Puberty: Physiology and Abnormalities, 1st ed. 2016

17. Growth Hormone and Steroid Assays’ Problems in Childhood and Puberty

Dobrin A. Svinarov 


Department of Therapeutic Drug Management and Clinical Pharmacology, Alexander Hospital, Medical University of Sofia, St. G. Sofiiski 1 Blvd, 1431 Sofia, Bulgaria

Dobrin A. Svinarov



Growth hormoneInsulin-like growth factorSteroid hormonesImmunoassaysMass spectrometryChildrenAdolescents

Growth Hormone Disorders: Analytical and Interpretative Challenges

Growth hormone (GH) and insulin-like growth factor-I (IGF-I) are the most important laboratory parameters used in the diagnosis of GH secretion disorders, assessment of children with short stature, and monitoring of replacement therapy with GH and IGF-I. Therefore, clinicians should be permanently updated by their laboratories on the analytical weaknesses and problems of GH and IGF-I assays and how they affect the accuracy of patient assessment, as well as of the ongoing progress for standardization and harmonization of these assays. GH, also known as somatotropin, is secreted by anterior pituitary somatotropic cells under the regulation of two hypothalamic factors: GH-releasing hormone (GHRH ) and somatotropin release-inhibiting hormone, with stimulation being predominant over suppression. GH has direct metabolic actions on protein metabolism: increased uptake of amino acids in tissues (anabolic effect), and on lipid metabolism: activation of lipolysis and elevation of free fatty acids (catabolic effect). GH effects on growth are indirect—it induces the secretion of IGF-I (known previously as somatomedin C) from GH responsive tissues—predominantly liver cells. Growth-promoting effects of IGF-I are mostly a result of autocrine and paracrine actions. Elevated IGF-I acts centrally to reduce hypothalamic release of GHRH, thus reducing GH secretion in a negative feedback mechanism. Apart from GH stimulation, IGF-I levels are influenced by thyroid hormones, sex steroids, chronic disease, and nutritional status. Therefore, growth can be impaired even with GH sufficiency in cases of hypothyroidism, sex steroid deficiency, chronic illness, and malnutrition. GH stimulates the production of two other proteins: IGF-binding protein-3 (IGFBP-3 ) and acid-labile subunit, which may form circulating complexes with IGF. There are six IGFBPs of which IGFBP-3 is the major serum carrier of IGF-I. In addition to GH and IGF-I, IGFBP-3 measurement is also considered for the diagnosis of GH disturbances in childhood and puberty [12].

Problems with GH Immunoassays (IAs)

Controversies in the diagnosis and management of GH-related disorders in childhood and puberty are multifactorial, and this chapter will address only those that are related to GH analysis in the clinical laboratory: heterogeneous nature of GH, its biological variability, problems with IAs, the lack of standardization and of definitive analytical technology, and the role of liquid chromatography coupled to mass spectrometry (LC–MS).

GH is a heterogenic molecular species. It is synthesized as a single-chain polypeptide composed of 191 amino acids with a molecular mass of 22 KDa but exists in the cells and in the circulation in a variety of forms: (1) GH1 (GH-N) and GH2 (GH-V) are full-size 22-KDa molecules encoded by two genes, the first produced primarily in pituitary and the second in placenta. GH1 represents 85–90 % of circulating GH and is its’ active form. (2) A 20-KDa GH isoform is derived by GH1 as a result of deletion of amino acids 32–46, has less biological activity, and shows a propensity to dimerize. It accounts for 5–10 % of GH in blood [2]. (3) “Big GH” is a dimeric moiety of the 20-KDa isoform. (4) “Big–big GH”: approximately 50 % of GH is bound to its protein transporter—GH-binding protein (GHBP) that is the cleaved N-terminal extracellular domain of the GH receptor (GHR) . The circulating complex (GH-GHBP) is sometimes called “Big–big GH” [1]. A small amount (5–8 %) of GH is carried by a low-affinity protein, and the rest is in free (unbound) state. (5) There are also oligomeric forms of GH, which could potentially interfere in some analytical platforms.

