Predicting persistent/recurrent disease in the cervix after excisional biopsy

Objective: To evaluate the ability of various factors to predict persistent/recurrent disease after excisional biopsy of the transformation zone (cold knife conization or loop electrosurgical excision procedure) with special attention to the endocervical curettage (ECC). Study Design: We reviewed the charts and histopathology findings of 152 women who underwent endocervical curettage at the time of conization (cold knife conization) or loop electrosurgical excision procedure (LEEP). Age, histopathologic findings on the cervical conization specimen, ectocervical margin, endocervical margin, and ECC specimens were assessed. These findings were analyzed for a relationship with the presence of cervical disease on subsequent follow-up (to include hysterectomy, repeat conization, colposcopically directed biopsies, endocervical curettage, and/or cytology). Results: Positive endocervical margin (odds ratio [OR], 9.168; 95% confidence interval [95% CI], 3.939, 23.488), positive ectocervical margin (OR, 3.561; 95% CI, 1.626, 7.799), positive specimens (OR, 17.683; 95% CI, 5.308, 58.912), and severity of disease (OR, 2.730; 95% CI 1.507, 4.947) on the conization were all individually significantly associated with the presence of persistent/recurrent disease. Age of the patient at the time of cervical conization was not statistically associated with the ability to predict persistent/recurrent disease. In the multivariate analysis, the endocervical curettage (OR, 8.710; 95% CI, 2.302, 32.958) and the endocervical margin status (OR, 9.170; 95% CI, 2.887, 29.125) together were significant predictors of persistent/ recurrent disease after adjusting for the other variable. However, when the degree of dysplasia and ectocervical margin status was included in the multivariate analysis, endocervical margin status (OR, 6.761; 95% CI, 2.657, 17.202) and severity of cervical disease (OR, 1.930; 95% CI, 1.038, 3.59) were the only statistically significant predictors of persistent/recurrent cervical neoplasia. Conclusion: In this retrospective analysis, positive endocervical or ectocervical margin, positive ECC specimens, and severity of cervical disease were all predictors of persistent/recurrent disease. However, on the multivariate stepwise logistic regression analysis, only endocervical margin status and severity of neoplasia significantly predicted the occurrence of persistent/recurrent disease. The results of the ECC, after adjustment for the degree of dysplasia and the endocervical margin status, do not add incremental value to the prediction of persistent/recurrent disease. At this time, ECC does not need to be routinely performed at the time of excisional biopsy of the cervical transformation zone.