Statistical Analysis
The combined effect of melatonin and cortisol levels on prostate cancer
could be measured at the same time to calculate the urinary
melatonin-sulfate/cortisol (M/C) ratio by dividing urinary
melatonin-sulfate levels by urinary cortisol levels. Differences between
the patient and control groups in urinary biomarkers such as cortisol,
melatonin-sulfate, and the M/C ratio were studied using the Wilcoxon
Rank-Sum Test. Furthermore, we dichotomized these urinary markers with
medians of 23.92 ng/mg creatinine for cortisol, 1.81 for the M/C ratio,
and 42.83 ng/mg creatinine for melatonin. 95% confidence intervals
(CIs) and odds ratios (OR) were estimated using the unconditional
logistic regression models for the correlation between urinary markers
and risk of prostate cancer. Family history of prostate cancer, habits
(smoking, alcohol), age at the time of the study (<65 vs. ≥ 65
years), PSA level (<10 vs.> 10 ng/mL) were
considered for all analyses.
The cases were also categorized as advanced (end-stage) cancer
(extraprostatic stage T3a or higher, N1/M1) and localized prostate
cancer (stage T2 or T1) 20. The risks of developing
advanced (end-stage) and localized prostate cancer based on the control
group were determined using the multi-category logistic regression
models. Polytomous logistic models containing categorical variables and
their cross-products were fitted to appraise multiplicative interaction.
The significance of multiplicative interaction was evaluated for each
pair of comparison using the Wald Z-tests for cross-product terms. All
two-sided p-values were considered significant if < 0.05.