11-Oxygenated Androgens Useful in the Setting of Discrepant Conventional Biomarkers in 21-Hydroxylase Deficiency

Abstract

Context

Serum 17-hydroxyprogesterone (17OHP) and androstenedione (A4) are the conventional biomarkers used to assess disease control in patients with 21-hydroxylase deficiency (21OHD). However, discrepancy between the two is not uncommon, limiting interpretation.

Objective

To evaluate 11-oxyandrogens in discriminating good versus poor disease control in 21OHD in the setting of discrepant 17OHP and A4.

Methods

Retrospective analysis of 2738 laboratory assessments obtained as part of Natural History Study of congenital adrenal hyperplasia (CAH) at the National Institutes Health Clinical Center. Patients with discrepant 17OHP and A4 and available sera were selected. A 15-steroid mass-spectrometry panel was performed in sera from patients with 21OHD and age- and sex-matched controls. Patients were categorized in “good” or “poor” control based on clinical assessment (bone age advancement, signs and symptoms of precocious puberty, menstrual irregularity, hirsutism, or hypogonadotrophic hypogonadism).

Results

Discrepant 17OHP and A4 was found in 469 (17%) laboratory assessments. Of these, 403 (86%) had elevated 17OHP with A4 in reference range. Of 46 patients with available sera, 30 (65%) were in good control. Median fold elevation relative to controls was higher in patients with poor versus good control for 11-hydroxytestosterone (median [interquartile range], 2.82 [1.25-5.43] vs 0.91 [0.49- 2.07], P = .003), and 11-ketotestosterone (3.57 [2.11-7.41] vs 1.76 [1.24-4.00], P = .047). Fold elevation of 11-hydroxytestosterone between 3.48 (sensitivity 97%, specificity 47%) and 3.88 (sensitivity 100%, specificity 40%) provided the best discrimination between poor vs good control.

Conclusion

11-Oxyandrogens, especially 11-hydroxytestosterone, may be useful in the management of CAH when conventional biomarkers are inconclusive.

Congenital adrenal hyperplasia (CAH) is a group of disorders resulting from defects in adrenal steroidogenic pathways. The most common form is 21-hydroxylase deficiency (21OHD), with its classic form accounting for more than 95% of all cases and affecting 1 in 14 000 to 1 in 18 000 newborns [1]. The defect leads to excess accumulation of 17-hydroxyprogesterone (17OHP), the main substrate for CYP21A2, and androstenedione (A4), a weak androgen formed by diversion of excess 17OHP. Patients with 21OHD need lifetime glucocorticoid and mineralocorticoid replacement at doses adequate to treat adrenal insufficiency and to avoid excess accumulation of androgens, which are produced from accumulation of the precursor products 17OHP and A4 [2].

Serum 17OHP and A4 are conventional biomarkers for monitoring therapy in CAH with 17OHP used more commonly in the management of children. In general, there is good linear correlation between 17OHP and A4 (r = 0.7-0.8) [3, 4]. However, the use of these conventional biomarkers has several limitations. First, discrepancies between the 2 biomarkers with disproportionate elevations of 1 or the other are not uncommon in clinical practice. The interpretation of the 2 can often be contradictory in these situations, with 17OHP suggesting poor control but not A4 or vice versa. This situation is further compounded by lack of published literature regarding discrepancies between the 2 conventional biomarkers. Second, symptoms and signs of hyperandrogenism do not always correlate with levels of serum 17OHP and A4 as these features can be slow in onset and regression. Third, dehydroepiandrosterone sulfate, the predominant adrenal androgen precursor in healthy individuals is suppressed in patients with CAH for reasons that are not quite clear [5, 6]. Fourth, the acceptable range of these conventional biomarkers may differ from the assay-based reference range. A serum 17OHP level of 1200 ng/dL is often used as the upper limit of acceptable range [7, 8]. This is because attempts to normalize 17OHP levels, the immediate precursor to 21-hydroxylase activity, is likely to result in overtreatment with features of Cushing syndrome [1]. Moreover, when the hypothalamic–pituitary–adrenal axis is not suppressed, large excursions in serum 17OHP levels may occur promptly in response to stress, for example, venipuncture, which is particularly relevant in children [9]. 17OHP also varies with menstrual phase in females and timing of the sample in relation to the last dose of glucocorticoid [10, 11]. A4 is typically targeted at the upper limit of the age, sex, and Tanner stage–based reference range [12]. Some studies have suggested that serum A4 has higher sensitivity and specificity than serum 17OHP in monitoring therapy [13, 14]. This may be due to the smaller diurnal variation of A4 than 17OHP [15-18] and rapid changes in serum levels of 17OHP after oral or intramuscular stress dose [11, 17, 19].

