CCA-13369; No of Pages 4 Clinica Chimica Acta xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Clinica Chimica Acta journal homepage: www.elsevier.com/locate/clinchim
The harmonisation of growth hormone measurements: Taking the next steps Gilbert E. Wieringa a,⁎, Catharine M. Sturgeon b, Peter J. Trainer c a b c
Department of Biochemistry, Bolton NHS Foundation Trust, Minerva Road, Farnworth, Bolton BL4 0JR, UK UK NEQAS [Edinburgh], Department of Laboratory Medicine, Royal Infirmary of Edinburgh, 51 Little France Crescent, Edinburgh EH16 4SA, UK Department of Endocrinology, The Christie NHS Foundation Trust, Wilmslow Road, Manchester M20 4BX, UK
a r t i c l e
i n f o
Article history: Received 6 September 2013 Received in revised form 8 January 2014 Accepted 8 January 2014 Available online xxxx Keywords: Growth hormone Method differences Harmonisation collaboratives
a b s t r a c t For over 20 years differences in results of growth hormone (GH) measurement have been recognised as being significant enough to lead to misdiagnosis and inappropriate management of patients with GH-related disorders. Whilst issues of method standardisation, variable antibody specificity, use of different reporting units with different conversion factors, and interference from GH binding protein have been acknowledged as contributing to the discrepancies, inconsistent approaches to method harmonisation have hampered opportunities to enhance the evidence base for GH measurements. Amongst the first steps to be taken, international collaboratives recommended the universal adoption of the International Standard 98/547 and the reporting of results in mass units. Whilst inter-method variability may have improved over the last 10 years, clinically significant differences remain. A more recently recognised issue contributing to the discrepancies may be the differences in the matrix materials used by kit manufacturers to assign values to their calibrants. The establishment of an international harmonisation oversight group is recommended: its key roles to include identification of a commutable matrix reference material, assessing the clinical significance of assay interferents, the evaluation of liquid chromatography– mass spectrometry as a reference measurement procedure and the provision of acceptance criteria for the clinical application of GH methods. © 2014 Elsevier B.V. All rights reserved.
1. Introduction The measurement of growth hormone (GH) has been the cornerstone to diagnosis and management of growth hormone related disorders for many years. In the United Kingdom, amongst its criteria for offering GH replacement to adults with GH deficiency, the National Institute for Health and Clinical Excellence (NICE) specifies inclusion of a peak growth hormone response of less than 9 mU/L during an insulin tolerance test or a similar low result in another reliable test [1]. In 2002, a mean integrated 24 hour GH level less than 2.5 μg/L and/or suppression of GH below 1 μg/L in an oral glucose tolerance test (2 mU/L) have been recommended as criteria for excluding a diagnosis of acromegaly [2]. However, it has long been recognised that the variability in GH results produced by commercial kits and in-house methods challenges the appropriateness of diagnostic criteria and affects patient outcomes by prejudicing access to appropriate management. [3]. Ellis et al. in reporting performance from the UK National External Quality Assessment Service (UK NEQAS) scheme for GH noted that the most positively biased method could typically report values twice that of
⁎ Corresponding author. Tel.: +44 1204 390172. E-mail address: [email protected] (G.E. Wieringa).
the most negatively biased [4]. Such differences were noted to cause 10% of laboratories to report the outcome of an insulin tolerance test as equivocal whilst 90% reported an adequate response. Arafat et al., in assessing the growth hormone suppression response during an oral glucose tolerance test, reported how results of one method were 2.3 times higher than those of a second and 6 times higher than those of a third [5]. In a retrospective study, Hauffa et al. reported that of 132 children who had been investigated for GH-related disorders, 36 would have been re-classified had the samples been measured by another method [6]. The differences are particularly relevant to the specialist endocrinologist to whom patients are referred with conflicting results from different centres which, given the low incidence of GH-related disorders (10 cases per million for isolated GH deficiency in children in the UK, 3 to 4 cases per million for acromegaly), is not an infrequent occurrence. The impact of the differences might not be so significant if method-specific diagnostic cut-offs were available. However, the evidence base for them also is poor, in part reflected by the low incidence of GH-related disorders but also affected by the limited collaboration between laboratories and their continuing willingness to transfer diagnostic cut-offs between methods despite the known biases [4,7]. These problems are compounded by the trend to apply international consensus criteria and guidelines to local practice without consideration, or even awareness, of the issues discussed here. (See Table 1.)
http://dx.doi.org/10.1016/j.cca.2014.01.014 0009-8981/© 2014 Elsevier B.V. All rights reserved.
Please cite this article as: Wieringa GE, et al, The harmonisation of growth hormone measurements: Taking the next steps, Clin Chim Acta (2014), http://dx.doi.org/10.1016/j.cca.2014.01.014
2
G.E. Wieringa et al. / Clinica Chimica Acta xxx (2014) xxx–xxx
Table 1 GH method performance reported by UK NEQAS before and after the recommended adoption of GH IS 98/574 and reporting of GH results in mass units (μg/L).
No. participants % labs. reporting GH in mass units (μg/L) GCV (%) across all methods VAR (%) for Siemens Immulite 2000 users No. participants using Siemens technology
2005
2006
2007
2008
2009
2010
2011
107 52% 21% 6.5% 69 (14/55)
109 68% 22% 6.1% 92 (14/78)
109 72% 15% 6.9% 94 (14/82)
112 79% 9.1% 7.2% 99 (13/86)
114 82% 15.5% 8.9% 99 (12/87)
109 85% 15.2% 7.3% 94 (12/82)
105 98% 14.8% 8.1% 87 (11/76)
GCV (%): Geometric coefficient of variation; VAR: Cumulative within method variability.
