Is there evidence of clustering in Down syndrome?

Abstract

Associations between environmental hazards and the occurrence of congenital anomalies may be detectable by seeking evidence of non-random occurrence of cases (clusters). There have been a number of anecdotal reports of occurrences of clusters of Down syndrome (DS). Data from a national register of cytogenetic diagnoses of Down syndrome births and legal terminations occurring between 1989 and 1995 were used to examine the possibility of clustering. Space-time clustering at Regional Health Authority (RHA) level was examined by comparing the expected monthly number of DS pregnancies given the maternal age distribution, with the observed numbers. Knox's method was used to determine if any clustering of RHA of unexpectedly high prevalence had occurred. Seasonality was also investigated by comparing monthly expected and observed numbers of DS pregnancies. Time clustering was examined by using the scan statistic to determine whether a statistically significant excess of pregnancies in any 3-month period occurred in any individual or adjacent groups of District Health Authority (DHA). The numbers of DS pregnancies were no higher than expected (P < 0.05) in the same RHA over consecutive months. There was no evidence of any seasonality of DS pregnancies (P > 0.5). Only two individual DHA and three pairs of adjacent DHA had significantly high scan statistics (P < 0.03), but as over 400 statistical tests had been completed 12 clusters would be expected to have occurred due to chance alone. There was no evidence of any space-time clustering in DS at DHA level.