Natural History of Anal Dysplasia in an HIV-Infected Clinical Care Cohort: Estimates Using Multi-State Markov Modeling

Objectives: (1) To model the natural history of anal neoplasia in HIV-infected patients using a 3-state Markov model of anal cancer pathogenesis, adjusting for cytology misclassification; and (2) to estimate the effects of selected time-varying covariates on transition probabilities. Design: A retrospective cytology-based inception screening cohort of HIV-infected adults was analyzed using a 3-state Markov model of clinical pathogenesis of anal neoplasia. Methods: Longitudinally ascertained cytology categories were adjusted for misclassification using estimates of cytology accuracy derived from the study cohort. Time-varying covariate effects were estimated as hazard ratios. Results: (1) There was a moderate to high probability of regression of the high grade squamous intraepithelial lesion (HSIL) state (27-62%) at 2 years after initial cytology screening; (2) the probability of developing invasive anal cancer (IAC) during the first 2 years after a baseline HSIL cytology is low (1.9-2.8%); (3) infrared coagulation (IRC) ablation of HSIL lesions is associated with a 2.2-4.2 fold increased probability of regression to = 350/ mm(3) are each associated with reduced probability of progression from