Home Report from the HarmoSter study: different LC-MS/MS androstenedione, DHEAS and testosterone methods compare well; however, unifying calibration is a double-edged sword
Article Open Access

Report from the HarmoSter study: different LC-MS/MS androstenedione, DHEAS and testosterone methods compare well; however, unifying calibration is a double-edged sword

  • Flaminia Fanelli ORCID logo EMAIL logo , Mirko Peitzsch , Stephen Bruce , Marco Cantù , Anastasia Temchenko , Marco Mezzullo ORCID logo , Johanna M. Lindner , James M. Hawley , Mariette T. Ackermans , Jody Van den Ouweland , Daniel Koeppl , Elena Nardi , Finlay MacKenzie , Pierre-Alain Binz ORCID logo , Manfred Rauh , Brian G. Keevil , Michael Vogeser , Graeme Eisenhofer ORCID logo , Annemieke C. Heijboer and Uberto Pagotto
Published/Copyright: January 12, 2024
Download Article (PDF)
Become an author with De Gruyter Brill
Submit Manuscript Author Information
Clinical Chemistry and Laboratory Medicine (CCLM)
From the journal Clinical Chemistry and Laboratory Medicine (CCLM) Volume 62 Issue 6

Abstract

Objectives

Current liquid chromatography-tandem mass spectrometry (LC-MS/MS) applications for circulating androgen measurements are technically diverse. Previously, variable results have been reported for testosterone. Data are scarce for androstenedione and absent for dehydroepiandrosterone sulfate (DHEAS). We assessed the agreement of androstenedione, DHEAS and testosterone LC-MS/MS measurements among nine European centers and explored benefits of calibration system unification.

Methods

Androgens were measured twice by laboratory-specific procedures in 78 patient samples and in EQA materials. Results were obtained by in-house and external calibration. Intra- and inter-laboratory performances were valued.

Results

Intra-laboratory CVs ranged between 4.2–13.2 % for androstenedione, 1.6–10.8 % for DHEAS, and 4.3–8.7 % and 2.6–7.1 % for female and male testosterone, respectively. Bias and trueness in EQA materials were within ±20 %. Median inter-laboratory CV with in-house vs. external calibration were 12.0 vs. 9.6 % for androstenedione (p<0.001), 7.2 vs. 4.9 % for DHEAS (p<0.001), 6.4 vs. 7.6 % for female testosterone (p<0.001) and 6.8 and 7.4 % for male testosterone (p=0.111). Median bias vs. all laboratory median with in-house and external calibration were −13.3 to 20.5 % and −4.9 to 18.7 % for androstenedione, −10.9 to 4.8 % and −3.4 to 3.5 % for DHEAS, −2.7 to 6.5 % and −11.3 to 6.6 % for testosterone in females, and −7.0 to 8.5 % and −7.5 to 11.8 % for testosterone in males, respectively.

Conclusions

Methods showed high intra-laboratory precision but variable bias and trueness. Inter-laboratory agreement was remarkably good. Calibration system unification improved agreement in androstenedione and DHEAS, but not in testosterone measurements. Multiple components, such as commutability of calibrators and EQA materials and internal standard choices, likely contribute to inter-laboratory variability.

Keywords: liquid chromatography-tandem mass spectrometry; harmonization; androstenedione; dehydroepiandrosterone sulfate; testosterone; inter-laboratory performance

Introduction

Circulating testosterone, androstenedione and dehydroepiandrosterone sulfate (DHEAS) are routinely evaluated to characterize the extent and the nature of androgen imbalance in diseases that include gonadal or adrenal insufficiency, tumors, genetic enzyme deficits, Cushing syndrome and polycystic ovary syndrome in women [1, 2]. Clinically relevant levels of these hormones span both well above and below the normal range. Moreover, therapies that impact androgen availability require monitoring for changes over time, even within the normal range. Consequently, androgen measurement requires optimal precision and accuracy over a wide interval of concentrations. Another requirement is to effectively distinguish the analyte from other steroidal species sharing similar structure and properties [3].

By virtue of intrinsic high specificity and sensitivity, liquid chromatography-tandem mass spectrometry (LC-MS/MS) is today implemented in routine and research laboratories for steroid measurements via laboratory developed tests (LDT) or commercial kits [3, 4]. Participation of LC-MS/MS users in External Quality Assessment (EQA) programs is increasing [4, 5]. Nevertheless, reproducibility among LC-MS/MS measurements in complex patient samples remains unclear [4, 6]. So far, only two studies analyzed the comparability of LC-MS/MS methods for androstenedione [7, 8], while none have investigated DHEAS. A few studies on testosterone demonstrated that LC-MS/MS measurements exhibit superior consistency compared to immunoassays [7], [8], [9], [10], [11], [12]. Nonetheless, these studies reported diverse degrees of agreement between LC-MS/MS methods. Unfortunately, no single cause has yet been identified for persistent variability [11, 13].

The HarmoSter consortium was created with the primary objective to investigate the harmonization status of LC-MS/MS measurements of 10 circulating steroids by nine European centers (Supplementary Table 1). Methods were tested on different sample matrices and calibration systems. Results for cortisol, 17OH-progesterone, aldosterone, corticosterone, 11-deoxycortisol and cortisone were recently published [14, 15], showing acceptable consistency among methods. Furthermore, we highlighted that the strategy of unifying calibration may sometimes improve overall inter-laboratory agreement, but not always [14, 15].

In this new study, we focus on LC-MS/MS measurements of androstenedione, DHEAS and testosterone in plasma and serum samples from patients and in EQA materials to assess the intra- and inter-laboratory performance, as well as any impact of unification of calibration systems.

Materials and methods

Consortium and methods

The HarmoSter study was approved by Bologna Ethics Committee (no. 141/2017/U/Tess) [14]. Nine laboratories with 10 LDTs overall (Laboratories B to I) and the MassChrom® kit (Chromsystems; Munich, Germany; https://chromsystems.com) (Laboratory L) participated. All measured testosterone; eight measured androstenedione and seven measured DHEAS (Table 1 and Supplementary Table 1). 6PLUS1® Multilevel Serum Calibrator set (Chromsystems) was used for in-house calibration by Laboratories D, E and F. Technical details and in-house measurement ranges are shown in Table 1 and Supplementary Table 1.

Table 1:

Assays’ pre-analytical, liquid chromatography and mass spectrometry features.

