Beyond GH stimulation tests: genetic heterogeneity and treatment response in children with diagnosed GH deficiency Authors: Lukas Plachy, Petra Dusatkova, Lukas Kavciak, Shenali Anne Amaratunga, Matvei Slavenko, Jana Drabova, Klara Maratova, Vit Neuman, Barbora Obermannova, Stanislava Kolouskova, Marta Snajderova, Zdenek Sumnik, Jan Lebl, Stepanka Pruhova Published: 2026 ABSTRACT AbstractContextGrowth hormone (GH) stimulation tests have limited reliability and may lead to false-positive results.ObjectiveThis work aimed to evaluate genetic causes of short stature in children clinically diagnosed with GH deficiency (GHD) and assess genotype-phenotype correlations.MethodsA retrospective single-center study was conducted including 233 children with clinically diagnosed primary GHD treated with GH for 5 years or more. Genetic testing was performed using a targeted next-generation sequencing panel of 398 growth-associated genes.ResultsA genetic cause was identified in 39 of 233 (17%) children. Only 13 of 39 (33%) findings confirmed GHD (genes CHD7, GH1, GHSR, GLI2, GNAO1, KMT2D, OTX2 [3], POU1F1, PROP1, SALL4, TBX3). The remaining 26 of 39 (67%) revealed alternative etiologies of growth failure: RASopathies (13/39; genes NF1 [2], PTPN11 [7], RAF1, SOS1 [2], SPRED1), growth plate disorders (8/39; genes ACAN, COL2A1, EXT2, FGFR3 [2], NPR2 [3]), and miscellaneous conditions (5/39; genes CDC42, LMNA, HMGA2, PMM2, RAI1). Genetically confirmed GHD patients presented with lower peak GH (median 1.8 vs 4.8 [alternative genetic etiology] and 5.9 μg/L [no genetic etiology]; P = .017), more frequent combined pituitary hormone deficiency (46% vs 4% and 12%; P = .001) and magnetic resonance imaging midbrain abnormalities (73% vs 14% vs 32%; P = .003). Conversely, the alternative etiology group demonstrated the least robust 5-year GH therapy response (median height gain 1.4 vs 2.3 [GHD group] and 1.8 [no genetic etiology]; P = .001).ConclusionThe population of children clinically diagnosed with GHD is genetically heterogeneous. A statistically significant proportion harbors defects in RAS-MAPK signaling or growth plate disorders, highlighting the limitations of current methods of GHD diagnostics. The diagnosis of growth hormone deficiency (GHD) relies on a combination of auxological assessment, biochemical testing, and radiological imaging. In children with suspected GHD, GH stimulation tests are performed to assess the secretory capacity of the pituitary gland. Currently, these tests remain the gold standard for confirming the diagnosis [1, 2]. However, a major limitation of this diagnostic approach lies in the limited reliability of GH stimulation testing [3, 4]. The results are highly variable, influenced by numerous physiological and technical factors [5-8], and may yield false-positive results [9]. Consequently, concerns have been raised regarding the validity of current GHD diagnostic criteria [4]. It has been proposed that children clinically diagnosed with GHD represent an etiologically heterogeneous group with diverse underlying causes of growth failure, many of which are unrelated to GH secretion or function [10]. This presumption lacks comprehensive evaluation in clinical studies. Although genetic approaches have demonstrated strong potential in clarifying the causes of short stature [11-16], studies focused on GHD cohorts have largely targeted genes involved in GH production or secretion [17-20], potentially overlooking other etiologies of impaired growth. The aim of this study was to comprehensively evaluate the genetic etiology of short stature in children clinically diagnosed with GHD and to assess correlations between genetic findings and clinical phenotypes including response to GH treatment. Materials and methods Study population Children with a clinical diagnosis of primary GHD were retrospectively identified from the institutional database of children treated with GH at the Centre of Pediatric Endocrinology of Motol University Hospital, Prague, Czech Republic (database search performed