Background: It is unclear whether the elevated rate of lung cancer among HIV-infected persons is due to biological effects of HIV, surveillance bias, or excess smoking. We compared the incidence of lung cancer between HIV-infected and demographically similar HIV-uninfected patients, accounting for smoking and stage of lung cancer at diagnosis. Design: Data from the Veterans Aging Cohort Study Virtual Cohort were linked to data from the Veterans Affairs Central Cancer Registry, resulting in an analytic cohort of 37 294 HIV-infected patients and 75 750 uninfected patients. Methods: We calculated incidence rates of pathologically confirmed lung cancer by dividing numbers of cases by numbers of person-years at risk. We used Poisson regression to determine incidence rate ratios (IRRs), adjusting for age, sex, race/ethnicity, smoking prevalence, previous bacterial pneumonia, and chronic obstructive pulmonary disease. Results: The incidence rate of lung cancer in HIV-infected patients was 204 cases per 100 000 person-years [95% confidence interval (CI) 167-249] and among uninfected patients was 119 cases per 100 000 person-years (95% CI 110-129). The IRR of lung cancer associated with HIV infection remained significant after multivariable adjustment (IRR 1.7; 95% CI 1.5-1.9). Lung cancer stage at presentation did not differ between HIV-infected and uninfected patients. Conclusion: In our cohort of demographically similar HIV-infected and uninfected patients, HIV infection was an independent risk factor for lung cancer after controlling for potential confounders including smoking. The similar stage distribution between the two groups indicated that surveillance bias was an unlikely explanation for this finding.