Background: High blood pressure (BP) is the leading risk factor for disease and mortality worldwide. However, risks associated with high BP in very old age (≥ 80 or ≥ 85 years) are not entirely understood, as the majority of scientific studies have been performed with younger populations and existing scientific knowledge about very old individuals is sometimes contradictory. Results of previous studies of very old individuals suggest that the associations of BP with mortality and stroke differ with levels of physical and cognitive function. More studies that are representative of very old individuals, including individuals with multimorbidity, that are of adequate size, involve proper adjustment, and investigate non-linear associations, are needed to investigate these issues.
Systolic blood pressure (SBP) decline is common among very old individuals and has been shown to precede adverse events. Previous studies have shown that SBP change is associated with baseline SBP, age, and health-related factors, but determinants of SBP change have not been investigated using comprehensive, multivariate models.
The three main aims of this thesis were to investigate, in a sample of individuals aged ≥ 85 years, 1) determinants of SBP change, 2) the association of BP with mortality risk and whether this association differs with respect to gait speed and/or Mini-Mental State Examination (MMSE) score, and 3) the association of BP with stroke risk and whether this association differs with respect to the Barthel Activities of Daily Living (ADL) index and/or MMSE score.
Methods: The studies conducted for this thesis were based on data from the population-based Umeå 85+/Gerontological regional database study, which provided cross-sectional and longitudinal data on socioeconomic factors, medical conditions, drug prescriptions, and health-related assessments from 2000 to 2015. Participants were aged 85, 90, and ≥ 95 years, and lived in Västerbotten, Sweden, and Österbotten/Pohjanmaa, Finland. Follow-up assessments were conducted after 5 years. Mortality data were collected after 2 and 5 years, and stroke data were collected after 5 years, from death certificates, medical records, population registers, and the inpatient diagnosis register. Comprehensive multivariate models were developed to investigate determinants of SBP change using multiple linear regression, and to investigate associations of mortality and stroke risks with BP using Cox proportional-hazard regression models.
Results: Average (± standard deviation) baseline SBP was 146 ± 23 mm Hg, and average diastolic blood pressure (DBP) was 74 ± 11 mm Hg. Within 5 years, 61% of participants had died and 10% had had incident strokes. Among participants followed for 5 years, the average annual SBP decline was 2.6 ± 5.4 mm Hg.
In a multivariate model, SBP decline was associated with later investigation year (p = .009), higher baseline SBP (p < .001), baseline antidepressant drug use (p = .011), incident acute myocardial infarction during follow-up (p = .003), use of a new diuretic drug during follow-up (p = .044), and declining Barthel ADL index scores during follow-up (p < .001).
In an age- and sex-adjusted analysis of the total sample, mortality risk was decreased in higher (vs. lower) BP categories (SBP ≥ 165 vs. ≤ 125 mm Hg: hazard ratio [HR] .50, p < .001; DBP 70–74 vs. 75–80 mm Hg: HR 1.32, p = .031). In a comprehensively adjusted analysis of the total sample, SBP was not associated significantly with mortality risk. The associations of SBP with mortality in the gait speed < .5 m/s subcohort corresponded with those found in the total sample. In comprehensively adjusted analyses in the gait speed ≥ .5 m/s subcohort, mortality risk increased independently with higher (vs. lower) BP (SBP ≥ 165 vs. 126–139 mm Hg: HR 2.13, p = .048; DBP > 80 vs. 75–80 mm Hg: HR 1.76, p = .026). In comprehensively adjusted analyses in the MMSE score subcohorts, SBP was associated significantly with mortality risk only in the 0–10 MMSE score subcohort; high and low SBP categories were associated independently with increased mortality risk, compared with an intermediary SBP category (SBP ≥ 165 vs. 126–139 mm Hg; HR 4.54, p = .007; SBP ≤ 125 vs. 126–139 mm Hg: HR 2.23, p = .023). Higher BP was associated significantly with increased stroke risk in multivariate models (SBP per 10 mm Hg increment: HR 1.19, p < .001; DBP per 10 mm Hg increment: HR 1.26, p = .013). SBP was not associated with stroke risk in participants with SBP < 140 mm Hg.
Interaction effects on the association with mortality were significant between SBP and gait speed (age- and sex-adjusted model: p = .031) but not between SBP and MMSE score. No interaction in the association with stroke was found between any BP measure and Barthel ADL index or MMSE score.
Conclusion: The decline in BP in very old age may be explained by health-related factors. Low BP may be a risk marker for short life expectancy, due to morbidity, in the general very old population and among very old individuals with low gait speeds. High BP seems to be an independent risk factor for mortality only in certain groups, which may be distinguished by high gait speed or very severe cognitive impairment. High SBP and DBP seem to increase stroke risk in very old age. These findings may contribute to a better understanding of the risks of adverse outcomes in very old individuals with different BP levels, the importance of comorbidity for these risks, and the etiology of SBP change.