Lab/method Ref In-house calibration Sample volume, µL Sample preparation Instrument Run time, min Purification column LC column; temperature Mobile phases (A; B) Source Androstenedione DHEAS Testosterone
Analyte (ion mode) mass transition; retention time, min IS IS (ion mode) mass transition; retention time, min Analyte (ion mode) mass transition; retention time, min IS IS (ion mode) mass transition; retention time, min Analyte (ion mode) mass transition; retention time, min IS IS (ion mode) mass transition; retention time, min
B [20] Cerilliant CRMa in 4 % BSA 600a PP: ZnSO4 in MeOH; SPE: C18 Series 200, Perkin Elmer; API4000 QTrap, Sciex 21 POROS R1/20 Luna RP-C8 100 × 4.6 mm, 5 µm; 40 °C 20 % MeOH in H2O; MeOH APCI (+) 287/97; 11.40 Androstenedione-[2,2,4,6,6-D5]a (+) 292/100; 11.06 (+) 289/97; 11.73 Testosterone-[3,4-13C2] (+) 291/111; 11.73
C [21] Gravimetric in charcoal stripped serum 50 PP: ZnTFA in MeOH Acquity UPLC; Xevo TQ-S, Waters 10 Kinetex Biphenyl 150 × 2.1 mm, 1.7 µm 0.2 mM NH4F in H2O; MeOH ESI (−) 367/97; 1.68 DHEAS-[2,2,3,4,4-D5] (−) 372/98; 1.66 (+) 289/97; 4.80 Testosterone-[2,3,4-13C3] (−) 292/97; 4.80
D [22] 6PLUS1®a,b, Chromsystems 500 SPE: Oasis HLB Acquity UPLC, Waters; API5500 Qtrap, Sciex 14 Kinetex C18 100 × 2.1 mm, 2.5 µm 5 mM NH4F in H2O; MeOH ESIa (+) 287/97; 6.90 Androstenedione-[2,2,4,6,6,16,16-D7] (+) 294/100; 6.85 (+) 271/255; 5.45 DHEAS-[16,16-D2] (+) 273/255; 5.43 (+) 289/97; 7.35 Testosterone-[2,2,4,6,6-D5] (+) 294/100; 7.32
E [23] 6PLUS1®a,b, Chromsystems 100 PP: ZnSO4 in MeOH 1,260 Infinity, Agilent; API6500 Qtrap, Sciexa 7.0 Chromolith SpeedRod C18 4.6 × 50 mm Chromolith performance C18, 100 × 2.1 mm, 30 °C 5 mM NH4HCO2 in H2O; 5 mM NH4HCO2 in MeOHa ESIa (+) 287/97; 4.62 17-OH-progesterone [2,3,4-13C3] (+) 334/112; 4.96 (+) 271/213; 4.07 DHEAS-[2,2,3,4,4-D5] (+) 276/218; 4.04 (+) 289/97; 4.87 17-OH-progesterone [2,3,4-13C3] (+) 334/112; 4.96
F [24] 6PLUS1®a,b, Chromsystems 50 SLE: Isolute 200 Acquity UPLC; TQS, Waters 4.0 Acquity T3 C18 50 × 2.1 mm, 50 °C 2 mM C2H7NO2 + 0.1 % FA in H2O; ACN ESI (+) 287/97; 2.30 Androstenedione-[2,3,4-13C3] (+) 290/100; 2.30 (+) 289/97; 2.00 Testosterone-[2,3,4-13C3] (+) 292/100; 2.00
G.1 [26] Cerilliant CRM in steroid free plasma 25 PP: ACN Acquity-Xevo TQ-S, Waters 9.9 HSS T3 2.1 × 100 mm, 1.8 μm 0.1 % FA in H2O; 0.1 % FA in ACN ESI (+) 287/97; 7.17 Androstenedione-[2,2,4,6,6,16,16-D7] (+) 294/100; 7.11 (+) 289/97; 6.56 Testosterone-[16,16,17-D3] (+) 292/97; 6.53
G.2 [25] Gravimetric in ethanol 25 PP: ACN Acquity-Xevo TQ-S Micro, Waters 4.7 Kinetex C18 75 × 2.1 mm, 2.6 µm 2.4 mM NH4HCO2 in H2O; ACN ESI (−) 367/97; 1.82 DHEAS-[2,3,4-13C3]a (−) 373/98; 1.82
H n.a. Cerilliant CRM in 10 % MeOH 1,000 PP: ZnSO4 in MeOH; SPE: C18 Acquity-Xevo TQ-S, Waters 12 BEH C18 2.1 × 100 mm, 1.7 μm 0.05 % FA in H2O; 0.05 % FA in MeOH ESI (+) 287/97; 6.53 Testosterone-[16,16,17-D3] (+) 292/97, 6.95 (+) 271/213; 7.32 DHEAS-[2,2,3,4,4-D5] (+) 276/218; 7.29 (+) 289/97; 6.97 Testosterone-[16,16,17-D3] (+) 292/97; 6.95
I.1 [27, 28] Gravimetric in charcoal stripped serum 100 PP: H3PO4; SPE: Oasis MCX I-class Acquity; TQS, Watersa 15a C18 Zorbax eclipse plus 2.1 × 50 mm, 1.8 μma H2O; MeOH ESI (+) 287/97; 5.70 Androstenedione-[2,2,4,6,6-D5] (+) 292/100; 5.60 (+) 289/97; 6.30 Testosterone-[16,16,17-D3] (+) 292/97; 6.20
I.2 [27, 28] Gravimetric in charcoal stripped serum 100 PP: H3PO4; SPE: Oasis MCX I-class Acquity; TQS, Watersa 5.0 C18 Zorbax eclipse plus 2.1 × 50 mm, 1.8 μma H2O; 0.0125 % NH4OH in ACN (−) 367/97; 2.00 DHEAS-[2,2,3,4,4-D5] (−) 372/98; 2.00
L n.a., c 6PLUS1®b, Chromsystems 500 Chromsystems’ intellectual property 1,290 series HPLC; 6,490, Agilent 11.5 Chromsystems’ proprietary column Chromsystems’ intellectual property ESI (+) 287/109; 7.49 Chromsystems’ intellectual property (+) 290/112; 7.48 (+) 369/99; 3.00 Chromsystems’ intellectual property (+) 375/100; 2.93 (+) 289/97; 5.95 Chromsystems’ intellectual property (+) 292/97; 5.89

  1. aModified from original publication. bLots different from the one distributed in the study. chttps://chromsystems.com/en/products/steroids/masschromr-steroids-in-serum-plasma-with-96-spe-well-plate-lc-ms-ms-72072-96%20&%2072072-480.html. IS, internal standard; CRM, certified reference material; PP, protein precipitation; ZnSO4, zinc sulfate; MeOH, methanol; SPE, solid phase extraction; RP, reversed phase; ZnTFA, zinc trifluoroacetate; HLB, hydrophilic-lipophilic balance; SLE, supported liquid extraction; FA, formic acid; ACN, acetonitrile; HSS, high strength silica; BEH, ethylene bridged hybrid; H3PO4, phosphoric acid; MCX, mixed-mode strong cation-exchange; n.a., not applicable.

Study samples

As previously detailed [14], the sample set included 78 serum and plasma samples from 26 volunteers (women/men: 13/13; age: 20–69 years) via gel-separator, bead clot activator and lithium-heparin containing tubes. The 6PLUS1® Multilevel Serum Calibrator set (Chromsystems; lot.5016, different from lots used for in-house calibration by Laboratories D, E, F and L) was measured within the study batch and used as the external calibration set (androstenedione: 0.64–50.6 nmol/L; DHEAS: 280–15,460 nmol/L; and testosterone: 0.173–39.9 nmol/L). Chromsystems’ calibrators were traceable to methanol certified reference materials for androstenedione and DHEAS, and to NIST SRM 972 for testosterone. EQA materials evaluated in this study included the Reference Institute for Bioanalytics (RfB; Bonn, Germany; www.rfb.bio) HM40121, HM40122, HM40123 and HM40124 (lyophilized human recalcified plasma spiked with steroids, no preservatives), assigned with target values determined by reference measurement procedure (RMP) for testosterone and by mean of all MS-methods for DHEAS; the United Kingdom National External Quality Assessment Service (UKNEQAS; Birmingham, UK; https://ukneqas.org.uk) materials (liquid off-the-clot minimally manipulated human serum) A382 (spiked with 10 nmol/L androstenedione standard) with androstenedione target assigned as mean of all MS-methods; the Instand e.V. (Düsseldorf, Germany; https://www.instand-ev.de) material pair N°302 (liquid human serum spiked with steroids, no additives) assigned with RMP-determined testosterone target value.

Running scheme and quantitation

Patient and EQA samples and external calibrators were measured in duplicate in two independent runs according to the protocols of each laboratory (Table 1). Each batch was quantified both by using in-house and external calibration sets. All calibration curves displayed R2>0.97.

Data analysis and statistics

Androstenedione, DHEAS and testosterone values are reported in nmol/L; to convert to ng/mL, multiply by 0.286, 0.368 and 0.288, respectively. Results were excluded if below the lower limit of quantification (LLOQ) or above the upper LOQ (ULOQ) of the in-house measurement range. Data were also excluded from the external calibration dataset when outside of the measurement range. Means and CVs of duplicate measurements were calculated (Supplementary Tables 3–6). Within – (CVi) and between – (CVg) subjects biological variabilities [16, 17] were used to assess the maximum allowable imprecision (MAI) and bias (MAB) and total allowable error (TAE) via the following equations [18]:

MAI = 0.5 × CVi;
MAB = 0.25 × ( CVi 2 + Cvg 2 ) 0.5 ;
TAE = 0.25 × ( CVi 2 + CVg 2 ) 0.5 + 1.65 × ( 0.5 × CVi ) ;

(androstenedione: 7.9, 10.5 and 23.5 %; DHEAS: 6.7, 8.1 and 19.0 %; testosterone: 6.3, 6.2 and 16.5 %; respectively) [16, 17].

