Use of the life-table technique to estimate effects of certain steroids on probability of tumor formation in a long-term study in rats

Abstract

An adaptation of the life-table technique directed toward a study of effects of two estrogens and a progestagen on tumor formation over two years in female rats of the Mead Johnson colony derived from the McCollum strain is described. The method estimates the probability of tumor incidence and then the “effective number” ([ N]) of animals exposed to risk throughout the study period. An alternate calculation, to estimate [ N] more directly and then the probability from the number of tumor-bearing rats/[ N], is described for use when the tumor incidence is small. Two groups of control rats, one of 60 and the other of 30, were used. Four groups of 30 rats each received steroid treatment ad libitum in their diets: ethynylestradiol (EE) in an average daily dose of 53 μg/kg, that dose of EE + megestrol acetate (MA) in an average daily dose of 2.63 mg/kg, 17α-chlorethynyl-3,17β-dimethoxyestra-1,3,5(10)-triene (BDH 2700) in the same dose as EE above, and EE + MA sequentially (30 μg EE/kg daily × 16 days, 30 μg EE + 1.5 mg MA/kg × 5 days, and 7 days' rest). Probabilities of five parameters of tumor incidence were estimated: any tumor, any mammary tumor, mammary adenocarcinoma, only benign tumor(s), and any malignant tumor(s) (occurring even in a rat also bearing a benign tumor). EE alone did not alter any of those five categories. BDH 2700 significantly increased the incidence of mammary adenocarcinoma ( p < 5%). The group that received both EE and MA daily exhibited a reduction in incidence of any tumor that was not quite significant at the 5% level. In the group that received EE and MA EE sequentially, the lower incidences of any tumor, a mammary tumor, and any malignant tumor were significant.