Labeled immunochemical analysis is by far the predominant analytical technology for measuring GH. As for most other hormones, IAs progressed from radioimmunoassay to non-isotopic methods utilizing enzyme or luminescent labels (fluorescent, chemiluminescent , and electrochemiluminescent ) and from the use of polyclonal to monoclonal antibodies. Nowadays, a variety of both earlier and recently introduced assay kits are commercially available, employing isotopic and non-isotopic labels, and all of them suffer from the same pitfalls. The presence of GH variants and the use of different antibodies, different calibration materials, different labels, and different assay reagents lead to significant discrepancies among the results given by the different IAs [35]. Some of the antibodies are quite specific for the 22-KDa GH, and others cross-react with the 20-KDa isoform and its dimers, while the interference of some GH oligomers is not clearly established. Measurement of the 20-KDa isoform has no clinical significance, but it has been shown to be useful for the assessment of GH doping in sports. Overall, affinity of most antibodies is high enough to compete with GHBP, provided the incubation time is sufficient to allow for the dissociation of GH-GHBP complexes. Therefore, those assays give a good approximation of the total GH concentration. The clinical value of free (unbound) GH is not established, although there are IAs developed to measure free GH [2]. Inter-assay variability of IAs was substantially improved over the last years, but still differences in results remain significant enough to lead to misdiagnosis and inappropriate patient management [46] and inability to create generally accepted harmonized clinical guidelines for diagnosis and treatment of GH-related disorders in childhood and puberty. Several international research and professional societies have organized an expert workshop to define criteria, strategies, and ways to implement harmonization of GH and IGF-I assays and created recommendations and requirements for improving assay comparability, which were published in year 2011 [7], and can be summarized as follows: (1) Preanalytical conditions should be strictly adhered to—serum is the preferred specimen for analysis; separation of serum from blood cells is recommended within 2 h of collection, stability at room temperature is proven for 8 h, and if specimens are not tested within that time, they should be stored at 2–8 °C during the same day or frozen at −20 °C for longer periods. Stability at −20 °C is proven for several weeks (no more than 30 days for IGF-I). (2) The most important requirement is by far the need for common calibration. All manufacturers should use a single universally accepted international reference recombinant human GH Standard, IS 98/574, and IS 02/254 WHO recombinant IGF-I reference standard, both of them available from National Institute of Biological Standards and Control (NIBSC) , and demonstrate traceability to those standards. Results should be reported in mass units, i.e., μg/L. (3) Commutability of calibrators and control materials should be achieved and validated—the matrix used should be identical to or as close as possible to nonpathologic human serum. (4) Antibodies used in GH assays should be of high affinity and specificity for the 22-KDa form of GH. These assays should have sufficient reproducibility and accuracy , especially at the lower end of the measured interval, with a lower limit of GH quantification (LLOQ) of 0.05 μg/L (CV of <20 %). (5) Interferences should be validated explicitly, including cross-reactivity with 20-KDa GH, GH 2, therapeutic GH analogs, and especially interference of GHBP. (6) Laboratories should use internal quality control (QC) materials from different manufacturer (independent from assay producer) and should participate in an accredited proficiency testing system/external quality assessment program at an international level. Further, development of a reference measurement procedure based on LC–MS is strongly needed and should be utilized for the establishment of acceptance criteria for all IAs intended for clinical use [6]. Manufacturers must inform laboratories for all of the above components and for every change in assay platforms that may influence clinical diagnostic and therapeutic decisions. Laboratories must communicate that information to clinicians and explain the way assay performance could affect patient care [7].

Problems with IGF-I and IGFBP-3 Immunoassays

Unlike GH, serum concentrations of IGF-I and IGFBP-3 are stable throughout the day, and therefore, they can be measured at any time. Total IGF-I is measured via IAs based on the same principles, showing the same historical development as described above for GH assays, but all techniques include a step for dissociation of IGF-I from its binding proteins prior to analysis [12]. Manufacturers should validate the efficiency of that process separately and demonstrate the influence of diseases affecting serum concentrations of IGFBP, such as diabetes, GH disorders, and hepatic or renal impairment, on the total IGF-I concentration measured [47]. Serum concentrations of IGF-I vary with age, sex, and degree of sexual maturation of the child. Unfortunately, there is an overlap of the normal range of serum IGF-I concentrations in young children with the levels found in children with GH deficiency (GHD ). In addition, children with malnutrition, hypothyroidism, chronic illness, renal failure, steroid deficiency, and diabetes also have low IGF-I concentrations. On the other hand, IGF-I levels rise dramatically during puberty. Therefore, reference intervals in childhood and puberty should be assessed for data normality, presented in percentiles (2.5–97.5 percentiles) after transformation, and reported as SD scores—the number of SDs a given result is deviated from the age-adjusted mean. Narrow age ranges (e.g., every 3 years) and Tanner stages should be considered, with sufficient number of individuals included and sex-specific reference values for ages between 6 and 18 years [78]. IGFBP-3 does not need a dissociation step from a binding protein, and its concentration is much higher compared to that of IFG-I or GH. Therefore, IAs are generally simpler than those for GH and IGF-I. In addition, IGFBP levels vary to a lesser extent with age, sex, and degree of sexual maturation [1] and are less affected by nutritional status of the child. Based on the above considerations, it was thought until recently that simultaneous measurement of IGF-I and IGFBP-3 could be potentially superior to the assessment of IGF-I alone. However, multiple studies reported a very poor diagnostic sensitivity of IGFBP (approximating 50 %), with no difference of its concentrations between GHD and non-GHD subjects and no advantage over measurement of IGF-I alone [4].

The Role of LC–MS in the Analysis of GH, IGF-I, and IGFBP-3

This technique allows recognition and quantification by mass rather than by epitope and provides the most sensitive and selective analytical results in biology and medicine. It has the potential to circumvent many of the problems associated with IAs for GH, IGF-I, and IGFBP-3 [89]. Several methods for ID-MS (isotope dilution mass spectrometry: MS using stable isotope labeled internal standards, which renders highest possible reliability) determination of these peptides have been described [1014], with excellent accuracy and precision and lack of interference by GHBP or other constituents of human serum. Application of LC–MS would allow the establishment of reproducible, method-independent, comparable, and unified reference intervals and cut off limits and would ultimately give a chance for the development of clinically unambiguous and generally recognized guidelines for the management of GH-related disorders in childhood and puberty. It was already noted that ID-MS is also the ultimate technique for the development of reference measurement procedures for GH and IGF-I—a key factor for the standardization and harmonization of newer generation IAs.

