Studie zur Vorbeugung von
Darmpolypen mit Grüntee-Extrakt

Sample size calculation

The sample size estimation was derived from the existing chemopreventive studies reporting a re-incidence of metachronous adenomas of about 45% at follow-up colonoscopy [13]. Assuming an effect similar to that observed for ASS, chemoprevention might lead to a relative risk reduction of 12% [13]. Therefore, if ECGC intake leads to a similar relative risk reduction of 11% and if there is a recurrence of adenomas in 45% of the patients in the placebo group, 40% of the patients taking green tea extract should have a recurrent adenoma after three years. Sample size estimation was performed using the Mantel-Haenszel test stratifying for the use of low-dose ASS, assuming that 20% of the study participants will be on low-dose (≤ 100 mg) ASS due to the frequent use of low-dose ASS in the age group of the participants (50 years and above). Assuming a recurrence rate of 38.3% in the combined verum green tea and low-dose ASS arm, and of 47.1% in the placebo/no ASS group, a sample size of n = 1203 (both in verum and placebo) is required for the intention to treat (ITT) sample with significance level of 5% and a power of 80% (stratified one-sided Mantel-Haenszel test, NQUERY 7.0 software, Statistical Solutions, Cork, Ireland).

Assuming a drop-out rate of 5% of patients for the control colonoscopy, 1267 randomized patients per arm will be required, resulting in randomization of 2534 patients after the run-in period.

A drop-out rate of 20% is assumed for regular capsule intake. Thus, 1014 patients will be evaluable for the per-protocol analysis of the green tea effect. For exploratory analysis of potential confounders, this sample size is sufficient to include 30-40 variables inclusive dummy-coding with around 20 events per variable into the model.

Nach oben

zum Seitenanfang