Loneliness and risk of mortality: a longitudinal investigation in Alameda County, California.

We investigated the prospective impact of self-reported loneliness on all-cause mortality, mortality from ischemic disease and mortality from other cardiovascular diseases. We tested these effects through GEE binomial regression models applied to longitudinal data from the Alameda County Study of persons aged 21 and over arranged into person-years. Controlling for age and gender, the chances of all-cause mortality were significantly higher among respondents reporting that they often feel lonely compared to those who report that they never feel lonely. Frequent loneliness was not significantly associated with mortality from ischemic heart disease but more than doubled the odds of mortality from other ailments of the circulatory system in models controlling for age and gender. Subsequent models showed that physical activity and depression may be important mediators of loneliness-mortality associations. Finally, we find support for the contention that chronic loneliness significantly increases risk of mortality but also find reason to believe that relatively recent changes in feelings of loneliness increase risk of mortality as well. Copyright (c) 2010 Elsevier Ltd. All rights reserved.