Clinical Significance

The following brief discussion will focus only on those interpretative challenges in the management of GH-related disorders of childhood and puberty, which are related to the analytical problems of GH, IGF-I, and IGFBP-3 assays. Clinically important states of GH excess or deficiency are rare and difficult to diagnose, with GHD being considerably more challenging and laboratory dependent, compared to GH excess. GH is stored in the pituitary cells after synthesis and is secreted in several (usually seven to eight) pulses per day with some “spikes” after meals or exercise, but most pulses (by number and intensity) occur at night after onset of sleep, with peak values reached during deepest sleep [12] and very low concentrations between pulses. This pattern of secretion and its short half-life makes random measurements of GH non-informative and misleading for assessment of sufficiency or insufficiency. Therefore, diagnosis of GHD requires a combination of provocative screening and stimulation tests and measurement of GH, IGF-I, and in some cases IGFBP-3. Although uncommon itself, GH excess in childhood (gigantism) is much less common than acromegaly, and its’ diagnosis is made predominantly by physical appearance (striking acceleration of linear growth) and magnetic resonance imaging [12]. In contrast to GH, IGF-I half-life is much longer, and its concentrations are much less variable during the day. Therefore, a random single measurement of IGF-I is considered to accurately reflect its production. As mentioned previously, IGF-I concentrations are influenced by age, sexual maturation, nutritional status, and several disease states.

There are two general approaches in diagnosing and management of GHD [1]. The first one is preferred by physicians, who believe that physiological variety of GH and reported discrepancies of the non-standardized GH measurements hamper the clinical value of its analysis. This approach relies solely on the measurement of IGF-I level, and diagnosis is considered proven if IGF-I is low and is combined with low growth velocity. The second, more traditional approach comprises a two-step testing of GH concentrations after pharmacologic stimulation with any of the following: clonidine, glucagon, L-dopa, insulin, arginine, pyridostigmine, and GHRH, applied alone or in combination. Physical stimuli of GH secretion, including sleep and exercise, are usually avoided, due to lack of reproducibility. The first step (screening) involves measurement of GH after a single pharmacologic challenge. If the GH rises above the cutoff value for GHD (usually around 7 μg/L, see billow), GH sufficiency is accepted with no need for further measurements. The second step is considered definitive for the diagnosis of GHD and requires a combination of two stimulatory tests, because after a single challenge only 80 % of normal children will respond sufficiently, while after a two-test provocation, this percentage rises over 95 %. The major pitfall in the traditional approach is the analytical uncertainty of GH IAs and method-dependent cutoff values, which differ significantly even with different batches of the same assay. Irrespective of the well-defined requirements for methodological improvement outlined above, there is no GH assay on the market that meets the criteria for standardization and harmonization. Therefore, the recent publication of Wagner et al. [15] plays a key role in the effort to establish comparative harmonized cutoff values for some of the most often used IAs. In this work, the cutoff concentrations for the diagnosis of GHD for six commercial IAs and for an ID-MS-based assay were reexamined, based on the retesting of serum samples of children with GHD and children with no GHD, all of which have undergone stimulation tests. The optimal cutoff value for one of the IAs was identified and converted via regression analysis to comparable cutoff values for the other five assays. An ID-MS-based cutoff concentration was independently derived by analysis of a subset of patients and regression comparison to the same IA. The calculated cutoff limits ranged from 4.32 to 7.77 μg/L. These huge differences demonstrate the unacceptable variability of results obtained by IAs for one sample and the ultimate need for assay-specific cutoffs for the diagnosis of GHD. In addition, comparison of the IAs to ID-MS provides long-term validity of the established cutoff values and helps to detect gaps in the traceability chain, which is a critical unmet requirement of assays’ manufacturers.