11-Oxygenated androgens (11-oxyandrogens) are 19-carbon steroids that originate primarily in the adrenal gland. Whereas 11-hydroxyandrostenedione (11OHA4) and 11-hydroxytestosterone (11OHT) are direct adrenal products via 11β-hydroxylase (CYP11B1) activity (Fig. 1), the synthesis of 11-ketoandrostenedione (11KA4) and 11-ketotestosterone (11KT) from 11OHA4 occurs primarily in the kidneys by 11β-hydroxysteroid dehydrogenase isoenzyme 2 (HSD11B2). Studies have shown that 11KT is a potent agonist of the human androgen receptor, with maximum androgenic activity equivalent to testosterone [20, 21]. The adrenal origin of 11-oxyandrogens is consistent with several studies, which show similar levels in both sexes with no evidence of decline following menopause [22], rise in levels during normal and premature adrenarche [20], higher concentrations in the adrenal vein than in the periphery, rise in levels after adrenocorticotropin stimulation [23], and detection of only trace amounts in patients with adrenal insufficiency [24].

Figure 1.

Steroidogenesis pathway in 21-hydroxylase deficiency. Genetic defect in CYP21A2 results in accumulation of androgenic precursors 17-hydroxyprogesterone and androstenedione (shown in green font). These precursors are then diverted towards the synthesis of androgens including testosterone and 11-oxygenated androgens (11OHA4, 11KA4, 11KT, and 11OHT) shown with yellow arrows. CYP11B1 and CY11B2 are exclusively expressed in the adrenals while HSD11β2 is expressed in the peripheral tissues, predominantly kidneys. The remaining enzymes are expressed in all steroidogenic tissues. Abbreviations: StAR, steroidogenic acute regulatory protein; DHEA, dehydroepiandrosterone; DHEA-S, dehydroepiandrosterone sulphate; 11OHA4, 11-hydroxyandrostenedione; 11KA4, 11-ketoandrostenedione; 11KT, 11-ketotestosterone; 11OHT, 11-hydroxytestosterone; HSD11β2, 11β-hydroxysteroid dehydrogenase isoenzyme 2.

The 11-oxyandrogens are clinically relevant in 21OHD, a condition with excess accumulation of A4. Elevated 11-oxyandrogens correlate positively with total adrenal volume and testicular adrenal rest tumors in patients with 21OHD, [25] and these steroids show a median elevation of 3- to 4-fold relative to age- and sex-matched controls in patients with 21OHD [24]. We hypothesized that in patients with 21OHD and discrepant conventional biomarkers, 11-oxyandrogens would be useful in discriminating the degree of disease control.

Patient and Methods

Patient cohort

In a retrospective review of adult and pediatric patients with classic 21OHD seen under Natural History Study of Patients with Excess Androgen (NCT00250159) at National Institutes of Health (NIH) Clinical Center, Bethesda, Maryland, from January, 2006 to July, 2019, we identified patients with at least 1 early morning (08:00 am) blood test obtained via venipuncture before morning glucocorticoid and mineralocorticoid with discordant results between serum 17OHP and A4. Of these identified patients, we selected those with available stored serum samples. Written informed consent had been obtained from these patients regarding use of their sera. Discrepancy between the conventional biomarkers was defined as 17OHP ≤1200 ng/dL with A4 ≥ laboratory-defined upper limit of reference, or 17OHP ≥1200 ng/dL with A4 within range for age- and sex-based reference range [12]. For children, the reference limit for A4 was referenced to their Tanner stage. For patients who had multiple laboratory assessments with discrepancy between the conventional biomarkers, samples from the visit with maximum discrepancy was used for analysis. Patients who needed stress dose glucocorticoids within 2 weeks preceding the encounter were excluded.

Patients were categorized as being in poor clinical control at the time of venipuncture if any of the following occurred (as applicable for age and sex): disproportionate bone age advancement, new signs and symptoms of precocious puberty, menstrual irregularity, hirsutism, or presence of hypogonadotrophic hypogonadism. Clinical control was determined by 2 clinicians (S.J. and D.P.M.) blinded to 17OHP and A4 values.

