Automated 22-kD Growth Hormone Specific Assay without Interference from Pegvisomant

Transcription

1 Clinical Chemistry 58: (2012) Endocrinology and Metabolism Automated 22-kD Growth Hormone Specific Assay without Interference from Pegvisomant Jenny Manolopoulou, 1 Younes Alami, 2 Stephan Petersenn, 3 Jochen Schopohl, 1 Zida Wu, 4 Christian J. Strasburger, 4 and Martin Bidlingmaier 1* BACKGROUND: Large variability exists among different growth hormone (GH) assays owing to differences in calibration, antibody specificity, isoform recognition, and interference from GH binding protein (GHBP). The GH receptor antagonist Pegvisomant presents a new challenge because Pegvisomant interferes with many GH assays. A recent consensus conference established criteria for standardization and evaluation of GH assays. Following consensus recommendations, we developed a new GH assay on an automated analyzer (IDS-iSYS, Immunodiagnostic Systems). METHODS: A monoclonal antibody not cross-reacting with Pegvisomant was combined with a monoclonal antibody specific for 22-kD GH. Isoform specificity and interference from GHBP was tested and compared to that seen in 2 existing automated GH assays (Siemens Immulite, Diasorin Liaison). We also compared GH concentrations measured by the 3 assays for healthy volunteers and patients with acromegaly receiving different treatments. Using the isys assay, we also established nadir GH values during oral glucose load and analyzed changes in endogenous GH during Pegvisomant treatment. RESULTS: Analytical and functional sensitivities were 0.01 g/l and 0.04 g/l, with a dynamic range from 0.04 to 100 g/l. Intraassay CVs were 2 4, whereas interassay CVs were 5 7 at GH concentrations between 1.7 and 27.5 g/l. The assay was specific for 22-kD GH and not affected by GHBP. The presence of Pegvisomant, which leads to a negative bias on the Immulite and dramatic overestimation of GH on the Liaison, had no impact on the isys GH assay. CONCLUSIONS: The new assay fulfils recent consensus recommendations and presents a useful new tool for reliable measurement of GH American Association for Clinical Chemistry Accurate determination of circulating growth hormone (GH) 5 concentrations is crucial in the diagnosis and monitoring of acromegaly and GH deficiency (1). Several factors, however, affect the recognition of GH in immunoassays and lead to inconsistencies between results from different assays (2). The nature of the GH molecule is a confounding factor in itself. Various isoforms exist in addition to the main 22-kD isoform, and heterodimers, homodimers, and multimers also occur (3). Depending on the epitopes recognized by the antibodies used, GH assays will measure only a certain spectrum of GH molecular isoforms. This explains why comparability of assay results worsened when assays based on polyclonal antisera were replaced by assays utilizing highly specific monoclonal antibodies (mabs) (4, 5). The question of measurement of solely 22-kD or a spectrum of isoforms is still under debate, although choosing 22-kD GH a defined analyte that also is most clinically relevant may present the best compromise to harmonize GH measurements (6, 7). The molecular heterogeneity is also reflected by the different standard preparations used to calibrate GH assays. Older pituitary-derived human GH (hgh) preparations contain a wide variety of isoforms, whereas the newer, recombinant WHO International Standards [Recombinant DNA-Derived Human Growth Hormone, NIBSC (National Institute for Biological Standards) codes 88/624 and 98/574] consist purely of 22-kD GH (8). The latter calibrators provide the advantage of unlimited supply and definition of 1 Endocrine Research Laboratories, Medizinische Klinik und Poliklinik IV, Ludwig Maximilians University, Munich, Germany; 2 IDS Ltd, Liège; Belgium; 3 ENDOC Center for Endocrine Tumors, Hamburg, and the Faculty of Medicine, University of Duisburg Essen, Essen, Germany; 4 Department of Medicine for Endocrinology, Diabetes and Nutritional Medicine, Charité Universitätsmedizin, Campus Mitte, Berlin, Germany. * Address correspondence to this author at: Medizinische Klinik und Poliklinik IV, Klinikum der Ludwig-Maximilians-University, Ziemssenstrasse 1, Munich, Germany. Fax ; martin.bidlingmaier@ med.uni-muenchen.de. Received April 13, 2012; accepted July 31, Previously published online at DOI: /clinchem Nonstandard abbreviations: GH, growth hormone; GHBP, GH binding protein; IDS, Immunodiagnostic Systems; mabs, monoclonal antibodies; hgh, human GH; LOD, Limits of detection; LOQ, limits of quantification; IGF-1, insulin-like growth factor 1; OGTT, oral glucose tolerance test. 1446