Intra-laboratory performance

The overall intra-laboratory imprecision was determined as the duplicate measurement CV, as follows: CV = 1 n ∑ i = 1 n D i 2  × 100, where D i = ( x i 1 − x i 2 ) / 2 X ‾ i , xi1 and xi2 are duplicate measures of each sample, X ‾ i is the arithmetic mean of the duplicates and n is the total number of duplicates [19]. Intra-laboratory CVs from each laboratory were then compared with the MAI. The impact of calibration within each method was evaluated using the Friedman test. Trueness and bias were estimated in EQA materials as % difference from target values (RMP-determined or mean of all MS methods in the survey, respectively) and compared with the MAB. In addition, the bias of laboratory measures of EQA materials and patient samples vs. all laboratory median were determined. For each laboratory, the Bland-Altman 95 % limits of agreement were calculated for the patient sample pool. Commutability of EQA materials for each method was considered acceptable when the % bias in EQA material measures was within the above agreement interval.

Inter-laboratory performance

Between-method reproducibility, valued by the inter-laboratory CV, was compared with the MAI. Passing-Bablok and Bland-Altman analyses were performed. The % bias was compared with the TAE. Wilcoxon and F tests were used to compare inter-laboratory CV, median bias and bias variance by in-house vs. external calibration. The Friedman test was used to compare the measurements between laboratories. Statistics were performed by SPSS (v.20, IBM Co., Somers, NY) and MedCalc (v.18.2.1; Mariakerke, Belgium).

Results

Androstenedione

Patient values ranged from 0.88 to 5.74 nmol/L (all laboratory median by in-house calibration) (Supplementary Table 3). The intra-laboratory CVs ranged from 4.2 to 13.2 %, which included six laboratories within the MAI (Supplementary Table 2). Cases with duplicate CVs exceeding the MAI were up to 65.4 % for Laboratory H (Supplementary Table 3). Replacing in-house with external calibration determined modest deviations of calculated concentrations, except for Laboratory D (−9.7 %) and H (−21.7 %) (Supplementary Table 2). Laboratories D, H and E showed a positive bias in the A382 material (target value: 10.35 nmol/L), which was corrected by using external calibration in Laboratories D and H but not in E (Figure 1). Median inter-laboratory CV was 12.0 % with in-house, and lowered to 9.6 % with external calibration (p<0.001). Most of the cases displayed an inter-laboratory CV above the MAI with either calibrations (Figure 2 and Supplementary Table 7). By Passing-Bablok analysis of individual laboratory measures vs. all laboratory median (Supplementary Figure 1 and Table 2), seven laboratories showed a significant slope coefficient (range [90 CI]: 0.851 [0.811–0.891] to 1.186 [1.133–1.252]). Using external calibration improved the slope in all laboratories, except E. By Bland-Altman analysis (Figure 3, Supplementary Table 8 and Supplementary Figure 2), median bias vs. all laboratory median ranged from −13.3 to 20.5 % using in-house calibration. Laboratories E and H showed the largest variance of bias (95.3 and 123.2 %, respectively), with several cases exceeding the TAE. When using external calibration, median bias reduced to within ±5 % in all except Laboratory E, whose median bias almost doubled to 18.7 % (p<0.001) (Figure 3, Supplementary Table 8 and Supplementary Figure 2). Supplementary Table 9 shows that androstenedione measurements by in-house calibration were not comparable between any laboratories, while, with external calibration, some laboratories showed non different measures.

Figure 1: Laboratory measurements in external quality assessment (EQA) materials as function of the calibration system. Black dots: in-house calibration; white dots: external calibration. Bars: range of the duplicate measurements. Dashed lines: zero ± maximum allowable bias (androstenedione: 10.5 %; DHEAS: 8.1 %; testosterone: 6.2 %). T, target value. Target values of A382 for androstenedione, and of HM40121, HM40123 and HM40124 for DHEAS were determined as mean/median of all MS methods participating in the EQA survey. Target values of HM40121, HM40122, HM40124 and N°302 pair for testosterone were determined by reference measurement procedure.
Figure 1:

Laboratory measurements in external quality assessment (EQA) materials as function of the calibration system. Black dots: in-house calibration; white dots: external calibration. Bars: range of the duplicate measurements. Dashed lines: zero ± maximum allowable bias (androstenedione: 10.5 %; DHEAS: 8.1 %; testosterone: 6.2 %). T, target value. Target values of A382 for androstenedione, and of HM40121, HM40123 and HM40124 for DHEAS were determined as mean/median of all MS methods participating in the EQA survey. Target values of HM40121, HM40122, HM40124 and N°302 pair for testosterone were determined by reference measurement procedure.

Figure 2: Inter-laboratory coefficient of variations according to the calibration system. Black dots: in-house calibration; white dots: external calibration. Horizontal lines: maximal allowable imprecision (androstenedione: 7.9 %; DHEAS: 6.7 %; testosterone: 6.3 %).
Figure 2:

Inter-laboratory coefficient of variations according to the calibration system. Black dots: in-house calibration; white dots: external calibration. Horizontal lines: maximal allowable imprecision (androstenedione: 7.9 %; DHEAS: 6.7 %; testosterone: 6.3 %).

Table 2:

Passing-Bablok analyses of steroid measurements from each laboratory vs. the medians of all laboratories as a function of the calibration system.

