Age, absolute CD4 count, and CD4 percentage in relation to HPV infection and the stage of cervical disease in HIV-1-positive women

A subgroup of women who are co-infected with human immunodeficiency virus type 1 (HIV-1) and human papillomavirus (HPV), progress rapidly to cervical disease. We characterized HPV genotypes within cervical tumor biopsies, assessed the relationships of cervical disease stage with age, HIV-1 status, absolute CD4 count, and CD4 percentage, and identified the predictive power of these variables for cervical disease stage in a cohort of South African women. We recruited 181 women who were histologically diagnosed with cervical disease; 87 were HIV-1-positive and 94 were HIV-1-seronegative. Colposcopy-directed tumor biopsies were confirmed by histology and used for genomic DNA extraction. The Roche Linear Array HPV genotyping test was used for HPV genotyping. Peripheral whole blood was used for HIV-1 rapid testing. Fully automated FC500MPL/CellMek with PanLeucogate (PLG) was used to determine absolute CD4 count, CD4 percentage, and CD45 count. Chi-squared test, a logistic regression model, parametric Pearson correlation, and ROC curves were used for statistical analyses. We used the Benjamini-Horchberg test to control for false discovery rate (FDR, q-value). All tests were significant when both P and q were <.05. Age was a significant predictor for invasive cervical cancer (ICC) in both HIV-1-seronegative (P < .0001, q < 0.0001) and HIV-1-positive women (P = .0003, q = 0.0003). Sixty eight percent (59/87) of HIV-1-positive women with different stages of cervical disease presented with a CD4 percentage equal or less than 28%, and a median absolute CD4 count of 400 cells/μl (IQR 300–500 cells/μl). Of the HIV-1-positive women, 75% (30/40) with ICC, possessed ≤28% CD4 cells vs 25% (10/40) who possessed >28% CD4 cells (both P < .001, q < 0.001). Furthermore, 70% (28/40) of women with ICC possessed CD4 count >350 compared to 30% (12/40) who possessed CD4 count ≤ 350 (both P < .001, q < 0.001). Age is an independent predictor for ICC. In turn, development of ICC in HIV-1-positive women is independent of the host CD4 cells and associates with low CD4 percentage regardless of absolute CD4 count that falls within the normal range. Thus, using CD4 percentage may add a better prognostic indicator of cervical disease stage than absolute CD4 count alone.