Interpretation of IGF-I results is greatly facilitated by the recent publication of Bidlingmaier et al. [16], presenting the results of a remarkable international project for the development and validation of a new IA conforming to international recommendations and its application to establish reference intervals for IGF-I from birth to senescence. The major results of this important work could be summarized in regard to assay characteristics and the reported IGF-I reference values. The new assay is calibrated against the recommended standard (02/254). It is very sensitive (limit of detection 4.4 μg/L) and has a broad dynamic range (10–1200 μg/L), with excellent accuracy (92–104 %) and precision (total analytical CV <9 %), and explicitly demonstrated lack of interference from the six IGFBPs and other matrix constituents. Sample material requirements, preanalytical, post-analytical, short-term, and long-term stability, and comparison with other assays complete the validation profile of this new, reliable IA technique. Generation of reference intervals is performed under strict adherence to the most rigorous international guidelines. Reference ranges in childhood and puberty are derived from the conducted multicenter study, with 4252 samples (1884 males, 2368 females) of newborns, children, and adolescents, obtained from seven pediatric cohorts from Canada, the United States, and Europe, with the significant contribution of the Canadian Laboratory Initiative on Pediatric Reference Interval Database (CALIPER and CALIPER new), providing a total of 1948 (1148 males) samples. The pediatric population was ethnically diverse and deemed to be metabolically stable. Pediatric reference values are calculated for each year of life from birth to the age of 20, separately for males and females. Additionally, assessment according to Tanner stages is presented, based on one of the largest cohorts (Danish Cohort, n = 854). This publication is a free-access paper, and all clinicians dealing with interpretation of IGF-I concentrations should use it. Briefly, IGF-I concentrations in cord blood correlated to birth weight but were not different between sexes. After that, concentrations declined and remained lower than at birth during the first year of life (from 57 to 77 μg/L, 2.5th percentile, to 126–157 μg/L, 97.5th percentile). Thereafter, IGF-I increased and reached a pubertal peak both in girls and boys at 15 years (from 127 to 152 μg/L, 2.5th percentile, to 554 μg/L, 97.5th percentile) and relatively fast decline until the age of 21–25 (from 93 to 115 μg/L, 2.5th percentile, to 342–355 μg/L, 97.5th percentile). IGF-I peak was similar in adolescents when data were stratified to chronological age but occurs slightly earlier in girls, when considering stratification according to Tanner stages. During puberty, reference interval was broader in girls, indicating a higher variability of IGF-I concentrations. In both sexes, the lower limit of reference interval (the 2.5th percentile) was lower when calculated on chronological age, compared to the calculation based on Tanner stage, which indicates that reference values adjusted for pubertal development can be useful, especially if IGF-I is low. Body mass index (BMI) extremities and comorbidities had no significant influence on the reference intervals.

Problems of Steroid Assays : Immunochemical Methods or LC–MS/MS

Steroid hormones are synthesized from cholesterol in the adrenal cortex, the gonads, and the placenta, and many of them are of great clinical importance [17]. Abnormal concentration of steroids in children and adolescents may be indicative of ambiguous genitalia in infants, hypogonadism, precocious puberty, oligomenorrhea, hirsutism in females, feminization in males, metabolic and electrolyte disturbances, and rarely, steroid hormone-producing neoplasms [18]. The adrenal gland is composed of the adrenal medulla (inner layer) and cortex (outer layer). The cortex is further composed of three layers: zona glomerulosa, which produces mineralocorticoids such as 11-desoxycorticosterone, corticosterone, and aldosterone, and zonae fasciculata and reticularis, which together are the source of glucocorticoids (11-desoxycortisol, cortisol) and of adrenal androgens dehydroepiandrosterone (DHEA) , DHEA-sulfate (DHEAS) , and androstenedione. Testosterone is produced by the Leydig cells of the testes, and female sex hormones, estradiol and progesterone, are synthesized in the ovaries. Estradiol could be derived from testosterone by aromatization of ring A (CYP19A1). The same aromatase catalyzes the transformation of androstenedione to estrone, which is also a precursor of estradiol. Estriol is derived from placenta, and its measurement is important in the assessment of fetal status [171920]. Biochemical pathways of steroid hormones are far more complex, and as a result, over 300 precursors and metabolites circulate in the body fluids together with the active hormones, in a variety of concentrations and biological activity but with very similar structure. Therefore, analytical methods for steroids need to pose sufficient selectivity and sensitivity to allow for measurement of the steroid of interest, without cross-reactivity and interference, a requirement that is very difficult to achieve.

Free Versus Total Steroid Measurement

Steroid hormones circulate in blood bound to carrier proteins, which regulate their activity. Total concentration of hormones that are highly protein bound depends on the concentration of the protein carrier and therefore may vary significantly, while their unbound (free) concentrations remain within physiologic limits. Measurement of free steroids is a much better indicator of their functional activity—it is the free hormone that binds to the steroid receptor and elicits the biochemical action, but methods that are able to separate free from bound fraction and to reliably measure the unbound hormone are very difficult to develop and validate. There is a need of a preanalytical phase to separate bound and free fraction, which utilizes two techniques—equilibrium dialysis (reference technique) or ultrafiltration. The first procedure is very slow and expensive, and both techniques should strictly be adhered to; otherwise, a disruption of the physiologic ratio between free and bound hormone could occur [1920]. After the introduction of very sensitive LC–MS instruments, the required high sensitivity for free steroid analysis is not an issue [17], but until recently, alternative indirect approaches for estimation of unbound concentrations were predominantly employed in the clinical laboratories. These included measurement of specific protein carriers [cortisol-binding globulin (CBG) , sex hormone-binding globulin (SHBG) , and albumin] and usage of mathematical algorithms for calculation of free concentrations: “selective” precipitation of the tightly bound form or direct (analog tracer) radioimmunoassay. In the past, measurement of urinary cortisol (also providing an estimate of free hormone concentration) has been used for as a screening test for Cushing disease [19]. All indirect methods for assessment of free steroid concentrations give approximation for the unbound fractions and not true values [2122]. Therefore, when alternative and new techniques are available, they should be preferred for routine work.

