Hypertension Prevalence and Stroke Mortality Across Populations

Blood pressure (BP) is the most consistent and powerful predictor of stroke, such that hypertension is causally involved in nearly 70% of all stroke cases.¹ Worldwide, stroke is second only to ischemic heart disease as a leading cause of death.² In a quantitative overview of 61 cohort studies with more than 1 million enrolled subjects,³ the Prospective Studies Collaboration demonstrated that small gradients in systolic or diastolic BP account for sizable differences in cardiovascular outcomes. Along similar lines, several recently published outcome trials in hypertensive patients or patients at high cardiovascular risk proved that reductions in systolic BP as small as 1 to 3 mm Hg decrease the relative risk of stroke by as much as 20% to 30%.⁴

Despite the consistency of findings among observational cohort studies³ and outcome trials in hypertension,⁴ the geographic heterogeneity in the patterns of stroke mortality remains poorly understood. Age-standardized rates vary more than 10-fold from countries with low incidence, such as Switzerland (25 stroke deaths per 100 000 population aged 40-69 years), to regions with a high stroke risk, such as Eastern Europe, Japan, China, or the "Stroke Belt" in the southeastern United States.⁵ In this respect, the study by Wolf-Maier et al⁶ reported in this issue of THE JOURNAL breaks new ground by raising the hypothesis that the average level of BP and its age-related increase might differ across populations and might go hand in hand with the observed differences in stroke mortality.

In their retrospective review of 8 surveys on hypertension, Wolf-Maier et al⁶ reported average BP differences between North America and Western Europe of 9 mm Hg systolic and 6 mm Hg diastolic with even larger contrasts between individual countries. As the authors acknowledge,⁶ various methodological issues may have biased their findings. Undetected selection in the enrollment of study participants, divergent lifestyle factors (including salt intake,⁷ smoking,⁸ or alcohol consumption⁹ ), secular trends in BP among surveys published over a time span exceeding 1 decade,¹⁰ and climatic influences on BP¹¹ are among the confounders unaccounted for in the assessment by Wolf-Maier et al.⁶

Furthermore, in all of the population studies reviewed by Wolf-Maier et al,⁶ BP was measured by auscultation of the Korotkoff sounds, except for 1 survey in England, in which an oscillometric device with doubtful precision¹² was used. This technique of BP measurement is prone to error and is difficult to standardize even among skilled observers. Blood pressure is nothing if variable and is characterized by large diurnal fluctuations. Single or multiple readings taken once or even several times during the day reflect a subject's usual BP only to a minor extent. The presence of an observer may arouse a patient and cause a transient increase in BP. This so-called white-coat effect usually diminishes with repeated BP measurement. In a screening program among poor residents of Dar es Salaam (Tanzania), in which Bovet and coworkers¹³ obtained 3 readings at each of 4 visits, BP decreased by an average of 15 mm Hg systolic and 8 mm Hg diastolic from the first reading of the initial visit to the last reading of the fourth visit. Thus, disparities both in the technique and in the conditions of BP measurement may have inflated the apparent between-country differences in the study by Wolf-Maier et al.⁶

In the European Project on Genes in Hypertension, which currently involves 5 Eastern and 2 Western European countries, investigators put a great deal of effort in the implementation of a quality assurance and quality control program for the measurement of the BP phenotypes.¹⁴ Blood pressure was measured by the Korotkoff technique according to the recommendations of the British Hypertension Society¹⁵ as well as by ambulatory monitoring. The database currently includes 4344 Caucasian participants recruited in Bucharest (Romania, 6.8%), Cracow (Poland, 7.5%), Hechtel-Eksel (Belgium, 58.7%), Novosibirsk (Russian Federation, 7.3%), Padova (Italy, 8.1%), Pilsen and Prague (Czech Republic, 8.4%), and Sofia (Bulgaria, 3.8%). Across countries, mean age varied from 35.5 to 43.3 years. In all participants combined, BP measured 5 times consecutively at the second contact, which always took place in the home environment, averaged 124.0 mm Hg systolic and 76.1 mm Hg diastolic. The corresponding values for the ambulatory measurements obtained between 10 AM and 8 PM in 2167 participants were 125.7 mm Hg and 77.7 mm Hg, respectively.

