The invasive potential of cervical intraepithelial neoplasia 3 (CIN3; also termed stage 0 carcinoma) has been poorly defined. At the National Women's Hospital, Auckland, New Zealand, treatment of CIN3 was withheld from a substantial number of women between 1965 and 1974 as part of an unethical clinical study. The resulting variation in management allows comparison of the long-term risk of invasive cancer of the cervix in women whose lesion was minimally disturbed with those who had adequate initial treatment followed by conventional management. We aimed to estimate the long-term risk of invasive cancer in these two groups of women. A judicial inquiry referred for independent clinical review in 1988 all women for whom there remained doubt about the adequacy of their management.

Between February, 2001, and December, 2004, medical records, cytology, and histopathology were reviewed for all women with CIN3 diagnosed between 1955 and 1976, whose treatment was reviewed by judicial inquiry and whose medical records could be located, and linkages were done with cancer and death registers and electoral rolls. To take into account the probability that the CIN3 lesion had been completely removed, we classified adequacy of treatment by type of procedure, presence of CIN3 at the excision margin, and subsequent cytology. The primary outcome was cumulative incidence of invasive cancer of the cervix or vaginal vault. Follow-up continued until death or Dec 31, 2000, whichever came first. Analyses accounted for procedures during follow-up.

1229 women whose treatment was reviewed by the judicial inquiry in 1987–88 were included. Of these, 48 records (4%) could not be located and 47 women (4%) did not meet the inclusion criteria. At histopathological review, a further 71 (6% of 1134) women were excluded because the review diagnosis was not CIN3. We identified outcomes in the remaining 1063 (86% of 1229) women diagnosed with CIN3 at the hospital in 1955–76. In 143 women managed only by punch or wedge biopsy, cumulative incidence of invasive cancer of the cervix or vaginal vault was 31·3% (95% CI 22·7–42·3) at 30 years, and 50·3% (37·3–64·9) in the subset of 92 such women who had persistent disease within 24 months. However, cancer risk at 30 years was only 0·7% (0·3–1·9) in 593 women whose initial treatment was deemed adequate or probably adequate, and whose treatment for recurrent disease was conventional.

This study provides the most valid direct estimates yet available of the rate of progression from CIN3 to invasive cancer. Women with untreated CIN3 are at high risk of cervical cancer, whereas the risk is very low in women treated conventionally throughout.

Cancer Society of New Zealand, Wellington, New Zealand.