Analyte Lab In-house calibration External calibration
n R (95 CI) Slope (95 CI) Intercept (95 CI) n R (95 CI) Slope (95 CI) Intercept (95 CI)
Androstenedione B 78 0.988 (0.981–0.992) 0.962 (0.938–0.988) 0.043 (−0.020 to 0.104) 78 0.988 (0.980–0.992) 0.994 (0.980–1.011) 0.043 (0.009–0.081)
D 78 0.988 (0.981–0.992) 1.074 (1.039–1.108) −0.053 (−0.136 to 0.036) 78 0.993 (0.988–0.995) 1.024 (1.000–1.046) −0.048 (−0.113 to 0.000)
E 78 0.973 (0.958–0.983) 1.104 (1.058–1.167) −0.019 (−0.192 to 0.096) 78 0.976 (0.962–0.985) 1.194 (1.138–1.263) −0.017 (−0.208 to 0.090)
F 78 0.995 (0.992–0.997) 0.978 (0.956–0.996) −0.025 (−0.082 to 0.019) 78 0.994 (0.991–0.996) 1.011 (0.987–1.032) −0.097 (−0.152 to −0.039)
G 78 0.972 (0.957–0.982) 0.851 (0.811–0.891) 0.058 (−0.063 to 0.150) 78 0.973 (0.958–0.983) 0.938 (0.897–0.985) 0.025 (−0.088 to 0.116)
H 78 0.974 (0.959–0.983) 1.186 (1.133–1.252) 0.038 (−0.148 to 0.158) 78 0.980 (0.969–0.987) 0.964 (0.926–1.009) 0.052 (−0.048 to 0.141)
I 75 0.986 (0.978–0.991) 0.941 (0.915–0.969) −0.070 (−0.142 to 0.004) 75 0.985 (0.976–0.990) 1.016 (0.993–1.040) −0.079 (−0.146 to −0.014)
L 78 0.992 (0.987–0.995) 1.016 (0.997–1.041) −0.022 (−0.089 to 0.038) 78 0.987 (0.980–0.992) 1.013 (0.989–1.040) −0.025 (−0.098 to 0.040)
DHEAS C 76 0.996 (0.994–0.998) 0.979 (0.964–0.997) −39.6 (−99.0 to 18.4) 76 0.996 (0.994–0.998) 1.004 (0.996–1.017) −8.7 (−63.4 to 14.5)
D 77 0.991 (0.987–0.995) 1.030 (0.994–1.053) 76.6 (−20.1 to 142.0) 77 0.992 (0.988–0.995) 0.929 (0.895–0.950) 120.1 (62.8–183.6)
E 69 0.995 (0.993–0.997) 0.998 (0.983–1.003) −49.4 (−63.6 to 12.5) 69 0.996 (0.994–0.998) 1.000 (0.998–1.010) 0.0 (−54.2 to 4.3)
G 75 0.997 (0.996–0.998) 1.000 (1.000–1.006) −0.0 (−15.3 to 0.0) 75 0.996 (0.993–0.997) 1.056 (1.049–1.064) −138.7 (−174.4 to −109.3)
H 78 0.998 (0.996–0.998) 0.825 (0.803–0.847) 201.2 (131.8–287.9) 78 0.998 (0.997–0.999) 0.972 (0.954–0.991) 192.0 (118.3–273.6)
I 75 0.998 (0.996–0.999) 1.087 (1.072–1.105) −108.1 (−173.4 to −69.7) 75 0.997 (0.996–0.998) 0.994 (0.980–1.006) −78.2 (−128.1 to −38.9)
L 78 0.997 (0.995–0.998) 1.005 (0.998–1.015) 47.1 (0.2–78.5) 78 0.996 (0.994–0.998) 1.009 (1.000–1.023) −22.2 (−58.0 to 0.0)
Testosterone – female range B 42 0.988 (0.977–0.993) 1.024 (0.980–1.077) −0.023 (−0.058 to 0.009) 42 0.985 (0.971–0.992) 1.081 (1.019–1.140) −0.037 (−0.075 to 0.010)
C 42 0.967 (0.939–0.982) 1.100 (1.030–1.177) −0.025 (−0.075 to 0.026) 42 0.962 (0.930–0.979) 1.000 (0.976–1.062) 0.000 (−0.043 to 0.025)
D 42 0.957 (0.921–0.977) 1.054 (1.000–1.125) −0.057 (−0.101 to 0.015) 42 0.951 (0.909–0.973) 1.000 (0.938–1.077) 0.040 (−0.006 to 0.093)
E 42 0.948 (0.905–0.972) 1.026 (0.991–1.094) −0.013 (−0.060 to 0.019) 42 0.965 (0.936–0.981) 1.019 (0.991–1.068) −0.021 (−0.060 to 0.005)
F 42 0.973 (0.949–0.985) 1.000 (0.973–1.000) 0.000 (0.000–0.021) 42 0.963 (0.932–0.980) 0.913 (0.875–0.944) −0.015 (−0.040 to 0.011)
G 42 0.931 (0.874–0.962) 1.025 (0.985–1.082) −0.022 (−0.064 to 0.012) 42 0.928 (0.869–0.961) 1.090 (1.019–1.143) −0.024 (−0.067 to 0.031)
H 42 0.966 (0.937–0.982) 0.979 (0.938–1.000) 0.016 (0.000–0.048) 42 0.961 (0.928–0.979) 1.000 (0.980–1.050) 0.000 (−0.038 to 0.015)
I 42 0.954 (0.915–0.975) 1.000 (0.972–1.022) −0.020 (−0.041 to 0.003) 42 0.958 (0.922–0.977) 1.000 (0.989–1.030) −0.010 (−0.040 to 0.004)
L 42 0.938 (0.887–0.967) 1.020 (0.973–1.107) −0.012 (−0.071 to 0.027) 42 0.934 (0.880–0.964) 1.016 (0.957–1.096) −0.043 (−0.109 to 0.009)
Testosterone – male range B 36 0.943 (0.890–0.971) 0.928 (0.877–0.979) 0.052 (−0.819 to 0.954) 36 0.971 (0.943–0.985) 0.994 (0.957–1.020) 0.105 (−0.366 to 0.723)
C 36 0.937 (0.880–0.968) 1.074 (1.031–1.140) −0.328 (−1.354 to 0.356) 36 0.966 (0.933–0.983) 1.000 (0.979–1.041) 0.000 (−0.761 to 0.386)
D 36 0.910 (0.829–0.953) 0.998 (0.927–1.046) 0.048 (−0.612 to 1.296) 36 0.905 (0.820–0.951) 0.943 (0.869–1.011) 0.606 (−0.581 to 1.832)
E 36 0.947 (0.897–0.973) 1.016 (0.942–1.095) 0.044 (−1.450 to 1.247) 36 0.929 (0.865–0.964) 0.996 (0.916–1.082) 0.364 (−1.172 to 1.716)
F 36 0.982 (0.966–0.991) 1.000 (0.998–1.028) −0.095 (−0.570 to −0.049) 36 0.983 (0.968–0.992) 0.940 (0.917–0.966) −0.189 (−0.637 to 0.148)
G 36 0.941 (0.886–0.969) 1.012 (0.971–1.060) −0.361 (−0.453 to 1.040) 36 0.927 (0.860–0.962) 1.089 (1.043–1.145) 0.466 (−0.454 to 1.201)
H 36 0.966 (0.934–0.983) 0.969 (0.917–1.022) −0.307 (−1.225 to 0.467) 36 0.984 (0.968–0.992) 1.000 (0.949–1.040) −0.015 (−0.741 to 0.721)
I 33 0.980 (0.959–0.990) 1.019 (0.978–1.070) −0.994 (−1.817 to 0.347) 33 0.973 (0.945–0.987) 1.017 (0.985–1.074) −0.469 (−1.467 to −0.030)
L 36 0.924 (0.855–0.961) 1.180 (1.097–1.316) −1.668 (−4.047 to 0.022) 36 0.930 (0.866–0.964) 1.161 (1.086–1.269) −1.464 (−3.299 to −0.265)

  1. n, number of samples; CI, confidence interval.

Figure 3: Laboratories % bias vs. median of all laboratories as function of the calibration system. % bias = ([laboratory value – median of all laboratories]/median of all laboratories) × 100. Segments: median; bars: 2.5 and 97.5 centiles; horizontal lines: zero ± total allowable error (androstenedione: 23.5 %; DHEAS: 19.0 %; testosterone: 16.5 %).
Figure 3:

Laboratories % bias vs. median of all laboratories as function of the calibration system. % bias = ([laboratory value – median of all laboratories]/median of all laboratories) × 100. Segments: median; bars: 2.5 and 97.5 centiles; horizontal lines: zero ± total allowable error (androstenedione: 23.5 %; DHEAS: 19.0 %; testosterone: 16.5 %).

DHEAS

Patient samples ranged from 541 to 14,372 nmol/L (all laboratory median by in-house calibration) (Supplementary Table 4). With in-house calibration, values were above the ULOQ of Laboratory D and C in one and two samples, respectively, and below the LLOQ of Laboratory E in nine samples. With external calibration, three samples were below the measurement range in Laboratory G. The intra-laboratory CV ranged from 1.6 to 10.8 % (Supplementary Table 2), within the MAI in all except Laboratory L, showing up to 73.1 % cases with duplicate CV above the threshold (Supplementary Table 4). Changing in-house with external calibration determined only modest deviations in measures within each laboratory, except for Laboratory H (18.9 %) (Supplementary Table 2). HM40122 target value (14,216 nmol/L) was out of range for most of the laboratories and was therefore not evaluated. Targets of other HMs ranged from 3,679 to 11,200 nmol/L. Bias were within the MAB in all except Laboratories H and I, showing respective under- and over-estimation which was corrected when applying external calibration (Figure 1). The median inter-laboratory CV was 7.2 % with in-house, and lowered to 4.9 % with external calibration (p<0.001). Similarly, the prevalence of cases with inter-laboratory CV above MAI reduced from 55.1 % with in-house to 15.4 % with external calibration; with the latter, all samples above 8,000 nmol/L were within the threshold (Figure 2 and Supplementary Table 7). At Passing-Bablok analysis, when using in-house calibration, laboratories showed modest slope deviations (from 0.979 [0.964–0.997] to 1.087 [1.071–1.105]), except for Laboratory H (0.825 [0.803–0.847]). External calibration improved the overall slope range (from 0.929 [0.895–0.950] to 1.056 [1.049–1.064]). However, large intercepts were noted for some methods with both calibration systems (Supplementary Figure 1 and Table 2). Median bias vs. all laboratory median was within ±4.8 % for all except Laboratory H (−10.9 %) (Figure 3, Supplementary Table 8 and Figure 3). Laboratory D and H exhibited large variance of bias and a few cases exceeding the TAE. Using external calibration overall reduced the median bias (within ±3.5 %) and cases exceeding the TAE (Figure 3, Supplementary Table 8 and Figure 3). Supplementary Table 9 shows that DHEAS measures by in-house calibration were mostly non comparable between laboratories, whereas some similarities were obtained by external calibration.

