To assess the performance of human papillomavirus (HPV) testing and colposcopy in detection of cervical pathology. A series of 389 women referred for colposcopy due to an abnormal Pap smear had cervical swabs analyzed for oncogenic (high-risk [HR]) HPV types using Hybrid Capture II (HC2) assay. Loop electrical excision procedure cone biopsy (88%) or colposcopic biopsy (11%) was used as the gold standard. Of the atypical squamous cells of undetermined significance (ASCUS) smears, 48% were positive for HR HPV, as compared to 76.3% of low-grade squamous intraepithelial lesions (LSIL) smears. HR HPV was detected in 66.7% and 90% of patients with cervical intraepithelial neoplasia (CIN) 1 and CIN2 (or higher), respectively. The sensitivity of the Pap smear using an ASCUS threshold in detecting high-grade CIN was 94.5% (95% confidence intervals (CI): 91-97%) and that of colposcopy 98.5% (95% CI: 95-99%). The respective specificities were 30% (95% CI: 17-28%) and 35.6% (CI: 29-42%). HC2 test had comparable sensitivity, 90% (95% CI: 85-93%), but higher specificity, 54.3% (95% CI: 47-61%). Combining HC2 test with Pap increased specificity, 66.7% and 41.3% for ASCUS and LSIL cutoff, respectively. The minor-abnormality threshold together with HC2 increased specificity of colposcopy with no changes in sensitivity. High viral load (>100 relative light unit/positive control) was associated with significant disease. HPV DNA testing improves the accuracy of colposcopy in the detection of high-grade CIN in women with ASCUS or LSIL smears.