Increased incidence of cervical and vaginal dysplasia in 3980 diethylstilbestrol-exposed young women: Experience of the national collaborative diethylstilbestrol adenosis project

The purpose of this article is to report the incident cases of intraepithelial neoplasia of the cervix and vagina among patients in the National Collaborative Diethylstilbestrol Ad-enosis Project. Incidence rates for dysplasia/carcinoma in situ have been calculated and analyses have been performed in an attempt to determine whether intrauterine diethylstilbestrol exposure is associated with increased rates of dysplasia/carcinoma in situ and, if so, to identify any other factors associated with its occurrence. The categories included diethylstilbestrol-exposed daughters identified through review of obstetric records ("prenatal record review"), nonexposed daughters identified through review of the same records ("controls"), exposed daughters who were referred to the project by other physicians ("referral"), and exposed daughters who themselves requested entry into the project ("walk-in"). Within the first two groups, exposed and unexposed offspring were paired for cohort control studies; the 744 pairs were either sisters (197 pairs) or were matched by date of birth (within 6 months), age of mother (within 5 years), race, and center (547 pairs). Comparison of the exposed and unexposed matched cohorts revealed that both groups were remarkably similar with regard to many variables, including number of sexual partners. Although both groups exhibited similar frequencies of several sexually transmitted diseases (syphilis, gonorrhea, and venereal "warts"), the exposed patients reported increased rates for genital herpes (11.8 vs. 6.3 per cent). Changes observable by colposcopy that extended into the vagina (vaginal epithelial changes) were identified more often in exposed women (44.2 vs. 8.0 per cent). In the pairs of record review exposed and control participants (matched cohorts), the incidence rates for dysplasia by either biopsy or cytological testing were significantly increased in exposed women: 15.7 cases per 1000 patient-years of follow-up were found in exposed women in contrast to 7.9 cases per 1000 in the controls (P < 0.01). The statistical association was considered significant only if the participants with biopsy-documented dysplasia or cytolog-ically detected dysplasia were tested alone. If cases of mild dysplasia were eliminated to adjust for the possibility that not all mild dysplasia may represent neoplasia (some cases are thought to reflect a papilloma-virus infection), the incidence rates were still higher in exposed females. In this case, however, the increase was statistically significant only in those participants with dysplasia detected by biopsy. These increases were also noted in the unmatched participant categories. The incidence rate of dysplasia and carcinoma in situ rose as the area of squamous metaplasia increased and extended to the outer half of the cervix or onto the vagina. The rates were similar in both the exposed and unexposed groups when the squamous metaplasia (defined as benign squamous epithelium with glycogen-poor cytoplasm) was confined to the os or the inner half of the cervix. The incidence rates of dysplasia in the three unmatched diethylstilbestrol-exposed cohorts (record review, self-referral (walk-in), and physician referral) were similar. All three cohorts had some cases of the more severe degrees of intraepithelial neoplasia (severe dysplasia and carcinoma in situ). Because of the marked differences between the incidence rates of dysplasia in the matched cohorts, additional features were examined in the larger number of participants in all three exposed cohorts to help identify factors that might explain the apparent association between diethylstilbestrol exposure and dysplasia. © Williams and Wilkins 1985.

Duke Scholars

Author Stanley J. Robboy Pathology