Despite significant differences in the age distributions and the small sample size in Bulgaria, the largest disparity in the conventional BP across countries averaged 5.5 mm Hg systolic and 7.6 mm Hg diastolic. Adjustment for age or for age and body mass index combined slightly increased these gradients to 6.8 and 6.3 mm Hg systolic and to 7.6 and 7.4 mm Hg diastolic, respectively. On daytime ambulatory monitoring, which excludes observer bias but incorporates diurnal BP variation, the maximal differences between any countries were 6.8 mm Hg systolic and 6.7 mm Hg diastolic without adjustment, 6.3 and 8.2 mm Hg after adjustment for age, and 6.3 and 8.0 mm Hg with cumulative adjustment for both age and body mass index. These highly standardized observations suggest that Wolf-Maier et al might have overestimated the between-country differences in BP, in particular systolic pressure.

Regardless of the inaccuracy inherent in the combination of data from different sources, the findings reported by Wolf-Maier et al⁶ should not be ignored. The authors noticed a steeper increase in BP with advancing age in Europe compared with North America. Different BP measurement techniques between studies cannot explain this observation, because within each country investigators used standardized epidemiologic methods. The study findings⁶ are also in line with observations in 41 populations of the World Health Organization Monitoring of Trends and Determinants in Cardiovascular Disease (WHO MONICA) Project.¹⁶ In men, the average age-standardized BP measurements among populations varied from 124 to 148 mm Hg systolic and from 75 to 93 mm Hg diastolic. For women, these ranges were 118 and 145 mm Hg and 74 and 90 mm Hg, respectively. More important, Wolf-Maier et al noticed a strong positive association between the age-adjusted prevalence of hypertension and stroke mortality across countries, which adds considerable support to the authors' interpretation that BP differences between countries and across continents are real.

These observations are consistent with an incident case-control study nested within a large Swedish survey.¹⁷ Multiple logistic regression revealed that among 129 cases of first-ever stroke and 2 randomly selected controls per stroke patient, untreated or poorly controlled hypertension was associated with an odds ratio of 4.3 (95% confidence interval, 1.7-10.5).¹⁷ In many countries of Western Europe and North America, age-adjusted stroke rates steadily declined over the past 50 years.¹⁰ Most experts believe that this favorable secular trend is due to more accurate primary and secondary prevention, which came with the recognition of the role of cardiovascular risk factors, the availability of more potent and well-tolerated antihypertensive drugs, and the increasing prosperity of industrialized nations.

Control rates of hypertension vary from one country to another, but in general are much lower than desirable. The National Health and Nutrition Examination Survey¹⁸ showed that the awareness of the hypertensive population in the United States improved from 50% in the 1970s to 70% in the 1990s, while during the same period the proportion of treated hypertensive patients with normalized BP increased from 10% to 29%. Data from 4 US centers participating in the Cardiovascular Health Study suggest that during the 1990s this favorable trend continued at least in the elderly.¹⁹ Recent studies demonstrated that in Europe¹⁷ and other parts of the world²⁰ the rule of halves still exists and that the fraction of hypertensive patients with properly controlled BP ranges from approximately 5% to 45%.^{21 - 22} For cardiovascular prevention in hypertensive patients, policy makers, the medical profession, and patients share responsibilities. Their short-term priorities should be respectively improving the accessibility to medical care, more rigorous adherence to the guidelines for diagnosis and treatment,²² and stricter compliance to therapy. However, the long-term perspective rests on the continuous nature of the relationship between stroke risk and BP and the notion that the majority of strokes occur in individuals currently categorized as normotensive.²³ Population-based strategies with the goal to produce a downward shift in the BP distribution of large populations are likely to lead to a substantial reduction of the global burden of stroke.²³

The key message of the report by Wolf-Maier et al⁶ lies in the observation that both the prevalence of hypertension and stroke mortality were lower in North America than in Europe. Although based on the retrospective combination of 8 surveys with differing methods, these observations⁶ underscore that BP control is central to the prevention of the cardiovascular complications of hypertension, especially stroke.⁴ High BP mostly arises as a complex quantitative trait that is under the influence of varying combinations of genetic and environmental factors. Further research based on standardized epidemiologic methods and on the profound integration of clinical, molecular, and genetic research should help to clarify which ecogenetic factors cause the age-related increase in BP and are amenable to intervention, so that in the not too distant future hypertension becomes a preventable disease and its complications can be eradicated.