Objective. To compare the histomorphologic and colposcopic results of cold
knife conization and loop excision.
Methods. Sixty-six women were randomly allocated to have the cone specimen
removed by cold knife excision (n = 38) or loop excision (n = 28). Subjects
eligible for inclusion were those who presented histologically verified gr
ade 3 cervical intraepithelial neoplasia (CIN) or grade 2 CIN with squamoco
lumnar junction not seen.
Results. The mean height of the cone specimens was greater in the cold knif
e group [18.9 mm (SD = 5.5) and 12.8 mm (SD = 4.3), respectively; P = 0.000
1], as was the frequency of clear margins (100 and 80%, respectively; P = 0
.001), In the loop excision group, thermal injuries were present in half of
the cone sections. The median (range) thickness of thermal injury was 0.98
mm (0-1.5 mm) in the ectocervix and 0.95 mm (0-1.75 mm) in the endocervix.
Histologic evaluation of the endocervical margins was not possible in 2 ca
ses (7%). At follow-up colposcopy, evaluation of the entire squamocolumnar
junction was possible in 15 (39%) and 20 (71%) women, respectively (P < 0.0
1). Four patients in the cold knife group and 6 in the loop group had histo
logically confirmed persistent dysplasia (P > 0.05), yielding success rates
of 90 and 79%, respectively (P > 0.05).
Conclusions. Loop excision provides a sample that is adequate for histologi
c evaluation in most cases, results in the same success rate as cold knife
conization, and allows optimal colposcopic surveillance in significantly mo
re cases than cold knife excision. (C) 1999 Academic Press.