Hormonal assays

Quantitation of 15 steroids in peripheral sera was performed by liquid chromatography/tandem mass spectrometry (LC-MS/MS) as previously described [22]. We measured 16 Δ ⁴ steroids (cortisol, cortisone, A4, 11OHA4, 11KA4, 11OHT and 11KT, 11-deoxycortisol, 21-deoxycortisol [21dF], 16α-hydroxyprogesterone [16OHP], 17OHP, A4, corticosterone, 11-deoxycorticosterone, testosterone, and progesterone). Discrepancy between conventional biomarkers originally performed using LC-MS/MS by Mayo Medical Laboratories, Rochester, MN, was further confirmed on repeat measurement for this study. Each patient was paired with an age- and sex-matched unaffected individuals, and fold elevation for metabolites was calculated by taking the ratio of patient and healthy control values.

Statistical analyses

Data are described as frequency distributions (percentage) and median (interquartile range [IQR]). All data and median fold elevations were assessed for distributional assumptions, and nonparametric tests were used when data were not approximately normally distributed. Binary data between groups were compared using Fisher’s exact tests. Continuous data between groups were compared using nonparametric Wilcoxon rank sum tests. Associations adjusting for age and sex utilized general linear models for either analysis of variance or multiple regression. Receiver operating characteristic (ROC) analyses were performed in logistic regression with corresponding ROC association statistics of curve area and its 95% CI, and optimal cut points. The strength of statistical evidence throughout the study was assessed using study statistics (eg, median), measures of uncertainty (IQR) and CIs, together with p values in drawing conclusions. This approach is based on a recent strong recommendation by the American Statistical Association [26] that use of P values alone can be misleading and their dichotomization into “statistically significant” or not, based on the traditional threshold of .05 is not sufficient justification for statistical conclusions. Data were analyzed using SAS v 9.4 (SAS Institute, Inc, Cary, NC).

Results

Frequency of discrepancy between conventional biomarkers in large cohort

The laboratory assessments denote testing performed on 142 adults and 208 children (the frequency and time interval of laboratory testing varied with an average of approximately 6 visits among adults and 9 visits among children). The majority (85%) of laboratory assessments were performed in children. Discrepant 17OHP and A4 was identified in 17% (469 of 2738) of laboratory assessments; elevated 17OHP with A4 within reference range was the most common discrepancy observed (86%) (Table 1). When discrepancy between the conventional biomarkers was observed, elevated 17OHP with A4 within reference range was more common in children than adults (92% versus 69%, P < .001), while elevated A4 with 17OHP ≤ 1200 ng/dL was more common in adults than children (31% versus 8%, P < .001.

Table 1.

Frequency of discrepancy between 17-hydroxyprogesterone and androstenedione in large cohort of patients with CAH due to 21-hydroxylase deficiency

Variable	Frequency (%)
Total laboratory assessments	2738
Assessments with discrepancy between conventional biomarkers	469 (17)
Elevated 17OHP + A4 within range	403/469 (86)
Elevated A4 + 17OHP within range	66/469 (14)
Frequency of discrepant biomarkers in children	347/1949 (18)
Elevated 17OHP + A4 within range	319/347 (92)
Elevated A4 + 17OHP within range	28/347 (8)
Frequency of discrepant biomarkers in adults	122/789 (15)
Elevated 17OHP + A4 within range	84/122 (69)
Elevated A4 + 17OHP within range	38/122 (31)

Clinical characteristics of patients with available sera

Forty-six patients (23 females, 50%) with at least 1 pair of discrepant 17OHP and A4, and available serum samples from the corresponding visit were identified (Table 2). Median age of selected patients was 13.5 years (8.0-22.0). Of these 46 patients, 28 (61%) were ≤18 years old. Forty-one patients (89%) had 17OHP ≥1200 ng/dL but A4 within range.

Table 2.

Clinical characteristics of patients with CAH due to 21-hydroxylase deficiency, discrepant conventional biomarkers and available sera

	No. (%)
Females	23 (50)
Children <18 years old	28 (61)
Patients in good clinical control	30 (65)
Patients with elevated 17OHP	41 (89)
Patients with elevated 17OHP (n = 41)
Females	21 (51)
Children <18 years old	27 (66)
Patients in good clinical control	28 (68)
Elevated A4 (n = 5)
Females	2 (40)
Children <18 years old	1 (20)
Patients in good clinical control	2 (40)

Patients were categorized as being in poor clinical control if any of the following occurred: incremental bone age advancement, new signs and symptoms of precocious puberty, menstrual irregularity, hirsutism, or presence of hypogonadotrophic hypogonadism.

