Background:  Screening efforts to detect anal intraepithelial neoplasia (AIN) may decrease the incidence of invasive anal cancer. We aimed to estimate the agreement between cytological and histopathological results for screening of anal dysplasia in HIV patients.

Methods:  A prospective study was carried out in a cohort of HIV⁺ patients, receiving Papanicolaou (Pap) anal screening as part of routine care. Patients with abnormal cytology (ThinPrep) were selected for high-resolution anoscopy (HRA) and anal biopsy. Reproducibility of screening measures was estimated by the weighted κ-statistic. Polymerase chain reaction (PCR) for human papillomavirus (HPV) DNA testing was available in many patients from a substudy that assessed the prevalence of anal HPV-types. We used χ² test to correlate types and number of different high-risk HPV, with histological results.

Results:  We included 212 HIV⁺ Caucasian males. Mean age 42 (SD = 8.6) years. By sexual behavior, 144 (68%) were men having sex with men (MSM) and the remaining defined themselves as heterosexual. Mean CD4 cell count 462 (SD = 286). Median (IQR) log₁₀ HIV plasma viral load 1.7 (1.7 to 3.05). Abnormal cytology was present in 106 (50%) cases (atypical squamous cells of undetermined significance [ASCUS] 39, low-grade squamous intraepithelial lesion [LSIL] 45, high-grade squamous intraepithelial lesion [HSIL] 22). To date, 59 high-resolution anoscopy examinations have been performed. In 3 patients high-resolution anoscopy did not detect any lesion. From the 56 biopsies obtained, 8 were technically unsatisfactory (14%). The table shows the agreement between 48 anal biopsies and their respective cytologies. PCR HPV DNA testing was available in 52 patients:  38 (73%) patients presented >1 high-risk HPV type, the most frequent subtypes being HPV-16 (27; 52%) and HPV-33 (15; 29%). According to anal biopsy, HPV-16 was detected significantly more often in advanced stages:  normal 3 of 11 (27%), AIN1  8 of 18 (44%), AIN2-3  10 of 12 (83%); p = 0.02. There was no difference when considering >1 high-risk HPV type:  normal 10 of 11 (91%), AIN1  11 of 18 (61%), AIN2-3  8 of 12 (67%); p = 0.56.

Conclusions:  These results show the low concordance between cytological and histological findings. The timing of follow-up should be decided according to both techniques, and the detection of high degree HPV types may be of interest for a closer follow-up.