Growth characteristics of children with 21-hydroxylase deficiency and the value of steroid hormones in height assessment

Abstract

Background

Impaired height is a common complication of 21-hydroxylase deficiency (21OHD), yet sensitive monitoring indicators remain limited. This study aims to elucidate growth characteristics and identify effective monitoring parameters for 21OHD children.

Methods

Cross-sectional data from 111 patients were categorized into four groups based on age and developmental stage: 0 – 2 years, 2 years old to pre-pubertal initiation, puberty initiation to pre-epiphyseal closure, and post-epiphyseal closure, named groups A to D, respectively. Each group was further stratified by phenotype and sex. Height standard deviation scores (HSDS), corrected for bone age (BA) and target height (HSDS – THSDS, H_BASDS – THSDS), were calculated. Steroid hormone levels and hydrocortisone (HC) doses were analyzed using statistical models to identify factors influencing height.

Results

The medians of HSDS – THSDS were > 0 in all subgroups of Group A. The medians of H_BASDS – THSDS were < 0 in all subgroups of Group B, and 17-hydroxyprogesterone (17OHP) and HC dose significantly positively influenced BA advancement. BA of patients in Group C was older than the calendar age(CA), while the medians of H_BASDS – THSDS in all subgroups except the non-classic females were all < 0, and 17OHP, 21-deoxycortisol(21DOF), and 11-oxy-androgen were significant influencing factors. The medians of final height (FHSDS – THSDS) of all subgroups in Group D were < 0, males with classic 21OHD significantly lower than females.

Conclusions

21OHD children exhibit accelerated bone maturation as early as childhood, worsening during adolescence, leading to severely impaired growth potential and final height. 17OHP, 21DOF, and 11-oxy-androgens are promising biomarkers for evaluating growth and bone maturity.

Background

21OHD is the most common type of congenital adrenocortical hyperplasia (CAH) and is caused by defects in the 21-hydroxylase enzyme encoded by the CYP21 A2 gene resulting in impaired cortisol synthesis. This disrupts the negative feedback loop to the hypothalamus and pituitary, causing excessive adrenocorticotropic hormone (ACTH) secretion, adrenal cortical hyperplasia, and overproduction of androgens. Classic 21OHD is categorized into salt-wasting (SW) and simple virilizing (SV) phenotypes, while the non-classic (NC) form presents with milder symptoms.

Chronic ACTH stimulation and steroid precursor accumulation elevate the activities of 17α-Hydroxylase/17,20-lyase (CYP17 A1, P450c17) and 11β-hydroxylase (CYP11B1, P450c11β). This drives the conversion of 17-hydroxyprogesterone (17OHP) to androstenedione (AD), which is further metabolized by CYP11B1 to 11-hydroxyandrostenedione (11OHAD), activating the 11-oxy-androgen pathway, generating 11 keto-androstenedione (11KAD), 11OH-testosterone (11OHT) and 11keto-testosterone (11 KT). Additionally, excess 17OHP is converted to 21-deoxycortisol (21DOF), a potential diagnostic marker for 21OHD, and a minor fraction is metabolized via the"backdoor"pathway to produce dihydrotestosterone (DHT), albeit with limited contribution [1, 2].

Treatment mainly consists of lifelong glucocorticoid replacement therapy to prevent adrenal crisis and to reduce androgen production. Several studies have consistently shown that 21OHD patients exhibit impaired FH [3,4,5,6], underscoring the importance of regular height and steroid hormone monitoring. While 17OHP, AD, and testosterone (T) are traditional biomarkers, which can fluctuate with glucocorticoid dosage and duration and lack adrenal specificity. In contrast, 11-oxy-androgens are strongly associated with disease control and specific comorbidities [7]. Combining traditional with adrenal-specific biomarkers might be an optimal monitoring protocol for comprehensive assessment of growth and skeletal maturation.

In this study, we analyzed the height growth and steroid hormone data of 21OHD patients to investigate the growth characteristics, identify key factors contributing to height impairment, and explore potential steroid hormone markers.

Materials and methods

We collected data from 111 children diagnosed with 21OHD based on clinical and biochemical criteria and/or newborn screening. The dataset included initial diagnostic information, CYP21 A2 gene variants (if available), height, Tanner stage, parental height, BA (for children over 2 years), glucocorticoid dosage, and steroid hormone levels at the latest follow-up.

Data collection and definition

• 1. Basic information: Sex, calendar age (CA)

• 2. Height assessment: Height standard deviation score (HSDS) was calculated using data from the Chinese Children's Physical Development Survey published in 2009 [8].

