Interlaboratory variability of urinary iodine measurements.

BACKGROUND: The iodine status of populations is usually assessed by median urinary iodine concentrations (UIC) in population-based studies, but it is unclear to which extent UIC are comparable across different laboratories. The aim of our study was to investigate the variability of UIC measurements across three well-established German laboratories with long-term clinical-chemical expertise in iodine measurements and to compare these results to the gold standard inductively coupled plasma mass spectrometry (ICP-MS).

METHODS: UIC levels were measured from 303 urine samples derived from the "Dortmund Nutritional and Anthropometric Longitudinally Designed Study" and from volunteers of the University Medicine Greifswald at four different German laboratories. Three of these laboratories used Sandell-Kolthoff reaction with different digestion methods for UIC measurement (Lab1-Lab3), whereas one laboratory used ICP-MS as gold standard.

RESULTS: Median UIC levels were significantly different across the four laboratories (ICP-MS: 77 μg/L; Lab1: 69 μg/L; Lab2: 73 μg/L; Lab3: 111 μg/L). Linear regressions associating UIC levels of Lab1-Lab3 with UIC levels of ICP-MS showed intercepts significantly different from 0 and slopes significantly different from 1. Intraclass correlations (ICC) in comparison to ICP-MS were 0.91 for Lab1, 0.98 for Lab2, and 0.69 for Lab3. Using the digestion method of Lab2 in Lab3 improved the comparison of UIC levels of Lab3 with those from the ICP-MS (ICC=0.89).

CONCLUSIONS: We have demonstrated larger interlaboratory variations across high-quality laboratories with long-lasting experience in iodine measurements indicating a relevant non-comparability of UIC measurements in iodine monitoring studies. Therefore, standardization of UIC measurements has to be expedited.