Because of the potential impact on patient care, the need to harmonise has been recognised for over 15 years [8–12]. Here we review the diversity of initiatives that have been undertaken, assess their contributions to a harmonised approach, and consider what further steps could be taken to improve methodologies for this heterogeneous polypeptide. 2. Why the discrepancies? The reasons for the discrepancies in GH results are numerous and have been well described. They include: • The use of calibrant materials whose values have been assigned in isolation of a defined international standard (IS). Until the availability of recombinant preparations, pituitary sourced IS 80/505 from the National Institute of Biological Standards and Control (NIBSC) with a nominally assigned bioactivity of 4.4 international units per ampoule was widely used [13]. Its advantage was that it reflected a physiological matrix that included GH moieties such as 22KDa and 20KDa monomers, dimers and hetero-isoforms thereof. Its main disadvantage was that the mass content was never defined although many users incorrectly used the approximate mass content (given only as a guide) as the assigned value. A further disadvantage was that an endless supply could not be maintained. As an IS it was widely used by kit manufacturers to then assign values to their (secondary) calibrators. The impact of individual practice in assigning those values was reflected in numerous reports highlighting the bias between methods albeit often in the face of correlation coefficients that could exceed 0.99. Since 1994, recombinant materials (IS 88/624 and its successor in 2001, IS 98/574) have been available [14]. IS 98/574 contains 22 kDa GH of N 95% purity with a defined specific activity of 3.0 IU/mg so that assay results could be reported in mass, molar or activity units. In practice, results continued to be expressed in mass or activity units, the detraction from molar units reflecting the molecular heterogeneity of GH in physiological fluids. A key factor favouring adoption of IS 98/547 is that it meets the requirement of EU Directive 98/79/EC (in vitro diagnostics) for values of commercial calibrators to be traceable to higher-order reference materials or methods, if available [15]. However, the reference material validity is also dependent on its integration in a traceability chain (further discussed below). Despite its ready availability from 2001 onwards, its adoption was variable and was often dictated by individual country and/or industry preference. One consequence has been to perpetuate not only intermethod differences but also to generate intra-method differences when manufacturers tailor a method to the calibration demands of their users. Thus, Meazza et al. reported how results could vary by 2 fold depending on whether the calibration of an otherwise identical kit had been based on IS 98/574 or IS 80/505 [16]. • The use of a plethora of conversion units between μg/L and mU/L. Pokrajac et al. assessed the impact of using 3 different conversion factors found in articles in one edition of a leading endocrine journal [17]. Adopting reported factors of 2.0, 2.6 and 3.0 for converting mU/L to μg/L across 14 GH methods participating in UKNEQAS, 11%, 55% and 100%, respectively, of submitted results would have been consistent with acromegaly if applied as a patient's GH nadir in an oral glucose tolerance test. Use of a variety of factors can also extend to conflict
amongst consensus statements when, for example, a conversion factor of 2.6 is used between μg/L and mU/L in the Growth Hormone Research Society's guideline for the diagnosis of GH deficiency [18] whilst a factor of 3.0 is used in the NICE guidance [1]. • Variable epitope specificity of antibodies used in commercial kits. Given the heterogeneity of GH, a key determinant of a reported value is the avidity and affinity of the method's antibody for the different GH moieties. The inter-individual physiological variation in the concentration of these moieties is thought to be considerable [19,20]. The clearest indication that antibody specificity impacts on variability has come from external quality assessment data during the transition from polyclonal (variably specific) to monoclonal (monospecific) antibodies in kit methods. From 1994 to 1998 the UK NEQAS-reported variability between methods increased from 17% to 30%, an increase which overlapped with the switch to the use of monoclonal antibodies by the scheme's participants. • The interference of GH binding protein (GHBP) which can show considerable inter- and intra-individual variability and which may complex up to 50% of GH to cause falsely low results depending on the antibody epitope specificity in the method being used [21,22]. Reports of the significance of GHBP interference stem largely from a time when polyclonal antibodies were in use in competitive immunoassays [23,24]. Data on GHBP interference in modern monoclonal antibody excess non-competitive immunoassays are more limited although interference leading to 40% reduction in values at low GH levels has been reported for one such method [25]. Decreases of 9.7%, 10.6% and 14.8% in GH levels have also been reported for 3 currently popular methods when spiked with 10 μg/L of GHBP and with interference reportedly above 30% in the presence of larger concentrations [26]. • The interference of pegvisomant, a GH competitor receptor antagonist licensed for the treatment of acromegaly which, depending on assay principle, has been reported to show positive interference (when both antibodies in a sandwich immunometric assay react), negative interference (when one of the antibodies in a sandwich immunometric assay reacts) [27] or, most recently, no interference in an assay using a monoclonal antibody that does not recognise pegvisomant [28]. The degree of interference has been investigated for some methods but deserves follow-up to quantitate the effects within and beyond pharmaceutical levels for all commercially available methods. 3. Taking one step at a time towards harmonisation In 2006 an International Growth Hormone Collaborative recommended that all manufacturers should adopt NIBSC's IS 98/574 and that mass units should be adopted as the reporting unit (1 mg corresponding to 3 international units somatropin), an initiative supported by 3 leading endocrine journals (Clinical Endocrinology, Growth Hormone & IGF Research, and European Journal of Endocrinology) that advised submitting authors in 2007 that only data reported in μg/L would be considered for publication [29]. In a second step, a 2011 consensus included recommendations that manufacturers should specify the degree of interference by GHBP in their methods, the identity and traceability to IS 98/574 of their calibrants, and the assay cross-reactivity characteristics [30]. The impact of the consensus statements on practice and
Please cite this article as: Wieringa GE, et al, The harmonisation of growth hormone measurements: Taking the next steps, Clin Chim Acta (2014), http://dx.doi.org/10.1016/j.cca.2014.01.014
G.E. Wieringa et al. / Clinica Chimica Acta xxx (2014) xxx–xxx
method performance is reflected in the annual UK NEQAS reports for GH in the period immediately before and after this scheme's 2007 recommendation to the diagnostics industry and laboratories to implement the changes. Fewer than 10% of manufacturers calibrated their methods against IS 98/574 in 2005 and only 52% of participants in the UK NEQAS scheme reported GH results in mass units. However by 2011 98% of participants reported results in mass units and almost all GH methods were calibrated against IS 98/574. [These figures are unlikely to change unless one method for which results are reported in mU/L of IS 80/505 is recalibrated.] Arguably the between method variability (reported here as GCV%) has reduced from approximately 21% to 15% but the significance of this is more difficult to determine when set against a background of increasing domination of Siemens-led methodology — 83% of all methods in 2011 compared to 65% in 2005. However, the within Siemens method group variability of approximately 6–9% in this period perhaps provides a measure of the optimum inter-laboratory performance that can be achieved within one technology. In tackling quantitative differences, Tanaka et al. observed correlation coefficients greater than 0.99 amongst 6 methods in use in Japan, and so factorised results to the mean of the most popular to reduce the inter-method variability from 35% to 18% [31]. Ross, in the Netherlands reduced inter-method variability from 25% to 15% by adjusting results to IS 88/624 [32], an IS that was superseded by IS 98/ 574. Müller et al., after excluding methods not calibrated to IS 98/574, calculated a coefficient of variation of 17.3% amongst methods used in Germany [33]. Whilst providing an indication of the impact of quantitative differences, the danger of such individual, empirical approaches to harmonisation is that they defy the unified approach that is required if the evidence base for the diagnosis and management of GH-related disorders is to be improved. Taking the above examples further, Tanaka et al. also reduced the diagnostic cut-off in Japan for provocative testing for GH deficiency from 10 μg/L to 6 μg/L — in the process by-passing the Port Stephens consensus guidelines for the diagnosis and management of growth hormone deficiency in adults [18]. Ross, although reporting a further reduction in between laboratory imprecision to 6.7% could be achieved when a ‘harmonisation serum’ was introduced provided little evidence of efficacy for its deployment [32]. In anticipating further reductions in variability, Müller et al. applied quantile transformation to their already adjusted results to reduce this to 11.4%. For Australasia, Davison has proposed that GH assays align their results with either those of the most popular assay or that an average value of the different standards in use could be used as the ‘standard’ [34]. The low incidence of GH-related disorders emphasises the need for consistent, systematic approaches for ensuring a universally supportable evidence base that in turn allows comparability of data published in international journals. 