Testosterone

Results were subdivided in female and male ranges. The former included 42 samples from 13 women and 1 severely hypogonadal male, with values between 0.43 and 1.71 nmol/L. The latter included 36 samples from 12 men, with values between 8.11 and 28.45 nmol/L (all laboratory median by in-house calibration) (Supplementary Tables 5 and 6). The intra-laboratory CV ranged 4.3–8.7 % for female and 2.6–7.1 % for male levels, which included four and eight laboratories within the MAI, respectively (Supplementary Table 2). Cases with duplicate CVs above the MAI were up to 65.4 % for Laboratory D at female levels (Supplementary Tables 5 and 6). Applying external instead of in-house calibration demonstrated modest deviation of measurements overall, except for Laboratories D (13.5 %), F (−9.0 %) and G (9.2 %) in the female range (Supplementary Table 2). Material HM40123 (target: 40.8 nmol/L) was above most of the measurement ranges, and was therefore not evaluated. Targets in other materials ranged from 4.68 to 25.05 nmol/L (Figure 1). When using in-house calibration, trueness ranged from −7.1 to 15.7 % in HM and in N°302 high materials, with six laboratories outside of the MAB in two materials or more. Employing external calibration determined worse (Laboratories B and G), similar (Laboratories D, E, H, I and L) or better (Laboratories C and F) trueness depending on the laboratory. Interestingly, some methods performed markedly differently with material N°302 low (overall trueness: −28.6 to 14.2 %) compared to other materials, such as Laboratories B, E and L (Figure 1). N°302 low also failed the commutability test with results either above (Laboratories G and I) and below (Laboratory L) the agreement interval (Supplementary Figure 4). Small deviations from commutability were also noted for materials HM40121 (Supplementary Figure 4). The median inter-laboratory CV was similar between female and male ranges when using in-house calibration (6.4 and 6.8 %, respectively). However, when applied to the female range, the external calibration significantly increased the inter-laboratory CV to 7.6 % (p<0.001) and cases above MAI from 50 to 81 % (Figure 2 and Supplementary Table 7). With Passing-Bablok analysis (Supplementary Figure 1 and Table 2), when using in-house calibration, methods’ slopes performed better in the female (from 0.979 [0.938–1.000] to 1.100 [1.030–1.177]) than in the male range (from 0.928 [0.877–0.979] to 1.180 [1.097–1.316]), and were mostly not significant. Using external calibration resulted in similar results for male levels, but poorer for female levels (from 0.913 [0.875–0.944] to 1.090 [1.019–1.143]). Large intercepts were also noted for Laboratories I and L at male levels, which were not corrected by external calibration. With in-house calibration, the median bias vs. all laboratory median ranged from −2.7 to 6.5 in the female, and −7.0 to 8.5 % in the male range (Figure 3, Supplementary Table 8, Supplementary Figures 5 and 6). External calibration overall enlarged the bias to −11.3 to 6.6 % in the female and −7.5 to 11.8 % in the male range. Deviations were mostly within the TAE for both levels and calibration systems. The largest bias variance, up to 67.6 %, was noted for Laboratory L in all tested conditions (Figure 3, Supplementary Table 8, Supplementary Figures 5 and 6). Supplementary Table 9 shows that testosterone measures in the female range were non different between various laboratories when using in-house calibration. However, comparability was observed only in a few cases when using external calibration and when moving to the male range.

Discussion

Our study contributes to harmonization/standardization knowledge and status of androgen measurement by LC-MS/MS. While some previous studies have reported on testosterone and androstenedione, none focused on DHEAS [7], [8], [9], [10], [11], [12]. We interpreted our results on precision, trueness, reproducibility and agreement by use of maximum allowable performances calculated from recently updated data on biological variability [16, 17]. The methods here compared were validated according to recommended guidelines [20], [21], [22], [23], [24], [25], [26], [27], [28]. They encompass different procedures for sample preparation, LC and calibration. Methods that included measurements of DHEAS adopted either positive or negative ionization and four different ion transitions. In contrast, the same ion transition was used by all methods for testosterone, and all but one for androstenedione. At least four different IS were used for DHEAS and five for testosterone and androstenedione quantitation.

According to our data, all except two methods demonstrated good precision, within or slightly above the MAI. Less than optimal precision for measurements of androstenedione by Laboratory H may be attributable to use of a non-matching isotopic IS (D3-testosterone). In contrast, sub-optimal precision for measurements of DHEAS by Laboratory L may have reflected initial inexperience with maintenance of the MS [14, 15]. In EQA materials, deviations of androstenedione and DHEAS were within ±20 % for all methods, while trueness for testosterone was within ±15 %. Of note, some methods exceeded the MAB thresholds.

Although the inter-laboratory performance was generally acceptable for all three hormones, agreement was less satisfactory for androstenedione than the other two androgens, both in absolute terms and according to allowable thresholds. Although the inter-laboratory CVs for DHEAS and testosterone were within or slightly above the MAI, those for androstenedione were mostly above the threshold. All laboratories displayed median bias largely within the TAE. Unexpectedly, inter-laboratory agreement was similar or better for testosterone in females than males. This contrasts with earlier studies [7, 8, 10, 11], but not with a more recent one [12].

Some methods displayed relevant random errors as indicated by bias variance. The largest were reported for androstenedione by Laboratory E and H, which may have reflected use of non-matching isotopic IS 13C3-17OH-progesterone and D3-testosterone in those respective methods. Laboratory E used the same 13C3-17OH-progesterone for testosterone measurement, and showed a relatively large bias variance also for this analyte. These data support the importance of a suitable IS for the reliability of quantitation. Non-matching isotopic IS may not efficiently cope for variability in pre-analytics and analytics [29] and should be avoided. In a study that focused on testosterone, it was even reported that different forms of isotopically-labeled testosterone may influence quantitation up to 15 % [13]. Differences in the retention times between the analyte and respective IS, though small in nature, have been suggested to play a role. Within the current study, methods using the same IS did not show closer inter-laboratory agreement compared to others (data not shown). This suggests that the random variability among LC-MS/MS methods, i.e., variability not explained by calibration bias, results from a complex interaction between individual matrix, extraction procedure, chromatographic separation, MS detection and the ability of the IS to correct for all confounders. Furthermore, several isobaric compounds exist that may impact the accuracy of LC-MS/MS quantitation of the three androgens [3]. Notably, only part of the methods verified the resolution between testosterone and epitestosterone.

Due to different designs among previous studies on androgen harmonization, direct comparison of those and our results is not possible. Most studies employed limited numbers of samples [7], [8], [9], [10], [11]. Some, similar to our study, were ring trials [7, 8, 11, 12], while others included a reference technique [9, 10]. Overall, our results compare favorably with most previous studies on testosterone [7, 8, 10, 11] and androstenedione [7, 8]. We can also confirm previous observations of higher inter-laboratory agreement for testosterone than androstenedione [7, 8]. In addition, our results compare similarly to a study measuring testosterone in 58 samples comparing four LC-MS/MS to one GC-MS method [9]. Recently, French et al. [12] compared testosterone measurements of 102 samples among four LC-MS/MS methods; they reported a mean bias between −4.9 and +3.7 %. Such agreement may reflect technical similarities among investigated methods. Indeed, all were calibrated using the NIST SRM 971; four used a C18 stationary phase and acetonitrile-based mobile phases; three used D3-testosterone as the IS; and three used the same sample extraction procedure [12].

In agreement with our previous reports, the strategy to unify calibration resulted in different findings depending on the analyte [14, 15]. Here, this strategy was beneficial for androstenedione and DHEAS comparability, as it reduced the CV for both hormones by about a third, and also enhanced correlation and agreement. Reasons are yet to be found for the correction of large calibration bias of Laboratories D and H for androstenedione, and of Laboratories H and I for DHEAS. Notably, improper calibration was not the reason for proportional overestimation demonstrated by Laboratory E for androstenedione. Once again, this may depend on the use of non-matched isotopic IS. As for testosterone, in line with previous observations [13], calibration was not a relevant source of error. Indeed, external calibration resulted in contrasting effects on testosterone trueness, and it worsened the inter-laboratory performance in the female range. Such findings are possibly due to a suboptimal commutability of the serum-derived matrix of the commercial set.

