White-Coat Hypertension as a Risk Factor for the Development of Home Hypertension:  The Ohasama Study FREE

White-coat hypertension (WCHT) is a condition characterized by an elevated blood pressure (BP) in a medical setting, combined with normal self-measured or ambulatory-recorded BP. In contrast, sustained hypertension is the presence of an elevated BP regardless of setting or circumstance.¹ However, the clinical relevance of WCHT has not been established, and it is controversial whether this condition involves an increased cardiovascular risk. Results from cross-sectional and prospective studies have been contradictory; some had found increased cardiovascular risk in patients with WCHT,^2- 6 whereas others did not.^7- 11

It is also unknown whether this condition represents a transient state in the development of hypertension outside medical settings. One small and short-term study¹² reported a similar rate of transition to ambulatory hypertension (high BP during ambulatory readings) in subjects with sustained normotension (SNT) (normal BP during office and ambulatory readings) and in subjects with WCHT. Although 2 other studies^13- 14 reported a high rate for the development of ambulatory hypertension in subjects with WCHT, these studies did not have normotensive control subjects for comparison. Thus, these studies^13- 14 failed to demonstrate that WCHT was truly a transient state compared with SNT or that it posed a greater risk for progression to hypertension outside medical settings.

Self-measurements of BP at home (home BP measurements) make it possible to obtain multiple measurements over a long observation period under relatively controlled conditions.^15- 17 It has been reported that multiple measurements eliminate observer bias and regression dilution bias; therefore, home BP measurements are more reliable than conventional BP measurements taken in medical settings (office BP).^15- 17 A few studies have also reported that home BP measurements are better predictors of cerebrovascular and cardiovascular events compared with office BP.^18- 20

We have followed home BP measurements in a general population sample in Japan since 1987.^18,20- 22 The present 8-year follow-up study, conducted with 777 normotensive subjects 40 years or older, aims to quantitatively determine the risk of transition to hypertension outside medical settings (at home) in subjects with WCHT and to compare the risk with that in subjects with SNT.

This study was a part of a longitudinal observational study of subjects who have been participating in a BP measurement project in Ohasama, Japan. The socioeconomic and demographic characteristics of this region and full details of the project have been described elsewhere.^18,20- 22 The study protocol was approved by the institutional review board of Tohoku University School of Medicine, Sendai, Japan, and by the Department of Health of the Ohasama Town Government.

Based on several guidelines,^15- 17 subjects with a home systolic BP of 135 mm Hg or higher and/or a home diastolic BP of 85 mm Hg or higher were classified as having high home BP, while others were classified as having normal home BP. Similarly, those who had office systolic/diastolic BP of 140/90 mm Hg or higher were classified as having high office BP. White-coat hypertension was defined as the occurrence of high office BP and normal home BP. Sustained normotension was defined as the occurrence of normal office BP and normal home BP. Development of home hypertension (hypertension based on home BP measurements) was defined as either progression to high home BP or start of treatment with antihypertensive medication.

The selection of subjects for this study has been reported previously.¹⁸ Briefly, the subjects were 40 years or older and were residents from 3 of 4 subdivisions of Ohasama (n = 2716). Hospitalized persons (n = 121) and persons with dementia or those who were bedridden (n = 31) were excluded from the study. Individuals who worked out of town (n = 575) were also excluded because the project involved consistent and routine ambulatory BP monitoring. Informed consent to participate in the study was given by 1957 of 1989 eligible individuals. Home BP measurements at baseline were conducted among 1913 subjects who collected their own BP data at least 3 days during the 4-week study period. This criterion was based on our previous observation that the average BP value obtained for the first 3 days was not significantly different from the values obtained for the entire study period.²¹ We had previously confirmed that these 1913 subjects were a representative sample of the general population.¹⁸ In the present analysis, to compare the risk for development of hypertension between WCHT and SNT, we excluded subjects who did not have office BP measured (n = 124) or those who regularly used antihypertensive medication (n = 582). We further excluded 235 subjects who had home hypertension at baseline. Thus, we followed up the remaining 972 participants.

