.29 6.6571.19 5.7471.14 a,b 6.4171.27 0.008 Diabetes mellitus (n (%)) 6 (4.4) 9 (10) 0 (0) 13 (12) 0.03 Use of BP-lowering drugs (n (%)) 19 (14) 37 (43) a 9 (29) 51 (49) a o0.001 Blood pressure (mmhg) Office Systolic 122.9710.8 153.9711.5 a 128.478.1 b 163.2716.7 a,b,c (o0.001) Diastolic 72.577.9 80.0710.4 a 75.279.3 82.7713.0 a,c (o0.001) Daytime Systolic 120.078.7 125.976.4 a 139.878.5 a,b 145.3710.1 a,b,c (o0.001) Diastolic 73.476.0 72.476.7 84.077.0 a,b 82.8710.2 a,b (o0.001) Results are for 359 patients in whom both office BP and daytime ambulatory BP were available. Values are means7s.d. or numbers (percentages). n: number of patients; NT: normotension; WCHT: white-coat hypertension; MHT: masked hypertension; SHT: sustained hypertension. Statistical analysis: *overall P-value; a,b,c refer to significant differences between groups (Pp0.05); a compared with NT; b compared with WCHT; c compared with MHT.

Table 5 Relative hazard rates for cardiovascular events with different methods of BP measurement, after additional consecutive adjustment for other BP measurements 805 BP measurement a Relative risks (95% CI) Adjusted for Office Home Daytime Night time Systolic BP Office 0.91 NS (0.66 1.26) 0.92 NS (0.72 1.34) 0.96 NS (0.72 1.29) Home 1.42* (1.05 1.91) 1.26 NS (0.95 1.67) 1.24 NS (0.95 1.63) Daytime 1.40** (1.07 1.82) 1.19 NS (0.90 1.56) 1.21 NS (0.84 1.51) Night time 1.43** (1.13 1.80) 1.30* (1.02 1.65) 1.32* (1.02 1.72) Diastolic BP Office 0.81 NS (0.62 1.07) 0.81 NS (0.61 1.08) 0.81 NS (0.60 1.07) Home 1.55*** (1.23 1.97) 1.37** (1.08 1.74) 1.37** (1.07 1.69) Daytime 1.44** (1.10 1.89) 1.09 NS (0.83 1.42) 1.12 NS (0.85 1.47) Night time 1.49** (1.16 1.92) 1.30* (1.03 1.65) 1.32* (1.01 1.72) Relative hazard rates are for each one standard deviation increment in BP with adjustment for age, gender, BMI, use of BP-lowering drugs, smoking, serum cholesterol and diabetes at baseline, and with additional consecutive adjustment for other BPs. *Pp0.05; **Pp0.01; ***Pp0.001; NS P40.05. a Number of subjects in analyses with simultaneous inclusion of office and home BP: 358; office and daytime BP: 359; office and night time BP: 352; home and daytime BP: 356; home and night time BP: 347; daytime and night time BP: 364. of office BP. Diastolic but not systolic home BP added prognostic precision to daytime and night time ambulatory BP. Finally, the predictive value of night time BP was independent of all other BP measurements. Whereas the night day ratio of systolic and diastolic BP predicted cardiovascular events before adjustment (Table 1), the RHR was only significant for systolic BP after adjustment (RHR for þ 1 s.d.: 1.24; 95% CI, 1.01 1.50; Po0.05). However, the RHR was no longer significant (P ¼ 0.16) after additional adjustment for mean 24-h BP. Figure 1 shows the RHRs for the different types of hypertension vs subjects with normal BP. In the analysis on all patients, with adjustment for covariates including antihypertensive treatment, the risk for cardiovascular events was similar in white-coat hypertension and true normotension (P ¼ 0.85). The risk of sustained hypertension was 2.54 times greater than that of normotension (P ¼ 0.005) and 2.16 times greater than in white-coat hypertensives (95% CI, 1.16 4.01; P ¼ 0.01). Results were similar in untreated subjects. The risk ratio of B1.5 of masked hypertension vs normotension did not reach statistical significance. Discussion The major findings of the present study are (1) that the superiority of out-of-the-office BP over office BP for the prediction of cardiovascular events also applies to older patients in general practice; (2) that the prognostic value of BP measured in the patient s home is at least equal to that of daytime ambulatory BP; (3) that night time BP is a better predictor of cardiovascular events than daytime Figure 1 Relative hazard rates (RHR) for cardiovascular events, and 95% confidence intervals (95% CI) of white-coat hypertension (WCHT), masked hypertension (MHT) and sustained hypertension (SHT) vs normotension (NT), in all patients (K) and in untreated patients (J), with adjustment for age, gender, BMI, smoking, serum total cholesterol, diabetes at baseline in all analyses, and antihypertensive treatment in the analyses on all patients. Results are for 359 patients in whom both office BP and daytime ambulatory BP were available. Ne: number of events; Np: number of patients. BP; (4) that the outcome of white-coat hypertensives is similar to that of normotensives; and (5) that the prognosis of sustained hypertension is worse than that of normotension and white-coat hypertension. The observation that the BP measured by the family physician in the office was not independently