in January 2024). Children were enrolled if they met the following criteria: 1) confirmed diagnosis of GHD (details discussed later); 2) absence of secondary causes of GHD (eg, brain tumor, neurosurgical intervention, cranial irradiation); 3) continuous GH treatment at our center for a minimum of 5 years; and 4) written informed consent for participation in the study (including genetic testing) provided by parents or legal guardians. The study protocol was approved by the institutional ethics committee of the Motol University Hospital and the Second Faculty of Medicine, Charles University, Prague (No. EK-753.3.5.21). Clinical evaluation Data regarding birth parameters, medical history, previous height development, GH treatment, and results of laboratory and radiological examinations were obtained from the medical records. Parental height was measured using a wall-mounted stadiometer to the nearest 1 mm. All anthropometric data were standardized using recent normative values [21, 22]. The clinical diagnosis of GHD was established in accordance with current international guidelines [1, 2]. GH stimulation testing was performed in children presenting with auxological and/or clinical features suggestive of GHD (height < −3 SD; and/or height < −2 SD with a decrease of height SD score [SDS] of >0.5 over 1 year in children age >2 years; and/or height <1.5 SD below midparent height; and/or height velocity < −2 SD sustained over 1 year or <−1.5 SD sustained over 2 years; and/or other pituitary hormone deficiency) [2] and a serum insulin-like growth factor type 1 (IGF-1) concentration below 0 SD for age and sex. Serum IGF-1 and GH concentrations were measured using automated immunoassays (Immunodiagnostic Systems; IGF-1: catalog No. IS-3900, RRID:AB_2861357; GH: catalog No. IS-3700, RRID:AB_2861356). GHD was confirmed in children with a peak GH concentration less than 10 μg/L in both clonidine and insulin tolerance tests. Sex-steroid priming (estradiol 2 mg administered 2 evenings preceding the test) was administered to all prepubertal children aged 9 years or older prior to stimulation testing. Genetic testing Genomic DNA was extracted from peripheral blood (QIAamp Blood Mini Kit, Qiagen). All girls had karyotype examination to exclude Turner syndrome. Children with a clinical phenotype suggestive of a specific genetic disorder underwent targeted genetic testing via Sanger sequencing or array comparative genomic hybridization. In children without identified causal variant in first-tier testing, or in cases in which no specific genetic disorder was suspected based on phenotype, next-generation sequencing (NGS) was performed using a custom-designed targeted panel comprising 398 growth-associated genes (for information about specific families, see supplementary data provided in the online repository [23]). All variants identified by NGS were confirmed by Sanger sequencing, as previously described [24]. Variant filtering was based primarily on low or absent frequency in control databases (eg, GnomAD, ExAc, and 1000 genomes). Functional impact was predicted using multiple in silico tools, including REVEL, CADD, DANN, SIFT, PolyPhen2, MutationTaster, and SpliceAI. The localization of variants within functionally relevant protein domains was also considered. Copy number variants were analyzed from NGS data using DECon software [25]. All variants with potential clinical importance were evaluated by American College of Medical Genetics and Genomics standards and guidelines [26]. Variant interpretation was supported by Franklin software (https://franklin.genoox.com version date May 3, 2024), which incorporates American College of Medical Genetics and Genomics criteria with rule strength (very strong, strong, moderate, or supporting). When applicable, rule strength was modified based on extended database investigation and detailed clinical evaluation. Familial segregation analysis was performed obtaining DNA samples and height data from available family members. The guidelines formulated by Jarvik and Browning [27] were followed to incorporate cosegregation data in the pathogenicity classification. In the end, all genetic variants were classified as pathogenic (P), likely pathogenic (LP), benign (B), likely benign (LB), or as variants of uncertain significance (VUS). Statistical evaluation Based on the analysis of the genes in which causal variants were found, children were stratified into 3 groups: those with genetic variants causing GHD (GHD group), variants accounting for short stature due to an alternative genetic etiology (alternative etiology group), and children with no causal genetic variant identified (no genetic etiology group). In addition, the children were stratified into 4 groups based on their peak stimulated GH concentration (<3.0 μg/L, 3.0-4.9 μg/L, 5.0-6.9 μg/L, and 7.0-9.9 μg/L). Clinical parameters were compared among groups that were formed based on genetic examination and peak stimulated GH concentrations. Comparisons among the groups were performed using the Kruskal-Walis test for continuous variables, and categorical variables were compared using the Pearson chi-square test. To identify differences between specific pairs of groups, post hoc analyses were conducted. For continuous variables, Dunn's test with Bonferroni correction for multiple comparisons was applied. For categorical variables, pairwise comparisons were performed using the Fisher exact test with Bonferroni correction. P values less than .05 were considered statistically significant. Results Of the 259 children diagnosed with primary GHD who received GH therapy at our center for a minimum duration of 5 years, parents/legal guardians of 233 (72 girls) provided informed consent for genetic testing, and these children were subsequently enrolled in the study. The characteristics of the study cohort are summarized in Table 1. Table 1Characteristics of the study cohortBirth weight adjusted for gestational age (SD)−0.8 (−1.6 to −0.2)Birth length adjusted for gestational age (SD)−1.5 (−2.1 to −0.9)Neonatal jaundice with phototherapy (n; %)43 (18%)Neonatal hypoglycemia (n; %)26 (11%)Cryptorchidism (n; % of boys)22 (14%)Age at GH therapy initiation, y4.7 (3.4 to 6.5)Height at GH therapy initiation (SD)−3.0 (−3.3 to −2.6)Pretreatment IGF-1 below detection limit 15 μg/L (n; %)25 (11%)Pretreatment IGF-1a (SD)−1.7 (−1.9 to −1.3)Peak stimulated GH, μg/L5.7 (3.9 to 7.4)Combined pituitary hormone deficiency (n; %)30 (13%)Central adrenal insufficiency (n; %)27 (12%)Central hypothyroidism (n; %)25 (11%)Hypogonadotropic hypogonadism (n; %)6 (3%)AVP deficiency (n; %)2 (1%)MRI performed (n; %)214 (92%)MRI midline abnormality (n; %b)69 (30%)Results are expressed as medians and interquartile ranges or as number and percentage of children.Abbreviations: AVP, arginine vasopressin; GH, growth hormone; IGF-1, insulin-like growth factor type 1; MRI, magnetic resonance imaging; n, number of children. a Only children with detectable pretreatment IGF-1 concentrations were included in the analysis. b Percentage of children with MRI performed. Characteristics of the study cohort Results are expressed as medians and interquartile ranges or as number and percentage of children. Abbreviations: AVP, arginine vasopressin; GH, growth hormone; IGF-1, insulin-like growth factor type 1; MRI, magnetic resonance imaging; n, number of children. a Only children with detectable pretreatment IGF-1 concentrations were included in the analysis. b Percentage of children with MRI performed. A genetic cause of short stature was identified in 39 out of 233 children (17%). In 13 of these children (33%), the genetic finding was consistent with a diagnosis of GHD. In the remaining 26 children (67%), the analysis revealed an alternative etiology of growth disorder. Specifically, a primary growth plate disorders were found in 8 children (21%), disruption of RAS-MAPK signaling pathway was detected in 13 children (33%), and miscellaneous genetic causes were identified in 5 children (13%). The identified genes associated with these diagnoses are listed in Fig. 1 