ObjectiveTo investigate whether the higher risks of certain cancers associated with high cardiorespiratory fitness can be explained by increased detection and unobserved confounders.DesignNationwide sibling-controlled cohort study of adolescents.SettingSweden.Participants1 124 049 men of which 477 453 were full siblings, who underwent mandatory military conscription examinations between 1972 and 1995 at a mean age of 18.3 years.Main outcome measuresHazard ratios (HR) and 95% confidence intervals (CI) of overall cancer diagnosis and cancer mortality, and 14 site-specific cancers (diagnosis or death), as recorded in the Swedish National Patient Register or Cause of Death Register until 31 December 2023, modelled using flexible parametric regressions.ResultsParticipants were followed until a median (maximum) age of 55.9 (73.5) years, during which 98 410 were diagnosed with cancer and 16 789 had a cancer-related death (41 293 and 6908 among full siblings respectively). The most common cancers were non-melanoma skin (27 105 diagnoses & 227 deaths) and prostate cancer (24 211 diagnoses & 869 deaths). In cohort analysis, those in the highest quartile of cardiorespiratory fitness had a higher risk of prostate (adjusted HR 1.10; 95% CI: 1.05 to 1.16) and skin cancer (e.g., non-melanoma HR 1.44; 1.37 to 1.50) compared to those in the lowest quartile, which led to a higher risk of any type of cancer diagnosis (HR 1.08; 1.06 to 1.11). However, those in the highest quartile had a lower risk of cancer mortality (HR 0.71; 0.67 to 0.76). When comparing full siblings, and thereby controlling for all behavioural, environmental, and genetic factors they share, the excess risk of prostate (HR 1.01; 0.90 to 1.13) and skin cancer (e.g., non-melanoma HR 1.09; 0.99 to 1.20) attenuated to the null. In contrast, the lower risk of overall cancer mortality was still statistically significant after control for such shared confounders (HR 0.78; 0.68 to 0.89). For other site-specific cancers, the influence of such confounding tended to vary, but none showed the same excess risk as prostate and non-melanoma skin cancer.ConclusionsThe association between high levels of adolescent cardiorespiratory fitness and excess risk of some cancers, such as prostate and non-melanoma skin cancer, appears to be fully explained by unobserved confounders shared between full siblings. However, the protective association with cancer mortality persists even after control for such confounding.Summary boxWhat is already known on this topicAdolescent physical activity and cardiorespiratory fitness are considered important factors for the prevention of cancer based on evidence from observational studies.Observational studies are, however, vulnerable to unobserved confounders and bias processes, including health-seeking behaviours and genetic and environmental confounders.These biases could explain why prior studies have found that high adolescent cardiorespiratory fitness is associated with higher risks of some cancers, typically low-mortality cancers such as prostate and non-melanoma skin cancer.What this study addsThis nationwide cohort study of 1.1 million male adolescents showed that while higher cardiorespiratory fitness was associated with excess risk of the most common cancers - prostate and non-melanoma skin - these associations attenuated to the null when accounting for behavioural, environmental, and genetic confounders shared between full siblings.In contrast, high adolescent cardiorespiratory fitness was associated with a lower risk of overall cancer mortality, which remained after controlling for unobserved confounders shared between full siblings.