Our aim was to assess the risks of directed biopsy in pregnant women and to compare its reliability to that of colposcopy. One hundred seventy six gravidas, whose final diagnosis could be ascertained either by the histological analysis of a surgical specimen or by a sufficiently informative postpartum follow-up, were examined by the same colposcopist. The reliability of colposcopy and directed biopsy were compared using the McNemar χ test for univariate analysis and the Mantel-Haenszel χ test for multivariate analysis. Directed biopsies were performed in 128 patients (72.7%), only 1 of whom experienced hemorrhage requiring vaginal packing. No other complications occurred, and premature labor was not more frequent after biopsy. The reliability of biopsy was significantly higher than that of colposcopy (82.6 vs 65.2%; p < .005). Among patients with accurate biopsy but discordant colposcopy, the final diagnosis was normal cervix, low-grade cervical intraepithelial neoplasia (CIN), high-grade CIN, and cancer in 48.4, 29.0, 22.6, and 0% of cases, respectively. When results were stratified by final diagnosis, biopsy was no more accurate than was colposcopy, regardless of cytological findings, colposcopic impression, gestational age, parity, or colposcopist's experience. One of the two occult invasive cancers correctly diagnosed by biopsy had an unsatisfactory colposcopy suggesting invasion. The difficulties of colposcopy during pregnancy and the minimal risks of directed biopsy justify a biopsy in gravidas when either cytology or colposcopy suggests at least a high-grade CIN, to ensure that no cancer has been overlooked.
Read full abstract