• 3. Puberty initiation: Tanner stage M2 in girls and a testicular volume of ≥ 4 mL in boys.

• 4. Bone age (BA): Assessed using the RUS-CHN method (Chinese 05 Standard) [9] from left hand and wrist radiographs. BA-corrected height (H_BASDS) and BA – CA (bone age advancement) were calculated.

• 5. Final height (FH): Defined as growth < 1 cm over 12 months or epiphyseal fusion (BA > 18 years). FHSDS was calculated using 2009 published data [8]. Males: FHSDS = (FH – 172.7)/6.1. Females: FHSDS = (FH – 160.6)/5.6.

• 6. Target height (TH): TH was calculated as median parental height ± 6.5 cm (boys/girls) [10]. THSDS was derived using 2009 survey data [8]. Males: FHSDS = (FH – 172.7)/6.1. Females: FHSDS = (FH – 160.6)/5.6. In the analyses, all height data were corrected for TH: HSDS – THSDS, H_BASDS – THSDS, and FHSDS – THSDS.

• 7. Height classification (HSDS'): Defined as HSDS – THSDS (< 2 years) or H_BASDS – THSDS (≥ 2 years), categorized as: Short: HSDS'< −2. Slightly short: −2 ≤ HSDS'< 0. Normal: HSDS'≥ 0 (included 4 cases with HSDS'≥ 2, (range 2.00 to 2.72)).

• 8. Glucocorticoid doses: Hydrocortisone (HC) was administered at 10 – 15 mg/m²/day for patients with open epiphyses. Some patients achieving FH were treated with dexamethasone for convenience, which was converted to hydrocortisone according to a 1:50 equivalence.

• 9. Steroid hormone measurements. Fasting blood samples collected at 8:00 AM (pre-medication) were analyzed for 17OHP, AD, 21DOF, 11KAD, 11OHT, and 11KT using LC–MS/MS at Guangzhou Golden Field Medical Laboratory.

• 10. Grouping: Participants were stratified by age and pubertal status: Group A: 0 – 2 years old. Group B: 2 years to pre-pubertal. Group C: puberty initiation to pre-epiphyseal closure. Group D: post-epiphyseal closure. Each group was further stratified by phenotype and sex.

Statistical analysis

Descriptive analyses were performed to test the normality of the data, and the mean ± SD was used to describe the data if normality was satisfied. Because of the small numbers in the subgroups by gender and phenotype, the data were presented as medians with ranges. Mann–Whitney U tests and T-tests were used among groups and Kruskal -Wallis Test within groups to test differences. The effects of steroid hormones (17OHP, AD, T, 21DOF, 11OHT, 11KAD and 11KT) and HC doses on HSDS – THSDS, H_BASDS – THSDS, and BA – CA were assessed using univariate and multivariate linear regression analysis. Multifactorial ordered logistic regression analysis was used to evaluate the effect of steroid hormones and HC dose on HSDS'. P-values < 0.05 are considered statistically significant. All statistical analyses were performed using IBM SPSS Statistics for Windows, version 22.0 (IBMCorp., Armonk, N.Y., USA).

Results

Patient characteristics

A total of 111 patients (62 males and 49 females) with 21OHD were enrolled in this study, 13 cases were categorized in Group A (SW-male [n = 8], SW-female [n = 4], and SV-male [n = 1]), 49 cases were classified into Group B (SW-male [n = 28], SW-female [n = 15], SV- male[n = 3], SV-female[n = 2], NC-male[n = 1]), 30 cases belonged to Group C (SW-male[n = 8], SW-female[n = 5], SV-male[n = 5], SV-female[n = 5], NC-male[n = 5], NC-female[n = 2]) and Group D contain the remaining 19 cases (SW-male[n = 1], SW-female[n = 3], SV-male[n = 2], SV-female[n = 9], NC-female[n = 4]). Steroid hormone levels for each subgroup are shown in the Supplementary material Table S1.

Height data

Group A analysis

HSDS – THSDS values: SW-males 1.00 (−0.79 to 3.22), SW-females 0.59 (−0.51 to 1.37), SV-male −0.28; overall 0.69 (−0.26,1.43) (Table 1).

Table 1 Height and BA-CA for Groups A to D

Group B analysis

• HSDS – THSDS: SW-males −0.24 (−1.83 to 2.56), SW-females 0.05 (−1.96 to 2.92), SV-males 1.77 (0.66 to 4.23), SV-females −0.50 (−0.60 to − 0.40), overall −0.19(−1.20,0.78) (Table 1).