and Table 2; specific genetic findings are summarized in data provided in an online repository as supplemental materials [12, 20, 23, 28-33] . Genetic etiology of growth disorders in children clinically diagnosed with growth hormone deficiency. Table 2Genetic etiology of a growth disorders stratified by peak stimulated growth hormone concentrationGHmax, μg/LGH deficiencyGrowth plate disorderRASopathyMiscellaneous13/39 (33%)8/39 (21%)13/39 (33%)5/39 (13%)<3.0 GLI2 — NF1 PMM2 OTX2 (2) LMNA PROP1 TBX3 POU1F1 GHSR KMT2D 3.0-4.9 — COL2A1 PTPN11 (2) RAI1 NPR2 (3) NF1 CDC42 FGFR3 EXT2 5.0-6.9 OTX2 FGFR3 PTPN11 (2) — SALL4 ACAN RAF1 CHD7 GNAO1 7.0-9.9 GH1 — PTPN11 (3) HMGA2 SOS1 (2) SPRED1 Abbreviations: GH, growth hormone; GHmax, peak stimulated GH concentration. Genetic etiology of a growth disorders stratified by peak stimulated growth hormone concentration Abbreviations: GH, growth hormone; GHmax, peak stimulated GH concentration. Analysis of clinical features revealed several differences among the evaluated groups. Specifically, the GHD group exhibited lower maximal stimulated GH concentrations (median 1.8 μg/L [interquartile range (IQR) 1.0-6.3 μg/L] vs 4.8 μg/L [4.0-7.1 μg/L]) compared to the alternative etiology group and 5.9 μg/L (4.1-7.6 μg/L) and the no genetic etiology group (P = .017). However, post hoc analysis indicated that the difference between the GHD group and alternative etiology group did not reach statistical significance. Furthermore, the GHD group showed the highest prevalence of combined pituitary hormone deficiency (CPHD; 46% vs 4% and 12%; P = .001) and magnetic resonance imaging (MRI) midbrain abnormities (73% vs 14% and 32%; P = .003). Importantly, children in the alternative etiology group demonstrated the least robust response to GH treatment. Despite receiving the highest GH doses, their median height gain after 5 years of therapy was only 1.4 SD (1.0-2.0 SD), compared to 2.3 SD (1.8-3.9 SD) in the GHD group and 1.8 SD (1.4-2.2 SD) in the no genetic etiology group (P = .001). Consequently, the alternative etiology group reached the lowest height SDS after 5 years of therapy (−1.8 SD [−2.4 to −1.1 SD] vs −0.8 SD [−2.0 to −0.4 SD] in the GHD group and −1.1 SD [−1.6 to −0.7 SD] in the no genetic etiology group; P = .006). A detailed comparison of the groups is provided in Table 3. Table 3Comparison between children with genetically proven growth hormone deficiency, those with different etiology of short stature elucidated, and those with no genetic etiology foundGHD group (n = 13)Alternative etiology group (n = 26)No genetic etiology group (n = 194) P Boys8 (62%)18 (69%)135 (70%).83Midparental height (SD)0.4 (−1.9 to +1.1)−1.0 (−1.5 to −0.3)−0.8 (−1.6 to +0.1).22Birth weight adjusted for gestational age (SD)−0.8 (−1.5 to −0.3)−0.8 (−1.7 to +0.4)−0.8 (−1.6 to −0.2).84Birth length adjusted for gestational age (SD)−1.2 (−1.4 to −0.9)−1.3 (−2.3 to −0.5)−1.5 (−2.2 to −0.9).32Neonatal jaundice requiring phototherapy4 (31%)7 (27%)32 (16%).52Neonatal hypoglycemia2 (15%)2 (8%)22 (11%).76Cryptorchidism2 (25%)7 (39%)13 (10%).06Pretreatment IGF-1 below detection limit 15 μg/L3 (23%)5 (19%)17 (9%).09Pretreatment IGF-1 (SD)−1.8 (−2.0 to −1.6)−1.6 (−2.1 to −1.2)−1.6 (−1.9 to −1.3).50Peak stimulated GH, μg/L1.8 (1.0 to 6.3)a4.8 (4.0 to 7.1)a,b5.9 (4.1 to 7.6)b .017 Combined pituitary hormone deficiency6 (46%)a1 (4%)b23 (12%)b .001 MRI performed11 (85%)22 (85%)181 (93%).2MRI midline abnormality8/11 (73%)a3/22 (14%)b58/181 (32%)b .003 Age at GH treatment initiation, y3.7 (2.3 to 6.7)4.0 (2.7 to 6.6)4.8 (3.5 to 6.5).24Height at GH treatment initiation (SD)−3.4 (−4.4 to −2.8)−3.0 (−3.3 to −2.8)−2.9 (−3.2 to −2.6).14GH dose at GH treatment initiation, μg/kg/d27 (25 to 32)29 (26 to 32)28 (25 to 30).20Height after 1 y of GH therapy (SD)−2.0 (−3.4 to −1.6)−2.4 (−3.1 to −2.0)−2.2 (−2.6 to −1.9).42Height gain after 1 y of GH therapy (SD)0.9 (0.7 to 1.4)0.7 (0.5 to 0.9)0.7 (0.5 to 0.9).06GH dose after 1 y of GH therapy, μg/kg/d25 (23 to 32)32 (29 to 35)31 (28 to 33).07Height after 5 y of GH