Loneliness and isolation tied to up to 5.8 fewer years free of depression or anxiety in 277,489 adults

Loneliness has been compared to smoking and to obesity so often that the comparisons have lost their force. They also describe only one outcome, which is death. For most of us, the years that precede it matter as much: how long a person goes before a diagnosis of heart disease, diabetes, dementia or depression changes what daily life is like.

A team at Tulane University has now put a number on that for lonely and isolated adults. Using records on 277,489 volunteers in the UK Biobank, Lu Qi and colleagues estimated how many years from age 50 a person could expect to live free of seven major conditions. Their paper, published in Nature Communications, reports that loneliness and social isolation are more strongly associated with life expectancy free of mental disorder than with life expectancy free of physical disease. For men who were both lonely and isolated, the gap in years free of depression or anxiety reached almost six.

Loneliness and social isolation are measured differently

The two words are used loosely in conversation, but researchers treat them as separate things. In Tulane’s description of the study, social isolation is an objective measure of the amount of social connection a person has, while loneliness is a subjective measure of a person’s dissatisfaction with the quality of their relationships. Someone can live alone, see few people and feel content. Someone else can be surrounded by family and feel that nobody knows them. Neither state is rare: the authors cite estimates that a third of middle-aged and elderly people in the UK and the US feel lonely.

UK Biobank asked about both when volunteers enrolled between 2006 and 2010. Loneliness was scored from two questions: whether a person often felt lonely, and how often they could confide in someone close. Isolation came from three: whether they lived alone, whether friends or family visited less than once a month, and whether they took part in any social activity each week. Anyone who gave the high-risk answer to both loneliness questions was classed as lonely, and those with at least two of the three isolation markers as isolated.

By those definitions the groups were small. About 3% of participants were lonely without being isolated, about 7% were isolated without being lonely, and fewer than 1% were both.

How the study counted years of life free of disease

The researchers began with adults aged 40 to 69 who had none of seven conditions: type 2 diabetes, cardiovascular disease, chronic lung disease, dementia or Parkinson’s disease, cancer, depression and anxiety. Anyone who already had one was left out, which removed about 189,000 volunteers and left a comparatively healthy group.

They then followed hospital and death records for a median of nearly 14 years. In that time a total of 16,356 events of death and 78,829 events of 7 major chronic diseases were recorded. From the rates at which participants moved from healthy to ill and from either state to death, the team built life tables, the same tool actuaries use, to estimate the years a 50-year-old in each group could expect to live before the first of those diagnoses.

The comparisons took account of age, ethnicity, neighborhood deprivation, education and household income. That matters, because lonely and isolated participants were poorer and less educated on average than the rest, and poverty is itself tied to fewer healthy years.

How many disease-free years went with loneliness and isolation?

A 50-year-old woman who was neither lonely nor isolated could expect about 25.5 more years free of all seven conditions. For a woman who was both, the figure was 22.6 years, about three years less. Men started lower, at 21.8 years, and those who were both lonely and isolated had 19.6, a gap of about two years. Being lonely only or isolated only went with smaller losses, of roughly one to two years.

Splitting the conditions into physical and mental changed the picture. For the five physical diseases, women who were lonely and isolated lost 2.4 years and men 1.7. For depression and anxiety, the gaps were wider: the abstract puts them at 3.7 and 5.8 fewer years free of mental disorder for women and men respectively.

The sexes also differed in what seemed to matter. Among women, loneliness by itself went with about as large a loss of years free of mental disorder as loneliness and isolation combined, and isolation by itself with about half as much. Among men, each went with a loss of about three years and the two together with almost six, the largest gap in the study. The authors suggest that men and women may draw support in different ways, with men depending more on the plain fact of contact.

“The findings weren’t surprising to me,” Qi said in a statement from Tulane. “Previous studies found that social isolation and loneliness are related to mortality, especially premature mortality, and this study provides additional evidence.”

