Objectives: To evaluate which indices of blood pressure (BP) homeostasis are the strongest predictors of mortality in older low-level-care residents in long-term health facilities.
Design: Prospective cohort study.
Setting: Eight long-term healthcare facilities in Canberra, Australia.
Participants: A total of 179 randomly selected semi-independent residents aged 65 and older (mean age±standard deviation 83.2±7.0; 80% women).
Measurements: Baseline BP levels taken while lying, after standing for 1 and 3 minutes, and sitting before and 1 hour after meal intake were recorded, as well as demographic information, chronic medical conditions, medications, and all-cause mortality during follow-up. Postprandial hypotension (PPH) was defined as a fall in systolic BP (SBP) of 20 mmHg or more 1 hour postmeal while sitting. Orthostatic hypotension (OH) was defined as a fall in SBP of 20 mmHg or more or in diastolic BP (DBP) of 10 mmHg or more within 3 minutes of standing from a supine position. Hypertension was defined as BP greater than 160/90 mmHg at commencement of the study. Mean arterial pressure (MAP) and pulse pressure (PP) were calculated.
Results: At baseline, 47% of participants had hypertension, 38% PPH, and 23% OH; PP was 70 mmHg or greater in 54%, and DBP was 65 mmHg or lower in 6%. Over 4.7 years, 97 (54%) participants died. Those who died were significantly older and more likely to have PPH (47% vs 28%) and atrial fibrillation (35% vs 17%) and a significantly greater decrease in BP after meal intake. Mortality rates in those with and without PPH were 145.0 and 98.5 per 1,000 person-years, respectively. Using multivariate Cox proportional hazards models after adjustment for age, sex, presence of atrial fibrillation, Parkinson's disease, and use of diuretics, PPH was the only BP parameter that significantly and independently predicted 4.7-year all-cause mortality (relative risk (RR)=1.79; 95% confidence interval (CI)=1.19−2.68; P=.005). Further adjustment for the presence of OH, hypertension, low resting BP, coronary artery disease, cerebrovascular disease, congestive heart failure, history of syncope, cognitive impairment, cancer, diabetes mellitus, chronic obstructive pulmonary disease, and history of smoking did not reveal any new statistically significant associations. There was a dose–response relationship between postprandial fall in SBP and mortality rates. Absolute postprandial SBP of 120 mmHg or less was also significantly associated with total mortality (RR=1.69, 95% CI=1.04−2.78; P=.04). Low DBP was also associated with increased mortality (RR=1.10, 95% CI=1.01−1.13; P=.03), although this association became nonsignificant in multivariate analysis.
Conclusion: In older low-level-care residents, PPH is an independent predictor of all-cause mortality with no added predictive value explained by other BP indices: OH, hypertension, PP, MAP.