Background: LC-MS/MS is replacing immunoassays for serum steroid measurement in clinical labs worldwide. The intrinsic high specificity of LC-MS/MS is supposed to guarantee accurate quantitation, however, differences in pre-analytical and analytical procedures, and overall performance may influence the result. A few studies so far investigated the reproducibility among different LC-MS/MS assays for sex steroids. No data are available for adrenal steroids.
Aim: To compare the LC-MS/MS measurement of 13 serum steroids among 10 European Centers.
Methods: Blood from 13 women and 13 men aged 2170 y was collected in fasting state, at 8:009:00 am, after 10 min saline infusion in three different vacuum tubes to obtain serum (gel separator, GS and beads-clot-activator) and plasma (Li-Heparin) (78 total samples). Aliquots were sent to each lab for LC-MS/MS measurement according to the labs own procedure. Intra-lab duplicate measurement CV%, inter-lab CV% and Bland-Altman analyses were performed.
Results: Preliminary results from six labs were obtained on cortisol (F), cortisone (E), corticosterone (B), 11 deoxycortisol (11S), 17OH-progesterone (17OHP4), androstenedione (A4), testosterone (T) and DHEAS. Intra-lab CV% ranged 3.19.9 (F), 2.713.8 (E), 3.38.0 (B), 3.814.4 (11S), 5.818.1 (17OHP4<1 ng/ml) 2.09.9 (17OHP4>1 ng/ml), 4.28.8 (A4), 4.37.6 (T) and 4.111.4 (DHEAS). Median inter-lab CV% were 5.0, 5.1, 6.8. 7.3, 8.9, 14.0 and 15.5 for DHEAS, E, T, A4, B, 11S and F, respectively, and 13.2 and 5.9 for 17OHP4 <1 ng/ml and >1 ng/ml, respectively. Cases showing inter-lab CV>15% were 0 for T; 1 (1.3%) for E and DHEAS; 3 (3.8%) for A4; 13 (16.7%) for B; 25 (32.1%) for 17OHP4; 36 (46.2%) for 11S and 42 (53.8%) for F. Among GS-sera, cases increased to 18 (69.2%) for both 11S and F. Results from each lab were compared to all-lab median values. Mean bias ranged −3.5 5.8%, -1.9 - 6.5%, −3.2 9.1%, -10.6 11.2%, −16.5 14.6% and −5.0 36.9% for DHEAS, E, A4, B, 11S, and F, respectively. Moreover, mean bias ranged −23.4 −7.5% and −6.7 − 4.7% for 17OHP4<1 ng/ml and >1 ng/ml, respectively, and −2.9 6.4% and −8.5 8.7% for T<1 ng/ml and >1 ng/ml, respectively.
Conclusions: Variable intra-lab performances were observed. Inter-lab reproducibility was good for E, B, DHEAS, A4 and T, but poor for F and 11S. The specimen type may influence the results. 17OHP4 reproducibility was higher at high than low levels, while the opposite was unexpectedly found for T. Further analyses will help to define and correct major sources of variability.
18 - 21 May 2019
European Society of Endocrinology