A healthy lifestyle narrowed the gap for isolation, less so for loneliness

The team had a second question. Lonely and isolated adults tend to smoke more, move less and eat worse, so perhaps healthy habits could offset part of the loss. Each participant was given a lifestyle score from zero to five, and the comparisons were repeated within unhealthy, medium and healthy groups.

For isolation the answer was yes. The paper reports that the loss of disease-free years associated with isolation narrowed substantially across higher levels of healthy lifestyles in both women and men. Among women, isolation went with 2.2 fewer years free of depression or anxiety in the unhealthy group, 1.1 in the middle group and 0.6 in the healthy group, a difference small enough to be chance. For loneliness there was no such pattern, and lonely women lost between about two and a half and four years free of mental disorder whatever their habits.

An earlier English study offers a reason. Lindsay Kobayashi and Andrew Steptoe followed 3,392 older adults for ten years and found that social isolation was associated with a range of health-related behaviors, including less exercise and more smoking, while loneliness was not. If isolation does part of its damage through habits, better habits would be expected to blunt it. Loneliness seems to travel by another route, and the Tulane authors point to stress and to evidence that there is substantial overlap between genes that influence loneliness and genes that influence depression.

How the life expectancy estimates sit beside earlier loneliness research

Most previous work counted deaths. In 2015 Julianne Holt-Lunstad of Brigham Young University pooled the studies then available and found that social isolation, loneliness and living alone went with a 29%, 26%, and 32% increased likelihood of mortality. A larger pooling in 2023 by Fan Wang and colleagues at Harbin Medical University covered 90 prospective cohort studies including 2,205,199 individuals and found a higher death rate with isolation, by about a third, and a smaller one with loneliness, by about a seventh.

Those reviews leave open how much of the excess is due to loneliness itself. A 2017 analysis led by Marko Elovainio at the University of Helsinki took that question to the same UK Biobank volunteers, looking for risk factors that might explain the increased mortality in socially isolated and lonely individuals. The isolated had a death rate 73% higher than others before detailed adjustment. Once income, education, habits, depressive symptoms and existing health were taken into account, the excess fell to 26%. For loneliness, an excess of 38% vanished altogether. On that analysis, loneliness predicted earlier death because the lonely were already poorer, less well and more often depressed.

The link between loneliness and later mental illness is firmer. Farhana Mann and colleagues at University College London pooled long-term studies in 2022 and found the odds of new depression more than doubled in adults who were often lonely compared with people who were not often lonely. That fits the Tulane result, in which the mental disorders account for most of the lost years.

What the new study adds is the unit. A percentage of extra risk is hard to picture. Three fewer healthy years out of twenty-five is easier to grasp, and it folds illness and death into one figure.

Limits of the loneliness and isolation measures

Five questions asked once are a thin record of a person’s social life. The authors acknowledge that because only the baseline measurement was used, they could not follow anyone who became lonelier or less so, and they cite a study in which nearly 50% of the participants experienced changes in their social isolation scores during a 4-year follow-up period.

Depression and loneliness are also hard to pull apart. Low mood prompts withdrawal and makes existing relationships feel emptier, so some participants who reported loneliness may have been in the early stages of a depression that had not yet been diagnosed. The team repeated the analysis without anyone who became ill or died in the first two years and the estimates barely moved, which eases that concern without removing it.

Illness was counted from hospital inpatient records, so conditions that never produced such a record were not included. The adjustments covered income and schooling but not baseline habits, which the authors treated as part of the pathway from isolation to illness. And the cohort itself is unusual: the paper notes that UK Biobank is not a representative sample, with a response rate of 5.5%.

Above all, the design is observational. In the authors’ words, “given the observational nature of this study, we cannot draw any conclusion with causality”.

Would easing loneliness or isolation return those years?

The study does not test any remedy, and the authors are frank about how little is known to work. They write that interventions such as social skills training and social support have been commonly applied for alleviating loneliness or isolation, but their effectiveness appears to be limited. Tulane says the group plans to examine other populations and to test causality in trials.

Until those are done, the estimate stands as a description: in this cohort, adults who were both lonely and isolated at midlife spent two to three fewer years free of major disease, and no one has yet shown that changing either would return those years.