Tamara van Donge

Reference intervals for kidney biomarkers in pediatrics 223 10 Supporting information - Determination of sample size Sample size estimation was performed applying a re-sampling method using uromodulin values of 267 healthy adults. Assumptions: • Variance of uromodulin values acquired from adults is the same for children in the mentioned age groups. • Mean uromodulin values increase by 18% (Θ) from age group 1 to age group 2, age group 2 to 3 and 3 to 4, respectively. • There are no differences related to sex. • After log transformation, values are normally distributed. n i (n i=1,...,21 = 40, ..., 200) uromodulin values were sampled with replacement and allocated to one of the age groups. Allocation was done with a ratio of 1:1:2:4 for age and groups 1, 2, 3, and 4. Subsequently, it was determined whether mean uromodulin values are different among age groups. An ANOVA was performed and tested whether was α < 0.05. The process was repeated R = 999 times for each possible Θ (15% ≤ Θ ≤ 25%) . Assuming an increase of the mean uromodulin plasma concentration from group to group of 18 %, a total of N = 117 children are required to allow identifying the effect in at least 80 % of the cases (i.e., power) and a type-I error rate of α = 5 %. See figure 1 below. Sample sizes for each age group are: • Age group 1: N = minimal 15 (maximal 30) • Age group 2: N = minimal 15 (maximal 30) • Age group 3: N = minimal 29 (maximal 40) • Age group 4: N = minimal 58 (maximal 60) The above-mentioned minimal sample size of 117 patients is the minimum for statistic results. As we have seen that recruitment can be difficult, we decided to calculate a minimum size and a maximum size. Therefore, recruitment will stop if the minimum sample size in all groups will be reached. Therefore, the minimum sample size is 117, the maximum sample size 160.