Abbreviations: 17OHP, 17-hydroxyprogesterone; A4, androstenedione.

Hormonal correlation with clinical parameters of disease control

Of the 46 patients with discrepant values of 17OHP and A4, 30 patients (65%) were categorized as being in good clinical control with the remaining 16 (35%) in poor control. There was no difference in the frequency of patients in good clinical control between patients with elevated 17OHP and those with elevated A4 (28 of 41[68%] vs 2 of 5 (40%); P = .32). We did not find a higher occurrence of elevated 17OHP versus elevated A4 in children (n = 16 of 27, 59% vs n = 0 of 1, 0%; P = .43) or adults (n = 12 of 14, 86% vs n = 2 of 4, 50%; P = .20); however, the small number of patients with elevated A4 limit the analysis.

Median fold elevation of 11OHT relative to controls was higher in patients with poor disease control (2.82 [1.25-5.43]) than in patients with good control (0.91 [0.49- 2.07]; P = .003). The median fold elevation of 11OHT in patients with poor versus good control remained higher after stratification by age: adults (5.76 [4.22-6.42] vs 0.91 [0.62-1.88]; P = .009) and children (1.97 [1.09-4.06] vs 0.99 [0.49-2.08]; P = .036) and among males (4.06 [1.97-5.43] versus 0.72 [0.48-1.48; P = .002]. However, median fold elevation of 11OHT in patients with poor versus good control among females was not different (1.96 [0.66-3.79] vs 1.14 [0.62-2.09]; P = .35) although it followed a similar pattern. Similarly, median fold elevation of 11KT was higher in patients in poor 3.57 (IQR 2.11-7.41) versus good control (1.76 [IQR 1.24-4.00]; P = .047) (Table 3). The difference between median fold elevation of 11KT between patients with poor versus good control was not observed after stratification by sex (females: 4.33 [1.41-9.64] vs 1.57 [1.24-3.54]; P = .38 and males: 3.57 [2.51-5.97] vs 2.21 [1.46-4.00]; P = .12) or age (children: 2.93 [2.05-5.97] vs 2.54 [1.49-5.36]; P = .58) and adults: 7.71 [4.15-9.98] vs 1.45 [0.94-2.66]; P = .050). The borderline significance of median fold 11KT elevation seen in adults in poor versus good control should be interpreted in the context of small sample size of those in poor control (n = 4). No significant difference in the median fold elevation of the remaining measured steroid metabolites was observed between patients with good versus poor clinical control.

Table 3.

Alternate steroid metabolites as potential biomarkers of clinical control in patients with CAH due to 21-hydroxylase deficiency and discrepant conventional biomarkers

	Good Control (n = 30)	Poor Control (n = 16)	P value
11KT	1.76 (1.24-4.00)	3.57 (2.11-7.41)	.047
11KA4	2.08 (1.14-4.33)	3.33 (1.63-6.15)	.39
11OHT	0.91 (0.49-2.07)	2.82 (1.25-5.43)	.003
11OHA4	1.27 (0.83-3.81)	2.70 (1.20-4.23)	.18
Cortisol	0.05 (0.03-0.10)	0.03 (0.02-0.10)	.25
Cortisone	0.05 (0.03-0.10)	0.05 (0.02-0.08)	.83
11dF	0.15 (0.08-0.34)	0.11 (0.06-0.49)	.93
21dF	293.82 (215.47-624.45)	351.04 (199.48-622.27)	.99
16OHP	36.26 (9.69-79.10)	43.07 (14.89-181.20)	.56
Corticosterone	0.17 (0.09-0.25)	0.10 (0.04-0.42)	.64
DOC	0.62 (0.40-1.03)	0.50 (0.22-1.18)	.93
Testosterone	1.19 (0.37-1.94)	1.60 (0.42-2.70)	.34
Progesterone	9.23 (4.10-28.86)	20.12 (8.10-52.46)	.17

Data are median (interquartile range: 25th-75th percentile) fold elevations relative to age- and sex-matched healthy controls. Patients were categorized as being in poor clinical control if any of the following occurred: incremental bone age advancement, new signs and symptoms of precocious puberty, menstrual irregularity, hirsutism, or presence of hypogonadotrophic hypogonadism.

