Application of the tukey trend test procedure to assess developmental and reproductive toxicity. I. Measurement data

Developmental and reproductive (DAR) toxicity studies typically include a series of increasing doses of a compound and a zero dose control. Given this framework, Tukey et al. (Biometrics, 41, 295-301, 1985) proposed a procedure (referred to as either the Tukey trend or TCH test procedure) for detecting a nonzero trend in response to increasing doses of the test compound. The procedure considers three candidate dosage scalings to ensure high power against relatively common dose-response patterns and appreciable power against most reasonable patterns. For toxicologic effects with near monotonic dose-response patterns, simulation studies have shown the TCH test to be overall more powerful than pairwise comparison procedures. The TCH test can be applied sequentially, eliminating the highest dose each time a statistically significant trend is observed, until a no-statistical-significance-of-trend dose is reached. This is the highest dose through which there is no statistically trustworthy evidence of the compound's impact on the response. Since DAR toxicity usually exhibits a progressive (monotonic) dose-response, we advocate routine use of Tukey's trend test for the evaluation of treatment effects in these studies. In this article, we discuss the procedure in detail and apply it to fetal body weight, a continuous measurement variable, from a developmental toxicity study.