Health Decline Is Associated with Reports of No Coffee Consumption Years After Reporting Coffee Consumption Among Older Adults in Spain

Abstract

ABSTRACT Background Part of the health benefits of coffee reported in observational studies might be due to health status influencing coffee intake rather than the opposite. Objective We examined whether changes in health influenced subsequent reports of no coffee consumption in older adults. Methods Data came from 718 coffee drinkers aged ≥60 y recruited in the Seniors-Estudio de Nutrición y Salud Cardiovascular en España (ENRICA) cohort in 2008–2010 (wave 0) and followed-up in 2012 (wave 1), 2015 (wave 2), and 2017 (wave 3). Health status was measured with a 52-item deficit accumulation index (DAI) with 4 domains: functionality, self-rated health/vitality, mental health, and morbidity/health services use. Coffee intake was estimated with a validated diet history. We examined how changes in health status over a 3-y period (wave 0 to wave 1) influenced reports of no coffee consumption during the subsequent 5 y (wave 1 to wave 3) by using logistic regression models. Results Health deterioration over 3 y was associated with a higher frequency of reports of no regular coffee consumption during the subsequent 5 y (fully adjusted OR: 1.48 per 1-SD increment in DAI; 95% CI: 1.17–1.87). Deteriorating function (OR: 1.38 per 1-SD increment; 95% CI: 1.06–1.81) and mental health (OR: 1.34 per 1-SD increment; 95% CI: 1.04–1.73) were the DAI domains associated with increased reports of no regular coffee consumption. Also, individuals with worsened perceived health or hypertension onset were more likely to report no regular coffee consumption. No associations were found for decaffeinated coffee. Conclusions Health deterioration was associated with reports of no regular coffee consumption years after reporting regular coffee consumption among older adults in Spain. A potential implication of this finding is that part of the beneficial effect of coffee consumption on health in observational studies might be due to reverse causation, which should be confirmed in future research.