4. Commutability of results across commercial methods The challenge of reducing quantitative differences between methods may relate to the commutability of reference materials (kit calibrators) across methods [35]. A recent study carried out at the European Commission's Institute for Reference Materials and Measurements (IRRM) has shown that different matrices used for kit calibrator preparation can significantly affect the commutability to patients' results [26] (A reference material may be deemed to be commutable when the same numerical relationship can be demonstrated for 2 or more measurement procedures for both the reference material and a representative panel of patient samples [36]). In a series of experiments in which IS 98/574 was added to matrices that included phosphate buffered saline (PBS), PBS/bovine serum albumin (PBS/BSA), GH depleted serum, charcoal stripped serum, sheep serum and SRM 971 male serum, the recoveries ranged from b1% to 145% across 5 commercial methods whose results were described as traceable to IS 98/574. The authors also compared results of serum samples with the IS 98/574-spiked matrices. On comparing the different matrices with the 95% prediction interval for
3
the serum samples using the CLSI/IFCC recommended definition of commutability [36], none of the preparations was commutable across all methods although PBS/BSA spiked with GH-depleted serum was commutable for 8 of the 10 combinations of methods. The authors concluded that statements requiring “traceability to IS 98/574” were not sufficiently demanding; the expectation for acceptability is for every step of the calibration pathway to be traceable. Thus, caution was expressed about a) attempts to harmonise based on a matrix reference material that is not commutable across all methods and b) use of correction factors given the difficulty in then maintaining standardisation and avoiding drift in values. In overcoming individual approaches a recently issued “road map” for harmonising measurands for which there is no reference measurement procedure (RMP) but for which there is a secondary (valueassigned) reference material perhaps provides a template for a systematic approach for GH from a harmonisation oversight group [37]. The road map's recommended milestones outline a systematic approach to harmonisation as follows: • Establishing a clinical need for harmonisation • Assessing current analytical performance of available methodologies • Identifying commutable reference materials for calibration traceability • Isolating methods that cannot meet set harmonisation objectives • Taking opportunities to identify a designated RMP. The sometimes reverse chronology adopted by clinicians, scientists and the diagnostics industry with GH perhaps highlights the piecemeal achievements that have not served patients well. Although many of the jigsaw pieces of the milestones are now in place perhaps the key to progress is a harmonised approach amongst the relevant stakeholders. The IRRM's 2013 work in identifying (commutability) gaps in the traceability chain of GH measurements [26] potentially makes an important contribution to developing criteria for clinically acceptable methods. The need to establish a set of reference samples consisting of human serum pools containing low, medium and high GH concentrations as well as QC sera with documented commutability and with concentrations relevant to diagnostic cut-offs are important in assessing agreement among methods [38]. The potential of liquid chromatography– mass spectrometry as a candidate method for providing reference values is important in that it is able to distinguish 22 kDa from 20 kDa fractions but is not vulnerable to interference from GHBP or dependent on antibody avidity/affinity for quantification of GH moieties [39]. 5. Taking the next steps Despite its limitations in not being representative of human GH's mix of isoforms and moieties, IS 98/574's well-defined physical and chemical characteristics favour it as the value-assigned reference material that most likely meets the requirement of EU Directive 98/79/EC (in vitro diagnostics). Taking the harmonisation “road map” guidance from Miller et al. [37] together with recommendations from international collaboratives for the already ongoing adoption of IS 98/574, this should be the material against which all results are derived provided that this can be used in a valid calibration step within the traceability chain. If the first steps have included recommendations to adopt IS 98/ 574, a switch to mass units for results reporting and for manufacturers to specify the degree of interference by GHBP in their methods, the next steps for a harmonisation oversight group are several: • To ensure that secondary (kit) calibrator values can be produced using IS 98/574 in a valid calibration step within the traceability chain • To identify a commutable matrix reference material that should be used by manufacturers for the preparation of method calibrants • To ensure the reporting of results in mass units • To establish a set of reference samples consisting of human serum pools containing low, medium and high GH concentrations as well
Please cite this article as: Wieringa GE, et al, The harmonisation of growth hormone measurements: Taking the next steps, Clin Chim Acta (2014), http://dx.doi.org/10.1016/j.cca.2014.01.014
4
• • • • •
•
G.E. Wieringa et al. / Clinica Chimica Acta xxx (2014) xxx–xxx
as QC sera with documented commutability and with concentrations relevant to diagnostic cut-offs To assess the clinical significance of GHBP and pegvisomant interference in immunoassays To determine the avidity and affinity of different GH moieties for antibodies used in immunoassays To further evaluate the potential of liquid chromatography–mass spectrometry as a reference measurement procedure for GH To establish acceptance criteria for the clinical application of GH methods To appraise the validity of current consensus statements and guidelines for the diagnosis and management of growth hormone related disorders in the light of harmonised methodologies To raise awareness of the assay performance issues amongst the clinical and scientific community, and the diagnostics industry.
Whilst standardisation can be achieved when a reference measurement procedure and/or a primary (pure substance) reference material are available, this is currently the unachievable goal for GH. Efforts to improve the outcomes of patients with GH-related disorders rely on harmonising the evidence base for GH measurements. Whilst there is recognition of the need for stepwise progression [39–41], Clemmons has highlighted that cooperative efforts and education to endorse these efforts are required amongst professional organisations, manufacturers, proficiency testing providers, and key opinion leaders [30]. A new engagement leap is required amongst all stakeholders if further progress is to be achieved. References [1] NICE Clinical Guideline TA64. Growth hormone deficiency (adults) - human growth hormone (TA64). See http://guidance.nice.org.uk/TA64. [Last accessed 8th January 2014]. [2] Giustina A, Barkan A, Casanueva FF, et al. Criteria for cure of acromegaly: a consensus statement. J Clin Endocrinol Metab 2000;85:526–9. [3] Strasburger CJ, Bidlingmaier M. How robust are laboratory measures of growth hormone status? Horm Res 2008;64(Suppl. 2):1–5. [4] Ellis A, Seth J, Al-Sadie R, Barth JH. An audit of the laboratory interpretation of growth hormone response to insulin-induced hypoglycaemia in the assessment of short stature in children. Ann Clin Biochem 2003;40:239–43. [5] Arafat AM, Mohlig M, Weickert MO, et al. Growth hormone response during oral glucose tolerance test: the impact of assay method on the estimation of reference values in patients with acromegaly and in healthy controls, and the role of gender, age, and body mass index. J Clin Endocrinol Metab 2008;93:1254–62. [6] Hauffa BP, Lehmann N, Bettendorf M, et al. Central laboratory reassessment of IGF-I, IGF-binding protein-3, and GH serum concentrations measured at local treatment centers in growth-impaired children: implications for the agreement between outpatient screening and the results of somatotropic axis functional testing. Eur J Endocrinol 2007;157:597–603. [7] Andersson AM, Orskov H, Ranke MB, Shalet S, Skakkebaek NE. Interpretation of growth hormone provocative tests: comparison of cut-off values in four European laboratories. Eur J Endocrinol 1995;132:340–3. [8] Bidlingmaier M, Freda PU. Measurement of human growth hormone by immunoassays: current status, unsolved problems and clinical consequences. Growth Horm IGF Res 2010;20:19–25. [9] Bidlingmaier M, Strasburger CJ. Growth hormone assays: current methodologies and their limitations. Pituitary 2007;10:115–9. [10] Seth J, Ellis A, Al-Sadie R. Serum growth hormone measurements in clinical practice: an audit of performance from the UK National External Quality Assessment scheme. Horm Res 1999;51(Suppl. 1):13–9. [11] Wieringa GE, Barth JH, Trainer PJ. Growth hormone assay standardization: a biased view? Clin Endocrinol 2004;60:538–9. [12] Wood P. Growth hormone: its measurement and the need for assay harmonization. Ann Clin Biochem 2001;38:471–82.