therapy (SD)−0.8 (−2.0 to −0.4)a−1.8 (−2.4 to −1.1)b−1.1 (−1.6 to −0.7)a .006 Height gain after 5 y of GH therapy (SD)2.3 (1.8 to 3.9)a1.4 (1.0 to 2.0)b1.8 (1.4 to 2.2)a .001 GH dose after 5 y of GH therapy, μg/kg/d31 (28 to 34)a35 (32 to 43)b32 (29 to 34)a .001 Continuous variables are expressed as medians (interquartile ranges). P values represent the overall difference among groups calculated by Kruskal-Wallis test. Values in bold signify statistical significance.Abbreviations: GH, growth hormone; GHD, growth hormone deficiency; IGF-1, insulin like growth factor type 1; MRI, magnetic resonance imaging. a,b Different superscript letters indicate a statistically significant difference (P < .05) between groups based on Dunn's post hoc test with Bonferroni correction for multiple comparisons. Groups sharing a common superscript letter (a,b) are not statistically different. Categorical variables are expressed as numbers (percentages). P values represent the overall difference among groups calculated using the Pearson chi-square test. a,b Different superscript letters indicate a statistically significant difference (P < .05) based on Fisher exact tests with Bonferroni correction for multiple comparisons. Comparison between children with genetically proven growth hormone deficiency, those with different etiology of short stature elucidated, and those with no genetic etiology found Continuous variables are expressed as medians (interquartile ranges). P values represent the overall difference among groups calculated by Kruskal-Wallis test. Values in bold signify statistical significance. Abbreviations: GH, growth hormone; GHD, growth hormone deficiency; IGF-1, insulin like growth factor type 1; MRI, magnetic resonance imaging. a,b Different superscript letters indicate a statistically significant difference (P < .05) between groups based on Dunn's post hoc test with Bonferroni correction for multiple comparisons. Groups sharing a common superscript letter (a,b) are not statistically different. Categorical variables are expressed as numbers (percentages). P values represent the overall difference among groups calculated using the Pearson chi-square test. a,b Different superscript letters indicate a statistically significant difference (P < .05) based on Fisher exact tests with Bonferroni correction for multiple comparisons. Among the children enrolled in the study, peak stimulated GH concentrations were distributed as follows: less than 3.0 μg/L (n = 41), 3.0 to 4.9 μg/L (n = 55), 5.0 to 6.9 μg/L (n = 62), and 7.0 to 9.9 μg/L (n = 75). The group with a peak GH less than 3.0 μg/L differed significantly from those with higher peaks across multiple clinical parameters. Specifically, this group presented with higher birth length SD, lower pretreatment IGF-1 concentrations, and a higher prevalence of neonatal hypoglycemia, CPHD, and MRI-detected midline anomalies. Furthermore, these children demonstrated a superior response to GH treatment despite receiving lower GH doses, and their clinical diagnosis of GHD was more frequently confirmed genetically. Importantly, no statistically significant differences regarding clinical parameters, treatment response, or genetic testing results were observed among the different groups with peak GH concentrations greater than 3.0 μg/L. Detailed data are available in Table 4. Table 4Comparison between children with different peak stimulated growth hormoneGH peak, μg/L<3.0 (n = 41)3.0-4.9 (n = 55)5.0-6.9 (n = 62)7.0-9.9 (n = 75) P Midparent height (SD)−0.1 (−1.2 to +0.5)a−1.2 (−1.6 to −0.4)b−1.1 (−1.8 to −0.2)b−0.8 (−1.5 to −0.2)a,b .02 Birth weight (SD)−0.3 (−1.3 to +0.1)−0.9 (−1.6 to −0.4)−0.8 (−1.7 to −0.1)−0.9 (−1.7 to −0.3).29Birth length (SD)−1.3 (−1.4 to −0.5)a−1.7 (−2.4 to −1.1)b−1.3 (−2.0 to −0.9)a,b−1.6 (−2.2 to −1.1)b .013 Neonatal jaundice requiring phototherapy13 (32%)10 (18%)9 (15%)11 (15%).18Neonatal hypoglycemia13 (32%)a,b6 (11%)a,b3 (5%)b4 (5%)b <.001 Cryptorchidism10 (30%)a6 (14%)a3 (7%)a3 (7%)a.031IGF-1 3.0 μg/L exhibited no significant phenotypic