Background: Blood pressure treatment strategies and targets are generally similar for most patients with hypertension, apart from individuals with certain specific comorbidities such as diabetes. We hypothesize that there is latent heterogeneity in the population of adults with hypertension and that this may be associated with differential risk of mortality and might suggest opportunities for more tailored treatment strategies. Our study of adults with hypertension in the US explores whether clusters of social and health-related characteristics exist, how trends in clusters have varied between 1999 and 2012, and how risk of mortality varies between clusters.
Methods: In a nationally representative sample of adults (> 18 years) with hypertension (physician diagnosed, systolic &ge 140 mmHg or diastolic &ge 90 mmHg, or taking hypertension -related medication) from the National Health and Nutrition Examination Survey from 1999 to 2012 (Total N=16,855), we used a hierarchical cluster analysis approach to identify hypertension sub-groups according to chronic diseases, social factors, and health behaviors.
Results: We found significant heterogeneity among hypertensive participants across time, according to 7 distinct clusters: Young Mexican-Americans with good health (55% of sample); Low-income non-White elderly with good health behaviors (7%); Elderly obese who smoke (18%); Men with poor health who drink and smoke (1%); Normal weight individuals with high prevalence of cancer (16%); Low-income and fit individuals with poor health (1%); and Obese and morbidly obese with high prevalence of chronic conditions (2%). The prevalence of each of these clusters remained relatively constant throughout the last decade, although there was an increase in the “Elderly obese who smoke” cluster from 14% to 18%. Mortality risk varied by subgroup. For example, the odds of dying for “Obese and morbidly obese with high prevalence of chronic conditions” were 1.73 times higher compared to “Young Mexican-Americans with good health” (p-value<0.001), adjusting for confounders over a 12 year follow-up interval.
Conclusions: Sociodemographic and clinical clusters among hypertensive individuals in the U.S. remained stable over a decade, with higher risk of mortality associated with specific clusters. Recognizing these clinically-identifiable phenotypes may highlight opportunities to develop and evaluate targeted interventions that account for complex and inter-related clinical and social factors. Further, there is an opportunity through clinical trials to determine whether these clusters merit different blood pressure treatment strategies.