High-risk human papilloma virus testing improves diagnostic performance to predict moderate-to-high grade anal intraepithelial neoplasia in HIV-infected men who have sex with men in low-to-absent cytological abnormalities.

Background. Screening methods for anal squamous intraepithelial lesions (SILs) are suboptimal. We aimed to determine the diagnostic performance of a composite endpoint comprising anal liquid-based cytology (aLBC) and high-risk human papillomavirus (HR-HPV) testing to predict histological high-grade SILs (hHSILs). Methods. From the SeVIHanal cohort, human immunodeficiency virus (HIV)-infected men who have sex with men (MSM) who had an aLBC with concomitant HR-HPV testing were included. hHSILs were determined by high-resolution anoscopy (HRA)-guided biopsy. Results. A total of 705 visits obtained from 426 patients were included. The prevalence of HR-HPV among aLBC results were 51.9% (133/215) normal, 87.9% (20/232) low-grade SILs (LSILs), and 90.9% (149/164) high-grade SILs; P (linear association) < .001. Low prevalence of hHSILs was only observed for the composite aLBC/HR-HPV testing endpoint "normal/noHR-HPV" (10%) and "LSIL/noHR-HPV" (4%). The prognostic values (95% confidence interval) for HR-HPV to predict hHSILs in normal cytology were positive predictive value (PPV), 29.3% (25.6%-33.3%); negative predictive value (NPV), 90.2% (82.8%-94.7%); sensitivity, 83% (69.2%-92.4%); and specificity, 44.1% (36.4%-51.9%). Corresponding figures for cytologic LSILs were PPV, 39.2% (37.4%-41.1%); NPV, 96.4% (78.9%-99.5%); sensitivity, 98.8% (93.3%-99.9%); and specificity, 17.9% (12.1%-24.9%). A positive interaction and a synergistic effect for the composite endpoint were observed (relative excess risk = 1.50, attributable proportion of histological results to interaction = 0.17, synergy index = 1.24). Conclusions. HRA should not be indicated in the setting of LSILs/noHR-HPV following aLBC-based screening. In contrast, HIV-infected MSM with normal aLBC/HR-HPV infection should be considered for HRA.