Clinical characteristics of isolated clinic hypertension

Abbreviations ABPM: ambulatory blood pressure monitoring; BMI: body mass index; ICH: Isolated clinic hypertension; ESH/ESC: European Society of Hypertension / European Society of Cardiology; ROC: receiving operator charactaristic; SBP: systolic blood pressure; SEH-LELHA: Sociedad Española de Hipertensión-Liga Española para la Lucha contra la Hipertensión Arterial (Spanish Hypertension Society).

Isolated clinic hypertension (ICH), or white-coat hypertension, is defined by persistently elevated blood pressure in the office, with values of ≥ 140/90 mmHg, and normal 24-h or diurnal ambulatory pressure values (under 125/80 mmHg). The diagnosis is based on the findings obtained by 24-h ambulatory blood pressure monitoring (ABPM) or self-measurement in the home (mean readings over several days) [1].

The prevalence of ICH varies according to the criteria employed. To date, different studies have been made taking as reference diurnal BP values of under 135/85 mmHg or under 130/80 mmHg [2-5], and discussion focuses on whether, in the future, systolic or diastolic BP should receive priority attention for establishing the diagnosis [6].

Doubts persist as to the long-term prognosis of ICH [7,8]. Agreement is lacking as to the cardiovascular risk of patients with ICH, which appears to be intermediate between the risk in normotensive individuals and in subjects with established hypertension. On the other hand, a proportion of patients with ICH are known to progress towards sustained hypertension [9,10].

The increase in ambulatory measurements, based on the utilization of ABPM according to international guidelines, will allow diagnostic optimization and avoid classification errors based on the BP magnitudes in the stratification of cardiovascular risk.

The present study analyzes the clinical characteristics of patients with ICH, diagnosed via ABPM, and compares them with those of hypertensive individuals without pharmacological treatment, in the context of the ABPM registry in Spain (Cardiorisc project). An evaluation is also made of the positive and negative predictive values in relation to the suspected diagnosis of ICH.

The present study forms part of a general project of the Spanish Hypertension Society (Sociedad Española de la Hipertensión, SEH-LELHA), known as the National ABPM Registry. The project design has been published elsewhere [11,12].

A total of 1126 investigators were trained in 50 seminars/workshops (each lasting 3 h), held throughout Spain. The participants received instruction via a webpage on the 24-h monitoring technique and uniform application of the protocol. The knowledge of the participants was upgraded online (www.cardiorisc.com ) and by different meetings held during congresses of the SEH-LELHA. The first patient was recruited in July 2004. Since then, a monthly average of 1500 ABPM recordings have been included, with a total in the year 2006 of over 40 000 ABPM in the central registry. Suspected ICH was recorded by the investigator in the case report form, prior to ABPM. The protocol was approved by the ethics committees of each participating center, and is formally supported by the European Society of Hypertension.

A cross-sectional, comparative multicenter descriptive study was conducted in primary care centers and hypertension units in the rural and urban settings in all Spanish Autonomous Communities. The study was carried out by the investigators of the Cardiorisc Network, with the sampling of consecutive cases. The present study included 6176 patients over 18 years of age and diagnosed with essential hypertension, without pharmacological antihypertensive treatment for at least 4 weeks before ABPM, prescribed by their personal physician. The study excluded patients with arrhythmias (atrial fibrillation, frequent ectopic ventricular beats) that could interfere with the oscillometric recording of blood pressure, as well as those ABPM (Spacelabs 90207, Richmond, Washington, USA) recordings failing to use appropriate cuff sizes, as established by arm circumference [13].

The ABPM monitors were new and recently validated by a Y-connector measurement in parallel to a mercury sphygmomanometer. The ABPM monitors were programmed to record pressure readings every 20 min over a 24-h period. The patients were advised to go about their normal activities during the recording period. The activity and sleeping periods were individualized and adjusted according to the indications provided by the patients. Obese patient cuffs were used where necessary. Recordings with valid reading percentages of under 80% were discarded, as were those recordings lasting under 24 h, recordings with periods of over 1 h without any pressure reading, or involving fewer than 14 measurements during the activity period and fewer than seven readings in the resting period [14].

In the consulting office, for each patient, two baseline sitting pressure readings were obtained after 5 min of rest, using a calibrated mercury sphygmomanometer or validated oscillometric equipment, along with heart rate, according to the standardized recommendations of the European Society of Hypertension (ESH)/European Society of Cardiology (ESC) 2003 guidelines [1]. A large cuff for obese patients was used when needed.

