RESUMEN
OBJECTIVE: To assess the role of additional biopsies performed with loop electrosurgical excisional procedure (LEEP) in predicting the likelihood of persistent high grade intraepithelial neoplasia. METHODS: Clinicopathologic data were abstracted from women who underwent excision of high grade intraepithelial lesions between 2001 and 2014. Persistent disease was defined as uninterrupted high grade intraepithelial neoplasia, whereas recurrent disease was defined as disease diagnosed ≥1year after treatment with intervening normal evaluation. Chi-square and Fisher's exact tests were used to examine associations between demographic and histologic parameters and clinical outcomes. RESULTS: A total of 606 women underwent LEEP for high grade intraepithelial neoplasia (HSIL), of whom, 178 (29%) were additionally evaluated by endocervical curettage, 80 (13%), top hat and 99 (16%), both procedures. With mean follow-up of 1.9±1.5years, persistent disease was identified in 87 women (14%) while recurrent disease was diagnosed in 20 (3%). After adjusting for age, HIV status and histologic grade of disease, the presence of disease at the endocervical margin (aOR=2.2, 95% CL 1.8-5.5, p<0.0001), with endocervical curettage (aOR=2.39, 95% CL 1.2-9.9, p=0.025) or on top hat (aOR=4.0, 95% CL 1.1-16.2, p=0.04) correlated with the likelihood of persistent but not recurrent disease. Only endocervical margin status remained predictive (p=0.03) of outcome after controlling for pre-procedure likelihood of endocervical disease. Sensitivity of endocervical margin status for persistent disease was 56.9% with specificity of 72.2%. Positive predictive value (PPV) was 24.9% and negative predictive value (NPV) 90.9%. CONCLUSIONS: Despite frequent use of additional procedures to sample the endocervix, these strategies do not improve the ability of endocervical margin status to predict persistent or recurrent dysplasia.