Document 313672

ANALYSIS OF SERUM TESTOSTERONE AND
ANDROSTENEDIONE FOR CLINICAL RESEARCH USING
EITHER MANUAL OR AUTOMATED EXTRACTION
Dominic Foley1, Brian Keevil2, Lisa Calton1
1
Waters Corporation, Stamford Avenue, Wilmslow, UK.
2
University Hospital of South Manchester, Wythenshawe, UK.
INTRODUCTION
RESULTS
Linearity
Chromatography

Following CLSI-EP6-A, the method was shown to have a
linear fit over the range of 0.05—15ng/mL (n=5).

Calibration lines using serum spiked with testosterone and
androstenedione
were
linear
with
coefficient
of
determinations > 0.994 (n=10).

The signal:noise ratios for the lowest calibrator (0.05ng/mL)
from the testosterone and androstenedione spiked serum
were >15:1.
Androstenedione
100
287.2 > 96.9
%
Here we evaluate a UPLC/MS/MS method used to
measure serum testosterone and androstenedione
enabling investigation of metabolic dysfunction for
clinical research purposes. An analytically selective
method was developed using a mixed-mode Solid
Phase Extraction (SPE) sorbent in 96-well plate
format. Either manual or automated extraction was
employed,
providing
flexibility
in
sample
preparation options depending on the laboratory
environment.
0
1.20
1.40
1.60
1.80
2.00
2.20
2.40
2.60
2.80
Testosterone
100
289.2 > 96.9
Analytical Bias

%
METHODS
Epitestosterone
Materials


Total precision was determined by extracting and
quantifying three replicates of tri-level QC material on two
occasions per day over five consecutive days (n=30).
Repeatability was determined by analyzing three replicates
at each QC level.
0
0

Sample supernatant was transferred to a Waters® Oasis®
MAX µElution plate, washed with 0.1% ammonia in 20%
methanol and eluted with methanol.


standard,
Automation was performed using the Offline Automated
Sample Prep Station (OASPS)*.
2.00
2.20
Protein Precipitation
1.20
1.40
1.60
100
%
All samples were pre-treated with internal
ammonia, zinc sulphate, methanol and water.
1.80
2.40
2.60
Testosterone
100
Time
2.80
0
Interf erence
1.80
2.00
2.20
2.40
2.60
Oasis HLB
1.20
1.40
289.2 > 109.0
Interf erence
1.60
1.80
2.00
2.20
2.40
2.60
Testosterone
Oasis MAX
1.20
1.40
2.80
Figure 3. A simple Linear Regression of the reported mean
values of the CDC HoSt samples against the reference values
2.80
289.2 > 109.0
No
Interf erence
1.60
1.80
2.00
2.20
2.40
2.60
2.80
Ion Suppression
The normalized matrix factors (using analyte:internal
standard response ratio) were 0.99 and 1.03 for
androstenedione and testosterone, respectively.
The analysis time per sample was approximately 4.0
minutes injection to injection.

Total precision and repeatability was evaluated at low
(0.15ng/mL), mid (1.0ng/mL) and high (10ng/mL) QC
concentrations.

Both manual and automated total precision
repeatability were ≤ 5.7% for both analytes.
Precision
*OASPS is under development
Testosterone
Cone Collision
(V)
(eV)
Total QC Precision
QC Repeatability
38
25 (28)
Testosterone – C3
292.2 > 99.9
38
25
Compound
Low
Mid
High
Low
Mid
High
Androstenedione
287.2 > 96.9 (109)
38
25 (28)
Testosterone
4.1%
3.0%
4.0%
3.2%
2.8%
4.0%
Androstenedione –13C3
290.2 > 99.9
38
25
Androstenedione 4.0%
2.8%
4.6%
3.5%
2.8%
4.6%
Table 1. MRM parameters used for the analysis of testosterone
and androstenedione and their Internal standards. Qualifier
ion parameters are shown in parentheses.

Comparison with samples previously analyzed by an
independent LC/MS/MS method (n= 35) was described by
the Deming equation y = 1.07x + 0.01 for testosterone and
y= 0.96x + 0.02 for androstenedione.
Manual vs Automated Bias

A comparison of samples extracted using manual and
automated extraction protocols (n=35) was described by
the Deming equation y = 1.03x + 0.00 for testosterone and
y= 0.98x + 0.04 for androstenedione.

No constant or proportional bias was observed, indicating
equivalency of the manual and automated methods.
CONCLUSION
and
289.2 > 96.9 (109)
13
The mean method bias was determined to be 3.3%.
Figure 2. Comparison of protein precipitation, Oasis HLB and
Oasis MAX SPE for the extraction of testosterone

MRM Transition
(m/z)

Time
Using a Waters ACQUITY UPLC® I-Class System, samples
were injected onto a 2.1x50mm Waters ACQUITY UPLC HSS
C18 SB column using a water/methanol/ammonium acetate
gradient and analyzed with a Waters Xevo® TQD Detector,
using MRM parameters in Table 1.
Compound
r2 = 0.9992
289.2 > 109.0
Qualifier
Testosterone
100


1.60
The additional selectivity of the Oasis MAX SPE removes a
an interference from the 289.2 > 109 trace for testosterone.

0
Methods
1.40
Extraction Selectivity
Analytical method bias for testosterone was determined
using Hormone Standardization (HoSt) samples obtained
from the CDC (Atlanta, GA) (n=40).
A
comparison
was
performed
using
anonymized
testosterone and androstenedione samples previously
analyzed using an independent LC/MS/MS method (n=35).
1.20
Figure 1. Separation of androstenedione, testosterone and its
epimer, epitestosterone
%

Certified testosterone and androstenedione reference
material purchased from Cerilliant (Round Rock, TX) was
used to prepare calibrators (0.05—15 ng/mL) and QC
material in MSG4000 pooled human serum obtained from
Golden West Biologicals (Temecula, CA).
%

Comparison with HoSt samples (n=40) was described by a
Co-efficient of Determination (r2) of 0.9992.
Table 2. Total precision and repeatability for the automated
extraction of testosterone and androstenedione

We have successfully quantified testosterone and
androstenedione using an anion exchange SPE
and UPLC/MS/MS for clinical research purposes.

The method demonstrates excellent linearity,
precision and accuracy with minimal matrix
effects.

The method has been fully automated using the
OASPS.
FOR CLINICAL RESEARCH USE ONLY. NOT FOR USE IN DIAGNOSTIC PROCEDURES.
©2014 Waters Corporation