What are the predictors of white-coat hypertension in Chinese adults?

Tammy K W Tam 譚嘉渭, Kwok-keung Ng 吳國強, Cheuk-man Lau 劉卓民

HK Pract 2007;29:411-418

Summary

Objective: To investigate the risk factors associated with white-coat hypertension (WCH) in Chinese adults and to predict the possibility of WCH in patients with elevated blood pressure in clinic but normal blood pressure at home in clinical practice.

Design: Retrospective case series.

Subjects: From three primary care clinics. They had 24-hour ambulatory blood pressure monitor (ABPM) performed from January 2001 to October 2006.

Main outcome measures: Percentage of subjects having WCH, relationship between WCH and variables including age, gender, smoking, office systolic blood pressure (SBP), office diastolic blood pressure (DBP), body mass index (BMI), diabetes, anxiety, occupation, education level, and family history of cardiovascular diseases.

Results: The percentage of WCH in the study population was 28.2%. The mean age of the 617 patients was 52.9+9.5 years, mean BMI 24.0+3.2 kg/m2, mean office SBP 153.0+11.8 mmHg, and DBP 89.8+7.6 mmHg. Advanced age and lower BMI characterized subjects with the development of WCH as opposed to those with sustained hypertension.

Conclusion: WCH is a common and important phenomenon. In selecting patients for ABPM, use of home BP monitor, supplemented with identification of features including advanced age and lower BMI would enable the practicing physicians to determine more accurately which subjects were likely to benefit from ABPM.

Keywords: hypertension, white-coat; risk factors; Chinese; general practice

摘要

目的： 研究中國籍成年人出現白大衣性高血壓white-coat hypertension (WCH) 的相關危險因素， 並預測病人在診所內有白大衣性高血壓，但當中血壓正常的機率。

設計： 回顧性研究。

研究對象： 2001年1月至2006年10月期間，3間基層診所，佩帶24小時流動血壓檢察儀器的病人。

主要測量內容： 診所發生白大衣性高血壓現象的百分比，以及這種現象，與其他變數．包括年齡、性別、吸煙、 診所內的收縮壓、診所內的舒張壓、體重指數、糖尿病、焦慮、職業、教育水平以及家族心血疾病史的關係。

結果： 被研究者發生白大衣性高血壓的百分比為28.2%。617名病人的平均年齡是52.9+9.5歲，平均體重指數為24.0+3.2 kg/m2，平均診所收縮壓153.0+11.8mmHg，平均舒張壓DBP 89.8+7.6mmHg。診所內引起高血壓現象的病人較持續有血壓高的病人更為年長及體重指數較低。

結論： 白大衣性高血壓是常見及重要現象。通過家中血壓觀察,觀察病人，是否年長或體重指數偏低，可以幫助醫生更準確地選擇哪些病人更適合流動血壓檢查。

主要詞彙： 高血壓，白大衣性高血壓，危險因素，中國籍，全科醫學。

Introduction

White-coat hypertension (WCH) refers to abnormally elevated blood pressures in the medical environment and normal blood pressures during regular daily life.^1-2 The phenomenon of WCH is common. In the analysis in the Progetto Ipertensione Umbria Monitoraggio Ambulatoriale (PIUMA) database, prevalence of WCH was 33.3% in Joint National Committee Stage I hypertension (systolic BP 140 to 159 mmHg or diastolic BP 90 to 99 mmHg), 11.8% in Stage II (systolic BP 160 to 179 mmHg or diastolic BP 100 to 109 mmHg) and 3% in Stage III (systolic BP > 180 mmHg or diastolic BP > 110 mmHg).³

Regarding the risk of cardiovascular complications from WCH, the results from past cohort studies to compare cardiovascular events among patients with WCH versus those with sustained hypertension had been inconclusive. In general, WCH is regarded as a point along the continuum of hypertensive disease and poses lower cardiovascular risks than sustained hypertension. Therefore, long-term hypertensive treatment is not justified unless the patient has sustained hypertension, evidence of cardiovascular disease, or signs of target organ injury.^4-5 However, these patients should be regularly monitored because of an increased risk to develop sustained hypertension, and they should be given intensive non-pharmacological treatment including lifestyle modification, moderate salt restriction, weight reduction, regular exercise, smoking cessation and correction of glucose and lipid abnormalities.^6-8

