Reproducibility of serum sex steroid assays in men by RIA and mass spectrometry

Abstract

There is an increasing trend to apply gas chromatography combined with mass spectrometry (GC-MS) or liquid chromatography tandem mass spectrometry (LC-MS/MS) assay methods to large-scale epidemiologic studies for the measurement of serum sex steroids. These methods are generally considered the gold standard for sex steroid measurements because of their accuracy, sensitivity, turnaround time, and ability to assess a more complete panel of steroid metabolites in the same run. In this report, we evaluated the precision, including within-batch (intra) and between-batch (inter) reproducibility, of steroid hormone measurements determined by GC-MS and LC-MS/MS assays and RIA and compared measurements among these methods. Specifically, 282 overnight fasting serum samples from 20 male volunteers were analyzed for 12 steroid metabolites by GC-MS or LC-MS/MS in one lab over a 4-month period. Six of the analytes were also measured by RIA in another lab. Unconjugated hormones, including testosterone, dihydrotestosterone, dehydroepiandrosterone, androstenedione, androst-5-ene-3beta,17beta-diol, estrone, and estradiol, were measured by GC-MS, whereas conjugated hormones, including DHEA sulfate, androsterone glucuronide, 5alpha-androstane-3alpha,17beta-diol 3-glucuronide, 5alpha-androstane-3alpha,17beta-diol 17-glucuronide, and estrone sulfate, were measured by LC-MS/MS. A subset of these hormones, including testosterone, dihydrotestosterone, androstenedione, 5alpha-androstane-3alpha,17beta-diol 17-glucuronide, estrone, and estradiol, were also measured by RIA following extraction and chromatography. We used the coefficient of variation (CV) and the intraclass correlation coefficient (ICC) to assess within- and between-batch assay variations. For the 12 analytes measured by GC-MS or LC-MS/MS, CVs and ICCs for within- and between-batch measurements were similar, with CVs ranging from 6.1% to 21.4% and ICCs ranging from 87.6% to 99.2%. The six analytes measured by RIA had good CVs and ICCs, with CVs <10% and ICCs >70% (range, 71.7-99.7%). For the six metabolites that were measured by both methods, the CVs were similar, whereas the ICCs were generally higher with the GC-MS method. The absolute values for each analyte measured by RIA and GC-MS differed, with RIAs usually yielding markedly higher levels than GC-MS, although the Pearson and Spearman correlation coefficients for these six analytes were near one and all were significant (P < 0.001). Our results show that RIA, GC-MS, and LC-MS/MS assays for androgens and estrogens in the two labs included in the study have good reproducibility, as measured by small CVs (<15%) and high ICCs (>80%), with the exception of estradiol (71.7%) when measured by RIA. Despite substantial differences in absolute measurements of sex steroid hormones by RIA and MS methods, correlations between the two assays for the six sex steroids measured in the two labs were high (>0.9). However, it is important for future large epidemiologic studies to incorporate MS with high reproducibility and specificity to measure a more complete profile of androgen and estrogen metabolites to clarify the role of sex steroids in prostate cancer.