[Statistical approach to cervical and vaginal cytology during contraception]

We analyzed 1432 smears in order to check if the usual methods of contraception influence the development of cervical lesions. Several observations have to be made. 1) Given the rarity of certain lesions, a much larger number of samples (x 50) should have been analyzed in order to obtain realistic statistical projections especially for dysplasias and cervical cancers. 2) Significant results could only be obtained when several morphological screening criteria were put together (ectropions, repairs or inflammation) provided that at least 10 cases/sample were observed in order to perform statistical analysis. Therefore, the rate of ectropions observed in a population without contreceptive treatment is 6% higher than in a population taking the pill (correlations coefficient 0.95) and more than 5% repairs are observed in a population using the IUD as compared to oral contraceptives. The rate of clean smears (class 1) and inflammatory smears (class 2 according to Papanicolaou) was found to be the same for all 3 populations studied. Comparing statistical data, it seems to be difficult to provide a correct answer about the influence of contraceptive methods on the development of cervical cancer. Indeed, our work and all other studies present the same difficulty in gathering sufficient samples which are qualitatively equivalent. Volume and present methodology of most of the anatomy and pathological cytology laboratories do not meet the requirements of statistical calculus. However, we think that smears should be routinely done on women taking the pill or wearing an IUD. That is not to say that contraception, regardless of method used, increases the risk of cancer or dysplasia, but that the smears represent a chance for screening and are thus considered preventive medicine. This would also enhance the physician's knowledge about precancerous lesions of the cervix and would, in the near future, boost the number of samples to allow for statistical analysis. (author's)