From the Epidemiologist to the Physician: A Complex Issue for a Labile Risk

In a recent study of elderly, rural-dwelling Tanzanians, the incorrect classification of more than a quarter of patients as hypertensive was attributed to the white-coat effect.1 Although the white-coat effect might contribute to the overestimation of the prevalence of hypertension (HTN), systolic blood pressure (SBP) variability is known to enhance cardiovascular (CV) risk. In addition, white-coat HTN is now regarded as an intermediate phenotype between normotension and HTN. On the other hand, relatively high nocturnal blood pressure (BP) resulted in nighttime HTN in more than two thirds of the cohort of elderly, rural-dwelling Tanzanians. Sub-Saharan Africans were already reported to have higher nighttime BPs than Caucasians with a blunted nocturnal decline, so that ambulatory BP monitoring (ABPM) might be especially useful in this ethnic group. As recently underlined by the Global Burden of Disease study,2 HTN results in more deaths than any other risk factors, including diabetes and cigarette smoking. High prevalence, inadequate awareness, suboptimal treatment, and low rates of guideline-recommended target BP control are key factors that lead to severe CV complications that impose a heavy socioeconomic burden, especially in developing countries. A systematic review and meta-analysis3 pooling data from 33 surveys published between 2000 and 2013, involving more than 110,414 participants with a mean age of 40 years, shows that in sub-Saharan Africa the predicted prevalence of HTN at mean participant ages of 30, 40, 50, and 60 years were 16%, 26%, 35%, and 44%, respectively, with a pooled prevalence of 30% (95% confidence interval [CI], 27–34). Most importantly, of those with HTN, only between 7% and 56% (pooled prevalence: 27%; 95% CI, 23–31) were aware of their hypertensive status before the surveys. Overall, 18% (95% CI, 14–22) of individuals with HTN were receiving treatment across the studies, and only 7% (95% CI, 5–8) had controlled BP.3 The HTN burden is likely to grow in the next few decades as the population ages and the prevalence of obesity and diabetes increases. These data clearly highlight the need for implementation of timely and appropriate strategies for diagnosis, control, and prevention. However, the direct transfer of data obtained in an epidemiological study in a real estimate of the costs necessary to the country for the treatment of CV risk hides some uncertainties. Although data from observational studies indicate that this risk is continuous, to have a precise estimate in terms of prevalence in population studies may not be straightforward. In the individual patient, the clinician performs the diagnosis of HTN on the basis of repeated BP values above 140/90 mm Hg at several visits.4-6 Although these criteria may be met in epidemiological surveys performed in high-income countries (in the third report of the National Health and Nutrition Examination Survey, nearly 80% of participants had up to six BP measurements on two occasions7), estimates of HTN prevalence in low- and middle-income countries (LMICs) are often based on data collected at a single visit, causing a potential systematic error.8-12 This discrepancy is mainly the result of the difficulty to bear the costs of more visits in an epidemiological study. A similar problem also arises in the case of diabetes. The strategy used in the clinic to make measurements of fasting blood glucose on separate days is not easily feasible in epidemiological studies so that attempts are made to obtain reliable estimates using different strategies based on glycated hemoglobin or the response to the oral glucose tolerance test. Comparisons between estimations of diabetes prevalence assessed with these different strategies are often given in the same survey,13 whereas only a few studies have assessed the impact of one or two visits on the estimation of HTN burden in LMICs.11 Taking BP multiple times can attenuate the influence of within-person variability because of physiological variation or measurement error. Another source of nonrandom variation is the phenomenon of “white-coat hypertension,” which may lead to an overestimation of HTN prevalence.14 The National Institute for Health and Care Excellence (NICE) recommends ABPM for all patients suspected of having HTN because of a previously elevated office BP measurement aiming to exclude patients with white-coat HTN.15 According to the study by Ivy and colleagues,1 the white-coat effect was responsible for an increase in recorded BP in more than two thirds of a cohort of elderly, rural-dwelling Tanzanians. Those data are in agreement with a recent study performed in a different world area.16 At least three points have to be considered. First, could the adoption of a strategy based on ABPM lead us to exclude all patients with the diagnosis of white-coat HTN from treatment? Although no prospective studies are available in LMICs, the answer is probably no. Visit-to-visit variability in SBP and maximum SBP were indeed found to be strong predictors of stroke, independent of mean SBP,17 and white-coat HTN is now regarded as an intermediate phenotype between normotension and HTN that requires regular follow-up.18 The World Health Organization proposed different 10-year total CV disease risk thresholds for intensive intervention that may be selected on the basis of the gross domestic product per head for the country. As the CV risk threshold for drug treatment is lowered, there is a concomitant increase in health benefits.19 However, although white-coat HTN could be important in CV risk stratification at the patient level, the adoption of ABPM may be much more expensive than treating all hypertensive patients diagnosed in the office. The panel of tests that can be included in surveys should be seen in the perspective of costs. Because the cost of ABPM and HTN management differ greatly from country to country and is dependent on the method of health-care delivery, the cost-effectiveness of ABPM may need to be evaluated at a national level. The ready provision of ABPM in primary care is dependent on reimbursement to physicians or other health-care providers by the national health-care systems or by private insurance and varies considerably from country to country, with most LMICs not providing any reimbursement. On this basis, the possible advantages of ABPM as a way to limit drug treatment in LMICs could be questionable. Second, the study restricts observations to the elderly population. However, the predictive value of visit-to-visit variability in SBP for stroke or coronary events was found to vary in relation to age, with the strongest association being in the youngest (≤57 years) quartile.17 In addition, misclassification of HTN at one or two visits was found to be more common in younger ages.11, 20 Therefore, the observation by Ivy and colleagues1 should be extended to the youngest cohorts, particularly affected by changes in cardiovascular risk linked to the epidemiological transition that is currently underway in LMICs.21 Last, the authors report that although masked HTN was not common, relatively high nocturnal BP resulted in nighttime HTN in more than two thirds of the cohort of elderly, rural-dwelling Tanzanians. This observation supports the opportunity to modify the traditional definition of white-coat HTN (now based on BP values recorded during the awake period) as recently suggested,22 because nocturnal BP is superior to daytime BP in predicting CV risk. This evidence might be especially important for some ethnic groups (sub-Saharan Africans) in which the importance of ABPM in identifying nocturnal HTN (nondippers) is well-known. In addition to the soaring increase in prevalence of HTN in Africans, they were reported to have higher nighttime BPs than Caucasians with a blunted nocturnal decline.23-27 This pattern was also proposed as a reason for the worse CV prognosis described in black patients.25, 26 Although the precise mechanism causing the nondipping pattern in blacks is still unknown,28 both salt sensitivity29, 30 and elevated blood glucose levels31 observed in the African population might be significant contributors. Notwithstanding these data, current guidelines for ABPM use do not consider the importance of ethnicity.32 The importance of reaching a common ground for epidemiologists and clinicians in the assessment of HTN burden in LMICs and to include ethnicity among factors relevant in the use of ABPM32 have to be stressed. The authors have no conflicts of interest to disclose.