Virologic versus cytologic triage of women with equivocal Pap smears: a meta-analysis of the accuracy to detect high-grade intraepithelial neoplasia.

BACKGROUND The appropriate management of women with minor cytologic lesions in their cervix is unclear. We performed a meta-analysis to assess the accuracy of human papillomavirus (HPV) DNA testing as an alternative to repeat cytology in women who had equivocal results on a previous Pap smear. METHODS Data were extracted from articles published between 1992 and 2002 that contained results of virologic and cytologic testing followed by colposcopically directed biopsy in women with an index smear showing atypical cells of undetermined significance (ASCUS). Fifteen studies were identified in which HPV triage and the histologic outcome (presence or absence of a cervical intraepithelial neoplasia of grade II or worse [CIN2+]) was documented. Nine, seven, and two studies also documented the accuracy of repeat cytology when the cutoff for abnormal cytology was set at a threshold of ASCUS or worse, low-grade squamous intraepithelial lesion (LSIL) or worse, or high-grade squamous intraepithelial lesion (HSIL) or worse, respectively. Random-effects models were used for pooling of accuracy parameters in case of interstudy heterogeneity. Differences in accuracy were assessed by pooling the ratio of the sensitivity (or specificity) of HPV testing to that of repeat cytology. RESULTS The sensitivity and specificity were 84.4% (95% confidence interval [CI] = 77.6% to 91.1%) and 72.9% (95% CI = 62.5% to 83.3%), respectively, for HPV testing overall and 94.8% (95% CI = 92.7% to 96.9%) and 67.3% (95% CI = 58.2% to 76.4%), respectively, for HPV testing in the eight studies that used the Hybrid Capture II assay. Sensitivity and specificity of repeat cytology at a threshold for abnormal cytology of ASCUS or worse was 81.8% (95% CI = 73.5% to 84.3%) and 57.6% (95% CI = 49.5% to 65.7%), respectively. Repeat cytology that used higher cytologic thresholds yielded substantially lower sensitivity but higher specificity than triage with the Hybrid Capture II assay. The ratio of the sensitivity of the Hybrid Capture II assay to that of repeat cytology at a threshold of ASCUS or worse pooled from the four studies that used both triage tests was 1.16 (95% CI = 1.04 to 1.29). The specificity ratio was not statistically different from unity. CONCLUSION The published literature indicates that the Hybrid Capture II assay has improved accuracy (higher sensitivity, similar specificity) than the repeat Pap smear using the threshold of ASCUS for an outcome of CIN2+ among women with equivocal cytologic results. The sensitivity of triage at higher cytologic cutoffs is poor.