The information collected included the indications for ABPM, age, sex, body weight and height, body mass index (BMI), abdominal circumference (abdominal obesity being defined as a circumference of ≥ 102 cm in males, or ≥ 88 cm in females), the duration of hypertension, smoking (regular daily consumption), diabetes (according to the criteria of the American Diabetes Association [15]), dyslipidemia [total cholesterol > 250 mg/dl, or low-density lipoprotein (LDL)-cholesterol > 155 mg/dl, or high-density lipoprotein (HDL)-cholesterol < 40 mg/dl in males, or < 48 mg/dl in females, or current treatment with lipid-lowering drugs] and family history of premature cardiovascular disease (< 55 years of age in males, or < 65 years in females). Information was also obtained on target organ damage, including the presence or absence of left ventricle hypertrophy (based on the electrocardiographic criteria of Sokolow and/or Cornell, or echocardiographic criteria, left ventricle mass index ≥ 125 g/m² in males or ≥ 110 g/m² in females), radiological signs of arteriosclerosis and impaired renal function (serum creatinine elevations of 1.3-1.5 mg/dl in males or 1.2-1.4 mg/dl in females, and/or microalbuminuria, albumin-creatinine index 22-300 mg/g in males or 31-300 mg/g in females, or albuminuria 30-300 mg/24 h). The presence of associated clinical conditions was also documented: coronary disease, cerebrovascular disease, heart failure, renal failure (serum creatinine and/or albuminuria/proteinuria above the previously defined limiting values) and peripheral vascular disease. The waking-sleep periods reported by the patients were also recorded.

The mean 24-h pressure values were registered, along with the waking and sleep periods, the circadian blood pressure pattern defined according to the variation in mean systolic blood pressure (SBP) during the activity and resting periods [extreme dipper: reduction of > 20% in SBP between resting and activity periods; dipper: reduction of ≥ 10% in SBP between resting and activity periods; nondipper: reduction of < 10% in SBP between resting and activity periods; riser: increase (> 0%) in SBP between resting and activity periods].

ICH was defined by blood pressure in the office ≥ 140 mmHg (systolic) or ≥ 90 mmHg (diastolic), with a daytime BP < 135 and < 85 mmHg (ICH1) or daytime BP < 130 and < 80 mmHg (ICH2) or 24-h BP < 125 and < 80 mmHg (ICH3).

The data were introduced in a database and processed using the statistical packages SPSS/PC version 14.0 (SPSS Inc., Chicago, Illinois, USA) and Stata 9.2 (StataCorp LP, College Station, Texas, USA). Descriptive analyses were made to summarize the global information. The qualitative variables are expressed as number (percentages), and the quantitative variables as means (SD). Statistical significance was considered for 5% in two-sided testing for all analyses.

The differences between groups were evaluated by means of the χ² test for qualitative variables and Student's t-test or analysis of variance (ANOVA) for quantitative variables - or the corresponding nonparametric tests, as required. The association between isolated clinic hypertension and the risk factors considered was analyzed using a nonconditional binary logistic regression model.

For selection of the variables of the final multivariate model, the methodology proposed by Hosmer and Lemeshow [16], Cobo [17] and Greenland [18] was used. The initial multivariate model incorporated the variables found to be significant in the bivariate analysis at a 25% level of significance, along with the variables considered to be clinically relevant.

Exploratory analysis (method of fractional polynomials) confirmed the appropriateness of modeling the continuous variable, age, as nonlinear. Possible confounding factors, interactions and collineality were studied. The collinearity of the maximal models were evaluated using the criteria proposed by Belsley [19].

The importance of each variable in the regression models was verified by the Wald statistic. In addition, all the models have been compared using the partial likelihood ratio test for nested models and the Akaike's Information Criterion (AIC) for non-nested models. Fitting of the model was evaluated according to the Hosmer-Lemeshow test and area under the receiving operator characteristic (ROC) curve. The following variables were introduced in all maximal models: age, sex, family antecedents of premature cardiovascular disease, the duration of hypertension (years), obesity, abdominal circumference, smoking, diabetes, dyslipidemia, target organ damage, ischemic heart disease, stroke, renal and heart failure, mean clinical blood pressure, heart rate and circadian profile.

A total of 6176 hypertensive individuals were included, with a mean age of 51.8 ± 14.1 years, 2842 (46%) females. ICH1, ICH2 and ICH3 criteria were met by 1807 (29.2%), 960 (15.5%) or 1133 (18.3%) subjects, respectively. Their baseline and comparative characteristics are shown in Table 1.