Our study highlighted a remarkable status of harmonization of androgen measurement by LC-MS/MS methods. Unfortunately, we were not able to include DHEA in this report, as only three laboratories measured this analyte, and levels were below the LOQ in several cases. One limitation is that no reference is available to assess true methods’ deviations in patient samples. In principle, our present and previous studies support the unification of the calibration material as a strategy to improve harmonization, provided that calibrators are traceable to higher order materials or RMP [14, 15]. However, we repeatedly observed issues with commutability of calibration matrix that need to be addressed before such a strategy can be introduced on a large scale [14, 15]. As reported here for N°302 and previously [14, 15], EQA materials are also not free from matrix effects. It is important that commutability is not to be taken for granted with LC-MS/MS, or the result of these precious surveys may be counterproductive [30]. Sources of variability other than calibration should be investigated. In these regards, our study design opens the discussion for the type of vacuum tube used to derive serum or plasma samples for steroid investigations.

Satisfactory results were observed for DHEAS and testosterone. Reasons are possibly found in the wider availability of EQA surveys for both hormones. Most probably, testosterone performance benefits from the availability of the RMP for assigning target values [31, 32]. Conversely, no RMP is available for DHEAS [32]. The less than ideal performance noted for androstenedione may relate to the scarce availability of surveys for this hormone. Recently, a RMP has been listed in the Joint Committee for Traceability in Laboratory Medicine [32, 33]. We hope that such an addition will soon be translated into benefits for EQA surveys. That underlines the importance of a reference measurement system which generally entails a RMP and higher order reference materials to improve measurement accuracy and reduce variability between laboratories. Another important limitation of EQA materials can be their usual high target values, sometimes even exceeding the calibration range. This implies that laboratories can only verify their trueness in severe hyperandrogenic conditions or, for testosterone, in the normal male range. Contrarily, to ensure effective disease comprehension and treatment, modern clinical practice and research need to rely on methods that perform optimally for both sexes and in physiologic and pathologic ranges.

In conclusion, our study highlights improvements in the inter-laboratory comparability of androgen measurements made possible by LC-MS/MS and supports the feasibility of achieving a successful harmonization and standardization. In these regards, we shed light on important critical points, such as commutability of calibrator and EQA materials and avoiding non-matching isotopic IS, that need to be addressed to achieve the desirable goal.

Corresponding author: Flaminia Fanelli, Senior Assistant Professor, Department of Medical and Surgical Sciences, Endocrinology Research Group, Center for Applied Biomedical Research, Alma Mater Studiorum University of Bologna, Via Massarenti 9, 40138 Bologna, Italy, Phone: +39 051 2143902, Fax: +39 051 2143902, E-mail:

Acknowledgments

We thank Dr. Oliver Midasch from Chromsystems for donating calibration materials used in the study.

  1. Research ethics: The study received approval by the local Ethics Committee (no. 141/2017/U/Tess).

  2. Informed consent: Informed consent was obtained from all individuals included in this study.

  3. Author contributions: All authors have accepted responsibility for the entire content of this manuscript and approved its submission. FF conceived, designed and coordinated the study, performed the statistical analysis and wrote the manuscript; MC, BGK, MV and UP contributed to the study design; MC, MM, JML, MP, JMH, SB, MTA, JVDO and DK carried out sample measurement and data exports; AT coordinated subject recruitment and sample management; EN performed the statistical analysis; UP conceived the study; MC, MP, JMH, SB, PAB, MTA, ACH, JVDO, FM, MR, GE, BGK, MV and UP contributed in result interpretation and in writing the manuscript.

  4. Competing interests: The authors state no conflict of interest.

  5. Research funding: This study was supported by the Emilia-Romagna Region, Alessandro Liberati Young Researcher Grants (project number PRUA 1-2012-004, granted to FF) and by Deutsche Forschungsgemeinschaft (DFG), German Research Foundation (project number 314061271-CRC/TRR 205, granted to MP and GE).

References

1. Pugeat, M, Plotton, I, de la Perrière, AB, Raverot, G, Déchaud, H, Raverot, V. Management of endocrine disease hyperandrogenic states in women: pitfalls in laboratory diagnosis. Eur J Endocrinol 2018;178:R141–54 https://doi.org/10.1530/eje-17-0776.Search in Google Scholar

2. Guzelce, EC, Galbiati, F, Goldman, AL, Gattu, AK, Basaria, S, Bhasin, S. Accurate measurement of total and free testosterone levels for the diagnosis of androgen disorders. Best Pract Res Clin Endocrinol Metab 2022;36:101683. https://doi.org/10.1016/j.beem.2022.101683.Search in Google Scholar PubMed

3. Wudy, SA, Schuler, G, Sánchez-Guijo, A, Hartmann, MF. The art of measuring steroids: principles and practice of current hormonal steroid analysis. J Steroid Biochem Mol Biol 2018;179:88–103. https://doi.org/10.1016/j.jsbmb.2017.09.003.Search in Google Scholar PubMed

4. French, D. Clinical utility of laboratory developed mass spectrometry assays for steroid hormone testing. J Mass Spectrom Adv Clin Lab 2023;28:13–19. https://doi.org/10.1016/j.jmsacl.2023.01.006.Search in Google Scholar PubMed PubMed Central

5. Lentjes, EGWM, Bui, HN, Ruhaak, LR, Kema, IP, Coene, KLM, van den Ouweland, JMW. LC-MS/MS in clinical chemistry: did it live up to its promise? Consideration from the Dutch EQAS organization. Clin Chim Acta 2023;546:117391. https://doi.org/10.1016/j.cca.2023.117391.Search in Google Scholar PubMed

6. Dirks, NF, Ackermans, MT, Martens, F, Cobbaert, CM, de Jonge, R, Heijboer, AC. We need to talk about the analytical performance of our laboratory developed clinical LC-MS/MS tests, and start separating the wheat from the chaff. Clin Chim Acta 2021;514:80–3. https://doi.org/10.1016/j.cca.2020.12.020.Search in Google Scholar PubMed

7. Büttler, RM, Martens, F, Fanelli, F, Pham, HT, Kushnir, MM, Janssen, MJ, et al.. Comparison of 7 published LC-MS/MS methods for the simultaneous measurement of testosterone, androstenedione, and dehydroepiandrosterone in serum. Clin Chem 2015;61:1475–83. https://doi.org/10.1373/clinchem.2015.242859.Search in Google Scholar PubMed

8. Büttler, RM, Martens, F, Ackermans, MT, Davison, AS, van Herwaarden, AE, Kortz, L, et al.. Comparison of eight routine unpublished LC-MS/MS methods for the simultaneous measurement of testosterone and androstenedione in serum. Clin Chim Acta 2016;454:112–8. https://doi.org/10.1016/j.cca.2016.01.002.Search in Google Scholar PubMed

9. Thienpont, LM, Van Uytfanghe, K, Blincko, S, Ramsay, CS, Xie, H, Doss, RC, et al.. State-of-the-art of serum testosterone measurement by isotope dilution-liquid chromatography-tandem mass spectrometry. Clin Chem 2008;54:1290–7. https://doi.org/10.1373/clinchem.2008.105841.Search in Google Scholar PubMed

10. Vesper, HW, Bhasin, S, Wang, C, Tai, SS, Dodge, LA, Singh, RJ, et al.. Interlaboratory comparison study of serum total testosterone [corrected] measurements performed by mass spectrometry methods. Steroids 2009;74:498–503. https://doi.org/10.1016/j.steroids.2009.01.004.Search in Google Scholar PubMed

11. Owen, LJ, MacDonald, PR, Keevil, BG. Is calibration the cause of variation in liquid chromatography tandem mass spectrometry testosterone measurement? Ann Clin Biochem 2013;50:368–70. https://doi.org/10.1177/0004563212474563.Search in Google Scholar PubMed

12. French, D, Drees, J, Stone, JA, Holmes, DT, van der Gugten, JG. Comparison of four clinically validated testosterone LC-MS/MS assays: harmonization is an attainable goal. Clin Mass Spectrom 2018;11:12–20. https://doi.org/10.1016/j.clinms.2018.11.005.Search in Google Scholar PubMed PubMed Central

