The authors list two criticisms in their letter. 1 The first is that the follow-up period in this study was not long enough to appreciate potential new infections that could result in high grade cervical intraepithelial neoplasia (CIN) despite the presumed excision of the squamocolumnar (SC) junction. The second is that reserve cells are important progenitors of preinvasive squamous lesions of the cervix and because of their wider distribution they cannot be excised surgically.Concerning the first criticism, the authors describe studies from their institution that reported a recurrent CIN3 or adenocarcinoma in situ in less than 4 women per 1,000 treated by excision for CIN3 with negative margins, in contrast to 22% of those with positive margins. 2,3 Four recurrences per 1,000 cases of CIN3 is an impressively low recurrence rate given that CIN3 is usually a large lesion and the likelihood that some lesions might have been multifocal or were otherwise incompletely excised. In fact, this low rate could be seen as an argument for the potential value of SC junction excision in preventing CIN3. Would we not expect the risk of recurrent or new CIN3 to be higher if conization had not imposed a significant reduction in recurrence risk? Coupled with the fact that most recurrences are observed in the ectocervix, this is indirect evidence suggesting that excision of the SC junction significantly reduces risk of a "new" CIN3 within the cervix. As for the delayed presentation of some lesions, it is impossible to determine whether they were new or persistent lesions, in as much as details regarding location and HPV types are not provided in these studies. 2,3 In our discussion, we cite the paper by Kocken et al. which calculated the long-term risk of recurrence following excision. They also reported recurrences; however, when they maximized the likelihood of complete excision by both a negative cytology and HPV test, the rate of new or recurrent CIN3 was zero. 4 The second criticism was that subcolumnar reserve cellswhich cannot be completely excised because they can be found more cephalad in the endocervix-can be infected following excision and produce new lesions. Although we would never completely exclude this possibility, the data do not support isolated reserve cells as a major site of HPV infection. First, if this were the case we would expect the risk of new CIN3s to be higher following excision. Second, we would expect to find CIN3s associated with two entities in the cervix-endocervical polyps and microglandular change-that are replete with reserve cells. In fact, they are rare and uncommon sites of CIN3 development, respectively. Finally, if reserve cells were targeted, primary and recurrent CIN3s would be much more likely to be identified further up the endocervix away from the SC junction, which is not our experience or seen in our study. 5 We have demonstrated previously that reserve cells arise from specialized columnar cells, including SC junction cells. 6-8 These cells in turn undergo squamous differentiation,...