The Pearson correlation between clinic BP and daytime BP was 0.513/0.503 (P < 0.0001), systolic and diastolic, respectively, and between daytime BP and 24-h BP it was 0.981/0.981 (P < 0.0001), systolic and diastolic, respectively.

In the comparative bivariate analysis between the patients with ICH1, ICH2 and ICH3 criteria and the rest of the patients with sustained hypertension, significant differences were found in terms of female sex (more prevalent in the ICH group). The patients with ICH were also older. Significant differences were also observed in other variables. Smoking was more prevalent in the patients with sustained hypertension. In contrast, BMI, increased waist circumference and systolic clinical blood pressure were significantly higher in the three groups of patients with ICH. The patients with ICH1 and ICH2, in turn, showed a greater prevalence of nondipper or riser circadian pressure profiles. They also had lower 24-h ambulatory and daytime pressures, as a consequence of the actual definition of ICH, although it must be stressed that the patients with ICH also presented lower pressure values during the nocturnal resting period (Table 1).

Variables associated with ICH were the female sex, age 55 or more, increased waist circumference or BMI, clinical blood pressure and a nondipper circadian profile. When we considered the less restrictive ICH definition (ICH1), dyslipidemia and heart failure were also associated with ICH (Tables 2 and 3). Other variables were associated with sustained hypertension: smoking, family premature cardiovascular disease and the duration of diagnosed hypertension. Considering the 24-h ICH definition (ICH3), mean office BP did not predict ICH, and dipping circadian profile was found to be predictive (Table 4).

In 2611 cases (42.2%), ABPM was prescribed due to suspected ICH. The sensitivity and specificity of the physician in relation to suspected ICH1, ICH2 and ICH3 were 48.7 and 60.4%, 52.9 and 59,7%, and 52.3 and 60.0%, respectively 

Isolated clinic hypertension (ICH) is considered to constitute an intermediate risk situation between normal blood pressure and sustained hypertension [20]. There is long-term prognostic evidence that white-coat hypertension is not an innocent condition [21]. However, in our study we identified a series of variables associated with ICH that suggest that the condition may pose greater cardiovascular risk than reported in the literature. In comparison with the sustained hypertensive individuals, the patients with ICH tended to be older, more obese, with a nondipper circadian profile, and with higher systolic blood pressure values in the consulting office. This latter observation is very difficult to explain, but may have been caused partially by errors in the measurement of clinical blood pressure, or by the known alert (or white-coat) reaction of patients when subjected to pressure measurement by healthcare personnel. Probably, if more baseline pressure readings had been obtained and the first readings of the analysis had been eliminated, the mean office blood pressure would have been lower, mainly in the group of patients with ICH. On the other hand, when 24-h ICH3 criteria were used, the association between clinic BP and diagnosis of ICH was lost (Table 4). One possible explanation to this discrepancy between analyses based on 24-h averages and those based on daytime/night-time averages is that daytime sleep would be included in some patients in the awake average. This would falsely lower the awake average, and then the subject would be falsely categorized as having ICH. Although nowadays daytime sleep is not as usual as it was in the past in Spain, we don't know how many patients slept during daytime. That could be a limitation of the study. Another possible explanation of this discrepancy is that reported going-to-bed and getting-up times reported by patients could have differed from the actual times.

As to nocturnal pressure, the individuals with ICH more commonly presented a nondipper, or even a riser, pressure profile compared to patients with sustained hypertension, except when we considered ICH3 criteria; however, the dipper/nondipper classification is arbitrary, with poor reproducibility [22-24]. On the other hand, there is no evidence that a nondipper circadian profile associated with low ambulatory blood pressure values has the same prognostic value as a nondipper profile associated with high ambulatory blood pressure values. The mean nocturnal pressures in patients with ICH were lower than in patients with sustained hypertension. This could explain the reduced nocturnal pressure decrease in white-coat hypertensive individuals: such individuals already start with lower pressures during the daytime period.

Another common risk factor for the three ICH definitions is obesity. Other authors concluded that the ambulatory white-coat effect was also significantly higher in women, in older subjects (65+), in obese subjects and in nonsmokers [25]. The white-coat effect may be increased in very elderly individuals [26]. Lurbe et al. [27] found that in severely obese young subjects, ICH was more prevalent than masked hypertension.