13. Owen, LJ, Keevil, BG. Testosterone measurement by liquid chromatography tandem mass spectrometry: the importance of internal standard choice. Ann Clin Biochem 2012;49:600–2. https://doi.org/10.1258/acb.2012.012037.Search in Google Scholar PubMed

14. Fanelli, F, Cantù, M, Temchenko, A, Mezzullo, M, Lindner, JM, Peitzsch, M, et al.. Report from the HarmoSter study: impact of calibration on comparability of LC-MS/MS measurement of circulating cortisol, 17OH-progesterone and aldosterone. Clin Chem Lab Med 2022;60:726–39. https://doi.org/10.1515/cclm-2021-1028.Search in Google Scholar PubMed

15. Fanelli, F, Bruce, S, Cantù, M, Temchenko, A, Mezzullo, M, Lindner, JM, et al.. Report from the HarmoSter study: inter-laboratory comparison of LC-MS/MS measurements of corticosterone, 11-deoxycortisol and cortisone. Clin Chem Lab Med 2022;61:67–77. https://doi.org/10.1515/cclm-2022-0242.Search in Google Scholar PubMed

16. Westgard, QC. Desirable biological variation database specifications. https://www.westgard.com/biodatabase1.htm [Accessed 28 Aug 2023].Search in Google Scholar

17. The European federation of clinical chemistry and laboratory medicine database. https://biologicalvariation.eu/meta_calculations [Accessed 28 Aug 2023].Search in Google Scholar

18. Oosterhuis, WP, Bayat, H, Armbruster, D, Coskun, A, Freeman, KP, Kallner, A, et al.. The use of error and uncertainty methods in the medical laboratory. Clin Chem Lab Med 2018;56:209–19. https://doi.org/10.1515/cclm-2017-0341.Search in Google Scholar PubMed

19. Hyslop, NP, White, WH. Estimating precision using duplicate measurements. J Air Waste Manag Assoc 2009;59:1032–9. https://doi.org/10.3155/1047-3289.59.9.1032.Search in Google Scholar PubMed

20. Fanelli, F, Belluomo, I, Di Lallo, VD, Cuomo, G, De Iasio, R, Baccini, M, et al.. Serum steroid profiling by isotopic dilution-liquid chromatography-mass spectrometry: comparison with current immunoassays and reference intervals in healthy adults. Steroids 2011;76:244–53. https://doi.org/10.1016/j.steroids.2010.11.005.Search in Google Scholar PubMed

21. Lindner, JM, Vogeser, M, Grimm, SH. Biphenyl based stationary phases for improved selectivity in complex steroid assays. J Pharm Biomed Anal 2017;142:66–73. https://doi.org/10.1016/j.jpba.2017.04.020.Search in Google Scholar PubMed

22. Peitzsch, M, Dekkers, T, Haase, M, Sweep, FC, Quack, I, Antoch, G, et al.. An LC-MS/MS method for steroid profiling during adrenal venous sampling for investigation of primary aldosteronism. J Steroid Biochem Mol Biol 2015;145:75–84. https://doi.org/10.1016/j.jsbmb.2014.10.006.Search in Google Scholar PubMed

23. Fahlbusch, FB, Heussner, K, Schmid, M, Schild, R, Ruebner, M, Huebner, H, et al.. Measurement of amniotic fluid steroids of midgestation via LC-MS/MS. J Steroid Biochem Mol Biol 2015;152:155–60. https://doi.org/10.1016/j.jsbmb.2015.05.014.Search in Google Scholar PubMed

24. Hawley, JM, Adaway, JE, Owen, LJ, Keevil, BG. Development of a total serum testosterone, androstenedione, 17-hydroxyprogesterone, 11β-hydroxyandrostenedione and 11-ketotestosterone LC-MS/MS assay and its application to evaluate pre-analytical sample stability. Clin Chem Lab Med 2020;58:741–52. https://doi.org/10.1515/cclm-2019-0959.Search in Google Scholar PubMed

25. Büttler, RM, Kruit, A, Blankenstein, MA, Heijboer, AC. Measurement of dehydroepiandrosterone sulphate (DHEAS): a comparison of isotope-dilution liquid chromatography tandem mass spectrometry (ID-LC-MS/MS) and seven currently available immunoassays. Clin Chim Acta 2013;424:22–6. https://doi.org/10.1016/j.cca.2013.04.028.Search in Google Scholar PubMed

26. Büttler, RM, Martens, F, Kushnir, MM, Ackermans, MT, Blankenstein, MA, Heijboer, AC. Simultaneous measurement of testosterone, androstenedione and dehydroepiandrosterone (DHEA) in serum and plasma using isotope-dilution 2-dimension ultra high performance liquid-chromatography tandem mass spectrometry (ID-LC-MS/MS). Clin Chim Acta 2015;438:157–9. https://doi.org/10.1016/j.cca.2014.08.023.Search in Google Scholar PubMed

27. Bruce, SJ, Rey, F, Béguin, A, Berthod, C, Werner, D, Henry, H. Discrepancy between radioimmunoassay and high performance liquid chromatography tandem-mass spectrometry for the analysis of androstenedione. Anal Biochem 2014;455:20–5. https://doi.org/10.1016/j.ab.2014.03.021.Search in Google Scholar PubMed

28. Laszlo, CF, Montoya, JP, Shamseddin, M, De Martino, F, Beguin, A, Nellen, R, et al.. A high resolution LC–MS targeted method for the concomitant analysis of 11 contraceptive progestins and 4 steroids. J Pharm Biomed Anal 2019;175:112756. https://doi.org/10.1016/j.jpba.2019.07.004.Search in Google Scholar PubMed

29. Vogeser, M, Seger, C. Pitfalls associated with the use of liquid chromatography-tandem mass spectrometry in the clinical laboratory. Clin Chem 2010;56:1234–44. https://doi.org/10.1373/clinchem.2009.138602.Search in Google Scholar PubMed

30. Miller, WG, Myers, GL, Rej, R. Why commutability matters. Clin Chem 2006;52:553–4. https://doi.org/10.1373/clinchem.2005.063511.Search in Google Scholar PubMed

31. Greaves, RF. The central role of external quality assurance in harmonisation and standardisation for laboratory medicine. Clin Chem Lab Med 2017;55:471–3. https://doi.org/10.1515/cclm-2016-0782.Search in Google Scholar PubMed

32. Joint Committee for Traceability in Laboratory Medicine (JCTLM). Database of higher-order reference materials, measurement methods/procedures and services. https://www.bipm.org/jctlm/ [Accessed 28 Aug, 2023].Search in Google Scholar

33. Gradl, K, Taibon, J, Singh, N, Albrecht, E, Geistanger, A, Pongratz, S, et al.. An isotope dilution LC-MS/MS-based candidate reference method for the quantification of androstenedione in human serum and plasma. Clin Mass Spectrom 2020;23:1–10. https://doi.org/10.1016/j.clinms.2020.01.003.Search in Google Scholar PubMed PubMed Central

Supplementary Material

This article contains supplementary material (https://doi.org/10.1515/cclm-2023-1138).

Received: 2023-10-12
Accepted: 2023-12-26
Published Online: 2024-01-12
Published in Print: 2024-05-27

© 2024 the author(s), published by De Gruyter, Berlin/Boston

This work is licensed under the Creative Commons Attribution 4.0 International License.

