Isotopic abundance yields bias in the assessment of testosterone in a new reference method procedure.
We read with interest the recent publication by Botelho and colleagues describing a candidate reference method procedure for measuring serum testosterone to support the CDC's Hormone Standardization (HoSt) Program (1). The aims of this program are to standardize serum testosterone measurements across laboratories and platforms and to support the generation of age- and sex-specific reference intervals for testosterone. After careful consideration, we have identified a potential error in the assay that could affect the success of the program. Specifically, the final testosterone concentration is determined by multiplying the response ratio of the analyte peak to the internal-standard peak by the concentration of the internal standard. By design, the response ratio is near 1.0, which has previously been demonstrated to reduce the imprecision of measurement (2).
The error in this approach lies in the use ofa testosterone internal standard labeled with the [sup.13][C.sub.3] stable isotope, which will lead to 2 important problems for the HoSt program. First, the reference intervals for testosterone will be inaccurate, and second, HoSt participants who use gravimetric preparation techniques for their calibration materials will observe that the program produces a bias from the expected results.
Mass spectrometry typically measures the dominant monoisotopic species of the analyte to be measured. Testosterone ([C.sub.19] [H.sub.28][O.sub.2]) has a monoisotopic mass of 288.209 with an expected protonated molecular ion [[M+H].sup.+] of m/z 289.2. Because 1.1% of the carbon is [sup.13]C, a mass spectrometer measuring the dominant monoisotopic mass of testosterone would measure approximately 77% of the molecule (at m/z 289.2), with approximately 21% occurring as the + 1 isotope (m/z 290.2) and approximately 2% as the +2 isotope (m/z 291.2) (3).
By enriching the internal standard with 3 [sup.13]C atoms, the internal standard has an altered proportion of the dominant monoisotopic species (m/z 292.2). More specifically, [[sup.13][C.sub.3]]testosterone consists of 80% dominant monoisotopic ion (m/z 292.2). Approximately 18% of the molecule would be measured as m/z 293.2 and approximately 2% would be measured as m/z 294.2.
Direct comparison of the peak areas of unlabeled testosterone and [[sup.13][C.sub.3]]testosterone introduced at the same concentration would yield a higher response for [[sup.13][C.sub.3]] testosterone owing to the discrepancy between monoisotopic abundances, thereby yielding a 3% underestimation of the circulating testosterone. For example, consider a sample containing an unknown amount of testosterone that is supplemented with approximately 100 nmol/L [[sup.13][C.sub.3]]testosterone. Analysis yields peak areas for the unlabeled analyte and the labeled internal standard that are exactly equivalent. Using the new reference method procedure's calculations would show the sample to contain 100 nmol/L endogenous testosterone. Applying the appropriate correction factor for isotopic abundance would yield an accurate testosterone concentration for the sample (103.1 nmol/L). Not correcting for isotopic abundance would underestimate the true testosterone concentration in samples measured with the reference method and thus overestimate the concentration (relative to the reference method) for laboratories that use gravimetry and isotope dilution with calibration curves for measurement.
The use of gravimetric mass measurements rather than molar measurements throughout the reference method partly corrects for a portion of the discrepancy. The report by Botelho et al. incorrectly describes the molarity of the internal standard solution preparation for calibration as calculated from nanograms per deciliter (i.e., 35 ng/dL [sup.13][C.sub.3]]testosterone is 1.20 nmol/L of the internal standard, erroneously stated as 1.21 nmol/L). This calculation error overestimates concentrations by 1.0%. The resulting bias in the reported result would be approximately 2%, or one third of the allowable inaccuracy.
A recent update to the performance criteria for testosterone measurement in support of the Partnership for the Accurate Testing of Hormones states that reducing bias errors is necessary to better diagnose patients when results are compared with a population-based reference interval (4). The shift in accurate concentrations due to the isotopic variation described has not been updated in reference intervals for many laboratories that participate in the HoSt program. Typically, assays that have a 2% bias in measurement accuracy would be of minimal consequence. The criteria, however, for the mean bias in the CDC HoSt program for testosterone is 6.4%; a 2% shift in accuracy will confound successful participation in the program.
Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data;(b) drafting or revising the article for intellectual content; and (c) final approval of the published article.
Authors' Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form. Disclosures and/or potential conflicts of interest:
Employment or Leadership: None declared. Consultant or Advisory Role: None declared.
Stock Ownership: B.A. Rappold, Laboratory Corporation of America and Quest Diagnostics; R.P. Grant, Laboratory Corporation of America.
Honoraria: None declared.
Research Funding: None declared.
Expert Testimony: None declared.
Patents: None declared.
(1). Botelho JC, Shacklady C, Cooper HC, Tai SSC, Uytfanghe KV, Thienpoint LM, Vesper HW. Isotope-dilution liquid chromatography/tandem mass spectrometry candidate reference method for total testosterone in human serum. Clin Chem 2013;59:372-80.
(2). Sargent M, Harte R, Harrington C. Guidelines for achieving high accuracy in isotope dilution mass spectrometry (IDMS). London: Royal Society of Chemistry; 2002.
(3). Rockwood AL, Haimi P. Efficient calculation of accurate masses of isotopic peaks. J Am Soc Mass Spec 2006;17:415-19.
(4). Yun YM, Botelho JC, Chandler DW, Katayev A, Roberts WL, Stanczyk FZ, et al. Performance criteria for testosterone measurements based on biological variation in adult males: recommendations from the Partnership for the Accurate Testing of Hormones. Clin Chem 2012;58:1703-10.
Brian A. Rappold  *
Russell P. Grant 
 Essential Testing, LLC Collinsville, IL
 Laboratory Corporation of America Holdings Burlington, NC
* Address correspondence to this author at: Essential Testing Laboratories 1616 Eastport Plaza Dr. Collinsville, IL 62234 Fax 618-623-0492 E-mail email@example.com
Previously published online at DOI: 10.1373/clinchem.2013.204206
Reference measurement procedures (RMPs) (1) are intended primarily to assign target values to reference materials that function as calibrators or trueness controls for routine measurement procedures (referred to as "routine assays" here). RMPs must be highly accurate and precise to keep the uncertainty of the target values small. This rigor minimizes the increase in measurement uncertainty occurring during the calibration of a routine assay and enables reliable determination of potential measurement bias. To achieve this high level of accuracy, RMPs commonly use special quantification procedures that are different from those used with routine assays. Because of these differences, certain considerations relevant to routine assays may not be applicable to RMPs. Not recognizing these differences can lead to incorrect conclusions about the measurement accuracy of RMPs.
In their letter, Rappold and Grant (1) point out a potential source of measurement bias, which can be referred to as bias due to "isotopic enhancement." This bias is different from the more commonly known bias caused by so-called spectral overlap. The RMP described in our article (2) is not affected by either of these sources of bias.
Bias due to isotopic enhancement is related to the naturally occurring carbon isotope [sup.13]C, which constitutes 1.1% of carbon and enhances the mass spectrometry response of the [sup.13]C-labeled internal standard (IS). Therefore, a solution of [sup.13]C-labeled IS will produce a higher instrument response for the monoisotopic peak than an equimolar, non-[sup.13]C-labeled analyte solution. This phenomenon can lead to a slight bias if the mass spectrometric signals from the IS are directly used for quantification without the use of a calibration curve. The RMP we described does not directly use the mass spectrometry signal from the IS for value assignment, as inferred by Rappold and Grant, but uses a bracketing technique involving calibration with standards prepared gravimetrically.
The bracketing procedure described for our assay (2) uses a 2-step process. First, we obtain a 1:1 match of mass concentration between the analyte and the IS and then bracket calibrators around this 1:1 mass ratio to create a calibration curve. Isotopic enhancement in the mass spectrometry signal of the IS is observed in the calibrators to the same extent as in the unknowns and therefore does not affect the quantification of testosterone in the unknown samples. Bracketing and creating calibration curves with brackets is a common technique used by other testosterone RMPs, such as those of NIST (3) and Ghent University (4).