2 A 22-kD GH-Specific Automated GH Assay content in mass units. The use of a single recombinant preparation in different GH assays has been demonstrated to greatly reduce the variability among assays (9). Formal demonstration of commutability is still lacking, but because monomeric 22 kd is the only GH form available in sufficient purity and quantity, its general adoption for assay calibration is recommended (1, 10). In circulation, approximately 50 of GH rapidly becomes associated with GH binding protein (GHBP) (5, 11). This presents a further source of variability between assays because GHBP can cover epitopes required for GH antibody interaction (12). It has been demonstrated that GHBP can induce a negative bias of up to 50 (13, 14), and the short incubation times employed in current automated assay setups might aggravate the problem. One important clinical application of GH measurements is the diagnosis and monitoring of GHsecreting pituitary adenomas (acromegaly). The introduction of the GH analog and receptor antagonist Pegvisomant (15) represents a new challenge for GH assays. Pegvisomant differs from wild-type GH by only 9 amino acid mutations. The high structural homology combined with 1000-fold higher concentrations required to achieve adequate antagonism lead to its interference in many conventional assays (16). Analysis of endogenous GH concentrations during Pegvisomant therapy therefore requires specific research assays (17 20). We report the development and the technical and initial clinical validation of a new GH assay for use on an automated system (IDS-iSYS, Immunodiagnostic Systems). The assay was designed to specifically measure 22-kD GH with no interference from Pegvisomant or GHBP. We achieved this goal by selecting highaffinity mabs after careful epitope mapping. In line with recent recommendations, the assay was calibrated with International Standard 98/574. Results obtained with the new isys GH assay in samples from healthy individuals and untreated and treated patients with acromegaly were compared to those from 2 other widely used automated methods (Siemens Immulite 2000; Diasorin Liaison). We also analyzed samples obtained from healthy individuals and patients with acromegaly undergoing oral glucose load with the scope to assess nadir values. Finally, using the new isys assay, we investigated changes in endogenous GH during treatment with Pegvisomant. Materials and Methods ANTIBODIES, ASSAY SETUP, AND CALIBRATORS Two mabs were chosen from a panel of antibodies on the basis of their affinity and specificity for 22-kD GH. The capture antibody (6C1) targeted an epitope inside receptor binding site 2 of the GH molecule, where the amino acid at position 120 was mutated in Pegvisomant. After a first incubation any Pegvisomant remaining in the sample was removed through a washing step consisting of 3 cycles with 500 L of IDS-iSYS Wash Solution (PBS containing detergent, Proclin 300 and 0.1 sodium azide). The detection mab (GH7) targeted an epitope in the loop connecting helix 1 and 2 of GH, which was missing in 20-kD GH, thereby conferring specificity of the assay for the 22-kD GH molecules. The method to identify the antigenic epitopes has been described previously and molecular models showing the proposed epitopes are presented in Figs. 3 and 4 in the Data Supplement that accompanies the online version of this article at clinchem.org/content/vol58/issue10 (21). Details of antibody conjugation, purification and the assay protocol are described in the online Supplemental Methods. A master curve ranging from 0.05 to 100 g/l of the International Standard code 98/574 (National Institute for Biological Standards and Control, Hertfordshire, UK) was prepared in heat-inactivated horse serum (Gibco Invitrogen). ASSAY VALIDATION AND CHARACTERIZATION Validation of the assay was performed according to the CLSI recommendations (22, 23). Limits of detection (LOD), limits of quantification (LOQ), imprecision, linearity, and recovery were determined. Details are given in the online Supplemental Data files. INTERFERENCE AND CROSS-REACTIVITY Interference was assessed by combining a serum pool free of the interferent (0 pool), and a pool containing a large concentration (100 pool). The low and high pools were mixed to create interferent concentrations of 25, 50, and 75. Samples were allowed to come to equilibrium with the interfering substances overnight at 4 C before measurement. Interference was tested for recombinant human GHBP (Protein Laboratories; Rehovot), biotin, triglycerides, and bilirubin (all from Sigma), hemoglobin (MP Biomed), red blood cells (Alba Bioscience), albumin solution from human serum (SCIPAC), human antimouse antibody (Hama 1, Hama 2; Roche), and rheumatoid factor (SCIPAC). Recombinant GHBP interference was tested more extensively in 4 samples containing low, medium, and high endogenous GH concentrations. Furthermore, to also asses the influence of endogenous GHBP, recovery of 5 g/l recombinant GH (98/574) supplemented into samples (n 41) with known GH ( g/l) Clinical Chemistry 58:10 (2012) 1447

