Does low blood pressure predict future depression in older adults?

Problem

Between 10% and 20% of older adults experience depression. Currently there are mixed findings as to whether low blood pressure (BP) is a predictor of future depression. Heterogeneity amongst existing study findings, due in part to differences in analytical methods and treatment of potential confounding variables, makes it difficult to draw clear conclusions as to whether low BP is implicated in the development of depression in older adults. There is also uncertainty over the impact of postural hypotension (PH) on development of future depression. We aimed to examine whether low BP and/or PH is associated with subsequent depression in a representative cohort of older adults.

Approach

Prospective cohort analysis of participants in the CARE75+ study; a representative UK cohort of community-dwelling older adults, aged over 75, recruited from general practices in England. We studied participants free of depression and antidepressant use at recruitment. Primary outcomes were Geriatric Depression Scale (GDS) scores at 6 and 12 months. We examined linear regression models adjusted for recruitment GDS scores, for associations between BP, PH and depression (defined as a GDS score ≥5 at 6 or 12 months). Analyses were adjusted for body mass index, tobacco and alcohol consumption, cognitive function, health-related quality of life, loneliness, living alone, frailty, ethnicity, highest attained educational level, and finally, also for antihypertensive use.

Findings

We studied 493 participants: 47.9% female; age range 76 to 100 years. After 6 months, participants with low systolic BP (<110 mmHg: 4.5% of cohort) had higher mean GDS scores than those with normal BP (systolic 110 to 130 mmHg; 23.7%): difference 0.856 (95% CI: 0.091 to 1.062; P=0.028). For diastolic BP GDS scores for low BP (<70 mmHg: 32.5%) were higher than with normal diastolic BP (70 to 80 mmHg: 38.5%): mean difference 0.419 (95% CI: 0.067 to 0.772; P=0.020). For each 20mmHg increment in systolic BP the GDS score was 0.089 points lower (95% CI: -0.241 to 0.061; P=0.244). After 12 months, there was no longer evidence that BP categories were associated with depression. Exploration of adjusted models taking account of demographic factors and antihypertensive drug use did not influence the findings. PH was present in 18.8% participants but was not a significant predictor of depression.

Consequences

We found evidence that low BP is associated with higher GDS scores after 6, but not 12, months. The magnitude of differences is too small to represent clinically meaningful differences. Few participants had low systolic BP however, substantial prevalences of low diastolic BP and of PH were noted in this population, emphasising the importance of regular medication review and assessment in older age