Measurement of saliva steroid concentrations is a practical and convenient way to assess the free steroid fraction. Saliva can be viewed as a natural plasma ultrafiltrate, where only unbound steroids can transfer from plasma. It is obtained noninvasively, which is a special advantage in childhood. In general, most steroids of clinical significance can be measured in saliva. For some steroids, such as cortisol, estriol, and progesterone, measurement of salivary concentration is documented to be a good indicator of the free plasma concentration [1920]. The best established and accepted application for hormone analysis in saliva is the use of salivary cortisol in the diagnosis of Cushing syndrome. In pediatric medicine, saliva has proven very beneficial for diagnosing and monitoring of congenital adrenal hyperplasia (most commonly caused by 21-hydrohylase deficiency) via the measurement of salivary 17α-hydroxypreogesterone and androstenedione. In addition to estrogens and gestagens, a special focus for assessment of male hypogonadism is testosterone analysis in saliva, which demonstrates a very high correlation with free testosterone in serum [2021]. The predominant advantages of saliva steroid analysis in childhood include the noninvasiveness of sampling procedure, ability for multiple collections, and, unlike venipuncture, availability of samples obtained without unwanted adrenal stress. At the same time, special attention is required regarding standardization of collection devices and procedures, patient compliance to collection procedures, preanalytical stability of steroids in saliva, and usage of standardized and validated assays [23]. Interpretation of salivary steroid concentrations requires the establishment of special reference intervals.

Immunochemical Assays ’ Problems for Steroid Hormones

The introduction of IAs nearly 50 years ago allowed for the first time the direct measurement of active hormones in human plasma instead of measuring their inactive metabolites by colorimetric methods in urine or application of cumbersome in vivo bioassays. Thus, IA platforms opened a new era in endocrine research, clinical diagnostics, and patient measurement [24]. Earlier, first-generation methods employed nonspecific polyclonal antibodies and sample pretreatment, such as solvent extraction, chromatographic purification, and pre-concentration. These “indirect” techniques were slow and tedious but provided relatively sufficient specificity and sensitivity—the pretreatment step compensated for the nonspecificity of polyclonal antibodies. Later, in favor of speed, automation, and simplification, newer generation, “direct” IAs, utilizing monoclonal antibodies and enzyme or luminescent labels, replaced the older techniques. Unfortunately, together with high capacity, high speed, and wide availability, the newer “direct” methods sacrificed the requirements for sensitivity and specificity of steroid measurements. Employment of monoclonal antibodies was not sufficient to assure the required reliability. Interference with chemically similar steroid precursors, metabolites, exogenous compounds, and protein carriers, in combination with nonstandard calibration, and lack of standardization and traceability introduced unacceptably high intra-assay and inter-assay variability of currently used IAs [171920]. The magnitude of this problem is exemplified with results from one of the external quality assessment programs, exhibiting a threefold to ninefold between-laboratory difference for a single challenge [17]. Therefore, general requirements to manufacturers, listed for GH and IGF-I, fully apply for steroid IAs—need for traceable calibrators, relevant validation for interferences, achievement of clinically needed selectivity and sensitivity, traceability of results to reference measurement procedures (ID-MS), etc. The majority of recent publications, textbooks, and guideline s demonstrate nonlinearity and lack of accuracy of the different IAs [12172426] and recommend their replacement by LC–MS where possible, especially for the following steroids: measurement of cortisol secretion in serum, 24-h urine sample, or midnight saliva [24]; total testosterone in women, in children, and in hypogonadal men; estradiol in clinical situations with expected low levels of less than 10 or 5 pmol/L (healthy men, children, postmenopausal women) [1171920242728]. In conclusion, there is no doubt that MS will be the routine method principle for quantification of all steroids in the near future, but the technological transfer will not happen in 1 day and in all places. IAs will continue to be part of the analytical arsenal of clinical laboratories, especially when manufacturers of these assays fulfill the previously mentioned requirements for adequate validation and traceability. For instance, current testosterone IAs are still (and will be) acceptable for determination of higher concentrations in men; IAs perform satisfactory for estradiol in the assessment and management of women with infertility problems [2024].