The association between ICH and autonomic dysregulation could partially explain these cardiovascular risk profiles in white-coat hypertensive patients [28]. Nevertheless, against this purported increased risk profile of ICH is the greater prevalence of nonsmokers among the female patients with ICH. This profile was confirmed in the case of ICH by the multivariate analysis, after correcting for potential confounding factors. Verdecchia et al. [29] also concluded that female sex and nonsmoking are associated with ICH.

In our study, white-coat hypertensive patients did not present an unfavorable lipid profile, except when we considered ICH1 criteria. In another study [30], total cholesterol was higher in the sustained hypertensive than in the normotensive subjects, but there was no difference between the sustained hypertensive and the white-coat hypertensive subjects or between the white-coat hypertensive and the normotensive subjects.

On the other hand, our clinical capacity to predict which hypertensive individual is ICH is limited. Sensitivity in this sense was a mere 48-52%. This finding, together with the observed prevalence of ICH, suggest that it would be very advisable to conduct a preliminary routine evaluation of ambulatory pressure among many of our untreated hypertensive subjects, before adopting other future therapeutic decisions. This initial evaluation should be based on ABPM [31].

The study sample included a small proportion of patients with a history of some cardiovascular event who were still without pharmacological antihypertensive treatment. As an explanation for this situation, some of these patients may be starting pharmacological treatment, some may have undergone drug withdrawal due to adverse effects, or (fundamentally) some may pose compliance problems.

Larger patient series would be needed to confirm whether heart failure is a predictive variable for ICH, when it is defined as mean daytime BP < 135/< 85 mmHg, as suggested in our study. No other associated clinical conditions (coronary disease, cerebrovascular disease, renal failure or peripheral vascular disease) predicted ICH.

In population-based studies with the inclusion of normotensive individuals, the prevalence of ICH ranges from 10 to 16% [32-36]. Studies in attended hypertensive populations show somewhat higher prevalences, although this depends on the criteria used. In our study, the prevalence varied from 15-29%, according to whether one or other criterion of ICH was applied. Nevertheless, our study is not a population-based study, and hypertensive patients were selected using a convenient sampling among those attending in different practices for hypertension or other medical conditions.

From a practical point of view, the decision to use one or other definition of ICH has clinical implications relating to the follow-up of hypertensive individuals. Regular ABPM of 29% of the untreated hypertensive individuals would not be the same as limiting such monitoring to only 15% of the patients. The design of the present study does not allow us to determine which of the three criteria is most cost-effective, although it would seem advisable to preferentially reserve the diurnal pressure cutoff value of 130/80 mmHg for diabetic hypertensive individuals, or those who already have associated vascular disease. It does not seem adequate to consider a diabetic patient as a white-coat hypertensive individual. The same applies to stroke patients, for example, with poor pressure control in the office and mean diurnal systolic pressures of between 130 and 135 mmHg, or diastolic pressures between 80 and 85 mmHg. Kikuya et al. [37] found that population-based outcome-driven thresholds for optimal and normal ambulatory BP are lower than those currently proposed by hypertension guidelines.

One of the limitations of the study could be interobserver variability in both data collection and patient selection; however, such variability coincides with the variability found in current clinical practice. On the other hand, the large number of individuals is an advantage of the study - making it possible to minimize some sources of bias. In addition, the methodology, protocol, ABPM system used and programming of the latter have been the same in all cases. Another limitation described by Cuspidi et al. [38] may be the reproducibility problems posed by ABPM for diagnosing ICH. These authors recommend performing ABPM more than once in order to diagnose ICH.

In conclusion, the prevalence of ICH is between 15 and 29%, depending on the defining criterion used, and the condition is more commonly observed in women, older patients, nonsmokers, more obese subjects, nondippers and patients with higher systolic BP in the office, except - in these two last cases - when we use the 24-h ICH criteria. The 24-h ICH criteria are not affected by awake/sleep definitions and other biases, so these criteria should be preferred.

On the other hand, the clinical capacity for predicting ICH is limited, in view of the characteristics of the associated variables and the low sensitivity and specificity of suspected ICH. It does not seem easy to establish a good predictive model for white-coat hypertension. Therefore, in the lack of a screening test or algorithm offering high sensitivity and allowing us to decide which patients to subject to ambulatory BP exploration, it seems advisable to increase the application of ABPM to the majority of hypertensive individuals. This would allow us to define the diagnosis (detection and assessment of white-coat hypertension, resistant hypertension and masked hypertension) and improve the quality of therapeutic intervention.