3 and GHBP concentrations ( nmol/l) was investigated. Cross-reactivity of 20-kD GH (Protein Laboratories; Rehovot), glycolsylated placental GH (24), Pegvisomant (Somavert, Pfizer Pharma), placental lactogen (code 73/545, NIBSC), and prolactin (Sigma) were analyzed on the isys in samples containing low and high endogenous GH concentrations. Cross-reactivity of 20-kD GH and interference from Pegvisomant and GHBP were also investigated in the Siemens Immulite and Diasorin Liaison. OTHER ASSAYS Two existing automated GH assays were used for comparison studies (Immulite 2000 hgh, Siemens Medical Solutions Diagnostics and Liaison hgh, Diasorin). For both assays, in our hands the LOQ was 0.05 g/l. According to the manufacturer s data the LODs were 0.01 g/l (Immulite) and 0.05 g/l (Liaison). Interassay CVs were 6.6 (Immulite) and 9.4 (Liaison) at concentrations between 0.3 and 17.0 g/l. All measurements on the Immulite were carried out with the recently recalibrated lots using IS 98/574, which is also the calibrator for the Liaison. GHBP was measured by a modification of a ligand immunofunctional assay (25, 26). Intraassay CVs were 9.4 and 6.1 at 115 pmol/l and 1550 pmol/l, and interassay CVs were 8.5 and 10.9, respectively. The LOQ was 69 pmol/l, with a linear range between 69 and 4000 pmol/l. Insulin-like growth factor 1 (IGF-I) was measured with the Immulite 2000 (Siemens Medical Solutions Diagnostics). According to the manufacturer the analytical sensitivity of the method is 20 g/l with intraand interassay CVs at 3.9 and 8.1 at IGF-I concentrations between 77 and 1358 g/l. PATIENT SAMPLES Blood samples were collected with S-Monovette (Sarstedt) with separator gel, centrifuged at 2500g, and the serum was separated and stored at 20 C until use. Samples were from male and female healthy volunteers and patients who were not taking any relevant medication (n 275) and compared to samples from Pegvisomant-treated patients with acromegaly (n 110). GH response to an oral glucose tolerance test (OGTT; 75 g glucose after overnight fast) was assessed in healthy volunteers (n 25, 18 females, 7 males), treatment-naïve patients with acromegaly (n 21, 8 females, 13 males) and patients treated with Pegvisomant (n 10, 3 females, 7 males) or somatostatin analogs (n 12, 6 females, 6 males). In 9 further patients with acromegaly (4 females, 5 males), GH (isys) and IGF-I (Immulite) concentrations Table 1. Assay characteristics and specificity for potentially cross-reacting agents. were measured before and after a mean of 9.4 months (range 1 18 months) treatment with Pegvisomant. All participants gave written informed consent. The studies were designed in agreement with the Declaration of Helsinki and approved by the Ethics Committee of the Medical Faculty of the Ludwig- Maximilians University, Munich. STATISTICAL METHODS Statistical analysis was carried out with StatView (Version 5, SAS Institute) and GraphPad Prism (Version 4.00, GraphPad Software). EP Evaluator (Data Innovations Europe) was used for analysis of the assay validation studies. Grouped data were expressed as mean (SD) or SE and median (range) as specified. GH concentrations during OGTT were compared by Wilcoxon matched-pairs test, and by Mann Whitney U for the unpaired group comparisons. For calculations, GH concentrations below the LOQ of the respective assays were arbitrarily set to the LOQ value (n 5 for the Immulite, n 4 for the isys). Overall differences between groups were assessed using the Kruskal Wallis nonparametric test followed by the Dunns Multiple Comparison posttest. Results given by the isys, Immulite, and Liaison instruments were compared with Deming regression analysis to obtain linear equations, S y x, and P values, and linear regression analysis was used to obtain R 2 values. Results Linear range ng/ml Analytical sensitivity 0.01 ng/ml Functional sensitivity 0.04 ng/ml Intraassay imprecision Interassay imprecision Interday imprecision Top dose tested Interference Pegvisomant ng/ml kD hgh 10 ng/ml 4 Placental GH 200 ng/ml 0.5 Human placental lactogen 10 g/ml 0.04 Prolactin ng/ml ASSAY VALIDATION AND CHARACTERIZATION On 6 different isys systems the LOD was 0.01 g/l with an mean of g/l (see online Supplemental Table 1), and measurements during 21 days on 2 dif Clinical Chemistry 58:10 (2012)