Steroid Analysis by Mass Spectrometry

MS has been used for steroid analysis for more than 70 years, and therefore, it is not correct to appreciate it as the “new guy” in that analytical field. In 1960s, gas chromatography–mass spectrometry (GC–MS) was the major tool for the investigation of steroid metabolism and assessment of steroid profiles in health and disease [24]. Despite of the significant complexity, GC–MS still remains the preeminent discovery technique of choice in the clinical research of steroid metabolome, even in the era of LC–MS/MS. Our current knowledge on inborn errors of steroid metabolism and the identification of nearly all steroid metabolite disorders is based on GC–MS [2429]. In the last 15 years, the great technological advance of triple quadrupole LC–MS/MS resulted in the introduction of methods with unbeatable sensitivity and selectivity and extended linearity range, which are much simpler to use in the routine clinical laboratories compared to GC–MS and at the same are based on the current reference (definitive) analytical principles [172932]. LC–MS/MS reference measurement procedure for testosterone, which will serve for standardization of routine methods (MS and immunoassays), has already been proposed, and such procedures for other steroid hormones are expected to appear in the near future [3334]. Further significant advantages of LC–MS/MS include relatively high throughput and the ability to perform panel steroid profiling with simultaneous measurement of precursors, active hormones, and metabolites in a single sample, thus amplifying enormously the informative value of laboratory results, with ultimate improvement of patient care [173035]. For instance, assessment of androgen status requires quantification of testosterone and dihydrotestosterone [36], but immunoassays, being single-analyte platforms, do not provide that. Simultaneous analysis of multiple steroids reduces significantly the required sample volume, compared to IAs. Small sample size is especially important in pediatric endocrinology—the ability to obtain more than ten quantitative results from 100 to 200 μL of plasma is easily achievable [1733]. However, it should be clearly understood that analysis by LC–MS/MS does not automatically mean reliable results and superiority over IAs. MS is in its early phase of clinical application, roughly at the same stage that IAs were many years ago, when labor-intensive manual testing dominated and automatic immunochemistry analyzers were not available [24]. Among the disadvantages of LC–MS/MS, which currently limit its wider use in routine clinical laboratories, are expensive equipment; fairly labor-intensive sample preparation (“indirect” technique) and requirement for a highly competent and robustly trained staff, compared to immunoassay platforms; lack of standardization (most laboratories develop their own “in-house” methods); and occasional interferences and matrix effects, which require rigorous validation [172933]. Manufacturers are trying to develop mass spectrometers that are like clinical chemistry analyzers—more user-friendly, technically, and methodologically standardized and thus appropriate to meet the requirements of the high-throughput medical laboratories. Until that happen, LC–MS/MS will be a preferred routine armamentarium in the larger laboratories, where the expertise and the larger sample workload will provide cost-effectiveness in the effort to overcome the problems of current IAs for steroid analysis.

Interpretation of Results and Reference Intervals

It is fair to admit that pediatric reference values for steroid hormones are not well defined—a situation that challenges diagnostic sensitivity and interpretation of results, especially in cases with mild to modest elevations. Current reference intervals are usually derived based on very small groups of children, which limits the establishment of adequate age-specific and sex-specific differences. In addition, the technological transfer from IAs to LC–MS/MS is underway but not completed: all data, based on IAs, are method dependent and hence only valid for the respective assay; reference intervals based on LC–MS/MS are method independent and could be used more universally, but with mentioned limitations—lack of wide availability of the technique and need for standardization. Of particular importance is also the lack of agreement and comparability between LC–MS/MS and IA results, the latter being significantly higher (often several fold). Therefore, pediatric endocrinologists should communicate closely with the laboratories, requiring information on the particular assays used for steroid analysis and how those assays refer to published reference interval data. This chapter will briefly summarize some of the most convincing and most recent studies presenting pediatric reference values for steroid hormones. Konforte et al. [18] published their results for seven fertility hormones in children and adolescents as part of the CALIPER study, utilizing a new IA platform. These authors present data obtained from 1234 recruited participants and correctly state the advantages and limitations of their work: well-defined assay characteristics with improved sensitivity, compared to other IAs; database still valid for the IA used only, with the need for validation for other IAs; assay used not traced to higher metrology class technique, i.e., ID-MS; and limited number of participants for some age subgroups and for Tanner stage-specific partition, which is based on self-reported Tanner stage. Notably, authors declare that the results for testosterone are not much different from those obtained by LC–MS/MS in the same laboratory. In addition, the results for estradiol, but not for progesterone, are comparable to previously published data, based on LC–MS/MS [20]. Their Tanner-specific results for estradiol could be summarized as follows: for males, upper limits are practically the same for Tanner stages I–III (<70 pmol/L) and for Tanner stages IV and V (<130 pmol/L); female reference values are fully stratified, <70 pmol/L for Tanner stage I, <90 pmol/L for Tanner stage II, <300 pmol/L for Tanner stage III, 50–520 pmol/L for Tanner stage IV, and 70–760 pmol/L for Tanner stage V. The same group has published another part of their study, based on LC–MS/MS method for simultaneous measurement of eight steroids, again as part of the CALIPER program, with reported pediatric reference intervals of a total of 337 specimens [37]. In this work, progesterone values were found to be significantly lower, compared to the previously mentioned immunoassay, and similar between boys and girls until the age of 12 years, with female levels began increasing thereafter. Table 17.1 compiles and summarizes their results for five of the steroids that did not require partitioning based on sex. Data have shown that in addition to the known diurnal variation, cortisol changed with age, increasing after the age of 7 years old. Slightly different, compared to other reports, were the findings of this group for corticosterone, which exhibited the highest concentrations between birth and 1 year of life, declining gradually thereafter. Elevation of 11-desoxycortisol, 17-hydroxyprogesterone, and 21-hydroxyprogesterone is of importance for the detection of the respective hydroxylase deficiencies. Kushnir et al. [38] have published pediatric and adult reference intervals for the major androgen steroids, testosterone, androstenedione, and DHEA, based on a reference population of over 2500 participants, utilizing a validated LC–MS/MS method. Their Tanner-specific stratification is derived from the largest number of observations in each subgroup—over 120 samples for each Tanner stage and results could be summarized as follows. At Tanner stage I, there were no clinically significant sex differences for the three steroids, with upper limits of 0.5–0.6 nmol/L for testosterone, 1.1–1.8 nmol/L for androstenedione, and 8.2–9.6 nmol/L for DHEA. In girls, the three androgens reached adult values at Tanner stage III (0.3–2.2 nmol/L for testosterone, 1.3–7.8 nmol/L for androstenedione, 3.7–27.3 nmol/L for DHEA), while in boys adult values (5.6–29.5 nmol/L for testosterone, 0.9–3.7 nmol/L for androstenedione, 4.3–23.4 nmol/L for DHEA) were reached later—at Tanner stages IV–V. The greatest increase was observed during the transition to Tanner stage II for females and to Tanner stage III for males. At Tanner stage II, the upper reference limit for testosterone was already significantly higher in boys, 10.5 nmol/L, compared 1.4 nmol/L for girls. Levels of androstenedione were approximately twice higher in females since Tanner stage II and thereafter, and for DHEA, the major sex differences were observed at Tanner stages II and III.

