Blood Pressure Variability and Mortality - Results From the Epidemiology of Hypertension in Croatia (Ehuh Study) - Mortality Data After 17 Years of Follow-Up

Objective: Our aim was to analyze association of visit-to-visit blood pressure variability (BPV) with comorbidity and mortality in random sample of Croatian population, an original cohort from the EHUH study. Design and method: EHUH study is nation-wide survey on prevalence, treatment and control of hypertension in Croatia. In 2001–2003 a random sample of 1267 subjects (mean age 53 ± 16; men 41%) were enrolled. Mortality data were analyzed after average period of 17 years. Office blood pressure (BP) was measured in sitting position using mercury sphygmomanometer and appropriate cuffs three times at two visits and once at home visit. Visits were organized within two months period. Standard deviation (SD) and coefficient of variability (CoV) (SD/individual mean SBP*100) were used as measures of visit-to-visit BPV. Results: Visit-to-visit BPV was higher in hypertensives (SD:5.5 ± 3.6 vs.4.1 ± 2.7; CoV:3.7 ± 2.4 vs.3.4 ± 2.1; for both p < 0.0001) diabetics (SD:6.1 ± 4.4 vs.4.5 ± 3.1;CoV:4.11 ± 3.8 vs.3.4 ± 2.2; for both p < 0.001) and obese subjects (SD:5.4 ± 3.7 vs.4.1 ± 2.7;CoV:3.7 ± 2.5 vs.3.3 ± 2.1; for both p < 0.0001) regardless we used SD or CoV. There were no gender differences but BPV increases with aging. It was higher in subjects with high normal vs. those with optimal BP (p = 0.003) as well as it was higher in stage 2 than stage 1 hypertension (p = 0.005). Higher BPV was observed in uncontrolled hypertensives than in non-treated and controlled ones (p < 0.001). All-cause mortality was associated with higher BPV (4.95 ± 3.5 vs.4.66 ± 3.2; p = 0.003), but we failed to find differences in BPV between subjects who died of CV disease or cancer. In multivariate linear regression model (age, gender SBP, SD, CoV) BPV was not found to be an independent predictor (R 2 = 0.221; B 0.22 SE 0.013; beta = 0.178;p = 0.08; CoV B 0.31 SE 0.018 beta = 0.173; p = 0.09). Conclusions: BPV was higher in hypertensive, diabetic, older and obese subjects. In normotensives it was significantly higher in prehypertensives than those with optimal and normal BP. In hypertensive subjects the highest BPV was found in uncontrolled group (compared to untreated and controlled), We failed to find positive predictive value of BPV for mortality probably the visit-to-visit period (2 months) was too short.