Loneliness is associated with an increased risk of developing depression, a longitudinal cohort study has found.¹

The English Longitudinal Study of Ageing (ELSA) recruited 11,391 participants aged 50 years and over who were living in a private household in 2002. Participants were followed up every two years until 2016/17. At each review, depressive symptoms were measured using a revised 7-item version of the Centre for Epidemiologic Studies Depression Scale (CES-D), with loneliness removed.

A baseline assessment of loneliness was made at the second follow-up appointment (2004-5) using the 3-item version of the University of California Loneliness Scale (R-UCLA). The scale asks respondents to rate how often they: feel they lack companionship; feel left out; feel isolated from others, using a 3-point Likert scale. The range of possible scores is 3-9.

Overall, 4,211 participants for whom there were complete data on exposure, outcome and confounders were included in the analysis. Results were adjusted for social variables, polygenic risk scores, sociodemographic factors, health indicators and baseline depressive symptoms. Participants had a mean age of 65.1 years and 55% were female. The mean baseline loneliness score was 4.12.

After full adjustment, for every 1-point increase in the baseline loneliness score there was a 0.16 increase in the depressive symptom severity score (95% CI: 0.13-0.19), averaged across all follow-ups.

Depression was defined as a CES-D score ≥ 2. The population attributable fraction (PAF) – the fraction of all cases of depression attributable to loneliness – was estimated to be 18% (95% CI: 12-24%) at one year of follow-up and 11% (95% CI: 3-19%) after 12 years of follow-up.

Loneliness is not the same as isolation: some people can be alone for long periods of time without feeling lonely whereas others can feel isolated even when surrounded by other people. According to the cognitive discrepancy model, loneliness results from a discrepancy between the social relationships someone needs (or desires) and the relationships he or she actually has.² Loneliness is characterised by feelings of social isolation, lack of companionship and rejection by peer groups.³ It has been suggested that social loneliness – the lack of a social network – should be distinguished from emotional loneliness – the lack of a close emotional attachment – but loneliness scales do not attempt to do this.²

The CES-D is a sensitive depression screening tool but it lacks specificity and ‘should not be used as an isolated diagnostic measure of depression.’⁴ Hence the PAF estimation can only be a rough approximation although the authors point out that measurement errors would be expected to be random with respect to their hypotheses.

Observational studies cannot demonstrate causation but a causal link between loneliness and depression is plausible, with a number of potential mediators. It has been suggested that perceived social isolation is ‘tantamount to feeling unsafe’ and this may lead to anxiety, insomnia, social distancing behaviour and low self-esteem. Loneliness has been linked to inactivity, obesity and alcohol abuse.⁵

It is of concern that the Covid lockdown has exacerbated loneliness. The Covid Social Survey has collected data from more than 70,000 people in the UK using the R-UCLA. The average UCLA score during the most stringent period of lockdown was 5, compared with a pre-lockdown average of 4.3.⁶

The authors of the present study recommend that ‘clinicians should be aware of loneliness as a possible risk factor for depression, and consider strategies to address this loneliness.’

Group interventions appear to be most effective and should focus on providing social support, increasing opportunities for social interaction, enhancing social skills and addressing maladaptive social cognition.^7,5