Table 17.1

Age-specific pediatric reference values for some serum steroids (method: LC–MS/MS, study as part of the CALIPER program)



Upper limit (nmol/L)


0–2 years


3–6 years


7–14 years


15–18 years



<1 month


<1 year


1–6 years


7–14 years


15–18 years



<1 year


1–6 years


7–18 years



0–14 days


<1 year


1–11 years


12–13 years


14–18 years



<1 year


1–2 years


2–11 years


12–18 years


Modified from Kyriakopoulou L, Yazdanpanah M, Colantonio DA, et al. A sensitive and rapid mass spectrometric method for the simultaneous measurement of eight steroid hormones and CALIPER pediatric reference intervals. Clin Biochem. 2013;46:642–651. With permission from Elsevier



Winter WE, Bazydlo LAL, Harris NS. Quick guide to endocrinology. Washington, DC: AACC Press; 2013.


Winter WE, Jialal I, Vance ML, et al. Pituitary function and pathophysiology. In: Burtis CA, Ashwood ER, Bruns DE, editors. Tietz textbook of clinical chemistry and molecular diagnostics. St. Louis: Elsevier Saunders; 2012. p. 1803–45.CrossRef


Boulo S, Hanisch K, Bidlingmaier M, et al. Gaps in the traceability chain of human growth hormone measurements. Clin Chem. 2013;59:1074–82.CrossRefPubMed


Murray PG, Dattani MT, Clayton PE. Controversies in the diagnosis and management of growth hormone deficiency in childhood and adolescence. Arch Dis Child. 2014. doi:10.1136/archdischild-2014-307228.PubMed


Van Helden J, Hermsen D, Von Ahsen N, et al. Performance evaluation of a fully automated immunoassay for the detection of human growth hormone on the Elecsys immunoassay system. Clin Lab. 2014;60:1641–51.PubMed


Wieringa GE, Sturgeon CM, Trainer PJ. The harmonisation of growth hormone measurements: taking the next steps. Clin Chim Acta. 2014;432:68–71.CrossRefPubMed


Clemmons DR. Consensus statement on the standardization and evaluation of growth hormone and insulin-like growth factor assays. Clin Chem. 2011;57:555–9.CrossRefPubMed


Gomez-Gomez C, Iglesias EM, Barallat J, et al. Lack of transferability between two automated immunoassays for serum IGF-I measurement. Clin Lab. 2014;60:1859–64.PubMed


Junnila RK, Strasburger CJ, Bidlingmaier M. Pitfalls of insulin-like growth factor-I and growth hormone assays. Endocrinol Metab Clin North Am. 2015;44:21–34.CrossRef


Cox HD, Lopes F, Woldemariam GA, et al. Interlaboratory agreement of insulin-like growth factor 1. Concentrations measured by mass spectrometry. Clin Chem. 2014;60:541–8.CrossRefPubMed


Such-Sanmartín G, Bache N, Bosch J, et al. Detection and differentiation of 22kDa and 20kDa growth hormone proteoforms in human plasma by LC-MS/MS. Biochim Biophys Acta. 1854;2015:284–90.


Oran PE, Trenchevska O, Nedelkov D, et al. Parallel workflow for high-throughput (>1,000 samples/day) quantitative analysis of human insulin-like growth factor 1 using mass spectrometric immunoassay. PLoS One. 2014;9:e92801.CrossRefPubMedPubMedCentral


Arsene CG, Henrion A, Diekmann N, et al. Quantification of growth hormone in serum by isotope dilution mass spectrometry. Anal Biochem. 2010;401:228–35.CrossRefPubMed


Barton C, Kay RG, Gentzer W, et al. Development of high-throughput chemical extraction techniques and quantitative HPLC-MS/MS (SRM) assays for clinically relevant plasma proteins. J Proteome Res. 2010;9:333–40.CrossRefPubMed


Wagner IV, Paetzold C, Gausche R, et al. Clinical evidence-based cutoff limits for GH stimulation tests in children with a backup of results with reference to mass spectrometry. Eur J Endocrinol. 2014;171:389–97.CrossRefPubMed


Bidlingmaier M, Friedrich N, Emeny RT, et al. Reference intervals for insulin-like growth factor-1 (IGF-I) from birth to senescence: results from a multicenter study using a new automated chemiluminescence IGF-I immunoassay conforming to recent international recommendations. J Clin Endocrinol Metab. 2014;99:1712–21.CrossRefPubMed


Soldin S, Soldin OP. Steroid hormone analysis by tandem mass spectrometry. Clin Chem. 2009;55:1061–6.CrossRefPubMedPubMedCentral


