AN ATTEMPT TO ESTIMATE THE INCIDENCE OF CERVICAL DYSPLASIA IN A GROUPOF NEW-ZEALAND WOMEN USING CONTRACEPTION

We attempted to estimate incidence rates of cervical dysplasia and car cinoma in a cohort of 7,199 New Zealand women using contraception. The cohort was followed prospectively with periodic cervical smears, inte nded to be annual, over a 6-year period. The principal outcome investi gated was a diagnosis of cervical dysplasia of any degree, from mild t o severe dysplasia and including carcinoma in situ or invasive carcino ma, made cytologically by a central study laboratory and confirmed by histology or deoxyribonucleic acid (DNA) evaluation. These diagnoses a re jointly referred to as ''dysplasia.'' Two successive negative (nond ysplastic) smears were required before a woman was considered eligible for the analysis of incidence. Even after these two negative smears, the estimated ''incidence'' of dysplasia declined markedly in each of the 5 years of the study, particularly among women who provided negati ve smears in each prior year. This suggests that prevalent cases were being diagnosed even after five or more negative smears. Assuming that nearly all of the prevalent cases were removed after five negative sm ears, our estimate of the annual incidence of cervical dysplasia in th is population during this time period would be of the order of 5 per 1 ,000 per year. We conclude that the sensitivity of cervical testing fo r identifying cervical dysplasia is quite low in this population but i s consistent with values reported from some other populations. Age at first intercourse, age at first pregnancy, number of sex partners, and current cigarette smoking were strongly associated with risk of dyspl asia. Our data are equivocal on the question of whether age at first i ntercourse is a risk factor independently of the closely associated va riable, number of sex partners.