Physicians and public health nurses instructed subjects on how to perform home BP measurements.²¹ Subjects were asked to measure their BP every morning within 1 hour of waking, while seated and rested for more than 2 minutes, for 4 weeks. The home BP of an individual was defined as the mean of all measurements obtained for that person. The mean (SD [range]) number of baseline home BP measurements was 23.2 (6.8 [3-60]). The clinical utility of those methods of home BP measurements in this project has been previously reported.^18,20- 22 The same procedure was used for the follow-up home BP measurements.

Annual health check-ups including BP measurements are available to all Japanese citizens 40 years or older once per year. Two consecutive measurements of BP are taken by a nurse or technician, using a semiautomatic device, after the subject has been seated at rest for at least 2 minutes.²¹ The office (screening) BP was defined as the average of the 2 readings and was obtained in the same year as the initiation of home BP measurement.

Home BP was measured with the HEM401C (Omron Healthcare Co Ltd, Kyoto, Japan), a semiautomatic device, based on the cuff-oscillometric method that generates a digital display of systolic and diastolic BP at baseline. We also used the HEM401C and HEM747ICN devices for measurement at follow-up measurement. The screening BP was measured with an USM-700F (UEDA Electronic Works Co Ltd, Tokyo, Japan), a fully automatic device, based on the Korotkoff sound technique (a microphone method). The circumference of the arm was less than 34 cm in most cases, so we used a standard arm-cuff for both BP measurements. All devices used in this study were validated^23- 24 and satisfied the criteria of the Association for the Advancement of Medical Instrumentation.²⁵

Residential status in the town of Ohasama on October 31, 1999, was confirmed using the residents’ registration cards, which were considered accurate and reliable because they are required for pension and social security benefits in Japan. Information on smoking status, obesity, family history of hypertension, and a history of hypercholesterolemia or diabetes mellitus was obtained from questionnaires sent to each household during the time of home BP measurements and from the medical records at Ohasama Hospital. Ohasama Hospital is the only hospital in the town, and more than 90% of the participants go there for regular checkups.

All data are given as mean (SD). The association between the baseline BP and the likelihood of progression to home hypertension was investigated using multiple logistic regression models, adjusted for age, sex, smoking status, obesity (body mass index [BMI; calculated as weight in kilograms divided by the square of height in meters] ≥25), family history of hypertension, and a history of hypercholesterolemia or diabetes mellitus. In all analyses, we treated the subjects with SNT as the reference group. Variables were compared using the unpaired, 2-tailed t test, χ² test, or analysis of variance, as appropriate. Differences of P<.05 were considered statistically significant. All analyses were conducted using the SAS package (version 8.2; SAS Institute Inc, Cary, NC).

Among the 972 subjects with WCHT or SNT, who did not take antihypertensive medication at the time of baseline survey, 60 died or moved away from the town before the follow-up. Of the remaining 912 subjects, 777 (85%) took part in the follow-up home BP measurements. Mean duration of the period between the baseline and the follow-up home BP measurements was 8.2 (2.0) years. The mean number of follow-up home BP measurements was 23.7 (5.6).

The baseline characteristics of the subjects in each group are presented in Table 1. The mean (SD) age of the 777 subjects was 56.0 (8.7) years and the proportion of men was 34.0%. Mean (SD) office and home systolic/diastolic BP values were 125.9 (15.8)/72.5 (10.3) mm Hg and 116.0 (9.1)/70.4 (7.2) mm Hg, respectively. Of the 777 subjects, 649 (83.5%) were classified as having SNT, while the remaining 128 (16.5%) were classified as having WCHT. Office and home BP values and the proportion of men were significantly higher in subjects with WCHT than among subjects with SNT. Subjects with WCHT tended to be older and have higher proportions of obesity and a more prominent family history of hypertension than those with SNT.

Development of home hypertension was defined as either progression to high home BP or the start of treatment with antihypertensive medication. At baseline, 649 subjects had SNT and 128 subjects had WCHT. At the time of follow-up measurements, 144 subjects (22.2%) with SNT and 60 (46.9%) with WCHT developed home hypertension. The rate was significantly higher in subjects with WCHT (P<.001). The significantly higher rate of development of home hypertension in subjects with WCHT was observed for both hypertension defined by a home BP of 135/85 mm Hg or higher (71 [10.9%] of 649 subjects with SNT and 31 [22.7%] of 128 subjects with WCHT [P<.001]) and hypertension defined by the start of treatment with antihypertensive medication (73 [11.2%] of 649 subjects with SNT and 29 [24.2%] of 128 subjects with WCHT [P<.001]).

