To assess the validity of colposcopy to correctly detect and grade squamous intraepithelial lesions (SIL) in Venezuelan women, we did a prospective, nonrandomized and cross sectional study on patients referred with low-grade and high-grade SIL. After a second cervical smear, they were colposcopically evaluated and biopsied. The outcome measures were interobserver variation, sensitivity, specificity, and predictive values. Ninety-nine patients were evaluable. Colposcopy had poor agreement with repeat cervical smears, and moderate to good agreement with conization biopsy (kappa = 0.55; 95% C.I.: 0.45 to 0.65), with a sensitivity of 0.87, a specificity of 0.69, a positive predictive value of 0.85 and a negative predictive value of 0.71 for high-grade SIL. The criterion of colposcopic vascular atypias was accurate enough to detect and grade SIL, showing good agreement with histopathology. Because of the disparity of results in previous reports, only a carefully designed, randomized study will settle this question.