You are here
CD4 Measures as Predictors of Lung Cancer Risk and Prognosis in HIV Infection
Keith Sigel1, Kristina Crothers2, Kirsha Gordon3, Sheldon Brown4, David Rimland5, Maria Rodriguez-Barradas6, Cynthia Gibert7, Matthew B. Goetz8, Roger Bedimo9, Robert Dubrow10
1 Division of General Internal Medicine, Icahn School of Medicine at Mount Sinai, New York, NY, United States. 2 Medicine, University of Washington School of Medicine, Seattle, WA, United States. 3 VA Connecticut Healthcare System and Yale University Schools of Medicine and Public Health, New Haven, CT, United States. 4 Medicine, James J. Peters VA Medical Center, Bronx, NY, United States. 5 Medicine, Atlanta VA Medical Center, Atlanta, GA, United States. 6 Medicine, Michael E. DeBakey Veterans Affairs Medical Center, Houston, TX, United States. 7 Medicine, Washington DC Veterans Affairs Medical Center, Washington, DC, United States. 8 Medicine, Los Angeles VA Medical Center, Los Angeles, CA, United States. 9 Medicine, Veterans Affairs North Texas Health Care System, Dallas, TX, United States. 10 Epidemiology, Yale University School of Public Health, New Haven, CT, United States.
Background: Immunodeficiency may adversely affect both lung cancer risk and outcomes in the setting of HIV infection. Using data from a large HIV cohort, we investigated relationships between 1) recent and cumulative measures of CD4 and CD8 count and lung cancer incidence and 2) CD4 measures and lung cancer prognosis.
Methods: We followed 26,065 HIV+ subjects from the Veterans Aging Cohort Study (VACS) for a minimum of 2 years, during 1999-2010. We linked VACS with the VA Central Cancer Registry to obtain incident, pathologically confirmed lung cancer cases. Our exposures of interest were longitudinal CD4 (<200 cells/mm3 [c/mm3], 200-500 c/mm3 or >500 c/mm3), CD4/CD8 (<0.4 or ≥0.4) and CD8 (≥850 c/mm3 or <850 c/mm3). We used Cox regression models to investigate the effect of time-updated CD4, CD4/CD8 ratio and CD8 measures on lung cancer risk, including values lagged 12 months, and 12- and 24-month simple moving averages. Models were adjusted for age, sex, race/ethnicity, smoking, and history of pneumonia and COPD. We then collected all non-small cell lung cancer cases from the full VACS (HIV+ and HIV- subjects from 1996-2010) and used conditional probability function regression (a competing risks method to account for higher risk of non-lung cancer death in HIV+) to compare lung cancer-specific survival in 3 groups: HIV- (n=679), HIV+ with CD4≥200 c/mm3 at cancer diagnosis (n=299) and HIV+ with CD4<200 c/mm3 at cancer diagnosis (n=113). These analyses were adjusted for demographics, comorbidity score, cancer stage and histology, cancer diagnosis year, and cancer treatment.
Results: We identified 325 (1.2%) cases of incident lung cancer in our cohort. In adjusted models (Table 1), a 12 month lagged CD4 count <200 c/mm3 as well as moving averages of both CD4<200 c/mm3 and CD4 200-500 c/mm3 were significantly associated with increased lung cancer incidence. In similar adjusted models, 12-month moving averages of CD4/CD8 ratio <0.4 were also significantly associated with increased risk of lung cancer. Among lung cancer cases, lung cancer-specific survival did not differ between either of the HIV+ groups and the HIV- group (p>0.05) after adjustment.
Conclusions: In our large HIV cohort, we found that several measures of recent and cumulative exposure to immunodeficiency were associated with increased lung cancer risk. CD4 count at time of cancer diagnosis was not associated with cancer-specific survival after accounting for competing risk of non-lung cancer death.