[13] Bangham DR, Gaines Das RE, Schulster D. The International Standard for Human Growth Hormone for Bioassay: calibration and characterization by international collaborative study. Mol Cell Endocrinol 1985;42:269–82. [14] Bristow AF, Jespersen AM. The Second International Standard for somatropin (recombinant DNA-derived human growth hormone): preparation and calibration in an international collaborative study. Biologicals 2001;29:97–106. [15] Directive 98/79/EC of the European Parliament and of the Council of 27 October 1998 on in vitro diagnostic medical devices. At http://ec.europa.eu/enterprise/ policies/european-standards/harmonised-standards/iv-diagnostic-medical-devices/. [Last accesssed on 30th August 2013]. [16] Meazza C, Albertini R, Pagani S, et al. Diagnosis of growth hormone deficiency is affected by calibrators used in GH immunoassays. Horm Metab Res 2012;44:900–3. [17] Pokrajac A, Wark G, Ellis AR, Wear J, Wieringa GE, Trainer PJ. Variation in GH and IGF-I assays limits the applicability of international consensus criteria to local practice. Clin Endocrinol 2007;67:65–70. [18] Consensus guidelines for the diagnosis and treatment of adults with growth hormone deficiency: summary statement of the Growth Hormone Research Society Workshop on Adult Growth Hormone Deficiency. J Clin Endocrinol Metab 1998;83: 379–81. [19] Baumann G. Growth hormone heterogeneity: genes, isohormones, variants, and binding proteins. Endocr Rev 1991;12:424–49. [20] Celniker AC, Chen AB, Wert Jr RM, Sherman BM. Variability in the quantitation of circulating growth hormone using commercial immunoassays. J Clin Endocrinol Metab 1989;68:469–76. [21] Amit T, Youdim MB, Hochberg Z. Clinical review 112: Does serum growth hormone (GH) binding protein reflect human GH receptor function? J Clin Endocrinol Metab 2000;85:927–32. [22] Baumann G. Growth hormone binding protein 2001. J Pediatr Endocrinol Metab 2001;14:355–75. [23] Ebdrup L, Fisker S, Sorensen HH, Ranke MB, Orskov H. Variety in growth hormone determinations due to use of different immunoassays and to the interference of growth hormone-binding protein. Horm Res 1999;51(Suppl. 1):20–6. [24] Jansson C, Boguszewski C, Rosberg S, Carlsson L, Albertsson-Wikland K. Growth hormone (GH) assays: influence of standard preparations, GH isoforms, assay characteristics, and GH-binding protein. Clin Chem 1997;43:950–6. [25] Hansen TK, Fisker S, Hansen B, et al. Impact of GHBP interference on estimates of GH and GH pharmacokinetics. Clin Endocrinol 2002;57:779–86. [26] Boulo S, Hanisch K, Bidlingmaier M, et al. Gaps in the traceability chain of human growth hormone measurements. Clin Chem 2013;59:1074–82. [27] Paisley AN, Hayden K, Ellis A, Anderson J, Wieringa G, Trainer PJ. Pegvisomant interference in GH assays results in underestimation of GH levels. Eur J Endocrinol 2007;156:315–9. [28] Manolopoulou J, Alami Y, Petersenn S, et al. Automated 22-kD growth hormonespecific assay without interference from Pegvisomant. Clin Chem 2012;58: 1446–56. [29] Trainer PJ, Barth J, Sturgeon C, Wieringa G. Consensus statement on the standardisation of GH assays. Eur J Endocrinol 2006;155:1–2. [30] Clemmons DR. Consensus statement on the standardization and evaluation of growth hormone and insulin-like growth factor assays. Clin Chem 2011;57:555–9. [31] Tanaka T, Tachibana K, Shimatsu A, et al. A nationwide attempt to standardize growth hormone assays. Horm Res 2005;64(Suppl. 2):6–11. [32] Ross HA. Reporting growth hormone assay results in terms of one consensus recombinant standard preparation offers less than optimal reduction of between-method variation. Clin Chem Lab Med 2008;46:1334–5. [33] Müller A, Scholz M, Blankenstein O, et al. Harmonization of growth hormone measurements with different immunoassays by data adjustment. Clin Chem Lab Med 2011;49:1135–42. [34] Davidson J. Harmonisation of growth hormone assays in Australasia. Clin Biochem Rev 2012;33:101–2. [35] Miller WG, Myers GL, Rej R. Why commutability matters. Clin Chem 2006;52:553–4. [36] (C53-A) Characterization and qualification of commutable reference materials for laboratory medicine; approved guideline. At http://www.ifcc.org/ifcc-communicationspublications-division-%28cpd%29/ifcc-publications/clsi-ifcc-joint-projects/ . [Last accessed 8th January 2014]. [37] Miller WG, Myers GL, Gantzer ML, et al. Roadmap for harmonization of clinical laboratory measurement procedures. Clin Chem 2011;57:1108–17. [38] Clemmons DR, Bruns DE. Clin Chem 2011;57:1463–4. [39] Arsene CG, Henrion A, Diekmann N, Manolopoulou J, Bidlingmaier M. Quantification of growth hormone in serum by isotope dilution mass spectrometry. Anal Biochem 2010;401:228–35. [40] Sheppard MC. Growth hormone assay standardization: an important clinical advance. Clin Endocrinol 2007;66:157–61. [41] Strasburger CJ. Taking one step at a time. Clin Endocrinol 2004;60:540.
