Ambulatory blood pressure monitoring and the taxonomy of hypertension

Introduction In this issue of the Journal of Hypertension, Owens et al. [1] report the prevalence of six pre-specified categories of noninvasively derived ambulatory blood pressure in 2092 patients with untreated essential hypertension. Their stated purpose was to describe the epidemiology of patterns of high blood pressure in a large population. Referred patients with the diagnosis of primary hypertension who met the authors' inclusion criteria for this analysis were classified on the basis of their clinic blood pressure as having either isolated systolic hypertension (9%), isolated diastolic hypertension (4%) or combined systolic and diastolic hypertension (87%), as they show in tables 1a and 2. When reclassified on the basis of ambulatory blood pressure data, subjects with elevated systolic and diastolic blood pressure remained in the majority (56%), but 13% were now labelled as having borderline hypertension, 11% with white-coat hypertension and 7% with nocturnal hypertension (definitions in table 1b). The prevalence of borderline hypertension and white-coat hypertension decreased with age, as did isolated diastolic hypertension, whereas the prevalence of isolated systolic hypertension increased. White-coat hypertension was diagnosed in 16% of women under the age of 65 years. The authors contend that these observations are important because they ‘put into perspective the variety and extent of hypertension in the hypertensive community’, and state that the identification of these ‘subforms’ by ambulatory blood pressure monitoring is important for the proper diagnosis and treatment of hypertension. Epidemiology of blood pressure patterns Do the observations outlined above truly describe the epidemiology of patterns of high blood pressure in a large population? This was a retrospective database survey of a highly selected group of patients referred by general practitioners and hospital doctors to a specialty blood pressure unit. Of the ambulatory recordings available for analysis, almost 80% were excluded because patients were either on treatment, involved in clinical trials or ‘participating in population prevalence studies’, or for technical reasons. Since the diagnosis and management of hypertension lies primarily within the clinical domain of the primary practitioner, there must have been some compelling rationale, related to practice patterns in this catchment area, for these patients to be referred before the initiation of treatment. Because the number of blood pressure measurements required for referral was not standardized, and clinic blood pressures were determined from a only single visit, white-coat hypertension may be overestimated in this sample. Indeed, selection bias becomes an important consideration when proposing this diagnosis [2]. The age distribution of these patients indicates that these data are biased against the elderly, and therefore would not be expected to describe the epidemiology of patterns of high blood pressure within the authors' community. Only 18% of these patients were aged 60 years or older, and less than 5% were over 70 years of age, while more than 10% were 30 years or younger. By comparison, in the United States, the prevalence of hypertension among non-Hispanic whites is more than 60% over the age of 60 years, and less than 3% under the age of 30 years [3,4]. Primary practitioners and hospital doctors in the authors' health district may be quite comfortable initiating treatment in their elderly patients, or may not refer such patients to hypertension clinics, whereas isolated diastolic hypertension in young patients may prompt early consultation for secondary causes of high blood pressure. Atrial fibrillation, usually considered a contra-indication to reliable ambulatory blood pressure recordings, is a common complication of hypertension, and one which increases in prevalence with age [5]. Whether these data were acquired from patients predominantly in sinus rhythm was not explicitly stated. The prevalence of other important cardiovascular risk factors that could influence the decision to treat blood pressure is not reported. Excluded are patients deemed normotensive by the clinic nurse. Some of these might have been classified as hypertensive had ambulatory monitoring been performed. Until the present observations are validated by determining the true prevalence of hypertension and each specific ‘subform’ within this community, the percentage distributions calculated from this sample should not be considered representative of either the hypertensive population in this region, or a population referred to hypertension clinics in other jurisdictions. Usefulness of the ‘subform’ concept Is the categorization of hypertensive patients on the basis of these proposed ‘subforms’ a reliable or useful exercise? The authors argue that ambulatory blood pressure monitoring is essential for the proper diagnosis of hypertension and for the identification of ‘subforms’ that are ‘separate aetiological conditions’ which differ with respect to prognosis and treatment. This is an important hypothesis, which the authors are in the position to test, prospectively, in an outcomes study. However, the present paper does not characterize the haemodynamic correlates of these ‘subforms’, nor inform us of the probability, based on age, sex and clinic blood pressures, that a particular patient will be reclassified into another ‘subform’ as a result of ambulatory monitoring. For example, clinic measurements may be sufficient for the proper diagnosis of most patients with isolated systolic or diastolic hypertension. More importantly, no data on the reproducibility of these six patterns in their patients, and therefore on the stability of these categories, are provided. The ‘new phenomenon’ of daytime normotension but nocturnal hypertension is the pattern least likely to be reproducible [5,6], and most likely to be caused by factors other than primary hypertension, such as absent or interrupted sleep, sleep-related breathing disorders, diabetes or autonomic insufficiency. These confounding factors, which could place such patients at greater risk for morbid events than the level of their night-time blood pressure, were not noted in the present survey. Obviously, these conditions do not require ambulatory monitoring for their identification or treatment. If the timing of sleeping and awakening is accounted for, and ambulatory recordings obtained using the intra-arterial method, the blood pressure and heart rate of untreated hypertensive patients decrease, on average, by 25% from periods of wakefulness to sleep [7]. Finally, the authors presume that the identification of nocturnal, borderline or white-coat hypertension should influence or preclude treatment. Two advances in clinical practice render this argument less relevant to the management of such patients: increasing recognition of altered blood pressure as only one aspect of a complex cardiovascular and metabolic disorder, and broader awareness of the importance of overall cardiovascular risk reduction. Conclusions This contribution by Owens et al. [1] is an important effort, but was primarily a descriptive exercise that did not attempt to acquire mechanistic, haemodynamic, outcome or other evidence that such categorization identifies ‘separate aetiological conditions’ or should influence clinical practice. Just as taxonomy of fauna on the basis of presumed physical similarities resulted in the misclassification of some species, the taxonomy of hypertension and recommendations for treatment should not arise from one-off ambulatory recordings, but from more fundamental information, such as multigenic predisposition, its expression as neural, humoral or paracrine mechanisms of high blood pressure, the risk to the patient of a cardiovascular event and the results of large outcome trials.