[Therapeutic results of conization with diathermy]

To assess the frequency, histological grade and time interval of residual and recurrent lesions after loop electrosurgical excision procedure and to analyse their associated risk factors.

288 patients treated by loop electrosurgical excision procedure between February 1991 and July 1995 were prospectively included in this study. They agreed to a postoperative cytological and colposcopic follow-up with a first visit between 3 and 6 months after excision, a second one between 9 and 15 months and a third one between 24 and 36 months. We used univariate and multivariate analyses to evaluate the association between clinical variables and residual or recurrent lesions.

The pathologic diagnosis of the specimens included 21 normal cervices (7.3%), 35 CIN 1 (12.2%), 43 CIN 2 (14.9%), 180 CIN 3 (62.5%), 1 adenocarcinoma in situ (0.3%) and 8 microinvasive cancers (2.8%). The pathologist considered the excision complete in 244 cases (85%). Among the other 44 patients, 27 had involved endocervical margins, 17 involved exocervical margins but none had both margins involved. The mean postoperative follow-up +/- standard deviation was 39 +/- 13 months (range 24-68 months). Treatment failure, defined as the persistence or recurrence of a cervical lesion, was observed in 20 patients (6.9%) having a squamous cancer stage IB (n = 1), a CIN 3 (n = 15), an adenocarcinoma in situ (n = 1) or a CIN 1 (n = 1). In 6 cases the second treatment was performed before the first cytological and colposcopic visit because of an excision considered incomplete by the pathologist. Overall 19 out of the 20 treatment failures were diagnosed within the first two years of the follow-up. The endocervical localisation of the initial lesion (adjusted RR 13.7; 95% CI 1.3, 150.1; P < 0.05) and incomplete excision (adjusted RR 9.1; 95% CI 3.0, 27.3; P < 0.001) were the only independent risk factors for treatment failure.

Residual or recurrent lesions occurs rarely after loop electrosurgical excision procedure. They are favoured by totally endocervical lesion and an incomplete excision. With the association of cytology and colposcopy, 95% of treatment failures can be detected in the first 2 postoperative years.