Please cite this article as: Wieringa GE, et al, The harmonisation of growth hormone measurements: Taking the next steps, Clin Chim Acta (2014), http://dx.doi.org/10.1016/j.cca.2014.01.014
Contents lists available at ScienceDirect
Clinica Chimica Acta journal homepage: www.elsevier.com/locate/clinchim
The harmonisation of growth hormone measurements: Taking the next steps Gilbert E. Wieringa a,⁎, Catharine M. Sturgeon b, Peter J. Trainer c a b c
Department of Biochemistry, Bolton NHS Foundation Trust, Minerva Road, Farnworth, Bolton BL4 0JR, UK UK NEQAS [Edinburgh], Department of Laboratory Medicine, Royal Infirmary of Edinburgh, 51 Little France Crescent, Edinburgh EH16 4SA, UK Department of Endocrinology, The Christie NHS Foundation Trust, Wilmslow Road, Manchester M20 4BX, UK
a r t i c l e
i n f o
Article history: Received 6 September 2013 Received in revised form 8 January 2014 Accepted 8 January 2014 Available online xxxx Keywords: Growth hormone Method differences Harmonisation collaboratives
a b s t r a c t For over 20 years differences in results of growth hormone (GH) measurement have been recognised as being significant enough to lead to misdiagnosis and inappropriate management of patients with GH-related disorders. Whilst issues of method standardisation, variable antibody specificity, use of different reporting units with different conversion factors, and interference from GH binding protein have been acknowledged as contributing to the discrepancies, inconsistent approaches to method harmonisation have hampered opportunities to enhance the evidence base for GH measurements. Amongst the first steps to be taken, international collaboratives recommended the universal adoption of the International Standard 98/547 and the reporting of results in mass units. Whilst inter-method variability may have improved over the last 10 years, clinically significant differences remain. A more recently recognised issue contributing to the discrepancies may be the differences in the matrix materials used by kit manufacturers to assign values to their calibrants. The establishment of an international harmonisation oversight group is recommended: its key roles to include identification of a commutable matrix reference material, assessing the clinical significance of assay interferents, the evaluation of liquid chromatography– mass spectrometry as a reference measurement procedure and the provision of acceptance criteria for the clinical application of GH methods. © 2014 Elsevier B.V. All rights reserved.
1. Introduction The measurement of growth hormone (GH) has been the cornerstone to diagnosis and management of growth hormone related disorders for many years. In the United Kingdom, amongst its criteria for offering GH replacement to adults with GH deficiency, the National Institute for Health and Clinical Excellence (NICE) specifies inclusion of a peak growth hormone response of less than 9 mU/L during an insulin tolerance test or a similar low result in another reliable test [1]. In 2002, a mean integrated 24 hour GH level less than 2.5 μg/L and/or suppression of GH below 1 μg/L in an oral glucose tolerance test (2 mU/L) have been recommended as criteria for excluding a diagnosis of acromegaly [2]. However, it has long been recognised that the variability in GH results produced by commercial kits and in-house methods challenges the appropriateness of diagnostic criteria and affects patient outcomes by prejudicing access to appropriate management. [3]. Ellis et al. in reporting performance from the UK National External Quality Assessment Service (UK NEQAS) scheme for GH noted that the most positively biased method could typically report values twice that of
⁎ Corresponding author. Tel.: +44 1204 390172. E-mail address: [email protected] (G.E. Wieringa).
the most negatively biased [4]. Such differences were noted to cause 10% of laboratories to report the outcome of an insulin tolerance test as equivocal whilst 90% reported an adequate response. Arafat et al., in assessing the growth hormone suppression response during an oral glucose tolerance test, reported how results of one method were 2.3 times higher than those of a second and 6 times higher than those of a third [5]. In a retrospective study, Hauffa et al. reported that of 132 children who had been investigated for GH-related disorders, 36 would have been re-classified had the samples been measured by another method [6]. The differences are particularly relevant to the specialist endocrinologist to whom patients are referred with conflicting results from different centres which, given the low incidence of GH-related disorders (10 cases per million for isolated GH deficiency in children in the UK, 3 to 4 cases per million for acromegaly), is not an infrequent occurrence. The impact of the differences might not be so significant if method-specific diagnostic cut-offs were available. However, the evidence base for them also is poor, in part reflected by the low incidence of GH-related disorders but also affected by the limited collaboration between laboratories and their continuing willingness to transfer diagnostic cut-offs between methods despite the known biases [4,7]. These problems are compounded by the trend to apply international consensus criteria and guidelines to local practice without consideration, or even awareness, of the issues discussed here. (See Table 1.)
http://dx.doi.org/10.1016/j.cca.2014.01.014 0009-8981/© 2014 Elsevier B.V. All rights reserved.
Please cite this article as: Wieringa GE, et al, The harmonisation of growth hormone measurements: Taking the next steps, Clin Chim Acta (2014), http://dx.doi.org/10.1016/j.cca.2014.01.014
2
G.E. Wieringa et al. / Clinica Chimica Acta xxx (2014) xxx–xxx
Table 1 GH method performance reported by UK NEQAS before and after the recommended adoption of GH IS 98/574 and reporting of GH results in mass units (μg/L).
No. participants % labs. reporting GH in mass units (μg/L) GCV (%) across all methods VAR (%) for Siemens Immulite 2000 users No. participants using Siemens technology
2005
2006
2007
2008
2009
2010
2011
107 52% 21% 6.5% 69 (14/55)
109 68% 22% 6.1% 92 (14/78)
109 72% 15% 6.9% 94 (14/82)
112 79% 9.1% 7.2% 99 (13/86)
114 82% 15.5% 8.9% 99 (12/87)
109 85% 15.2% 7.3% 94 (12/82)
105 98% 14.8% 8.1% 87 (11/76)
GCV (%): Geometric coefficient of variation; VAR: Cumulative within method variability.