Konforte D, Shea JL, Kyriakopoulou L, et al. Complex biological pattern of fertility hormones in children and adolescents: a study of healthy children from the CALIPER cohort and establishment of pediatric reference intervals. Clin Chem. 2013;59:1215–27.CrossRefPubMed


Bertholf RL, Jialal I, Winter WE. The adrenal cortex. In: Burtis CA, Ashwood ER, Bruns DE, editors. Tietz textbook of clinical chemistry and molecular diagnostics. St. Louis: Elsevier Saunders; 2012. p. 1847–904.CrossRef


Isbell TS, Jungheim E, Gronowski AM. Reproductive endocrinology and related disorders. In: Burtis CA, Ashwood ER, Bruns DE, editors. Tietz textbook of clinical chemistry and molecular diagnostics. St. Louis: Elsevier Saunders; 2012. p. 1945–90.CrossRef


Turpeinen U, Hämäläinen E, Haanpää M, et al. Determination of salivary testosterone and androstenedione by liquid chromatography-tandem mass spectrometry. Clin Chim Acta. 2012;413:594–9.CrossRefPubMed


Fritz KS, McKean AJS, Nelson JC, et al. Analog-based free testosterone test results linked to total testosterone concentrations, not free testosterone concentrations. Clin Chem. 2008;54:512–6.CrossRefPubMed


Gröschl M. Current status of salivary hormone analysis. Clin Chem. 2008;54:1759–69.CrossRefPubMed


Taylor A, Keevil B, Huhtaniemi IT. Mass spectrometry and immunoassay; how to measure steroid hormones today and tomorrow. Eur J Endocrinol. 2015;173:D1–12. European Society of Endocrinology, Manuscript EJE-15-0338.CrossRefPubMed


Benton SC, Nuttal M, Nardo L, et al. Measured dehydroepiandrosterone sulfate positively influences testosterone measurement in unextracted female serum: comparison of 2 immunoassays with testosterone measured by LC-MS. Clin Chem. 2011;57:1174–83.CrossRef


Huang X, Spink DC, Schneider E, et al. Measurement of unconjugated estriol in serum by liquid chromatography–tandem mass spectrometry and assessment of the accuracy of chemiluminescent immunoassays. Clin Chem. 2014;60:260–8.CrossRefPubMed


Stanczyk FZ, Lee JS, Santen RJ. Standardization of steroid hormone assays: why, how, and when. Cancer Epidemiol Biomarkers Prev. 2007;16:1713–9.CrossRefPubMed


Handelsman DJ, Newman JD, Jimenez M, et al. Performance of direct estradiol immunoassays with human male serum samples. Clin Chem. 2014;60:510–7.CrossRefPubMed


Krone N, Hughes BA, Lavery GG. Gas chromatography/mass spectrometry (GC/MS) remains a pre-eminent discovery tool in clinical steroid investigations even in the era of fast liquid chromatography tandem mass spectrometry (LC/MS/MS). J Steroid Biochem Mol Biol. 2010;121:496–504.CrossRefPubMedPubMedCentral


Handelsman DJ, Wartofsky L. Requirement for mass spectrometry sex steroid assays in the Journal of Clinical Endocrinology and Metabolism. J Clin Endocrinol Metab. 2013;98:3971–3.CrossRefPubMed


Kushnir MM, Rockwood AL, Roberts WL, et al. Liquid chromatography tandem mass spectrometry for analysis of steroids in clinical laboratories. Clin Biochem. 2011;44:77–88.CrossRefPubMed


Ketha H, Kaur S, Grebe S, Sihgh RJ. Clinical applications of LC-MS sex steroid assays: evolution of methodologies in 21st century. Curr Opin Endocrinol Diabetes Obes. 2014;21:217–26.CrossRefPubMed


Botelho JK, Shaklady C, Cooper HC, et al. Isotope-dilution liquid chromatography–tandem mass spectrometry candidate reference method for total testosterone in human serum. Clin Chem. 2013;59:372–80.CrossRefPubMed


Rosner W. Another reference measurement procedure for total testosterone—what’s the fuss? Clin Chem. 2013;59:338–9.CrossRefPubMed


Koal T, Schmiederer D, Pham-Tuan H, et al. Standardized LC-MS/MS based steroid hormone profile-analysis. J Steroid Biochem Mol Biol. 2012;129:129–38.CrossRefPubMed


Shiraishi S, Lee PWN, Leund A, et al. Simultaneous measurement of serum testosterone and dihydrotestosterone by liquid chromatography–tandem mass spectrometry. Clin Chem. 2008;54:1855–63.CrossRefPubMed


Kyriakopoulou L, Yazdanpanah M, Colantonio DA, et al. A sensitive and rapid mass spectrometric method for the simultaneous measurement of eight steroid hormones and CALIPER pediatric reference intervals. Clin Biochem. 2013;46:642–51.CrossRefPubMed


Kushnir MM, Blamires T, Rockwood LA, et al. Liquid chromatography–tandem mass spectrometry assay for androstenedione, dehydroepiandrosterone, and testosterone with pediatric and adult reference intervals. Clin Chem. 2010;56:1138–47.CrossRefPubMed