Articles in the same Issue

  1. Frontmatter
  2. Editorial
  3. SARS-CoV-2 is here to stay: do not lower our guard
  4. Reviews
  5. SARS-CoV-2 subgenomic RNA: formation process and rapid molecular diagnostic methods
  6. Prognostic value of anti-SARS-CoV-2 antibodies: a systematic review
  7. Presence of SARS-CoV-2 RNA in COVID-19 survivors with post-COVID symptoms: a systematic review of the literature
  8. Opinion Papers
  9. Harmonizing the post-analytical phase: focus on the laboratory report
  10. Blood-based biomarkers in Alzheimer’s disease – moving towards a new era of diagnostics
  11. A comprehensive review on PFAS including survey results from the EFLM Member Societies
  12. General Clinical Chemistry and Laboratory Medicine
  13. Report from the HarmoSter study: different LC-MS/MS androstenedione, DHEAS and testosterone methods compare well; however, unifying calibration is a double-edged sword
  14. An LC–MS/MS method for serum cystatin C quantification and its comparison with two commercial immunoassays
  15. CX3CL1/Fractalkine as a biomarker for early pregnancy prediction of preterm premature rupture of membranes
  16. Elevated S100B urine levels predict seizures in infants complicated by perinatal asphyxia and undergoing therapeutic hypothermia
  17. The correlation of urea and creatinine concentrations in sweat and saliva with plasma during hemodialysis: an observational cohort study
  18. Tubular phosphate transport: a comparison between different methods of urine sample collection in FGF23-dependent hypophosphatemic syndromes
  19. Reference Values and Biological Variations
  20. Monocyte distribution width (MDW): study of reference values in blood donors
  21. Data mining of reference intervals for serum creatinine: an improvement in glomerular filtration rate estimating equations based on Q-values
  22. Hematology and Coagulation
  23. MALDI-MS in first-line screening of newborns for sickle cell disease: results from a prospective study in comparison to HPLC
  24. Cardiovascular Diseases
  25. To rule-in, or not to falsely rule-out, that is the question: evaluation of hs-cTnT EQA performance in light of the ESC-2020 guideline
  26. Temporal biomarker concentration patterns during the early course of acute coronary syndrome
  27. Diabetes
  28. Proteomic analysis of diabetic retinopathy identifies potential plasma-protein biomarkers for diagnosis and prognosis
  29. Infectious Diseases
  30. Serum biomarkers of inflammation and vascular damage upon SARS-Cov-2 mRNA vaccine in patients with thymic epithelial tumors
  31. A high throughput immuno-affinity mass spectrometry method for detection and quantitation of SARS-CoV-2 nucleoprotein in human saliva and its comparison with RT-PCR, RT-LAMP, and lateral flow rapid antigen test
  32. Evaluation of inflammatory biomarkers and vitamins in hospitalized patients with SARS-CoV-2 infection and post-COVID syndrome
  33. The CoLab score is associated with SARS-CoV-2 viral load during admission in individuals admitted to the intensive care unit: the CoLaIC cohort study
  34. Development and evaluation of a CRISPR-Cas13a system-based diagnostic for hepatitis E virus
  35. Letters to the Editor
  36. Crioplast® is a reliable device to ensure pre-analytical stability of adrenocorticotrophin (ACTH)
  37. Falsely decreased Abbott Alinity-c gamma-glutamyl transferase-2 result from paraprotein and heparin interference: case report and subsequent laboratory experiments
  38. Impact of hemolysis on uracilemia in the context of dihydropyrimidine dehydrogenase deficiency testing
  39. Value of plasma neurofilament light chain for monitoring efficacy in children with later-onset spinal muscular atrophy under nusinersen treatment
  40. Analytical evaluation of the Snibe β-isomerized C-terminal telopeptide of type I collagen (β-CTX-I) automated method
  41. Acute myeloid leukemia with blue-green neutrophilic inclusions have different outcomes: two cases and review of the literature
  42. Congress Abstracts
  43. The 10+1 Santorini Conference
  44. 14th National Congress of the Portuguese Society of Clinical Chemistry, Genetics and Laboratory Medicine
  45. 15th National Congress of the Portuguese Society of Clinical Chemistry, Genetics and Laboratory Medicine
  46. ISMD2024 Thirteenth International Symposium on Molecular Diagnostics
https://doi.org/10.1515/cclm-2023-1138
Keywords for this article
liquid chromatography-tandem mass spectrometry; harmonization; androstenedione; dehydroepiandrosterone sulfate; testosterone; inter-laboratory performance
Creative Commons
BY 4.0

Articles in the same Issue

  1. Frontmatter
  2. Editorial
  3. SARS-CoV-2 is here to stay: do not lower our guard
  4. Reviews
  5. SARS-CoV-2 subgenomic RNA: formation process and rapid molecular diagnostic methods
  6. Prognostic value of anti-SARS-CoV-2 antibodies: a systematic review
  7. Presence of SARS-CoV-2 RNA in COVID-19 survivors with post-COVID symptoms: a systematic review of the literature
  8. Opinion Papers
  9. Harmonizing the post-analytical phase: focus on the laboratory report
  10. Blood-based biomarkers in Alzheimer’s disease – moving towards a new era of diagnostics
  11. A comprehensive review on PFAS including survey results from the EFLM Member Societies
  12. General Clinical Chemistry and Laboratory Medicine
  13. Report from the HarmoSter study: different LC-MS/MS androstenedione, DHEAS and testosterone methods compare well; however, unifying calibration is a double-edged sword
  14. An LC–MS/MS method for serum cystatin C quantification and its comparison with two commercial immunoassays
  15. CX3CL1/Fractalkine as a biomarker for early pregnancy prediction of preterm premature rupture of membranes
  16. Elevated S100B urine levels predict seizures in infants complicated by perinatal asphyxia and undergoing therapeutic hypothermia
  17. The correlation of urea and creatinine concentrations in sweat and saliva with plasma during hemodialysis: an observational cohort study
  18. Tubular phosphate transport: a comparison between different methods of urine sample collection in FGF23-dependent hypophosphatemic syndromes
  19. Reference Values and Biological Variations
  20. Monocyte distribution width (MDW): study of reference values in blood donors
  21. Data mining of reference intervals for serum creatinine: an improvement in glomerular filtration rate estimating equations based on Q-values
  22. Hematology and Coagulation
  23. MALDI-MS in first-line screening of newborns for sickle cell disease: results from a prospective study in comparison to HPLC
  24. Cardiovascular Diseases
  25. To rule-in, or not to falsely rule-out, that is the question: evaluation of hs-cTnT EQA performance in light of the ESC-2020 guideline
  26. Temporal biomarker concentration patterns during the early course of acute coronary syndrome
  27. Diabetes
  28. Proteomic analysis of diabetic retinopathy identifies potential plasma-protein biomarkers for diagnosis and prognosis
  29. Infectious Diseases
  30. Serum biomarkers of inflammation and vascular damage upon SARS-Cov-2 mRNA vaccine in patients with thymic epithelial tumors
  31. A high throughput immuno-affinity mass spectrometry method for detection and quantitation of SARS-CoV-2 nucleoprotein in human saliva and its comparison with RT-PCR, RT-LAMP, and lateral flow rapid antigen test
  32. Evaluation of inflammatory biomarkers and vitamins in hospitalized patients with SARS-CoV-2 infection and post-COVID syndrome
  33. The CoLab score is associated with SARS-CoV-2 viral load during admission in individuals admitted to the intensive care unit: the CoLaIC cohort study
  34. Development and evaluation of a CRISPR-Cas13a system-based diagnostic for hepatitis E virus
  35. Letters to the Editor
  36. Crioplast® is a reliable device to ensure pre-analytical stability of adrenocorticotrophin (ACTH)
  37. Falsely decreased Abbott Alinity-c gamma-glutamyl transferase-2 result from paraprotein and heparin interference: case report and subsequent laboratory experiments
  38. Impact of hemolysis on uracilemia in the context of dihydropyrimidine dehydrogenase deficiency testing
  39. Value of plasma neurofilament light chain for monitoring efficacy in children with later-onset spinal muscular atrophy under nusinersen treatment
  40. Analytical evaluation of the Snibe β-isomerized C-terminal telopeptide of type I collagen (β-CTX-I) automated method
  41. Acute myeloid leukemia with blue-green neutrophilic inclusions have different outcomes: two cases and review of the literature
  42. Congress Abstracts
  43. The 10+1 Santorini Conference
  44. 14th National Congress of the Portuguese Society of Clinical Chemistry, Genetics and Laboratory Medicine
  45. 15th National Congress of the Portuguese Society of Clinical Chemistry, Genetics and Laboratory Medicine
  46. ISMD2024 Thirteenth International Symposium on Molecular Diagnostics
Sign up now to receive a 20% welcome discount
Institutional Access
How does access work?
Have an idea on how to improve our website?
Please write us.
© 2025 De Gruyter Brill
Downloaded on 11.9.2025 from https://www.degruyterbrill.com/document/doi/10.1515/cclm-2023-1138/html
Scroll to top button