Spectral overlap due to natural occurring isotopes in the analyte and/or insufficient labeling of the IS is another potential source of bias that is absent from our RMP because of our use of bracketing involving the calibration curves described for our assay. Our RMP is not significantly biased by spectral overlap or isotopic enhancement. This feature is reflected in the highly linear calibrator brackets ([r.sup.2] > 0.999), excellent agreement with no statistically significant difference between the RMPs of NIST and Ghent University, and, finally, excellent matching of the certified values for the NIST reference materials with a bias of <0.3%. The NIST RMP is not significantly affected by isotopic enhancement because it uses a deuterated IS (3) and the naturally frequency of deuterium is only 0.0156%.
We would like to clarify that all calibrator and IS solutions described in our article have mass concentrations (in nanograms per gram). Gravimetric measurements of testosterone, [[sup.13][C.sub.3]]testosterone, and serum were used to prepare calibrators and samples. Volume concentrations (in nmoles per liter and nanograms per liter) were presented to ease readability and were calculated by using solvent density and molecular weight. Therefore, the molarity reported in the article does not affect the final calculated concentration reported by our RMP and does not introduce a bias as suggested by Rappold and Grant.
Our RMP for testosterone is highly accurate and suitable for assigning target values to materials used as calibrators and trueness controls. It has been reviewed by the Joint Committee for Traceability in Laboratory Medicine (JCTLM) and is now listed as an RMP in the JCTLM database.
Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article.
Authors' Disclosures or Potential Conflicts of Interest: No authors declared any potential conflicts of interest.
(1.) Rappold BA, Grant RP. Isotopic abundance yields bias in the assessment of testosterone in a new reference measurement procedure. Clin Chem 2013;59:1129-30.
(2.) Botelho JC, Shacklady C, Cooper HC, Tai SSC, Van Uytfanghe K, Thienpont LM, Vesper HW. Isotope-dilution liquid chromatography/tandem mass spectrometry candidate reference method for total testosterone in human serum. Clin Chem 2013;59:372-80.
(3.) Tai SSC, Xu B, Welch MJ, Phinney KW. Development and evaluation of a candidate reference measurement procedure in the determination of testosterone in human serum using isotope dilution liquid chromatography/tandem mass spectrometry. Anal Bioanal Chem 2007;388: 1087-94.
(4.) Thienpont LM, Van Nieuwenhove B, Stockl D, De Leenheer A. Calibration for isotope dilution mass spectrometry - description of an alternative to the bracketing procedure. J Mass Spectrom 1996;31:1119-25.
Hubert W. Vesper  *
Julianne Cook Botelho 
Susan S.-C. Tai 
Katleen Van Uytfanghe 
Linda M. Thienpont 
 Nonstandard abbreviations: RMP, reference measurement procedure; IS, internal standard; JCTLM, Joint Committee for Traceability in Laboratory Medicine.
 Division of Laboratory Sciences National Center for Environmental Health Centers for Disease Control and Prevention Atlanta, GA
 Analytical Chemistry Division National Institute of Standards and Technology Gaithersburg, MD
 Laboratory for Analytical Chemistry Faculty of Pharmaceutical Sciences Ghent University Ghent, Belgium
* Address correspondence to this author at: Centers for Disease Control and Prevention 4770 Buford Hwy. NE, MS-F25 Atlanta, GA 30341 Fax 770-488-7030 E-mail firstname.lastname@example.org
Previously published online at DOI: 10.1373/clinchem.2013.205856
|Printer friendly Cite/link Email Feedback|
|Author:||Rappold, Brian A.; Grant, Russell P.|
|Date:||Jul 1, 2013|
|Previous Article:||Falsely increased hCG in patients with high leukocyte counts.|
|Next Article:||It's not that crystal clear.|