Because of the potential impact on patient care, the need to harmonise has been recognised for over 15 years [8–12]. Here we review the diversity of initiatives that have been undertaken, assess their contributions to a harmonised approach, and consider what further steps could be taken to improve methodologies for this heterogeneous polypeptide. 2. Why the discrepancies? The reasons for the discrepancies in GH results are numerous and have been well described. They include: • The use of calibrant materials whose values have been assigned in isolation of a defined international standard (IS). Until the availability of recombinant preparations, pituitary sourced IS 80/505 from the National Institute of Biological Standards and Control (NIBSC) with a nominally assigned bioactivity of 4.4 international units per ampoule was widely used [13]. Its advantage was that it reflected a physiological matrix that included GH moieties such as 22KDa and 20KDa monomers, dimers and hetero-isoforms thereof. Its main disadvantage was that the mass content was never defined although many users incorrectly used the approximate mass content (given only as a guide) as the assigned value. A further disadvantage was that an endless supply could not be maintained. As an IS it was widely used by kit manufacturers to then assign values to their (secondary) calibrators. The impact of individual practice in assigning those values was reflected in numerous reports highlighting the bias between methods albeit often in the face of correlation coefficients that could exceed 0.99. Since 1994, recombinant materials (IS 88/624 and its successor in 2001, IS 98/574) have been available [14]. IS 98/574 contains 22 kDa GH of N 95% purity with a defined specific activity of 3.0 IU/mg so that assay results could be reported in mass, molar or activity units. In practice, results continued to be expressed in mass or activity units, the detraction from molar units reflecting the molecular heterogeneity of GH in physiological fluids. A key factor favouring adoption of IS 98/547 is that it meets the requirement of EU Directive 98/79/EC (in vitro diagnostics) for values of commercial calibrators to be traceable to higher-order reference materials or methods, if available [15]. However, the reference material validity is also dependent on its integration in a traceability chain (further discussed below). Despite its ready availability from 2001 onwards, its adoption was variable and was often dictated by individual country and/or industry preference. One consequence has been to perpetuate not only intermethod differences but also to generate intra-method differences when manufacturers tailor a method to the calibration demands of their users. Thus, Meazza et al. reported how results could vary by 2 fold depending on whether the calibration of an otherwise identical kit had been based on IS 98/574 or IS 80/505 [16]. • The use of a plethora of conversion units between μg/L and mU/L. Pokrajac et al. assessed the impact of using 3 different conversion factors found in articles in one edition of a leading endocrine journal [17]. Adopting reported factors of 2.0, 2.6 and 3.0 for converting mU/L to μg/L across 14 GH methods participating in UKNEQAS, 11%, 55% and 100%, respectively, of submitted results would have been consistent with acromegaly if applied as a patient's GH nadir in an oral glucose tolerance test. Use of a variety of factors can also extend to conflict
amongst consensus statements when, for example, a conversion factor of 2.6 is used between μg/L and mU/L in the Growth Hormone Research Society's guideline for the diagnosis of GH deficiency [18] whilst a factor of 3.0 is used in the NICE guidance [1]. • Variable epitope specificity of antibodies used in commercial kits. Given the heterogeneity of GH, a key determinant of a reported value is the avidity and affinity of the method's antibody for the different GH moieties. The inter-individual physiological variation in the concentration of these moieties is thought to be considerable [19,20]. The clearest indication that antibody specificity impacts on variability has come from external quality assessment data during the transition from polyclonal (variably specific) to monoclonal (monospecific) antibodies in kit methods. From 1994 to 1998 the UK NEQAS-reported variability between methods increased from 17% to 30%, an increase which overlapped with the switch to the use of monoclonal antibodies by the scheme's participants. • The interference of GH binding protein (GHBP) which can show considerable inter- and intra-individual variability and which may complex up to 50% of GH to cause falsely low results depending on the antibody epitope specificity in the method being used [21,22]. Reports of the significance of GHBP interference stem largely from a time when polyclonal antibodies were in use in competitive immunoassays [23,24]. Data on GHBP interference in modern monoclonal antibody excess non-competitive immunoassays are more limited although interference leading to 40% reduction in values at low GH levels has been reported for one such method [25]. Decreases of 9.7%, 10.6% and 14.8% in GH levels have also been reported for 3 currently popular methods when spiked with 10 μg/L of GHBP and with interference reportedly above 30% in the presence of larger concentrations [26]. • The interference of pegvisomant, a GH competitor receptor antagonist licensed for the treatment of acromegaly which, depending on assay principle, has been reported to show positive interference (when both antibodies in a sandwich immunometric assay react), negative interference (when one of the antibodies in a sandwich immunometric assay reacts) [27] or, most recently, no interference in an assay using a monoclonal antibody that does not recognise pegvisomant [28]. The degree of interference has been investigated for some methods but deserves follow-up to quantitate the effects within and beyond pharmaceutical levels for all commercially available methods. 3. Taking one step at a time towards harmonisation In 2006 an International Growth Hormone Collaborative recommended that all manufacturers should adopt NIBSC's IS 98/574 and that mass units should be adopted as the reporting unit (1 mg corresponding to 3 international units somatropin), an initiative supported by 3 leading endocrine journals (Clinical Endocrinology, Growth Hormone & IGF Research, and European Journal of Endocrinology) that advised submitting authors in 2007 that only data reported in μg/L would be considered for publication [29]. In a second step, a 2011 consensus included recommendations that manufacturers should specify the degree of interference by GHBP in their methods, the identity and traceability to IS 98/574 of their calibrants, and the assay cross-reactivity characteristics [30]. The impact of the consensus statements on practice and
Please cite this article as: Wieringa GE, et al, The harmonisation of growth hormone measurements: Taking the next steps, Clin Chim Acta (2014), http://dx.doi.org/10.1016/j.cca.2014.01.014
G.E. Wieringa et al. / Clinica Chimica Acta xxx (2014) xxx–xxx
method performance is reflected in the annual UK NEQAS reports for GH in the period immediately before and after this scheme's 2007 recommendation to the diagnostics industry and laboratories to implement the changes. Fewer than 10% of manufacturers calibrated their methods against IS 98/574 in 2005 and only 52% of participants in the UK NEQAS scheme reported GH results in mass units. However by 2011 98% of participants reported results in mass units and almost all GH methods were calibrated against IS 98/574. [These figures are unlikely to change unless one method for which results are reported in mU/L of IS 80/505 is recalibrated.] Arguably the between method variability (reported here as GCV%) has reduced from approximately 21% to 15% but the significance of this is more difficult to determine when set against a background of increasing domination of Siemens-led methodology — 83% of all methods in 2011 compared to 65% in 2005. However, the within Siemens method group variability of approximately 6–9% in this period perhaps provides a measure of the optimum inter-laboratory performance that can be achieved within one technology. In tackling quantitative differences, Tanaka et al. observed correlation coefficients greater than 0.99 amongst 6 methods in use in Japan, and so factorised results to the mean of the most popular to reduce the inter-method variability from 35% to 18% [31]. Ross, in the Netherlands reduced inter-method variability from 25% to 15% by adjusting results to IS 88/624 [32], an IS that was superseded by IS 98/ 574. Müller et al., after excluding methods not calibrated to IS 98/574, calculated a coefficient of variation of 17.3% amongst methods used in Germany [33]. Whilst providing an indication of the impact of quantitative differences, the danger of such individual, empirical approaches to harmonisation is that they defy the unified approach that is required if the evidence base for the diagnosis and management of GH-related disorders is to be improved. Taking the above examples further, Tanaka et al. also reduced the diagnostic cut-off in Japan for provocative testing for GH deficiency from 10 μg/L to 6 μg/L — in the process by-passing the Port Stephens consensus guidelines for the diagnosis and management of growth hormone deficiency in adults [18]. Ross, although reporting a further reduction in between laboratory imprecision to 6.7% could be achieved when a ‘harmonisation serum’ was introduced provided little evidence of efficacy for its deployment [32]. In anticipating further reductions in variability, Müller et al. applied quantile transformation to their already adjusted results to reduce this to 11.4%. For Australasia, Davison has proposed that GH assays align their results with either those of the most popular assay or that an average value of the different standards in use could be used as the ‘standard’ [34]. The low incidence of GH-related disorders emphasises the need for consistent, systematic approaches for ensuring a universally supportable evidence base that in turn allows comparability of data published in international journals. 