Obstructive sleep apnoea (OSA) has been linked to cancer in several clinical and community-based cohorts. The effect in community-based studies free of clinical referral bias needs to be replicated.

We pooled data from three community-based prospective cohorts (n(total)=1,467). All cohorts had objective data on OSA along with registry linkage data on cancer and cancer mortality. Analyses for different measures of OSA (Apnea-Hypopnea-Index (AHI), Oxygen Desaturation Index (ODI), and minimal saturation) as risk factors for cancer incidence (all cancers) were performed using Cox proportional hazards models.

In this pooled community-based cohort we did not find an overall increased risk of cancer after adjustment for age, sex and BMI (HR for AHI (95%CI) =1.00 (0.98; 1.01) and HR for ODI (95%CI) =0.99 (0.97; 1.01)). A post-hoc sub-analysis, showed an increased risk of cancer with increasing AHI (1.02 (1.001; 1.04)) in people with BMI?30kg/m2. We did not observe an increased cancer risk in sex or age stratified analysis and stratifying by daytime sleepiness did not influence the association. Cancer mortality was not significantly associated with OSA.

In conclusion, we did not observe an overall increased risk of cancer or cancer mortality in relation to OSA, however our confidence limits remain quite wide for important diagnostic categories of sleep apnea severity. The relationship between OSA and cancer needs further investigation in a comprehensive multi-cohort approach with greater statistical precision.