An analysis for temporal variation in Down syndrome births in Ohio, 1970-1979.

Our study investigates the epidemiology of Down syndrome (DS) in the state of Ohio during the 1970s. The occurrence of DS births was examined to learn if statistically significant temporal variation was present among these data. Both monthly and annual numbers of DS births, adjusted for changing numbers of live births, were tested for such variation; furthermore, the data were analyzed for cyclic variation by attempting to fit simple trigonometric functions to the data. Individuals with DS were ascertained using the records of cytogenetics laboratories and birth certificate records. Demographic data such as race, date of birth, and maternal age were collected on these individuals using their birth certificates as the data source. Appropriate parallel live-birth data were obtained from the Ohio Department of Health. The total number of affected individuals ascertained was 1,364, 66.7% of the total estimated population size. The data analysis was restricted to whites only (1,203 individuals) because they represented a more homogeneous sample than the total.Monthly and annual variation in the numbers of live births was removed by producing single-year maternal-age adjusted numbers of DS births using the total Ohio white live births as the reference population. Analysis of covariance using single-year maternal ages </= 16 and >/= 45 as the covariate was used to analyze the adjusted numbers of DS births for temporal variation. No significant differences were detected among the annual adjusted numbers of DS births (P = .24), nor were there differences among the monthly adjusted numbers of DS (P = .37). The modes of ascertainment were tested to learn if there were annual or monthly differences in the method of ascertainment. No significant differences were detected for these data (P = .82 and P = .85, respectively). Furthermore, the data were separated into the maternal-age categories < 35 and >/= 35, and annual and monthly adjusted DS births to these two maternal-age categories were examined for temporal variation. No significant differences were found among these data, P > .10 for all four of the tests. No simple cyclic functions were found to fit either the annual or monthly data. The Ohio study reported here showed that through the use of a large sample, controlling for variation in the numbers of live births, and the use of detailed statistical tests, no significant temporal variation in the occurrence of DS births existed during the 1970s.