Table 2 gives the rates of development of home hypertension, home BP values, and the magnitude of changes for home BP from baseline to follow-up. (Categories were divided according to baseline home systolic and diastolic BP values.) In subjects with WCHT, as well as those with SNT, rates for development of home hypertension showed a significant trend (P<.001), as the baseline home BP values increased across the categories (Table 2). Across a broad range of baseline home BP levels, compared with subjects with SNT, rates of development of home hypertension were higher in most subjects with WCHT (Table 2). In both groups, subjects with a baseline home systolic BP of 108 mm Hg or lower or a diastolic BP of 52 mm Hg or lower had a less than 5% chance of developing home hypertension (Table 2). Similar tendencies were observed regarding the rate for those who developed home hypertension defined by the start of treatment with antihypertensive medication (Table 2).

The odds ratio (OR) for WCHT to progress to home hypertension, adjusted for other factors, was significantly higher than that of SNT (OR, 2.86; P<.001) (Table 3). In this multivariate model, older age, male sex, and obesity also significantly predicted the development of home hypertension (Table 3). The significant OR of WCHT for progression to home hypertension was similarly observed for both hypertension defined by a home BP of 135/85 mm Hg or higher (OR, 3.09; 95% confidence interval [CI], 1.83-5.23 [P<.001]) and hypertension defined by the start of treatment with antihypertensive medication (OR, 2.71; 95% CI, 1.61-4.56 [P<.001]).

Because higher home BP values at baseline were associated with a significantly higher rate of development for home hypertension (Table 2), we adjusted for home systolic/diastolic BP values separately, in addition to the previously mentioned factors. Although higher home BP levels were significantly and independently associated with the risk of home hypertension in the multivariate model (home systolic BP: OR (per 1–mm Hg increase), 1.11; 95% CI, 1.08-1.14 [P<.001]; home diastolic BP: OR (per 1–mm Hg increase), 1.10; 95% CI, 1.07-1.13 [P<.001]), WCHT remained a significant predictor of home hypertension (adjusted home systolic BP: OR, 1.81; 95% CI, 1.16-2.82 [P = .009]; adjusted home diastolic BP: OR, 2.36; 95% CI, 1.55-3.61 [P<.001]). Subgroup analysis of home BP levels at baseline also showed similar results (in subjects with a home BP lower than 125/80 mm Hg²⁶: OR of WCHT, 2.24; 95% CI, 1.25-4.01 [P = .007]; in subjects with a home BP of 125/80 mm Hg or higher and lower than 135/85 mm Hg: OR of WCHT, 1.84; 95% CI, 0.94-3.60 [P = .08]). There was no significant interaction between the subgroup of home BP levels and the presence of WCHT on the risk of development of home hypertension (P = .90).

This 8-year follow-up study demonstrated that WCHT was a significant predictor of the development of home hypertension, independent of other confounding factors and baseline home BP levels. The risk of developing home hypertension was consistently higher in subjects with WCHT than in those with SNT, starting at a threshold of 108/58 mm Hg and across a broad range of baseline home BP levels. Our results indicate that WCHT could pose a greater risk for progression to hypertension outside medical settings even if home BP values were within completely normal range.

Some studies reported that WCHT is associated with hyperreactivity to stress and higher sympathetic nerve activity.^27- 29 Because stress has been reported to be an independent risk factor in the development of hypertension,^30- 31 it is possible that a higher reactivity to stress in medical environments leads to WCHT, which in turn contributes to higher rates of progression to home hypertension in subjects with WCHT.

Only 1 study has ever compared the risk of developing hypertension outside medical settings in subjects with WCHT with those with SNT¹²; results showed that the transition to ambulatory hypertension occurred in a similar way in subjects with SNT and in subjects with WCHT. Those results were inconsistent with our findings; however, in that study the follow-up period was shorter (3.5 years) and the sample size was too small (36 subjects with WCHT and 56 subjects with SNT) to reliably determine the risk of WCHT. In the present study, we followed 649 subjects with SNT and 128 subjects with WCHT for 8-years and found high risk in subjects with WCHT.