806 predictive for future cardiovascular events, whereas the BP measured by the same investigator in the patient s usual living situation by the same methodology was predictive, suggests that the person who measured the BP and the number of measurements were of less importance than the circumstances of the measurement to explain the difference in prognostic power between home and office BP. Average diastolic BP was significantly lower at home than in the office. However, systolic BP was similar in the two conditions, suggesting that the visit to the office influenced BP to variable extents in individual patients with consequent loss of predictive power. The fact that office BP did not predict the incidence of cardiovascular events after control for a number of covariates is compatible with most reports on out-of-the-office BP 3,5,7,11,14 with few exceptions. 12 The significant prognostic value of home BP in elderly patients registered in primary care extends the findings of Ohkubo et al 5 and of Bobrie et al. 14 These authors observed in a Japanese rural population 5 and in treated elderly hypertensives in general practice in France, 14 that selfmeasured home BP, by use of an automated device, significantly predicted cardiovascular mortality and/or morbidity, independent of a number of confounders and screening 5 or office 14 BP, which were not significant predictors by themselves. The present study confirms in the primary care setting that ambulatory BP is a better predictor of cardiovascular events than office BP. 1 4,6 13 However, the prognostic value of BP measured by the physician during one home visit was at least similar to that of daytime ambulatory BP. We also considered the prognostic significance of various types of hypertension. The classification was based on office BP and daytime ambulatory BP, for which normal values have been established. 19 The results confirm that patients with hypertension in the office and out-of-the office have a greater incidence of cardiovascular events than true normotensives and/or white-coat hypertensives. 2,4,6,9 11,14 It is, however, less well known if the prognosis of white-coat hypertensives differs from that of true normotensives, in which the selection of the study populations in previous studies has to be taken into account. Studies that recruited white-coat hypertensives and normotensives from different populations found that the incidence of cardiovascular events or stroke was similar in these two groups, 2,10,15 or that the risk was higher in white-coat hypertension. 13 In our study and in that of Bobrie et al, 14 normotensives and hypertensives were drawn from the same study population and both studies concluded that prognosis was similar in the two groups. Only a small number of our patients had a normal office BP and an elevated ambulatory BP, and the contention that patients with so-called masked hypertension have a worse prognosis than normotensive subjects 14,21 cannot be confirmed nor refuted in the current study. Studies that reported on daytime and night time BP separately found that both BPs carried significant prognostic information. 3,7,8,12 Staessen et al 8 directly compared the prognostic value of the two BPs and found that night time BP was a better predictor than daytime BP in older patients with isolated systolic hypertension. We confirmed this finding in an elderly population in primary care. Results on the nocturnal fall of BP or on the night day ratio are not consistent. Some studies observed a significantly better prognosis in patients with a greater decline in night time BP 2,8,22 but this was not confirmed by others. 12,23,24 We found that a lower night day ratio of systolic BP was indeed associated with a lower incidence of cardiovascular events, but this was no longer significant after additional adjustment for the average 24-h BP. Several limitations have to be considered. In the current study, home BP was the BP measured by the investigator in the patient s home, whereas other studies on the prognostic significance of home BP reported on self-measured BP by use of an automated device. 5,14 Office BP was the average of three BPs during one visit, whereas guidelines recommend BP measurements at several visits, at least for the initial patient evaulation. 25 In addition, Bovet et al 26 showed that triplicate readings on two visits could be a minimal strategy for assessing adequately the prevalence of hypertension at the population level. However, it is not certain if the results of population studies apply to the primary care situation of our study, in which registered patients were familiar with the physician, the technique and the circumstances of the BP measurement. Moreover, also home BP consisted of three BP measurements during one single visit, and did predict the incidence of cardiovascular events. A number of patients were on antihypertensive treatment at baseline but it is unlikely that this influenced the results. Use of BP-lowering drugs was included as a covariate in the analyses, but never reached statistical significance in the multivariate models, and separate analyses in untreated patients yielded What is known on this topic and what this study adds What is known on this topic Out-of-the-office BP had greater prognostic value than office BP in different populations, but there are few data in the primary care setting. Whereas most studies used ambulatory BP monitoring only two studies used BP measurements in the patient s home. Prognosis of white-coat hypertension is worse than that of sustained hypertension, but it is not well known if prognosis differs from that of true normotensives. What this study adds Both ambulatory BP and home BP have greater prognostic value than office BP in older patients registered in general practice. The prognostic value of BP measured in the patients home is at least equal to that of ambulatory BP. The outcome of white-coat hypertensives is similar to that of true normotensives.