4. Commutability of results across commercial methods The challenge of reducing quantitative differences between methods may relate to the commutability of reference materials (kit calibrators) across methods [35]. A recent study carried out at the European Commission's Institute for Reference Materials and Measurements (IRRM) has shown that different matrices used for kit calibrator preparation can significantly affect the commutability to patients' results [26] (A reference material may be deemed to be commutable when the same numerical relationship can be demonstrated for 2 or more measurement procedures for both the reference material and a representative panel of patient samples [36]). In a series of experiments in which IS 98/574 was added to matrices that included phosphate buffered saline (PBS), PBS/bovine serum albumin (PBS/BSA), GH depleted serum, charcoal stripped serum, sheep serum and SRM 971 male serum, the recoveries ranged from b1% to 145% across 5 commercial methods whose results were described as traceable to IS 98/574. The authors also compared results of serum samples with the IS 98/574-spiked matrices. On comparing the different matrices with the 95% prediction interval for
3
the serum samples using the CLSI/IFCC recommended definition of commutability [36], none of the preparations was commutable across all methods although PBS/BSA spiked with GH-depleted serum was commutable for 8 of the 10 combinations of methods. The authors concluded that statements requiring “traceability to IS 98/574” were not sufficiently demanding; the expectation for acceptability is for every step of the calibration pathway to be traceable. Thus, caution was expressed about a) attempts to harmonise based on a matrix reference material that is not commutable across all methods and b) use of correction factors given the difficulty in then maintaining standardisation and avoiding drift in values. In overcoming individual approaches a recently issued “road map” for harmonising measurands for which there is no reference measurement procedure (RMP) but for which there is a secondary (valueassigned) reference material perhaps provides a template for a systematic approach for GH from a harmonisation oversight group [37]. The road map's recommended milestones outline a systematic approach to harmonisation as follows: • Establishing a clinical need for harmonisation • Assessing current analytical performance of available methodologies • Identifying commutable reference materials for calibration traceability • Isolating methods that cannot meet set harmonisation objectives • Taking opportunities to identify a designated RMP. The sometimes reverse chronology adopted by clinicians, scientists and the diagnostics industry with GH perhaps highlights the piecemeal achievements that have not served patients well. Although many of the jigsaw pieces of the milestones are now in place perhaps the key to progress is a harmonised approach amongst the relevant stakeholders. The IRRM's 2013 work in identifying (commutability) gaps in the traceability chain of GH measurements [26] potentially makes an important contribution to developing criteria for clinically acceptable methods. The need to establish a set of reference samples consisting of human serum pools containing low, medium and high GH concentrations as well as QC sera with documented commutability and with concentrations relevant to diagnostic cut-offs are important in assessing agreement among methods [38]. The potential of liquid chromatography– mass spectrometry as a candidate method for providing reference values is important in that it is able to distinguish 22 kDa from 20 kDa fractions but is not vulnerable to interference from GHBP or dependent on antibody avidity/affinity for quantification of GH moieties [39]. 5. Taking the next steps Despite its limitations in not being representative of human GH's mix of isoforms and moieties, IS 98/574's well-defined physical and chemical characteristics favour it as the value-assigned reference material that most likely meets the requirement of EU Directive 98/79/EC (in vitro diagnostics). Taking the harmonisation “road map” guidance from Miller et al. [37] together with recommendations from international collaboratives for the already ongoing adoption of IS 98/574, this should be the material against which all results are derived provided that this can be used in a valid calibration step within the traceability chain. If the first steps have included recommendations to adopt IS 98/ 574, a switch to mass units for results reporting and for manufacturers to specify the degree of interference by GHBP in their methods, the next steps for a harmonisation oversight group are several: • To ensure that secondary (kit) calibrator values can be produced using IS 98/574 in a valid calibration step within the traceability chain • To identify a commutable matrix reference material that should be used by manufacturers for the preparation of method calibrants • To ensure the reporting of results in mass units • To establish a set of reference samples consisting of human serum pools containing low, medium and high GH concentrations as well
Please cite this article as: Wieringa GE, et al, The harmonisation of growth hormone measurements: Taking the next steps, Clin Chim Acta (2014), http://dx.doi.org/10.1016/j.cca.2014.01.014
4
• • • • •
•
G.E. Wieringa et al. / Clinica Chimica Acta xxx (2014) xxx–xxx
as QC sera with documented commutability and with concentrations relevant to diagnostic cut-offs To assess the clinical significance of GHBP and pegvisomant interference in immunoassays To determine the avidity and affinity of different GH moieties for antibodies used in immunoassays To further evaluate the potential of liquid chromatography–mass spectrometry as a reference measurement procedure for GH To establish acceptance criteria for the clinical application of GH methods To appraise the validity of current consensus statements and guidelines for the diagnosis and management of growth hormone related disorders in the light of harmonised methodologies To raise awareness of the assay performance issues amongst the clinical and scientific community, and the diagnostics industry.
Whilst standardisation can be achieved when a reference measurement procedure and/or a primary (pure substance) reference material are available, this is currently the unachievable goal for GH. Efforts to improve the outcomes of patients with GH-related disorders rely on harmonising the evidence base for GH measurements. Whilst there is recognition of the need for stepwise progression [39–41], Clemmons has highlighted that cooperative efforts and education to endorse these efforts are required amongst professional organisations, manufacturers, proficiency testing providers, and key opinion leaders [30]. A new engagement leap is required amongst all stakeholders if further progress is to be achieved. References [1] NICE Clinical Guideline TA64. Growth hormone deficiency (adults) - human growth hormone (TA64). See http://guidance.nice.org.uk/TA64. [Last accessed 8th January 2014]. [2] Giustina A, Barkan A, Casanueva FF, et al. Criteria for cure of acromegaly: a consensus statement. J Clin Endocrinol Metab 2000;85:526–9. [3] Strasburger CJ, Bidlingmaier M. How robust are laboratory measures of growth hormone status? Horm Res 2008;64(Suppl. 2):1–5. [4] Ellis A, Seth J, Al-Sadie R, Barth JH. An audit of the laboratory interpretation of growth hormone response to insulin-induced hypoglycaemia in the assessment of short stature in children. Ann Clin Biochem 2003;40:239–43. [5] Arafat AM, Mohlig M, Weickert MO, et al. Growth hormone response during oral glucose tolerance test: the impact of assay method on the estimation of reference values in patients with acromegaly and in healthy controls, and the role of gender, age, and body mass index. J Clin Endocrinol Metab 2008;93:1254–62. [6] Hauffa BP, Lehmann N, Bettendorf M, et al. Central laboratory reassessment of IGF-I, IGF-binding protein-3, and GH serum concentrations measured at local treatment centers in growth-impaired children: implications for the agreement between outpatient screening and the results of somatotropic axis functional testing. Eur J Endocrinol 2007;157:597–603. [7] Andersson AM, Orskov H, Ranke MB, Shalet S, Skakkebaek NE. Interpretation of growth hormone provocative tests: comparison of cut-off values in four European laboratories. Eur J Endocrinol 1995;132:340–3. [8] Bidlingmaier M, Freda PU. Measurement of human growth hormone by immunoassays: current status, unsolved problems and clinical consequences. Growth Horm IGF Res 2010;20:19–25. [9] Bidlingmaier M, Strasburger CJ. Growth hormone assays: current methodologies and their limitations. Pituitary 2007;10:115–9. [10] Seth J, Ellis A, Al-Sadie R. Serum growth hormone measurements in clinical practice: an audit of performance from the UK National External Quality Assessment scheme. Horm Res 1999;51(Suppl. 1):13–9. [11] Wieringa GE, Barth JH, Trainer PJ. Growth hormone assay standardization: a biased view? Clin Endocrinol 2004;60:538–9. [12] Wood P. Growth hormone: its measurement and the need for assay harmonization. Ann Clin Biochem 2001;38:471–82.