Effective recognition of patients with WCH is important in clinical practice. The condition is common and failure to recognize it will lead to over-treatment and inappropriate use of medications and investigations, causing unnecessary side effects to patients and excessive financial burden to the healthcare system. Several hypertension guidelines even stipulate that all suspected WCH should have ABPM before implementing antihypertensive therapy.^9-11 In view of this, some overseas studies have attempted to find out the clinical characteristics of subjects who are susceptible to the development of WCH, so that more data are available for clinicians to estimate the probability of WCH when evaluating patients with elevated office blood pressure, particularly for those without home blood pressure measurements.^2,3,12-14 These data have offered additional help to the practicing physician in determining which subjects are more likely to benefit from ABPM. It is not certain whether similar results can be applied for Chinese population, as there may be racial differences for the factors involved, and similar studies for our locality have not been available. With an ageing population and rising prevalence of essential hypertension in our society, the implication of identifying WCH precisely and effectively in daily practice is expected to be great. Furthermore, WCH is believed to be secondary to an apparent stressor response during cuff measurement by a physician, even with multiple visits to the same physician.^15-17 Previous studies have not examined the effect of anxiety and other social factors on WCH. This study sought to define the clinical features of Chinese population that would predict the presence of WCH, and further, to investigate the relationship between anxiety, occupation and educational level and WCH.

Methods

Study design and selection criteria of subjects

This was a retrospective study conducted in three primary care clinics, and included subjects with suspected WCH, based on abnormally elevated blood pressures in clinic but normal blood pressures at home and had 24-hour ABPM performed, in the period from January 2001 to October 2006. Exclusion criteria were subjects already receiving anti-hypertensive treatment or those who had developed target organ damages such as stroke, transient ischaemic attack, peripheral vascular disease, hypertensive retinopathy, electrocardiographic evidence of left ventricular hypertrophy, raised serum creatinine and proteinuria. Moreover, recruited subjects should not be taking non-steroidal anti-inflammatory medications, sympathomimetics, or liquorice at time of monitoring; and they should not be having acute intercurrent illnesses, acute stressful events or unstable psychiatric conditions.

Procedure

The ambulatory blood pressure monitor used in this study was model TM-2420. It was validated by the British Hypertension Society and the US Association for the Advancement of Medical Instrumentation. Before referring a patient for 24-hour ABPM, there should be at least 3 abnormal blood pressure readings recorded in 3 separate clinic visits. The referring doctor had to complete a standardized referral form for the subject from which clinical information would be collected. These data included patient"s age and gender, body mass index (BMI), mean office blood pressure over previous three clinic visits, occupation, educational level, smoking history (ever or never smoked), presence of diabetes mellitus (American Diabetic Association diagnostic criteria), presence of anxiety or other psychological problems (indicated by presence of International Classification of Primary Care, second edition [ICPC-2] diagnostic codes of P01, P02, P04, P25, P27, P29, P74, P75, P76, P78, P79, P82, and P99 in case record), and family history of cardiovascular diseases.

Subjects were diagnosed sustained hypertension if the 24-hour ABPM showed systolic blood pressure > 140 mmHg or diastolic blood pressure > 90 mmHg, or WCH if systolic blood pressure < 135 mmHg and diastolic blood pressure < 85 mmHg, or inconclusive if the blood pressure readings were in-between these two defined levels. On the day of ambulatory monitor, at least 14 blood pressure readings from the time interval 9:00 to 21:00 should be obtained before the results were regarded as valid and interpretable. Otherwise, the same procedure would be repeated for patients with insufficient number of clinical data. The final result was determined by one of the doctors in the "Central panel of WCH" to ensure standardization of data interpretation.

Statistical analysis

Data analysis was performed with the Statistical Package for the Social Sciences version 10.0 (SPSS Inc. Chicago [IL], United States). Descriptive information for each explanatory variable was derived. Bivariate association of each variable with WCH was assessed by t-test for continuous variables and Chi-squared test for categorical variables. Multivariate logistic regression was applied for adjusting confounding relationship between different variables. A p value of < 0.05 was considered statistically significant.