or therapeutic differences. Adopting the stricter 7.0 μg/L threshold recommended by recent guidelines [1] would substantially (by 32% in our study cohort) reduce the number of children diagnosed with GHD. However, given the lack of distinction between these children and those with peak stimulated GH concentration of 3.0 to 6.9 μg/L, our data suggest that this adjustment would not optimize diagnostic accuracy or GH treatment efficacy. Given that idiopathic short stature is not currently an approved indication for GH treatment in Europe, the diagnosis of GHD carries considerable importance, often serving as the sole access to the therapy. Importantly, a presumed false-positive GHD diagnosis does not preclude a therapeutic benefit. Although children with an identified alternative genetic etiology exhibited a less robust response to GH therapy despite receiving higher GH doses, they achieved substantial growth improvement (a median of 1.4 SD after 5 years). Consequently, refining the current indication criteria for GH therapy represents a major challenge for contemporary pediatric endocrinology. We suggest that until comprehensive data regarding the efficacy of GH in specific genetic disorders become available, the debate regarding the approval of idiopathic short stature as an indication in Europe warrants reinitiation. Recent international guidelines for the genetic evaluation of short stature [66] recommend a 2-tier approach for children with severe GHD: an initial targeted panel for GHD-causing genes, followed by an expanded evaluation if first-line testing is negative. Our data support this strategy, as 70% (7/10) of the children in our cohort with a peak GH < 3.0 μg/L and an identified genetic etiology harbored variants directly associated with GHD. Although current guidelines do not recommend routine genetic testing in children with a peak GH > 3.0 μg/L in the absence of additional phenotypic features, our findings indicate that a genetic etiology of short stature can still be identified in a considerable proportion of this group at a rate comparable to that of children with isolated short stature [67]. Given the predominance of variants in genes not affecting GH secretion in this cohort, a comprehensive approach encompassing a broader array of short stature genes should be implemented from the outset. Our study has several strengths. Most important, the relatively large cohort size, combined with a high enrollment rate, mitigates the risk of potential selection bias and enhances the robustness of our findings. However, we must acknowledge several limitations. First, functional validations of identified novel variants were not performed. Second, noncoding variants (except for intron-exon boundaries) were not captured by NGS. Finally, although our NGS panel covered a comprehensive set of 398 growth disorder–associated genes, causative variants in genes not included in this panel could have remained undetected. Detailed anthropometric measurements, including the evaluation of body proportionality or facial recognition software that could potentially help optimize phenotypic evaluation and candidate gene selection, were not used. Finally, although a comprehensive clinical evaluation was regularly performed in all patients, we recognize that IGF-1 levels can be influenced by non–GH-dependent factors such as nutritional status or illness. To conclude, the clinical diagnosis of GHD frequently masks a wide spectrum of underlying genetic diagnoses, ranging from typical GHD genes to primary growth plate disorders and disruption of fundamental intracellular processes, particularly the RAS-MAPK signaling pathway. Our results highlight the limitations of current GH stimulation tests and support the integration of broad NGS panels into the genetic investigation of children diagnosed with GHD. To improve the cost-effectiveness and accessibility of genetic testing, further studies are needed to identify specific clinical features predictive of a monogenic etiology in children with diagnosed GHD.