similar results. In fact, the prognostic significance of the out-of-the-office BP has not only been shown in untreated subjects but also in hypertensive patients under treatment. 6,12,14 Finally, as in the other studies except in one controlled trial, 8,9 the management of hypertension and other clinical conditions was not standardized during follow-up. A number of patients were institutionalized, treated in other general practices or moved out of the area, and some were unavailable for follow-up. Acknowledgements We gratefully acknowledge the secretarial assistance of N Ausseloos. References 1 Perloff D, Sokolow M, Cowan RM, Juster RP. Prognostic value of ambulatory BP measurements: further analysis. J Hypertens 1989; 7(Suppl 3): S3 S10. 2 Verdecchia P et al. Ambulatory BP: an independent predictor of prognosis in essential hypertension. Hypertension 1994; 24: 793 801. 3 Ohkubo T et al. Prediction of mortality by ambulatory BP monitoring versus screening BP measurements: a pilot study in Ohasama. J Hypertens 1997; 15: 357 364. 4 Khattar RS, Senior R, Lahiri A. Cardiovascular outcome in white-coat versus sustained mild hypertension: a 10-year follow-up study. Circulation 1998; 98: 1892 1897. 5 Ohkubo T et al. Home BP measurement has a stronger predictive power for mortality than does screening BP measurement: a population-based observation in Ohasama, Japan. J Hypertens 1998; 16: 971 975. 6 Redon J et al. Prognostic value of ambulatory blood pressure monitoring in refractory hypertension: a prospective study. Hypertension 1998; 31: 712 718. 7 Khattar RS et al. Prediction of coronary and cerebrovascular morbidity and mortality by direct continuous ambulatory BP monitoring in essential hypertension. Circulation 1999; 100: 1071 1076. 8 Staessen JA et al. Predicting cardiovascular risk using conventional and ambulatory blood pressure in older patients with systolic hypertension. JAMA 1999; 282: 539 546. 9 Fagard RH et al. Response to antihypertensive therapy in older patients with sustained and nonsustained systolic hypertension. Circulation 2000; 102: 1139 1144. 10 Kario K et al. Silent and clinically overt stroke in older Japanese subjects with white-coat and sustained hypertension. J Am Coll Cardiol 2001; 38: 238 245. 11 Celis H et al. Cardiovascular risk in white-coat and sustained hypertensive patients. Blood Pressure 2002; 11: 352 356. 12 Clement DL et al. Prognostic value of ambulatory blood pressure recordings in patients with treated hypertension. N Engl J Med 2003; 348: 2407 2415. 13 Gustavsen PH, Hoegholm A, Bang LE, Kristensen KS. White-coat hypertension is a cardiovascular risk factor: a 10-year follow-up study. J Hum Hypertens 2003; 17: 811 817. 14 Bobrie G et al. Cardiovascular prognosis of masked hypertension detected by blood pressure self-measurement in elderly treated hypertensive patients. JAMA 2004; 17: 1342 1349. 15 Pierdomenico SD et al. Cardiovascular and renal events in uncomplicated mild hypertensive patients with sustained and white-coat hypertension. Am J Hypertens 2004; 17: 876 881. 16 Fagard RH, Celis H. Prognostic significance of various characteristics of out-of-the-office blood pressure. J Hypertens 2004; 22: 1663 1666. 17 Fagard RH, Van den Enden M, Leeman M, Warling X. Survey on treatment of hypertension and implementation of World Health Organization/International Society of Hypertension risk stratification in primary care in Belgium. J Hypertens 2002; 20: 1297 1302. 18 Fagard R, Brguljan J, Thijs L, Staessen J. Prediction of the actual awake and asleep blood pressures by various methods of 24-h pressure analysis. J Hypertens 1996; 14: 557 563. 19 O Brien E et al. European Society of Hypertension recommendations for conventional, ambulatory and home blood pressure measurement. J Hypertens 2003; 21: 821 848. 20 Staessen JA et al. Randomized double-blind comparison of placebo and active treatment for older patients with isolated systolic hypertension. Lancet 1997; 350: 757 764. 21 Bjorklund K et al. Isolated ambulatory hypertension predicts cardiovascular morbidity in elderly men. Circulation 2003; 107: 1297 1302. 22 Ohkubo T et al. Prognostic significance of the nocturnal decline in blood pressure in individuals with and without high 24-h blood pressure: the Ohasama study. J Hypertens 2002; 20: 2183 2189. 23 Khattar RS et al. Effect of aging on the prognostic significance of ambulatory systolic, diastolic, and pulse pressure in essential hypertension. Circulation 2001; 104: 783 789. 24 Bjorklund K et al. Prognostic significance of 24-h ambulatory blood pressure characteristics for cardiovascular morbidity in a population of elderly men. J Hypertens 2004; 22: 1691 1697. 25 Guidelines Committee. 2003 European Society of Hyypertension European Society of Cardiology guidelines for the management of arterial hypertension. J Hypertens 2003; 21: 1011 1053. 26 Bovet P et al. Assessisng the prevalence of hypertension in populations: are we doing it right? J Hypertens 2003; 21: 509 517. 807