Although 3 prospective studies reported the prognostic significance of WCHT compared with normotensive control subjects, the results were controversial: 2 studies^7,10 with short duration follow-up periods (mean of <5 years) showed similar lower cardiovascular risk for subjects with WCHT compared with normotensive control subjects, but 1 study² with a longer follow-up (10 years) demonstrated higher risk in subjects with WCHT compared with normotensive controls. In a recent 3-year follow-up study¹⁹ of hypertensive patients taking antihypertensive medication, cases of isolated uncontrolled hypertension at the office had a similar risk of cardiovascular events compared with subjects with sustained controlled hypertension. These results and the present 8-year follow-up results suggest that WCHT could potentially represent a cardiovascular risk after 10 or more years. We are pursuing follow-up with research subjects to find the answer to this hypothesis.

To define WCHT, BP information obtained outside medical settings (home BP measurements or ambulatory BP monitoring) is necessary. Ambulatory BP monitoring provides a wide variety of BP information outside medical settings, and thus it may offer a more reliable definition of WCHT. However, because ambulatory BP monitoring is not easily achieved in typical clinical settings, it is not necessarily practical as a method to determine WCHT. On the other hand, home BP measurements are now widely recommended by professionals who practice in clinical settings in most developed countries^15- 17 (in Japan, 30 million devices for self BP measurement at home have been distributed³²). For these reasons, we based the definition of WCHT on actual and practical home BP measurements. However, applicability of the present findings to WCHT using a definition based on ambulatory BP remains to be investigated.

Development of home hypertension was defined as either progression to high home BP or start of treatment with antihypertensive medication at follow-up. Because subjects with WCHT had a higher office BP level, it is possible that the higher risk of developing hypertension might be attributable to receiving antihypertensive medication according to high office BP levels. However, WCHT showed a significant risk for progression to home hypertension by the definition of home BP levels as well as the start of treatment with antihypertensive medication. Furthermore, the rate of developing home hypertension, defined by the start of treatment with antihypertensive medication, was consistently higher in subjects with WCHT than in those with SNT, across a broad range of baseline home BP levels. Therefore, these findings suggest that WCHT is a transitional condition leading to home hypertension irrespective of high office BP levels. The objective of this study was to investigate home BP, and therefore no office BP data were collected at the time of follow-up. We recommend that the evaluation of subjects with SNT who develop WCHT is a topic for future research.

The present 8-year follow-up study based on home BP measurements demonstrated that subjects with WCHT had an approximately 2-fold higher risk of eventually manifesting home hypertension compared with those who had SNT. Although the prognostic significance of WCHT remains unclear, these results suggest that WCHT is not a totally benign condition. Further follow-up studies targeting cardiovascular outcomes are needed to clarify whether WCHT is a potentially dangerous condition. In the meantime, patients with WCHT should be carefully monitored.

Correspondence: Yutaka Imai, MD, PhD, Department of Clinical Pharmacology and Therapeutics, Tohoku University Hospital, 1-1 Seiryo-machi, Aoba-ku, Sendai 980-8574, Japan (imai@tinet-i.ne.jp).

Accepted for Publication: February 15, 2005.

Financial Disclosure: None.

Funding/Support: This work was supported by Grants for Scientific Research (12877163, 13470085, 13671095, 15790293, and 14010301) from the Ministry of Education, Culture, Sports, Science, and Technology, Tokyo, Japan; Health Science Research Grants on Health Services (13170201, 13072101, and H12-Medical Care-002) and grant H15-Gan Yobou-039 from the Ministry of Health, Labor, and Welfare, Tokyo; research grants from Junkanki-byo Itaku Kenkyu 11C-5, Tokyo (1999-2000); the Japan Atherosclerosis Prevention Fund, Tokyo (2000-2003); the Uehara Memorial Foundation, Tokyo (2002); a grant from the Japan Cardiovascular Research Foundation, Tokyo (2002); and the Takeda Medical Research Foundation, Osaka, Japan (2003).