Evaluating variability in 24-hour urine free cortisol measurements: impact of sample collection errors.

OBJECTIVES

24 hour urine free cortisol measurement is widely used for screening Cushing's syndrome, but its reliability is limited by variability due to inherent biological variation and variation introduced by collection errors. This study aimed to determine the proportion of inaccurately collected samples and evaluate the impact on urine free cortisol variability.

METHODS

A retrospective analysis was conducted using paired 24-h urine samples from subjects screened for hypercortisolism at two academic hospitals in the Netherlands. Paired 24-h urine samples from 552 subjects were evaluated. Collection accuracy was determined by comparing creatinine and urine volume differences between consecutive samples. Samples were classified as inaccurate if the creatinine difference exceeded 4 mmol/24 h or the volume difference exceeded 55 %.

RESULTS

19 % of paired urine collections were classified as inaccurate. Accurate collections showed lower variability in urine free cortisol concentrations (coefficient of variation 17 %) compared to inaccurate collections (coefficient of variation 33 %).

CONCLUSIONS

This study shows that many routine 24-h urine samples are collected incorrectly. Collection errors contribute considerably to the total variability of 24-h urine free cortisol measurements. This highlights the importance of patient instruction and adherence to collection protocols for reproducibility of urine free cortisol measurements.