4 A 22-kD GH-Specific Automated GH Assay ferent instruments established the LOQ at 0.04 g/l (see online Supplemental Table 2a, b). On 3 different instruments the intraassay imprecision varied from 1.1 to 3.4, interassay imprecision from 0.9 to 3.8, and interday imprecision from 2.7 to 6.1 in samples containing 1.8, 10.9, and 27.5 g/l (see online Supplemental Table 3). Results from 50 samples analyzed on 2 different instruments in different locations showed excellent correlation [y 0.963x 0.008, r (95 CI: )]. Linearity was demonstrated in samples containing endogenous GH concentrations 20 g/l and in samples supplemented to high GH concentrations (100 g/l) (see online Supplemental Table 4a, b). Recovery ranged from 100 to 110 at concentrations of g/l (see online Supplemental Table 5a, b). INTERFERENCE No interference was observed for biotin, triglycerides, hemoglobin, bilirubin, red blood cells, human serum albumin, human antimouse antibody, and rheumatoid factor (see online Supplemental Table 6). When increasing concentrations of recombinant GHBP were supplemented into samples of low GH [n 4, mean (SD) 0.86 (0.76) g/l], medium GH [8.16 (2.18) g/l], and high GH [18.31 (9.33) g/l], recoveries were , , and (see online Supplemental Table 7). With the Immulite assay, in the same samples recoveries were , , and At the maximum added concentration (5000 pmol/l, mimicking high endogenous GHBP concentrations) recovery was about 10 lower on the Immulite compared to the isys (see online Supplemental Fig. 1). Samples of known endogenous GHBP content (n 41; range: pmol/l) were analyzed on the isys before and after addition of recombinant GH. Recovery of added GH did not change across the range of GHBP concentrations (P 0.743, r , S y x 0.3; also see online Supplemental Fig. 2). CROSS-REACTIVITY Recombinant 20-kD GH up to 10 g/l added to samples with low and high total GH concentrations did not affect results in the isys assay (Table 2), whereas there was a large positive bias on the Liaison (106 and 14, respectively). The highest cross-reaction with 20-kD GH was observed on the Immulite, for which the addition of 10 g/l introduced an overestimation of 1047 in the low and 51 in the high GH sample. Cross-reactivity with Pegvisomant (up to g/l) was negligible on the isys (Table 3). In contrast, there was a consistent negative bias of about 50 on the Immulite. The strongest interference was observed with the Liaison assay. Addition of moderate Pegvisomant concentrations led to a dramatic overestimation of GH, whereas addition of very high concentrations led to a dramatic underestimation (Table 3). CORRELATION OF IMMULITE VS ISYS In samples from healthy volunteers, Deming linear regression analysis gave an excellent correlation between the methods [isys (0.85 Immulite) 0.007, R ]. Bland Altman plots (27) confirmed the somewhat lower concentrations measured on the isys (up to1or2 g/l) in the untreated group (Fig. 1B). Because of the negative interference from Pegvisomant in the Immulite assay, samples from patients on Pegvisomant give approximately 2-fold higher GH concentrations on the isys [isys (2.20 Immulite) 0.035, R ]. As shown in the Bland Altman plot in Fig. 1C, the negative interference from Pegvisomant gave rise to a large increase in bias that was greater with increasing mean concentrations of GH. CORRELATION OF LIAISON VS isys In samples from healthy participants and patients not on Pegvisomant, the correlation between results from the 2 instruments was excellent [isys (1.04 Liaison) 0.014, R ]. The good agreement was also shown by the Bland Altman analysis. As expected, the high cross-reaction with Pegvisomant on the Liaison led to spurious GH concentrations in samples from treated patients with no correlation with the isys [y (0.0007), R ]. NADIR GH DURING OGTTS In healthy participants, baseline GH values were higher in females than in males in both the isys [males, median (range), 0.05 ( ) g/l, females 1.09 ( ) g/l; P ] and the Immulite [males 0.11 ( ) g/l, females 1.31 ( ) g/l; P 0.001]. Postglucose GH nadirs (120 min) were not significantly different between sexes [isys: males 0.04( ) g/l, females 0.07 ( ) g/l; Immulite: males 0.08 ( ) g/l, females 0.11 ( ) g/l]. However, after we excluded 2 study participants who exhibited the highest GH nadir values (1 male, 1 female), males tended to have lower GH nadir values compared to females [isys males (mean (SD), 0.05 g/l (0.02), maximum 0.09; females 0.11 g/l (0.09), maximum 0.34]. Combined data for males and females are shown in Fig. 2A. As expected, no nadir value was observed in the treatment-naïve patients with acromegaly (Fig. 2D), with similar mean GH concentrations in males and females for both the isys and the Immulite. In somatostatin-treated patients (Fig. 2C), a slight decrease in GH values was observed with measure- Clinical Chemistry 58:10 (2012) 1449

