Endocrine abnormalities are either self-defined disorders or occur as part of other diseases. Many diseases that were historically considered as non-endocrine are increasingly being found to be associated with endocrine abnormalities. Optimal use of valid laboratory tests and correct diagnosis of endocrine diseases or the endocrine components of non-endocrine diseases obviously are important both medically and economically. Both forms of vitamin D (ergocalciferol or D2 and cholecalciferol or D3) are now considered prohormones. During the past year, we recognized and studied important medication-related analytical interferences with high-performance liquid chromatography (HPLC) testing for the biologically active 25-hydroxyvitamin D (25OH-vitD) and completed evaluation of one such interference. Although HPLC with UV detection generally is considered the gold standard, this methodology also is subject to interference. For 25OH-vitD quantification we are using a recently described, selective validated HPLC method (Clin Chem 2006;52:1120-6) with a Waters HPLC system. This method involves the use of laurophenone as internal standard and is able to separate and detect both forms (D2 and D3) of 25OH-vitD. Analyzing more than 2,500 serum specimens over an 18-month period, we observed unusually high (arbitrarily defined as greater than 5 standard deviations above run mean) and usually asymmetric internal standard peaks in 9 specimens of 8 patients. Except for one specimen, repeat testing of the specimens reproduced the abnormally high and distorted internal standard peaks. In one case, a split internal standard peak was obtained on repeat: the first peak corresponded to the expected retention time of laurophenone, while the second later peak apparently represented an interfering substance. Medication history of the patients revealed only one common drug: the anti-fungal itraconazole. Addition of itraconazole (1 mg/ml) to a serum specimen, previously shown to be free of interference for the internal standard laurophenone, indeed mimicked the observed interference and a heightened HPLC peak appeared at the usual chromatographic position of the internal standard. Since falsely high internal standard peaks produce falsely low 25OH-vitD results and repeat testing does not eliminate the problem, we attempted to correct the interference by using the mean internal standard peak height of the HPLC run or the two neighbouring HPLC chromatograms for calculations. The two approaches gave virtually identical results that were confirmed to be valid by testing the specimens by liquid chromatography-mass spectrometry/mass spectrometry (LC-MS/MS) at Mayo Medical Laboratories in Rochester, MN. Thus even in the absence of relevant medication history, the internal standard interference described here can be recognized by an unusually high and/or distorted internal standard peak in the HPLC chromatogram and corrected for by either of our two proposed approaches.
Showing the most recent 10 out of 11 publications
