Jennifer Fahse (Presenter)
Mayo Clinic
Authorship: Jennifer H. Fahse, Eric T. Bro, Brandon J.N. Hill, Robert L. Taylor, Sandra C. Bryant, Kendall W. Cradic, Ravinder J. Singh, Stefan K. Grebe
Clinical Mass Spectrometry Development Lab, Mayo Clinic, 200 1st St. SW., Rochester, MN 55905
| Short Abstract Isotope dilution equilibrium dialysis (IDED) has long been considered the gold-standard method for free testosterone measurement. However, IDED requires a high sample volume and is labor intensive, making it a challenge for clinical laboratories processing hundreds of samples daily. We combined two necessary components of comprehensive testosterone testing, the measurement of the total testosterone concentration by LC-MS/MS and free testosterone by isotope dilution equilibrium dialysis, into a single automated method with the use of a microdialysis plate and robotic liquid handler. |
Long Abstract
Introduction
Isotope dilution equilibrium dialysis (IDED) has long been considered the gold-standard method for free testosterone measurement. However, this method not only requires a large amount of sample volume (1.2 mL), but is also labor intensive, making it a challenge for clinical laboratories processing up to hundreds of samples daily. We combined two necessary components of comprehensive testosterone testing, the measurement of the total testosterone concentration by LC-MS/MS and free testosterone by IDED, into a single automated method with the use of a microdialysis plate and robotic liquid handler. In doing so, our laboratory has reduced our minimum sample volume requirements three-fold and has removed the laborious manual process of using equilibrium dialysis tubing by introduction of automation. This method also allowed our group to gather a large amount of normative data in a relatively short amount of time. We were able to distinguish significant age and sex relationships between free testosterone from early childhood to advanced age.
Methods
A Hamilton Star automated liquid handling system was used to transfer 350µL of patient serum to each, (i) a 96-well plate (for total testosterone measurement) and (ii) 96 tubes containing dried down purified tritiated testosterone (for free testosterone measurement).
Total Testosterone
Deuterated testosterone internal standard and acetonitrile were added to the total testosterone plate prior to centrifugation of the plate at 3,000 rpm for 10 min. After centrifugation, the supernatants were analyzed on an AB Sciex API 5000 tandem mass spectrometer operating in positive APCI mode.
Free Testosterone
Once spiked with tritiated testosterone, the free testosterone samples were automatically transferred to a Pierce® 96-well microdialysis plate (10 KDa molecular weight cut-off), using a Hamilton Star liquid handler and incubated for 23 hours at 37°C to reach equilibrium. Serum, dialysate and scintillation cocktail were then transferred to scintillation vials and analyzed using a Taurus Beta Scintillation Counter. The ratio of unbound to bound testosterone in each sample was used to calculate the percent free testosterone.
Results
Total Testosterone
Inter-assay precision of quality control samples at average concentrations of 29, 361 and 1029 ng/dL were 6.6, 5.7 and 5.6% CV. Inter-assay precision of patient pools at average concentrations of 203, 311 and 679 ng/dL were 5.4, 5.3 and 4.9% CV with the intra-assay precision of the same pools at 1.1, 2.2 and 1.3% CV respectively. A linear regression analysis conducted on 1360 samples comparing microdialysis plate vs. tubing total testosterone results produced a trend line equation of y = 1.0585x + 4.739 with an R2 value of 0.9902.
Free Testosterone
Inter-assay precision of quality control samples at average counts per minute (CPM) of 1061, 1854 and 2608 were 12.5, 7.2 and 6.0% CV. Inter-assay precision of patient pools at average CPM of 493, 1261 and 2165 were 10.6, 6.8 and 6.2% CV, while the intra-assay precision of the same pools were 5.8, 5.7 and 4.1%. Eighteen additional patient pools containing free testosterone percentages of 0.2% - 3.6% in 0.2% increments were used to obtain a detailed inter-assay precision profile across our analytical measuring range. The inter-assay precision for pools between 0.2 – 0.6% were <15% CV and <10% for pools between 0.8 – 3.6%. A linear regression analysis conducted on 1360 samples comparing microdialysis plate vs. tubing free testosterone results produced a trend line equation of y = 0.734x – 0.0366 with an R2 value of 0.8465.
Normative Data
Free testosterone values were obtained from 3,142 patients, 1,573 (50.06%) male, 1,569 (49.94%) female, aged 1 to 100 years, using this method. Relationships of the mid-95th percentile (2.5th & 97.5th percentile) with age and sex were evaluated using quantile regression (Koenker & Hallock, 2001). To assess the age relationship, we evaluated the models with terms for age; age and age2; and age, age2 and age3. We assessed sex as a categorical variable. We estimated standard errors of the quantile regression parameters using a bootstrap resampling procedure with replacement with 10,000 replicates, from which a p-value was determined. Significant relationships with age and sex are defined as a p-value <0.05. Once the appropriate relationship with age and/or sex was determined, we estimated the appropriate two-sided reference ranges with their confidence intervals using the same quantile regression methods.
The distributions of free testosterone were first compared by gender using non-parametric Wilcoxon rank sum and Kolmogorov-Smirnov tests. Differences between males and females were detected (adult p-value < 0.0001; child p-value < 0.0001). The data was then assessed to determine if any age or sex relationships existed using the quantile regression method described above; for adult samples a linear relationship between age and free testosterone at both the 2.5th and 97.5th percentiles and a linear relationship was found at the 97.5th percentile for women; for children (samples from individuals less than 18 years old at time of collection) a cubic relationship was found for boys at both the 2.5th and 97.5th percentiles and for girls there was a cubic relationship at the 97.5th percentile and a linear relationship at the 2.5th percentile.
Conclusions
The development of our automated method for the evaluation of both total testosterone and free testosterone was a significant process improvement for our high-volume clinical laboratory workload while still preserving the highest-quality patient testing practices. As a part of the validation this also allowed us to improve our previous reference ranges with up-to-date comprehensive and highly granular new ranges.
References & Acknowledgements:
| Description | Y/N | Source |
| Grants | no | |
| Salary | no | |
| Board Member | no | |
| Stock | no | |
| Expenses | no |
IP Royalty: no
| Planning to mention or discuss specific products or technology of the company(ies) listed above: | no |