The Fifteen-Cigarettes Claim: Where It Comes From and Whether It Has Been Replicated

The U.S. Surgeon General’s May 2023 advisory on loneliness opens with a comparison that has by now appeared in hundreds of news articles, op-eds, and policy documents: lacking social connection carries a mortality risk comparable to smoking up to fifteen cigarettes a day. It is a striking image, precise-sounding, and quotable in a way that “elevated hazard ratio” is not. It has also been repeated so often, in so many venues, that it has begun to sound like an independently established fact rather than what it actually is: one statistical comparison, made once, in one meta-analysis, and then carried forward by citation rather than by further testing.

The origin is 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. To make that figure legible to a general audience, the paper’s authors situated it alongside other well-known mortality risk factors, including smoking. That comparison — not a separate study of smokers and lonely people side by side, but a rhetorical anchoring of one effect size against known effect sizes for others — is the entire empirical basis for the fifteen-cigarette claim as it circulates today.

What “comparable to” does and does not mean

The 2010 paper’s core finding is a relative measure: people with weaker social ties had roughly 50% higher mortality risk than those with stronger ties, across a heterogeneous set of studies using different definitions of social connection, different follow-up periods, and different populations. Framing that as equivalent to a specific number of cigarettes requires converting one risk estimate into the units of another, using published data on smoking-attributable mortality. It is a reasonable back-of-envelope translation for public communication. It is not a finding that lonely people and smokers were compared directly in the same dataset, with the same outcome measure, over the same period.

Holt-Lunstad’s own 2015 follow-up in Perspectives on Psychological Science, which separated the isolation-mortality literature into distinct estimates — social isolation at an odds ratio of 1.29, loneliness at 1.26, living alone at 1.32 — did not restate the cigarette comparison. It reported the risk ratios directly and noted that these effects held after adjusting for health status, and were in fact more predictive of death in samples averaging under 65 than in older samples. That is a more cautious and more useful paper for anyone trying to gauge the actual size of the effect, but it received a fraction of the attention the cigarette framing did, because odds ratios do not travel the way an image of a lit cigarette does.

Every subsequent citation traces back to the same source

The question this piece asks is not whether social disconnection is associated with elevated mortality — the association is well established across multiple independent meta-analyses, including the American Heart Association’s 2022 scientific statement, which found isolation and loneliness associated with roughly a 30% increased risk of heart attack, stroke, or death from either. The question is narrower: has the specific fifteen-cigarettes comparison itself been tested again, by anyone, with different data?

It has not, as far as the sources available here show. The 2020 National Academies consensus report on isolation in older adults does not repeat the smoking comparison; it works instead in terms of prevalence (roughly one-quarter of adults 65 and older are socially isolated) and calls for routine clinical assessment. Holt-Lunstad’s 2021 review in the American Journal of Lifestyle Medicine, which explicitly argues that social connection belongs alongside diet, exercise, and smoking cessation in preventive health frameworks, treats the comparison as a policy argument for parity of attention rather than as a re-derived statistic. The 2023 Surgeon General advisory restates the original figure. The AHA’s statement, notably, does not use the cigarette framing at all, and explicitly names the absence of intervention evidence — not the absence of association evidence — as the field’s central gap.

So the fifteen-cigarette figure is not fabricated, and it is not obviously wrong as a rough translation of a real relative-risk estimate. But it has a single point of origin, has been restated rather than re-estimated across a decade of subsequent literature, and has migrated from a specific mortality endpoint in specific study populations into a general-purpose soundbite about loneliness as such — collapsing the isolation/loneliness distinction the 2015 paper had been careful to keep separate. The BMC Public Health review of the state of the field, published in June 2023, lists inconsistent measurement as a persistent barrier to comparing findings across studies in this literature; a statistic that has never been measured a second way is a case in point.

What would actually settle it

A genuine test of the comparison would require a study designed from the outset to estimate mortality risk from social disconnection and smoking status within the same cohort, using consistent follow-up periods and adjustment for the same confounders, then reporting the two hazard ratios on a common scale. No source reviewed here does this. Until one does, the fifteen-cigarette figure should be understood for what it is: a persuasive translation of a real association, made once in 2010, never independently re-derived, and now doing a great deal of rhetorical work on the strength of a single calculation.