• All subgroups showed H_BASDS – THSDS medians < 0 and BA – CA medians > 0 (Table 1).

• No significant differences in HSDS – THSDS (P = 0.220), H_BASDS – THSDS (P = 0.679), or BA – CA (P = 0.348) distributions (Table 1).

• Multivariate analysis confirmed 17OHP (b = 0.014, P = 0.003) and HC (b = 0.104, P = 0.037) as positive from 2 years to pre-pubertal BA – CA predictors, with 17OHP having a stronger effect (0.401 vs. 0.274) (Table 2, Table S2).

Group C analysis

• Median HSDS – THSDS > 0 across subgroups; no distribution differences (P = 0.169).

• H_BASDS – THSDS medians < 0 in all but NC-females; SW-males significantly lower than Group B (P = 0.009).

• BA – CA > 0 (0.19 to 6.44) across all patients; SW-males and SW-females significantly higher than Group B (both P < 0.05) (Table 1).

• Univariate and multivariate analysis for SW-males from age 2 years to pre-epiphysis closure identified 11OHT (b = −4.731, P = 0.042) was the sole significant H_BASDS-THSDS influencer (Table 2, Table S3).

• Multivariate analysis for SW-males highlighted 17OHP (b = 0.032, P < 0.001) and 21DOF (b = −0.096, P = 0.023) as significant factors, with 17OHP having a stronger effect (0.825 vs. −0.423) (Table 2, Table S4).

• Multivariate analysis for SW-females highlighted 11 KT (b = 0.456, P = 0.007) was the sole significant BA-CA factor (Table 2, Table S5).

Table 2 Summary of Results of Multivariate Analysis of Height and Steroid Hormones and HC Dose

Group D analysis

• FHSDS – THSDS medians < 0 across all subgroups; SW-males worst (−4.84), followed by SV-males (−2.82; −4.03 to −1.61) and SW-females (−2.01; −2.69 to −0.48) (Table 1).

• No significant distribution differences among subgroups (P = 0.062) (Fig. 1A).

• Classic 21OHD FHSDS – THSDS: −1.48 ± 1.54 (SW: −2.50 ± 1.81; SV: −1.11 ± 1.34); non-classic: −0.03 ± 0.96. No significant difference between Classic and non-classic (P = 0.094), and between SW and SV (P = 0.216) (Fig. 1B-C).

• Males with classic 21OHD significantly lower than females (−3.49 ± 1.68 vs. −0.98 ± 1.07, P = 0.006) (Fig. 1D).

Fig. 1

Comparison of FHSDS – THSDS among subgroups. A: FHSDS – THSDS by subgroup. SW-male (n = 1): −4.84, SW-female (n = 3): −2.01 (−2.69 to − 0.48), SV-male (n = 2): −2.82 (−4.03 to −1.61), SV-female (n = 9): −0.79 (-2.43 to −0.80), NC-female (n = 4): −0.40(−0.70 to −1.39). No significant difference in overall distribution among subgroups (P = 0.062) B: Classic vs. Non-classic 21OHD. Classic 21OHD(n = 15): −1.48 ± 1.54; Non-classic 21OHD (n = 4): −0.03 ± 0.96. No statistically significant difference (P = 0.094). C: SW phenotype vs. SV phenotype. SW phenotype(n = 4): −2.50 ± 1.81; SV phenotype (n = 11): −1.11 ± 1.34, No statistically significant difference (P = 0.216). D: Male vs. Female Classic 21OHD. Males (n = 3): −3.49 ± 1.68;Females (n = 12): −0.98 ± 1.07. FHSDS – THSDS was significantly lower in males than in females (P = 0.006)

Comparative Analysis of Groups B and C

• Group C HSDS – THSDS (1.43(0.34,2.38) vs. −0.19(−1.20,0.78), P < 0.001) and BA – CA (2.66 (1.89,4.19) vs. 0.28(−0.32,1.00), P < 0.001) significantly higher than Group B; no H_BASDS – THSDS difference (−1.21(−2.29, −0.46) vs. −0.82(−1.70,0.29), P = 1.00).

• Multivariate analysis confirmed AD (b = 0.278, P = 0.001) as the sole significant HSDS – THSDS predictor (Table 2, Table S6).

• Multivariate analysis confirmed 17OHP (b = 0.016, P < 0.001) as the sole significant BA – CA factor (Table 2, Table S7).

