Yield of Screening Colposcopy Among Human Immunodeficiency Virus-Infected Women

Abstract

Women infected with human immunodeficiency virus (HIV) are at greater risk of developing cervical neoplasia and of having rapidly progressing disease. This study was designed to assess the prevalence of and to identify risk factors for cervical intraepithelial neoplasia (CIN) in women with HIV who were enrolled in the Women’s Interagency HIV Study in Chicago. All patients who presented to the clinic over a 1-year period completed a questionnaire aimed at assessing risk factors for CIN. A subsequent clinical examination included cervical cytology and testing for HIV. Colposcopy was performed routinely at two of the study sites as part of all initial examinations, but in the other locations, colposcopy was performed only when a cytologic abnormality was found. Seventy-six of 103 women fulfilled inclusion criteria for this study. Forty-three were white, 23 African American, and 10 Hispanic. Twenty-eight had never smoked cigarettes, and 31 were current smokers. Fifty-six women reported fewer than 10 male sexual partners, and 20 said they had had at least 11 partners, with the numbers ranging above 100 for three of these women. Thirty-two reported previous abnormal cervical cytology. A cytologic diagnosis of a squamous intraepithelial lesion (SIL) was made in 10 patients, eight low grade and two high grade. Thirteen cases of CIN1 were identified by colposcopically directed biopsy (Table 1). No disease more serious than CIN1 was seen. Table 1: Distribution of colposcopic findings according to cytological resultWhen a broad definition of abnormal cytology including atypical squamous cell of undetermined significance and high- and low-grade SIL was used, the sensitivity of cytology for predicting SIL was 65 percent, the specificity was 68 percent, the positive predictive value was 37 percent, and the negative predictive value was 89 percent. Corresponding values for the ability of cytology to predict CIN1 were a sensitivity of 62 percent, specificity of 65 percent, positive predictive value of 27 percent, and negative predictive value of 89 percent. In all, 35 women had human papilloma virus detected on the cervix. The correlation of cytology with human papilloma virus detection was significant (P < 0.001), but neither colposcopy nor histology had a significant association with human papilloma virus findings. A multivariate logistical regression analysis found that a cytology result of atypical squamous cell of undetermined significance or SIL was an independent predictor of biopsy-proven CIN (odds ratio versus benign smear, 6.6; P < 0.04). The only behavioral characteristic that reached significance was a history of trading sex for money or drugs (odds ratio versus no such history, 18.4; P < 0.02). J Lower Genital Tract Dis 1999;3:185–190