Results

We recruited 795 subjects for 24-hour ABPM in the study; 617 (77.6%) subjects successfully completed the procedure and produced conclusive results. During the study period, there was no report of major complication which had required premature termination of the procedure.

We included 617 subjects with conclusive results for statistical analysis. Of this sample population, 23 (3.7%) were required to repeat the procedure once or more before conclusive results were obtained. The mean age of the subjects was 52.9+9.5 years, BMI 24.0+3.2 kg/m², mean systolic blood pressure in office 153.0+11.8 mmHg, and mean diastolic blood pressure in office 89.8+7.6 mmHg. Majority (91.7%) never smoked. Over 90% of them had achieved secondary education of above. Presence of anxiety was noted in 13.3% of subjects. (Table 1)

The percentage of WCH in our study sample was 28.2%. In bivariate analysis, we found that WCH was significantly associated with advanced age (p < 0.0001), female gender (p = 0.01), lower BMI (p = 0.0006), lower mean diastolic blood pressure in office (p < 0.0001), educational level (p = 0.02), of which university education was negatively associated with WCH, and occupation (p = 0.02), of which more labour workers were found to have WCH. Subjects who were non-smokers (p = 0.28) and anxious (p = 0.31) were not significantly more likely to be diagnosed WCH. To adjust for confounding variables, we applied multivariate logistic regression by selecting explanatory variables with p < 0.05 in bivariate analysis to enter the model. The logistic regression model showed that only advanced age and lower BMI were statistically associated with WCH. However, the area under receiver operating characteristics (ROC) was 0.64 only.

Discussion

WCH is a common condition in the hypertensive populations and in the population at large. The phenomenon of WCH may reflect an abnormally vigorous sympathetic response to the environment of the blood pressure measurement, especially in the presence of a nurse or physician.^18-19 Depending on definitions, different studies on Caucasians have variably reported that 21-58% of hypertensive patients without target organ injuries have WCH.^2,3,20 In the present study, we adopted the definition laid down by the European Society of Hypertension and diagnosed the condition of WCH when ambulatory blood pressure showed systolic BP < 135 mmHg and diastolic BP < 85 mmHg during the daytime period.²¹ With this definition, the percentage of WCH in our study population was 28.2%. In other words, as in Caucasians, the phenomenon of WCH also represented a significant entity in the diagnosis of hypertension in Chinese, because more than one fourth of the patients with elevated blood pressure in clinic would be inappropriately treated for hypertension if the possibility of WCH was overlooked by the case physician. Consequently, this would lead to overuse of antihypertensive agents in a substantial number of patients and render them to sustain unnecessary side effects of medications, not yet to mention the financial cost added on our already-strained healthcare economic structure by the unnecessary drug prescriptions.

Managing patients with episodic elevation of blood pressure in office is the bread and butter in general practice. In facing the diagnostic challenge as posed by the high prevalence of WCH, ABPM would play an essential role in establishing the diagnosis of hypertension before implementation of anti-hypertensive therapy. The technique is non-invasive, and the recording monitor reliable, accurate and convenient. They typically take readings every 15 to 30 minutes throughout the day and night while the patients go about their normal daily activities, and store the readings, which can then be downloaded onto a personal computer for analysis. There is also literature on its clinical utility and body of evidence showing ambulatory blood pressure levels are more closely correlated with cardiovascular morbidity.²² In our 5-year experience of performing 24-hour ABPM for patients with suspected WCH, general feedbacks from patients were positive. Out of our study population, most of them (617/795 or 77.6%) had successfully completed the procedure with conclusive results, and none of them had reported any major complications that had caused premature termination of the procedure. We also observed that for subjects documented to have sustained hypertension from ABPM, they were more willing to follow medical advice and comply with the prescribed anti-hypertensive treatment.