5 Table 2. Interference from the 20-kD GH isoform tested in a low and high sample on 3 different GH assays (observed and expected concentrations of GH in ng/ml). 20-kD, isys a 10 Low sample High sample Native kD, Liaison 10 Low sample High sample Native kD, Immulite 10 Low sample High sample Native a, observed/expected. ments from both instruments [isys: males 0.68 g/l ( ), females 1.05 g/l ( ); Immulite: males 0.75 g/l ( ), females 1.18 g/l ( )]. In the subset of Pegvisomant-treated patients (Fig. 2B), baseline values were higher on the isys [4.04 g/l ( ) compared to the Immulite (2.24 g/l ( )], which was due to the negative interference from Pegvisomant in the latter assay. A decrease in endogenous GH was observed with the isys (P 0.047) compared to baseline, but not in the GH concentrations measured with the Immulite (P 0.144). GH AND IGF-I RESPONSE TO PEGVISOMANT TREATMENT In 9 patients with acromegaly, endogenous GH concentrations as measured by the isys increase during treatment [before, mean (95 CI): 4.32 ( ) g/l; after, mean 7.60 ( ) g/l]. IGF-I concentrations decreased from 501 ( ) g/l to 261 ( ) g/l (Fig. 3). Discussion We present a new automated immunometric GH assay with unique specificity. The assay combines a capture antibody targeting an epitope on binding site 2 of GH where amino acid 120 is mutated in the drug Pegvisomant (15) and a detection antibody targeting an epitope that is deleted in 20-kD GH. The new isys GH assay thereby allows specific measurement of 22-kD GH concentrations with no interference from Pegvisomant. Following recent recommendations (1) we calibrated the assay using IS 98/574. Our results demonstrated excellent analytical sensitivity (LOD 0.01 g/l, LOQ 0.04 g/l), good linearity and recovery, and a range up to 100 g/l Clinical Chemistry 58:10 (2012)

6 A 22-kD GH-Specific Automated GH Assay Table 3. Interference from Pegvisomant (Pegv) tested in a low-, medium-, and high-gh concentration samples in the 3 different assays (observed and expected concentrations of GH in ng/ml). xaln;rowisys Pegv, a Low sample Medium sample High sample Native Liaison Pegv, Low sample Medium sample High sample Native Immulite Pegv, Low sample Medium sample High sample Native a, observed/expected. Clinical Chemistry 58:10 (2012) 1451

7 Fig. 1. Correlation to existing GH assays. Immulite vs isys: (A), Deming linear regression correlation in healthy individuals and other patients not treated with Pegvisomant and patients on Pegvisomant treatment. (B), Bland Altman analysis of the results from samples of healthy individuals and other patients not treated with Pegvisomant. (C), Bland Altman analysis of the results from samples of patients on Pegvisomant treatment. Liaison vs isys: (D), Deming linear regression correlation in healthy individuals and other patients not treated with Pegvisomant and patients on Pegvisomant treatment; (E) Bland Altman analysis of the results from samples of healthy individuals and other patients not treated with Pegvisomant. The adoption of a uniform calibrator consisting of only the 22-kD isoform of GH in combination with antibodies detecting only the 22-kD isoform has repeatedly been proposed as a step toward more consistent interlaboratory measurements of GH (7, 10). Our studies demonstrate that, in contrast to the isys assay, the Immulite GH assay shows strong cross-reactivity with 20-kD GH, whereas the Liaison GH assay shows a moderate recognition of 20-kD GH. This finding highlights the large differences in isoform specificity among currently available assays. One could argue that, in theory, by measuring only 22-kD GH, information about other isoforms contributing to total bioactive GH might be missed (5, 28, 29). However, 22-kD GH is the most abundant, bioactive form of GH, and no evidence exists to suggest isolated changes in other isoforms during specific clinical situations. Its sole measurement also provides the advantage of traceability to the welldefined calibrator IS 98/574 (8). Therefore, the recent interdisciplinary consensus statement on GH and IGF assays calls for GH assays to specifically measure the 22-kD GH isoform (1). GHBP interference can be an issue in GH assays (5) and was therefore systematically investigated in the new isys assay. When up to 5 nmol/l recombinant GHBP corresponding to high endogenous concentrations was added to serum samples, measured GH concentrations in the isys and the Immulite GH assay decreased [mean recoveries (isys); (Immulite)]. Recombinant GHBP is produced in Escherichia coli and is not glycosylated. Therefore, its affinity to GH might differ from that of endogenous GHBP. Consequently, with the isys assay we also analyzed recovery of recombinant GH supplemented into samples with known endogenous GHBP content. Over a broad range of GHBP concentrations, 1452 Clinical Chemistry 58:10 (2012)