Height classification analysis

Our study revealed distinct height distribution patterns among 111 children: 58 (52.25%) were short, 23 (20.72%) slightly short, and 30 (27.03%) normal. Ordered logistic regression analysis demonstrated significant associations between steroid metabolites and height status: 17OHP (OR = 0.981) and 11KAD (OR = 0.136) showed negative correlations with HSDS', while 21DOF (OR = 1.150) exhibited a positive correlation. Specifically, each 1 ng/ml increase in 17OHP and 11KAD reduced the probability of higher HSDS' classification by 1.9% and 86.4%, respectively, whereas each 1 ng/ml increase in 21DOF increased this probability by 15.0% (Table 3).

Table 3 Ordered multiple logistic regression of factors influencing height classification HSDS'

Steroid hormones and HC correlation analysis

Correlations between steroid hormone concentrations of 17OHP, AD, T, 21DOF, 11KAD, 11KT, 11OHT, and HC doses were analyzed (Fig. 2). The correlations between steroid hormones were moderate or strong (r' range 0.43 to 0.91, all P < 0.001), while HC doses were only weakly correlated with AD (r'= 0.26, P < 0.01).

Fig. 2

Steroid hormones and HC correlation heat map. There was a moderate or strong correlation between 17OHP, AD, T, 11KAD, 11OHT, and 11KT (r' range 0.43 to 0.91, all P < 0.001), and only a weak correlation between HC dose and AD (r'= 0.26, P < 0.01). *P ≤ 0.05 **P ≤ 0.01 ***P ≤ 0.001

Discussion

This single-center study analyzed 111 21OHD patients across diverse ages and pubertal stages, examining steroid hormone profiles, glucocorticoid dosage, and growth parameters. While Wasniewska et al. [11] concluded that FH and growth trajectory in non-classic 21OHD patients were influenced by the severity of phenotype at diagnosis, their study lacked in-depth analysis of steroid hormones' impact on height and bone age progression, limiting its clinical utility for treatment monitoring and regimen optimization. Distinct from previous age-stratified approaches [5, 6], our study implemented a dual stratification system incorporating both chronological age and developmental status, enabling more precise characterization of growth patterns and their determinants in 21OHD patients.

Adolescent patients exhibited significantly higher HSDS – THSDS and exacerbated BA – CA progression compared to preadolescents. Multivariate analyses demonstrated positive effect between HSDS – THSDS and 17OHP, and between H_BASDS – THSDS and AD, indicating that inadequate steroid hormone control drives apparent accelerated growth at the expense of premature growth potential depletion, ultimately impairing final height. These findings align with Matsubara et al. [12], who reported severe pubertal growth impairment in 21OHD patients. While standard growth charts fail to capture growth potential loss, BA and TH corrected height (H_BASDS – THSDS) provides a more accurate assessment of height status.

Notably, our study revealed the HSDS-THSDS could be elevated up to 3.22 in 0 – 2 years, contrasting with previous reports of unaffected growth in untreated infants in the first year of life when may be insensitive to androgens [13]. BA advancement begins in preadolescence, driven by adrenal-derived hormones prior to gonadal activation. Multivariate analysis identified 17OHP and HC as positive predictors of BA – CA, underscoring the importance of controlling BA progression, with 17OHP serving as a key monitoring biomarker.

After puberty initiation, BA – CA in SW-phenotype patients worsened, and the H_BASDS – THSDS significantly decreased in SW-males, indicating further reduction in growth potential and diminished pubertal growth, consistent with prior findings [14]. Multivariate linear regression analysis demonstrated that 17OHP and 11KT significantly and positively influenced BA – CA in SW-males and SW-females, respectively, while 11OHT negatively impacted the H_BASDS – THSDS in SW-males. Preventing further loss of growth potential is critical in adolescence, with 17OHP and 11-oxy-androgens serving as key biomarkers. Notably, 11KT and 11OHT are predominantly adrenal-derived, and 11KT activates the human androgen receptor with potency comparable to testosterone [15]. Previous studies have identified 11-oxy-androgens as the primary androgen source in patients on conventional oral glucocorticoid therapy [16], with reductions observed in modified-release hydrocortisone and continuous subcutaneous hydrocortisone infusion regimens [2, 17], but these treatments are in the clinical trial phase and lack data on long-term data on BA maturation and height in children.

