European Journal of Preventive Cardiology | 2019
What can be learnt from an atypical population?
Abstract
Hypertension affects one in three of people throughout the world and is the leading cause of preventable deaths world-wide. However, a wide variety of genetic and environmental factors may account for the prevalence of hypertension in different regions of the world. Intranational surveys of the prevalence of hypertension might be expected to reduce the influence of both disparate genetic and environmental factors in the genesis and outcomes of hypertension locally. There are limited studies in Europe. The current study (CONSTANCES) is of the largest population cohort in France and includes participants from the different regions of mainland France. The CONSTANCES cohort is a self-selected sample of French adults derived from the mandatory French insurance scheme for salaried workers. The sample of 62,467 adults aged 18–69 years (7.4%) of the total insured population volunteered to undergo a free health check-up which included self-administered questionnaires on demographic characteristics, socio-economic status, lifestyle, and personal and family medical history. This volunteer population was intensively analysed and is the basis of this paper. The basic reference model considered only the ageand sex-based differences in the prevalence of hypertension between different centres (Model 0). Model 1 added a family history of hypertension as a proxy for the heritability of hypertension. Model 2 then added body mass index (BMI) as a summary of life-long dietary and exercise habits. The addition of socio-economic status contributed to Model 3 and, finally, the addition of neighbourhood level socio-economic indicators made up Model 4. The prevalence of hypertension (defined as a blood pressure of greater than 140/90 and/or taking antihypertensive medications) in the CONSTANCES sample was 30.1% and was higher in men than in women. The cohort was aged 47.9 years and had a normal BMI (25 kg/m). The prevalence rates of hypertension decreased in a north–east–south gradient in men and women. BMI was the highest contributor to the differences between the centres, followed by socio-economic variables and a family history of hypertension. There was an almost two-fold difference in the prevalence of hypertension between the highest and lowest centres. More than 90% of the population studied in this paper had either graduated from high school, obtained professional qualifications or had tertiary education. Only 10.5% were blue collar workers. This study was therefore that of self-selected, middle-class, white-collar participants who were not typical of the bulk of the French National Insurance Scheme. Table 3 is complex and the impacts of various factors were separated by gender. Just considering age alone (Model 0), the centre-level variance in men was 0.073. Adding a family history of hypertension reduced the centre-level variance to 0.069 (Model 1). The further addition of BMI reduced the centre-level variance to 0.057 (Model 2). The addition of education and occupation (Model 3) and residential deprivation index (Model 4) reduced the centre-level variance to 0.055 and 0.053 for Models 3 and 4 serially. Does this mean that the levels of education, occupation and residential deprivation are relatively unimportant factors in the prevalence of hypertension? Or does this have something to do with this study being on a volunteer group of 7.4% of the total French National Insurance Scheme? The CONSTANCES investigators have acknowledged that their carefully studied volunteer population were probably healthier, higher educated and more health conscious than the non-volunteers who were not analysed. We submit that selection bias has accounted for the apparent lack of importance of education, occupation and residential deprivation in explaining the differences of these factors in the centre-level variance in the prevalence of hypertension. Furthermore, the hierarchy used in the modelling has accentuated the apparent lack of influence of education, occupation and residential deprivation. This study of affluent middle-class French people has confirmed the impact of age, gender, family history and BMI on the prevalence of hypertension. But