In most practice settings where there is resource limitation, it is not possible to admit all patients with elevated blood pressure in clinic to have ABPM to exclude WCH. To enable a more sensitive case selection, overall data from overseas studies indicate that the probability of WCH increases in subjects with lower office systolic BP 140 to 159 mmHg or diastolic BP 90 to 99 mmHg, female gender, lower BMI, nonsmokers, hypertension of recent onset, limited number of BP measurements in the office and small left ventricular mass.^{2,3,12-14,23,24} In line with some of these findings, our results also demonstrated that WCH was significantly associated with lower BMI (p = 0.0006) and suggested that female patient was more likely to develop WCH, though the relationship was not significant in logistic regression model. For the relationship with office blood pressure, we found that lower office diastolic BP was associated significantly with WCH in bivariate analysis, but this factor was excluded in the logistic regression model after adjusting for confounding variables. We might need to accumulate more clinical data in the near future to further establish the relationship between WCH and office BP. In previous studies, the role of age as an independent predictor of WCH is controversial.^2,12 In our study, advanced age was significantly associated with the development of WCH. The implication of detecting WCH in elderly patients is great because they are usually more susceptible to side effects and drug interactions of medications than younger subjects.

We were not aware of any previous research investigating the relationship between WCH and anxiety, education level and job nature. The present study demonstrated that subjects with anxiety were not more likely to develop WCH than those without. However, the diagnosis of this condition was based only on the presence of certain diagnostic codes in the case records, rather than on individual clinical assessment. Further study with more valid and standardized measurement of anxiety might be necessary to confirm its relationship with WCH. Nevertheless, persons prone to anxiety or stress might be more susceptible to development of sustained hypertension instead of WCH, as suggested by a study establishing that higher levels of self-reported occupational stress are predictive of greater ambulatory blood pressure among British doctors.²⁵ For education level and occupation, we found that they were significantly associated with WCH in bivariate analysis; their exclusion by the logistic regression model might be related to the insufficient sample size due to missing data. Interestingly, we observed subjects with university qualifications, or working in disciplinary forces were more likely to be diagnosed sustained hypertension; while those attaining only primary education or below, or working as labour workers were more likely to have WCH. These findings might have reiterated the role of stress in development of sustained hypertension.

Nonetheless, the discriminating power of our logistic regression model was not strong, as judged by its low ROC of 0.64 only. We proposed that when evaluating Chinese subjects with elevated BP in primary care setting, the presence of characteristics such as advanced age and low BMI would increase the pretest probability of WCH but were by no means able to replace the established indications for ABPM.²⁶

The major limitation of this study was that we recruited only subjects with home blood pressure readings available. In this regard, the actual prevalence of WCH in our population might be different from what we had estimated in the study. Also, patients with self blood pressure measurements might be more health conscious and socially advantageous than those without, and this discrepancy might have given rise to a different spectrum of clinical and social characteristics so derived. However, according to our estimation in a previous study, about 70% of hypertensive patients in our study setting did have home BP monitor.²⁷ The results of the current project might have already represented the majority of our patient population. In actual fact, the reference from home BP readings had enabled us to detect most, if not all, of our patients with WCH in the past. We would like to suggest here the importance of home blood pressure monitor in uncovering the condition of WCH in a heterogeneous group of patients presented with elevated blood pressure in clinic environment.

Conclusion

WCH is a common and important condition to be recognized for both Chinese and Caucasians who presented with elevated blood pressures in the medical environment. ABPM is an effective and reliable tool for detecting the phenomenon of WCH. Under resource limitations where ABPM could not be universally performed for patients with elevated blood pressures in clinics, the use of home BP monitor, supplemented with identification of features including advanced age and lower BMI would enable the practicing physicians to determine which subjects were more likely to benefit from ABPM.

Acknowledgement

We would like to express our sincere thanks to Dr Tsang Chiu Yee, Luke for his support on this study and establishment of ABPM in service. We also thank Dr Lau Kam Tong, Dr Lau Man Wai, Dominic and Dr Lai Wing Yiu, Stephen for their contribution in data entry in this project.

Key messages

1. White-coat hypertension is a common and important condition.
2. Ambulatory blood pressure monitoring is the standard diagnostic tool for detecting white-coat hypertension in clinical practice.
3. Home blood pressure monitoring serves important role in guiding clinician to diagnose white-coat hypertension.
4. Patients with advanced age and low BMI were more likely to have white-coat hypertension.