Elevated Incidence of Lung Cancer Among HIV-Infected Individuals

PURPOSE: People with HIV infection in the United States frequently smoke tobacco. We sought to characterize lung cancer incidence among HIV-infected individuals, examine whether cancer risk was related to HIV-induced immunosuppression, and assess whether the high prevalence of smoking explained elevated risk. METHODS: We conducted a retrospective cohort study at an HIV specialty clinic in Baltimore, MD (1989-2003). Incident lung cancers were identified using hospital records. We used negative binomial regression to compare incidence across subgroups defined by demographics, use of highly active antiretroviral therapy (HAART), and HIV markers. Standardized incidence ratios (SIRs) compared incidence with an urban reference population (Detroit, MI). We adjusted SIRs for the effect of smoking, using smoking prevalences estimated from part of the cohort and the general population. 95% CIs and P values were two sided. RESULTS: Thirty-three lung cancers were observed among 5,238 HIV-infected patients (incidence: 170 per 100,000 person-years). Incidence increased with age (P < .0001), but did not differ by sex, race, or CD4 count. Incidence tended to increase with calendar year (P = .09) and HAART use (P = .10), and was inversely related to HIV viral load (P = .03), but these associations were attenuated with age adjustment. The SIR was 4.7 (95% CI, 3.2 to 6.5) versus the general population. Twenty-eight lung cancer patients (85%) and 69% of the cohort were smokers. After smoking adjustment, risk remained elevated (SIR, 2.5; 95% CI, 1.6 to 3.5). CONCLUSION: Lung cancer risk was substantially elevated in HIV-infected individuals. Incidence was unrelated to HIV-induced immunosuppression. Notably, incidence remained high after adjustment for smoking, suggesting the involvement of additional factors.