In Tuvalu, secondhand smoke tracked with sleep apnoea more closely than smoking did

Sleep apnoea research is overwhelmingly conducted in high-income countries with sleep laboratories. A study published in the Journal of Global Health on 21 August reports data from Tuvalu — a Pacific Island nation of roughly a dozen inhabited islands — on how tobacco exposure tracks with sleep apnoea risk and body measurements [s1].

The data

COMmunity-based Behaviour and Attitude surveys were conducted in Tuvalu between 2020 and 2025 among 4,066 participants drawn from Funafuti, the capital atoll, and the outer islands [s1]. Smoking status was assessed in all survey years; second-hand smoke exposure was assessed in 2022, 2023 and 2025 [s1].

The measurements taken were body mass index, waist circumference and neck circumference, along with obstructive sleep apnoea-related outcomes captured through STOP-Bang scores [s1]. Associations were assessed using multivariable linear and logistic regression, with analyses stratified by age group and island of residence [s1].

A note on what STOP-Bang is, because it matters for reading the result: it is an eight-item screening questionnaire, not a diagnosis. A score of 3 or above indicates elevated risk of obstructive sleep apnoea and is a trigger for further investigation. Nobody in this study underwent polysomnography, and the paper reports OSA risk, not confirmed OSA.

The prevalence figures

Among participants, 26.1% currently smoked and 60.7% reported second-hand smoke exposure [s1].

Mean BMI was 32.1 kg/m² [s1] — a population mean, not a subgroup mean, sitting in the range conventionally classified as obesity. And 23.7% of participants had a STOP-Bang score of 3 or above [s1], meaning close to a quarter of the surveyed population screened as elevated risk for obstructive sleep apnoea.

The associations

Current smoking was associated with obesity at an odds ratio of 1.80 (95% CI 1.30–2.48) [s1]. Second-hand smoke exposure was associated with obesity at an odds ratio of 2.15 (95% CI 1.54–3.02) [s1]. Both were associated with higher STOP-Bang scores and increased odds of a STOP-Bang score of 3 or above, with the size of the association varying by island of residence [s1].

The second-hand smoke odds ratio being larger than the active smoking one is the counterintuitive part, and it should be read carefully rather than dramatically. This is a cross-sectional association, not a dose-response comparison of inhaled exposure, and the two exposures are not independent — households where someone smokes are households where others report second-hand exposure. The finding is that reported second-hand exposure identified a group with higher odds of obesity than reported current smoking did, in this population, on these measures. It does not establish that passive exposure is biologically more harmful than active smoking.

Nor does the design establish direction. Cross-sectional data cannot say whether tobacco exposure contributed to obesity, whether the two share upstream causes, or some combination. Obesity is itself a component of the STOP-Bang instrument, which means part of the smoking–STOP-Bang association is likely running through the smoking–obesity association rather than being independent of it. The paper reports associations; it does not claim a causal chain.

The island variation

That the associations varied by island of residence [s1] is a finding worth flagging, because it implies the relationship is not uniform across a population that is often treated as homogeneous in Pacific health statistics. Funafuti and the outer islands differ in food supply, imported goods availability, health service access and household density — any of which could plausibly modify the relationship. The published abstract reports the heterogeneity without resolving its cause.

Why this study exists

The authors' framing is that tobacco use and second-hand smoke exposure remain highly prevalent in Pacific Island nations while evidence on their health consequences in those settings is limited [s1]. That is the gap the study addresses: not discovering that smoking is harmful, but producing population-level Pacific data where almost none existed.

Their conclusion is a tobacco control argument rather than a sleep medicine one — that the findings indicate the importance of strengthening implementation and enforcement of comprehensive tobacco control for sleep health and non-communicable disease prevention in Pacific Island populations [s1].

The specific contribution is connecting the two. Tobacco control in the Pacific is usually justified on cancer and cardiovascular grounds. Framing it as a sleep health and NCD measure, in a population where roughly a quarter screen positive for OSA risk and three in five report second-hand exposure [s1], is a different argument aimed at a different set of policy levers — household and indoor smoking restrictions rather than cessation programmes alone.

What to watch

The measurement that would strengthen this substantially is objective sleep testing in a Pacific setting, which would convert screening risk into diagnosed prevalence. Until then, the honest description of what Tuvalu's data shows is a high-risk population by questionnaire, strongly associated with an exposure that most of the population reports.

Sources

• [s1] Associations of smoking and second-hand smoke exposure with obesity and obstructive sleep apnoea risk in Tuvalu, Journal of Global Health, published online 21 August 2026. https://doi.org/10.7189/jogh.16.04250