[13] Bangham DR, Gaines Das RE, Schulster D. The International Standard for Human Growth Hormone for Bioassay: calibration and characterization by international collaborative study. Mol Cell Endocrinol 1985;42:269–82. [14] Bristow AF, Jespersen AM. The Second International Standard for somatropin (recombinant DNA-derived human growth hormone): preparation and calibration in an international collaborative study. Biologicals 2001;29:97–106. [15] Directive 98/79/EC of the European Parliament and of the Council of 27 October 1998 on in vitro diagnostic medical devices. At http://ec.europa.eu/enterprise/ policies/european-standards/harmonised-standards/iv-diagnostic-medical-devices/. [Last accesssed on 30th August 2013]. [16] Meazza C, Albertini R, Pagani S, et al. Diagnosis of growth hormone deficiency is affected by calibrators used in GH immunoassays. Horm Metab Res 2012;44:900–3. [17] Pokrajac A, Wark G, Ellis AR, Wear J, Wieringa GE, Trainer PJ. Variation in GH and IGF-I assays limits the applicability of international consensus criteria to local practice. Clin Endocrinol 2007;67:65–70. [18] Consensus guidelines for the diagnosis and treatment of adults with growth hormone deficiency: summary statement of the Growth Hormone Research Society Workshop on Adult Growth Hormone Deficiency. J Clin Endocrinol Metab 1998;83: 379–81. [19] Baumann G. Growth hormone heterogeneity: genes, isohormones, variants, and binding proteins. Endocr Rev 1991;12:424–49. [20] Celniker AC, Chen AB, Wert Jr RM, Sherman BM. Variability in the quantitation of circulating growth hormone using commercial immunoassays. J Clin Endocrinol Metab 1989;68:469–76. [21] Amit T, Youdim MB, Hochberg Z. Clinical review 112: Does serum growth hormone (GH) binding protein reflect human GH receptor function? J Clin Endocrinol Metab 2000;85:927–32. [22] Baumann G. Growth hormone binding protein 2001. J Pediatr Endocrinol Metab 2001;14:355–75. [23] Ebdrup L, Fisker S, Sorensen HH, Ranke MB, Orskov H. Variety in growth hormone determinations due to use of different immunoassays and to the interference of growth hormone-binding protein. Horm Res 1999;51(Suppl. 1):20–6. [24] Jansson C, Boguszewski C, Rosberg S, Carlsson L, Albertsson-Wikland K. Growth hormone (GH) assays: influence of standard preparations, GH isoforms, assay characteristics, and GH-binding protein. Clin Chem 1997;43:950–6. [25] Hansen TK, Fisker S, Hansen B, et al. Impact of GHBP interference on estimates of GH and GH pharmacokinetics. Clin Endocrinol 2002;57:779–86. [26] Boulo S, Hanisch K, Bidlingmaier M, et al. Gaps in the traceability chain of human growth hormone measurements. Clin Chem 2013;59:1074–82. [27] Paisley AN, Hayden K, Ellis A, Anderson J, Wieringa G, Trainer PJ. Pegvisomant interference in GH assays results in underestimation of GH levels. Eur J Endocrinol 2007;156:315–9. [28] Manolopoulou J, Alami Y, Petersenn S, et al. Automated 22-kD growth hormonespecific assay without interference from Pegvisomant. Clin Chem 2012;58: 1446–56. [29] Trainer PJ, Barth J, Sturgeon C, Wieringa G. Consensus statement on the standardisation of GH assays. Eur J Endocrinol 2006;155:1–2. [30] Clemmons DR. Consensus statement on the standardization and evaluation of growth hormone and insulin-like growth factor assays. Clin Chem 2011;57:555–9. [31] Tanaka T, Tachibana K, Shimatsu A, et al. A nationwide attempt to standardize growth hormone assays. Horm Res 2005;64(Suppl. 2):6–11. [32] Ross HA. Reporting growth hormone assay results in terms of one consensus recombinant standard preparation offers less than optimal reduction of between-method variation. Clin Chem Lab Med 2008;46:1334–5. [33] Müller A, Scholz M, Blankenstein O, et al. Harmonization of growth hormone measurements with different immunoassays by data adjustment. Clin Chem Lab Med 2011;49:1135–42. [34] Davidson J. Harmonisation of growth hormone assays in Australasia. Clin Biochem Rev 2012;33:101–2. [35] Miller WG, Myers GL, Rej R. Why commutability matters. Clin Chem 2006;52:553–4. [36] (C53-A) Characterization and qualification of commutable reference materials for laboratory medicine; approved guideline. At http://www.ifcc.org/ifcc-communicationspublications-division-%28cpd%29/ifcc-publications/clsi-ifcc-joint-projects/ . [Last accessed 8th January 2014]. [37] Miller WG, Myers GL, Gantzer ML, et al. Roadmap for harmonization of clinical laboratory measurement procedures. Clin Chem 2011;57:1108–17. [38] Clemmons DR, Bruns DE. Clin Chem 2011;57:1463–4. [39] Arsene CG, Henrion A, Diekmann N, Manolopoulou J, Bidlingmaier M. Quantification of growth hormone in serum by isotope dilution mass spectrometry. Anal Biochem 2010;401:228–35. [40] Sheppard MC. Growth hormone assay standardization: an important clinical advance. Clin Endocrinol 2007;66:157–61. [41] Strasburger CJ. Taking one step at a time. Clin Endocrinol 2004;60:540.
Please cite this article as: Wieringa GE, et al, The harmonisation of growth hormone measurements: Taking the next steps, Clin Chim Acta (2014), http://dx.doi.org/10.1016/j.cca.2014.01.014
