A network for students interested in evidence-based health care

Health in the Media: Loneliness and Smoking

Posted on 26th June 2020 by

Evidence Reviews

Social distancing has inevitably led to social isolation. According to the Office of National Statistics, 7.4 million people in the UK said their wellbeing had been affected in the first month of lockdown through feeling lonely. Research has pointed to an association between loneliness and smoking, but is there a causal relationship?

A recent article on MailOnline is titled: Lonely people are more likely to start smoking, smoke more and find it harder to give up, study finds. Asking Could feeling lonely make it harder to quit smoking?, Yahoo!News also covered the same study carried out by researchers from the University of Bristol and the University of Amsterdam. Published in the journal Addiction, this study used Mendelian randomisation to determine whether there are bidirectional causal effects among loneliness, smoking behaviours and alcohol (mis)use.

Disentangling cause and effect?

Differentiating between correlation and causation – and ultimately, efforts to establish the latter – remains a research challenge. The authors apply Mendelian randomisation (MR) to circumvent confounding effects of the myriad of lifestyle behaviour choices. MR groups participants according to their genes, variants of which will predispose to behaviours throughout life, but which remain immune to influence by confounding factors. By ruling out sources of bias including residual confounding and reverse causation, this method addresses the issue of confounding factors to present evidence for cause-effect relationships.

An important strength to this study was the application of multiple different MR sensitivity methods. These each make different assumptions about the nature of pleiotropy, when one genetic variant affects multiple, apparently unrelated, traits. This was done to overcome any individual limitation of a specific method.

Bidirectional MR was applied to data from genome-wide association studies (GWAS). The authors highlight that there may be selection bias, since the UKBiobank cohort (on which the loneliness GWAS was predominately based) may not be representative of the general population. If any of the investigated variables of smoking, alcohol use, and/or loneliness reduce the likelihood of participating in the UKBiobank, the results reported may actually underestimate the association.

Looking at the evidence: what do the results show?

Given that this was the first time MR had been applied to address the study’s aims, the authors highlight that evidence is still tentative. However, bidirectional effects did emerge between loneliness and smoking behaviour. Prolonged loneliness appears to increase the likelihood of starting to smoke, heavier smoking (more cigarettes-per-day) once started, as well as reducing the success of quit attempts.

This effect was bidirectional, with strong evidence for smoking initiation increasing the odds of experiencing loneliness. The authors posit that a potential mechanism behind this relationship is that nicotine from cigarette smoke interferes with neurotransmitters such as dopamine and serotonin, which are involved in the aetiology of depression. The MailOnline article writes that cigarettes may be ‘a source of comfort or alleviator of anxiety, or [provide] a familiar activity that can fill long periods of time’, suggestions for explanations that are not included in the original study.

There was no clear evidence for a causal relationship between loneliness and alcohol use, measured via the outcome variables of drinks per week and alcohol dependence. Fuller assessment of a link could also examine the quantities of alcohol consumption; there may be further complexities such that loneliness is associated with extremes of drinking frequency rather than moderate drinking. Similarly, but perhaps to a more fundamental extent, there are also issues with the judgement of loneliness as a nominal variable. What about those who, in place of prolonged periods of loneliness, experience intermittent but intense loneliness? Or those who enjoy or benefit from solitude and limited social connectedness? Considering these limitations, the authors recommend for a stronger genetic instrument for loneliness to be used in future analyses.

Looking Ahead

Although results should be considered exploratory, this study has been the first to apply MR to explore the question of causality between loneliness, smoking and alcohol use.

Despite uncertainty over how the social landscape will evolve post-pandemic, there has been an undeniable impact on loneliness, a problem that has been referred to as a ‘pandemic within the pandemic’. YouGov’s COVID-19 tracker suggests that 2.2 million people are smoking more than before lockdown; additional smoking cessation campaigns such as #QuitforCovid have been created to help those trying to stop smoking during this challenging time. Smoking has already been established as a major risk factor for non-communicable diseases, the adverse effects of which have not been halted, but rather exacerbated, by the pandemic. As society adjusts to a new normal, this study shows that there is a need to support smokers suffering from loneliness in their attempts to quit.

References (pdf)


Ji Yun Stephanie Yeung

Stephanie is a final year dental student at Barts and The London School of Medicine and Dentistry. Her interests include health communication, interprofessional education and community health promotion. You can find her over on Twitter @jiyunyeung View more posts from Ji Yun Stephanie

Leave a Reply

Your email address will not be published. Required fields are marked *

Subscribe to our newsletter

You will receive our monthly newsletter and free access to Trip Premium.