8 A 22-kD GH-Specific Automated GH Assay Fig. 2. OGTTs on the isys and Immulite. Combined data for males and females are shown for (A) healthy volunteers, (B) patients with acromegaly on Pegvisomant treatment, (C) patients with acromegaly on somatostatin analogs, and (D) treatment-naive patients with acromegaly. Data shown here are mean of all the male and female data with 95 CI plotted for clarity. we did not detect any marked reduction in GH recovery, indicating that the impact of endogenous GHBP in the isys GH assay is negligible. There is also no cross-reactivity of the drug Pegvisomant with the isys GH assay, even at the extremely high concentrations occurring in patients following injection. On the Immulite, the addition of Pegvisomant to samples caused a consistent 50 reduction in measured GH. Most likely, Pegvisomant is recognized by only 1 of the antibodies in the assay, thereby displacing endogenous GH and preventing sandwich formation. In contrast, in the Liaison GH assay the presence of moderate Pegvisomant concentrations leads to gross overestimation of GH, whereas at extremely high concentrations, measured GH concentrations drop. This concentration-dependent type of interference from Pegvisomant, which was also described in other simultaneous GH assays (16), could be explained by recognition of the drug by both antibodies, combined with a high-dose hook effect. Obviously, any type of cross-reaction of Pegvisomant prohibits the use of those assays to study GH concentrations during treatment. In contrast, the new isys GH assay can be used in such studies. Our observation of increased endogenous GH during treatment is in line with previous reports on the use of specific manual Pegvisomant-excluding GH assays (19, 30, 31). Interestingly, there was also no crossreactivity with placental GH or placental lactogen in the isys GH assay. Therefore, this routine assay Clinical Chemistry 58:10 (2012) 1453

9 Fig. 3. GH and IGF-I response to Pegvisomant treatment in samples from a series of patients with acromegaly (n 9) before and after a mean of 15 months treatment with Pegvisomant. could also be used to investigate pituitary-derived GH during pregnancy, which previously has required specific research assays (18, 20, 24). There was excellent correlation but only moderate agreement between the isys and Immulite assays in samples from healthy volunteers and random patients [isys (0.85 Immulite) 0.007, R ], while in the same samples both correlation and agreement between isys and Liaison were excellent [isys (1.04 Liaison) 0.014, R ]. In part, the higher concentrations on the Immulite can be explained by the fact that the other assays partially (Liaison) or completely (isys) exclude 20-kD GH from the measurement. As expected from the supplementation experiments, a different outcome was observed when samples from Pegvisomant-treated patients were measured: Compared to the isys, reported GH was approximately 50 lower on the Immulite, whereas the strong interference from Pegvisomant resulted in meaningless GH concentrations on the Liaison. Several authors have emphasized that assayspecific cut-offs for dynamic tests of GH secretion are mandatory (32 36). A formal establishment of cut-offs goes beyond the scope of this report. We do present, however, GH concentrations obtained during OGTT in healthy individuals and patients with acromegaly, measured by both the isys and the Immulite assay, to provide a comparison with a widely used method. In both assays, healthy females had higher baseline GH than males, and in agreement with previous studies (37, 38) females also exhibited higher GH nadir values than the males. Nadir GH concentrations were slightly lower on the isys compared to the Immulite. Our data also support the notion that the traditional cutoff of 1 g/l (39) might be too high and require downward adjustment in an assay-specific manner (33, 40). In our cohort of 25 healthy individuals, however, we found 2 who exhibited GH nadir concentrations that, despite being 1 g/l, were higher than those in all other individuals, regardless of the GH assay used. Both individuals had IGF-I within reference intervals and no clinical suspicion of acromegaly, so the clinical relevance remains unclear. In all other individuals, the mean GH nadir for the isys was 0.05 g/l (maximum 0.09) in males and 0.11 g/l (maximum 0.34) in females. Analysis of GH during OGTT in treatment-naïve patients with acromegaly revealed comparable results with the Immulite and the isys, both indicating increased tonic GH secretion. In somatostatin-treated patients GH concentrations were lower, and both assays indicated a slight but noticeable decline in GH concentrations during OGTT. In contrast, in the Pegvisomant-treated group GH values reported from the 2 assays differed markedly. Only with the use of the Pegvisomant-insensitive isys assay was a suppression of endogenous GH detectable, suggesting that glucosedependent downregulation of GH is maintained under Pegvisomant treatment. Finally, we used the isys to analyze samples from patients