A large meta-analysis revealed an FHSDS of −1.38 (range −1.56 to −1.20) and a TH-corrected FHSDS of −1.03 (range −1.20 to −0.86) in patients with classic 21OHD [18], and the FHSDS – THSDS of the classic 21OHD patients in our study was even lower at −1.48 ± 1.54 (range −4.84 to 0.80), potentially attributable to the absence of newborn screening and early treatment. Although not statistically significant, the impairment in FH was more pronounced in SW patients compared to SV patients. Non-classic patients generally fell within the normal population height range. However, no significant difference was observed between classic and non-classic patients, possibly due to the limited sample size. Interestingly, males with classic 21OHD had significantly lower FHSDS – THSDS than females. Previous studies have demonstrated that SW males with early pubertal onset tend to have the poorest lifelong height outcomes [19], and SV males experienced greater TH loss than SV females during puberty [20]. In addition a study found that newborn screening improved near-adult height in men with CAH [21], and none of the patients who reached lifetime height in this study had received newborn screening. However, some studies suggest that FH is independent of gender [12, 18]. Given the cross-sectional design and limited subgroup sample size in this study, further research is necessary to clarify the relationship between gender and height in 21OHD patients.

Ordered logistic regression identified 21DOF as a protective factor of height class, while multivariate linear regression analysis revealed its significant negative correlation with BA – CA in SW-males. In vitro experiments found 21DOF, corticosterone, 17OHP, and progesterone bound to the glucocorticoid receptor with an affinity of 24% ~ 43% of cortisol. Notably, 21DOF exhibited 8.5% − 17% of cortisol's transcriptional activation activity, substantially higher than progesterone and 17OHP (0.2% − 0.8%) [22, 23]. These findings suggest that 21DOF may exert clinically relevant glucocorticoid receptor agonistic effects, potentially mitigating cortisol deficiency and associated clinical phenotypes, including height impairment, in 21OHD patients.

Current guidelines recommend hydrocortisone (HC) dosing at 10 – 15 mg/m²/day for pediatric patients, with dose adjustments during puberty and caution advised when exceeding 17 mg/m²/day [14, 24]. In this study, patients with open epiphyses received HC within the recommended range. While glucocorticoid dosage did not demonstrate significant correlations with most steroid hormone levels, multivariate analysis revealed a significant positive effect between HC and BA – CA, and ordered logistic regression identified HC as a negative predictor for height classification, the cross-sectional study design and potential confounding factors, including treatment adherence, disease severity, and developmental stage — all of which may independently affect growth — complicate the interpretation of HC's impact on height. These findings underscore the difficulty and importance of maintaining a therapeutic balance between disease control and optimal growth in pediatric patients receiving glucocorticoid therapy.

Several small nonrandomized studies suggested growth hormone therapy alone or combined with gonadotropin-releasing hormone

agonist improves predicted adult height in classic 21OHD patients [25, 26]. However, our data indicated early skeletal maturation, begin in childhood, is the primary factor impairing height. This suggests therapies delaying epiphyseal maturation may be more effective. A 2-year randomized study (n = 28) demonstrated that aromatase inhibitors maintain normal growth and skeletal maturation despite elevated androgens [27], though final height outcomes and long-term safety require further investigation. Abiraterone acetate, a potent CYP17 A1 inhibitor [28], is undergoing clinical trials for prepubertal CAH treatment (NCT02574910) [29].

This study has the following limitations: First, the cross-sectional design lacks longitudinal data (such as annualized steroid hormone levels, pubertal onset timing, and body composition indicators), which may affect the control of confounding factors in growth assessment. Second, the insufficient sample size across subgroups (particularly in the extreme age groups) limits statistical power. Notably, 11KT exhibited paradoxical effects, accelerating BA advancement in SW females while exerting a protective effect on height classification. Additionally, the relationship between glucocorticoid dosage and height has not been fully elucidated. These findings highlight the need for large-scale prospective cohort studies with standardized data collection and long-term follow-up to clarify the interaction mechanisms between steroid hormones metabolism and growth.

Conclusions

Our findings demonstrate that 21OHD patients exhibit progressive skeletal maturation acceleration, initiating in childhood and exacerbating during puberty, ultimately leading to compromised FH. These results underscore the critical importance of controlling BA advancement as a primary therapeutic target. We propose an innovative monitoring protocol combining traditional 17OHP measurement with adrenal-specific biomarkers (21DOF and 11-oxy-androgens) for comprehensive assessment of growth and skeletal maturation. This biomarker-based approach facilitates targeted steroid hormone monitoring and timely treatment adjustments, potentially mitigating long-term growth impairment in 21OHD patients.