Has Anyone Replicated the 'Loneliness Is Like Smoking 15 Cigarettes a Day' Claim?

The 2023 U.S. Surgeon General advisory on loneliness states that the mortality risk of social disconnection is comparable to smoking up to 15 cigarettes a day. The line has since appeared in newspaper coverage, conference keynotes, and policy documents on several continents, usually presented as an established empirical fact. It is worth asking what study actually produced that number, and whether anyone has since checked it.

Where the comparison originates

The figure does not come from a study that measured loneliness and smoking side by side in the same cohort and computed a shared risk metric. It traces back to Julianne Holt-Lunstad’s 2010 meta-analysis in PLoS Medicine, which pooled 148 studies covering 308,849 participants and found that stronger social relationships were associated with a 50% increased likelihood of survival over the follow-up periods studied. The paper’s contribution was to note that this effect size was comparable in magnitude to other well-established mortality risk factors, obesity among them, and, by extension, smoking. That comparison was a statement about the relative size of an odds ratio, not a direct calculation of cigarette-equivalents.

The “15 cigarettes” framing sharpened over the following decade as the comparison was repeated, first in academic review pieces and then in advocacy and policy writing, including the UK’s 2018 loneliness strategy, which cited the broader body of Holt-Lunstad’s work in making the case for a national strategy. Holt-Lunstad’s 2015 meta-analysis in Perspectives on Psychological Science refined the picture considerably: social isolation carried an odds ratio of 1.29, loneliness 1.26, and living alone 1.32 for early mortality, with the deficits proving more predictive of death in samples averaging under 65 than in older cohorts. Those numbers are more precise than the original 50% figure, and they are not restated in cigarette terms anywhere in that paper. The specific “15 cigarettes” phrasing appears to have calcified from the original 2010 comparison rather than from the 2015 refinement, even though the 2015 study is the more rigorous of the two.

What would count as replication

A genuine replication of the smoking comparison would need a study that models loneliness or isolation and smoking as competing or comparable exposures within the same dataset, using a consistent outcome measure and adjustment strategy for both. That is a demanding design, because smoking dose-response curves are themselves built from decades of cohort data using standardized measures of cigarettes per day, while loneliness and isolation are measured with instruments that vary widely across studies, from single-item self-report questions to the 20-item UCLA Loneliness Scale. The 2023 BMC Public Health review of the loneliness and isolation literature specifically names inconsistent measurement as a barrier to comparing findings across studies. That is not a minor caveat. It means that even a study attempting the direct comparison would be measuring loneliness on a different scale than the studies that established the smoking risk curve, and translating between them requires an extra layer of statistical assumption that the original 2010 comparison did not spell out.

Nothing in the record reviewed here does this. The American Heart Association’s 2022 scientific statement on isolation and cardiovascular and brain health, one of the more careful recent syntheses, finds social isolation and loneliness associated with roughly a 30% increased risk of heart attack, stroke, or death from either, close to Holt-Lunstad’s earlier estimates but expressed in outcome-specific terms rather than as an all-cause mortality equivalent to a specific behavior. Notably, the AHA statement explicitly identifies the absence of intervention evidence as the field’s central gap, and it does not restate or attempt to verify the smoking comparison. If the most recent large scientific statement on the topic declines to repeat the figure, that is itself informative.

Why the comparison persists anyway

The claim survives for reasons that have little to do with whether it has been checked. It is vivid, it maps an unfamiliar risk onto a familiar one, and it serves an argumentative purpose that both public health officials and advocacy organizations need: establishing that loneliness belongs in the same category of concern as long-recognized behavioral risk factors. Holt-Lunstad’s own later writing, including a 2021 review in the American Journal of Lifestyle Medicine, explicitly argues for treating social connection as a modifiable protective factor that belongs alongside diet, exercise, and smoking cessation in preventive health frameworks. That is a reasonable policy argument. It is a different claim from asserting a specific, checked, cigarette-equivalent dose-response relationship, and the distinction has been lost in most of the citations that followed.

What this does and does not mean for the underlying finding

None of this undermines the broader finding that social disconnection is associated with meaningfully elevated mortality risk. That result has been replicated repeatedly, across the 2010 and 2015 Holt-Lunstad meta-analyses and the 2022 AHA statement, using different samples, different outcome definitions, and different research teams. The association is one of the more consistently observed findings in this literature. What has not been replicated is the specific rhetorical device used to communicate its size. The “15 cigarettes” figure functions as an illustration lifted from a single early paper’s discussion section, repeated across a decade of policy and media writing, and never independently re-derived.

A stronger version of the claim would require a cohort study measuring both smoking intensity and a validated loneliness or isolation instrument in the same population, modeling both against a common mortality outcome with comparable adjustment for confounders, and reporting the resulting risk ratios on the same scale. Until that exists, the comparison should be treated as a useful analogy for explaining why loneliness research got the public health attention it did, not as a finding that has itself been tested and confirmed.