with acromegaly before and after treatment with Pegvisomant. In agreement with previous studies in which a manual assay specific for endogenous GH was used (19, 31, 41), we found that the decrease in IGF-I was accompanied by an increase in endogenous GH. Changes in endogenous GH during Pegvisomant treatment could be related to treatment efficacy or potential growth of the adenoma (42). Because of the lack of interference from Pegvisomant such questions now can be addressed with the isys GH assay. In summary, we developed a new highly specific automated GH assay that, following recent consensus recommendations, addresses known problems of existing GH immunoassays including heterogeneity of measured GH isoforms and interferences from GHBP and GH analogs. Comparison of results from the isys and 2 other established GH assays demonstrates that despite good overall correlation between methods a large discordance in values can be present. Such differences require assayspecific decision limits and also can become particularly important in specific clinical situations such as pharmacotherapy. The assay described here is, to our knowledge, the first assay specifically designed to measure the 22-kD isoform while completely excluding interference from Pegvisomant, presenting a useful tool for reliable measurement of GH secretion Clinical Chemistry 58:10 (2012)

10 A 22-kD GH-Specific Automated GH Assay 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: J. Manolopoulou, IDS; Y. Alami, IDS. References Consultant or Advisory Role: C.J. Strasburger, Pfizer; M. Bidlingmaier, IDS. Stock Ownership: None declared. Honoraria: J. Schopohl, Ipsen, Novartis, and Pfizer; C.J. Strasburger, Pfizer; M. Bidlingmaier, Pfizer, IDS, Siemens, and Diasorin. Research Funding: C.J. Strasburger, Pfizer; M. Bidlingmaier, Pfizer, IDS, Siemens, and Diasorin. Expert Testimony: None declared. Role of Sponsor: The funding organizations played no role in the design of study, choice of enrolled patients, review and interpretation of data, or preparation or approval of manuscript. 1. Clemmons DR. Consensus statement on the standardization and evaluation of growth hormone and insulin-like growth factor assays. Clin Chem 2011;57: Bidlingmaier M, Freda PU. Measurement of human growth hormone by immunoassays: current status, unsolved problems and clinical consequences. Growth Horm IGF Res 2010;20: Baumann GP. Growth hormone isoforms. Growth Horm IGF Res 2009;19: 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: Popii V, Baumann G. Laboratory measurement of growth hormone. Clin Chim Acta 2004;350: Wieringa GE, Trainer PJ. Commentary: Harmonizing harmonizing growth hormone measurements: learning lessons for the future. J Clin Endocrinol Metab 2007;92: Trainer PJ, Barth J, Sturgeon C, Wieringaon G. Consensus statement on the standardisation of GH assays. Eur J Endocrinol 2006;155: 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: Tanaka T, Tachibana K, Shimatsu A, Katsumata N, Tsushima T, Hizuka N, et al. A nationwide attempt to standardize growth hormone assays. Horm Res ;(Suppl 2): Ranke MB, Orskov H, Bristow AF, Seth J, Baumann G. Consensus on how to measure growth hormone in serum. Horm Res 1999;51: Baumann G. Growth hormone heterogeneity: genes, isohormones, variants, and binding proteins. Endocr Rev 1991;12: Jan T, Shaw MA, Baumann G. Effects of growth hormone-binding proteins on serum growth hormone measurements. J Clin Endocrinol Metab 1991;72: 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: Hansen TK, Fisker S, Hansen B, Sorensen HH, Christiansen JS, Jorgensen JO, Orskov H. Impact of GHBP interference on estimates of GH and GH pharmacokinetics. Clin Endocrinol 2002;57: Trainer PJ, Drake WM, Katznelson L, Freda PU, Herman-Bonert V, van der Lely AJ, et al. Treatment of acromegaly with the growth hormonereceptor antagonist pegvisomant. N Engl J Med 2000;342: 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: Higham CE, Atkinson AB, Aylwin S, Bidlingmaier M, Drake WM, Lewis A, et al. Effective combination treatment with cabergoline and low-dose pegvisomant in active acromegaly: a prospective clinical trial. J Clin Endocrinol Metab 2012;97: Brian SR, Bidlingmaier M, Wajnrajch MP, Weinzimer SA, Inzucchi SE. Treatment of acromegaly with pegvisomant during pregnancy: maternal and fetal effects. J Clin Endocrinol Metab 2007; 92: Veldhuis JD, Bidlingmaier M, Anderson SM, Wu Z, Strasburger CJ. Lowering total plasma insulin-like growth factor I concentrations by way of a novel, potent, and selective growth hormone (GH) receptor antagonist, pegvisomant (B2036-peg), augments the amplitude of GH secretory bursts and elevates basal/nonpulsatile GH release in healthy women and men. J Clin Endocrinol Metab 2001;86: Thorner MO, Strasburger CJ, Wu Z, Straume M, Bidlingmaier M, Pezzoli SS, et al. Growth hormone (GH) receptor blockade with a pegmodified GH (B2036-peg) lowers serum insulinlike growth factor-i but does not acutely stimulate serum GH. J Clin Endocrinol Metab 1999;84: Strasburger CJ, Kostyo J, Vogel T, Barnard GJ, Kohen F. The antigenic epitopes of human growth hormone as mapped by monoclonal antibodies. Endocrinology 1989;124: NCCLS. Evaluation of precision performance of quantitative measurement methods; approved guideline second edition NCCLS document EP5 A2 (ISBN ). Wayne (PA): CLSI; NCCLS. Protocols for determination of limits of detection and limits of quantitation; approved guideline NCCLS document EP17-A (ISBN ). Wayne (PA): CLSI; Wu Z, Bidlingmaier M, Friess SC, Kirk SE, Buchinger P, Schiessl B, Strasburger CJ. A new nonisotopic, highly sensitive assay for the measurement of human placental growth hormone: development and clinical implications. J Clin Endocrinol Metab 2003;88: Carlsson LM, Rowland AM, Clark RG, Gesundheit N, Wong WL. Ligand-mediated immunofunctional assay for quantitation of growth hormonebinding protein in human blood. J Clin Endocrinol Metab 1991;73: Doehner W, Pflaum CD, Rauchhaus M, Godsland IF, Egerer K, Cicoira M, et al. Leptin, insulin sensitivity and growth hormone binding protein in chronic heart failure with and without cardiac cachexia. Eur J Endocrinol 2001;145: Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res 1999;8: Wood P. Growth hormone: its measurement and the need for assay harmonization. Ann Clin Biochem 2001;38: Hayakawa M, Shimazaki Y, Tsushima T, Kato Y, Takano K, Chihara K, et al. Metabolic effects of 20-kilodalton human growth hormone (20k-hGH) for adults with growth hormone deficiency: results of an exploratory uncontrolled multicenter clinical trial of 20k-hGH. J Clin Endocrinol Metab 2004;89: Muller AF, Janssen JA, Hofland LJ, Lamberts SW, Bidlingmaier M, Strasburger CJ, van der Lely AJ. Blockade of the growth hormone (GH) receptor unmasks rapid GH-releasing peptide-6-mediated tissue-specific insulin resistance. J Clin Endocrinol Metab 2001;86: Veldhuis JD, Bidlingmaier M, Anderson SM, Evans WS, Wu Z, Strasburger CJ. Impact of experimental blockade of peripheral growth hormone (GH) receptors on the kinetics of endogenous and exogenous GH removal in healthy women and men. J Clin Endocrinol Metab 2002;87: Freda PU. Current concepts in the biochemical assessment of the patient with acromegaly. Growth Horm IGF Res 2003;13: Freda PU, Post KD, Powell JS, Wardlaw SL. Evaluation of disease status with sensitive measures of growth hormone secretion in 60 postoperative patients with acromegaly. J Clin Endocrinol Metab 1998;83: Costa AC, Rossi A, Martinelli CE Jr, Machado HR, Moreira AC. Assessment of disease activity in treated acromegalic patients using a sensitive GH assay: should we achieve strict normal GH levels for a biochemical cure? J Clin Endocrinol Metab 2002;87: Arafat AM, Mohlig M, Weickert MO, Perschel FH, Purschwitz J, Spranger J, 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, Clinical Chemistry 58:10 (2012) 1455

11 and body mass index. J Clin Endocrinol Metab 2008;93: Rakover Y, Lavi I, Masalah R, Issam T, Weiner E, Ben-Shlomo I. Comparison between four immunoassays for growth hormone (GH) measurement as guides to clinical decisions following GH provocative tests. J Pediatr Endocrinol Metab 2000;13: Chapman IM, Hartman ML, Straume M, Johnson ML, Veldhuis JD, Thorner MO. Enhanced sensitivity growth hormone (GH) chemiluminescence assay reveals lower postglucose nadir GH concentrations in men than women. J Clin Endocrinol Metab 1994;78: Markkanen H, Pekkarinen T, Valimaki MJ, Alfthan H, Kauppinen-Makelin R, Sane T, Stenman UH. Effect of sex and assay method on serum concentrations of growth hormone in patients with acromegaly and in healthy controls. Clin Chem 2006;52: Giustina A, Barkan A, Casanueva FF, Cavagnini F, Frohman L, Ho K, et al. Criteria for cure of acromegaly: a consensus statement. J Clin Endocrinol Metab 2000;85: Freda PU, Nuruzzaman AT, Reyes CM, Sundeen RE, Post KD. Significance of abnormal nadir growth hormone levels after oral glucose in postoperative patients with acromegaly in remission with normal insulin-like growth factor-i levels. J Clin Endocrinol Metab 2004;89: Roemmler J, Steffin B, Gutt B, Sievers C, Bidlingmaier M, Schopohl J. The effect of acute application of pegvisomant alone and in combination with octreotide on endogenous GH levels during a 6-h test in patients with acromegaly on constant pegvisomant treatment. Growth Horm IGF Res 2009;19: Buchfelder M, Weigel D, Droste M, Mann K, Saller B, Brubach K, et al. Pituitary tumor size in acromegaly during pegvisomant treatment: experience from MR re-evaluations of the German Pegvisomant Observational Study. Eur J Endocrinol 2009;161: Clinical Chemistry 58:10 (2012)