Abbreviations: 11KT, 11-ketotestosterone; 11KA4, 11-ketoandrostenedione; 11OHT, 11-hydroxytestosterone; 11OHA4, 11-hydroxyandrostenedione; 11dF, 11-deoxycortisol; 21dF: 21-deoxycortisol; 16OHP, 16α-hydroxyprogesterone; DOC, 11-deoxycorticosterone.

ROC analysis of 11OHT using the clinical control model showed good accuracy (ROC area 0.78, 95% CI 0.62-0.93, Fig. 2), supporting that fold elevation of 11OHT relative to age- and sex-matched controls provides an informative model for differentiating good from poor clinical control. Based on this, a 3.48 (sensitivity 97%, specificity 47%) fold elevation of 11OHT provided the best discrimination between poor and good clinical control in this cohort with discrepancy between conventional biomarkers of 21OHD. When stratified by age group, the model had better discriminatory ability in adults (ROC area 0.95, 95% CI 0.83-1.00) than in pediatric patients (ROC area 0.69, 95% CI 0.47-0.90).

Figure 2.

Performance of 11-hydroxytestosterone elevation in discriminating disease control in 21-hydroxylase deficiency. Receiver operating characteristic (ROC) curve for fold elevation of 11-hydroxytestosterone (11OHT) relative to age- and sex-matched controls in patients with 21-hydroxylase deficiency and discrepant conventional biomarkers (17-hydroxyprogesterone and androstenedione) shows area under curve of 0.78 (95% CI 0.62-0.93). Cut-off fold elevations of 11OHT (red arrows) from left to right provide the smallest distance to perfect distinguishing capability: 0.96, 1.20, 2.75, and 3.48 median folds. Of these, 3.48 provides the highest correct classification rate (80%) with 97% sensitivity and 47% specificity.

Patients with elevated 17OHP had several other steroids that were substantially different relative to controls in comparison to patients with elevated A4, including 21dF (359.36 [237.05-688.84] vs 77.38 [64.18-175.29]; P = .002), 16OHP (44.5 [17.08-84.23] vs 1.84 [1.16-2.25]; P < .001), and corticosterone (0.19 [0.10-0.39] vs 0.02 [0.0-0.03]; P = .002). However, these steroids were not useful in discriminating between patients in good vs poor control (Table 3). We did not find any differences in the serum levels of 11-oxyandrogens between patients with elevated 17OHP and those with elevated A4.

Discussion

In our large cohort of patients with CAH due to 21OHD, we found discrepancy between the 2 conventional biomarkers, 17OHP and A4, to be common, with an overall prevalence of 17% of laboratory assessments. Elevated 17OHP with A4 in reference range was more frequent (86%) in comparison to elevated A4 but 17OHP in range. Median fold elevation of 11OHT and marginally 11KT was higher in patients in poor versus good control, indicating potential advantage of these 11-oxyandrogens in assessing disease control.

Our study is the first to systematically evaluate the prevalence of discrepancy between the 2 conventional biomarkers of 21OHD. Elevated serum 17OHP with A4 within range was the most frequent finding (86%) among the laboratory assessments with discrepancy between conventional biomarkers, with higher prevalence in children (92%). This high frequency of discrepancy is important, as physicians, especially pediatricians, may use serum 17OHP as the sole biomarker for monitoring therapy in 21OHD. Greater temporal variability of 17OHP [11, 17, 19] might explain higher prevalence of elevated 17OHP with A4 within reference range; nevertheless, 28% of these patients had evidence of poor disease control. Conversely, 31% of laboratory assessments in adults and 8% in children had elevated A4 but 17OHP within acceptable range. Of the patients with elevated A4 but 17OHP in range on LC-MS/MS assays, 2 of 5 patients (40%) appeared to be in good clinical control. Thus, neither of the 2 conventional biomarkers in 21OHD are ideal for monitoring therapy. These findings highlight the need for new biomarkers for monitoring therapy in the disease.

Our findings suggest that 11OHT and 11KT may have a role as biomarkers of disease control in 21OHD in patients with discrepancy between conventional biomarkers. This discriminatory potential of 11OHT and 11KT but not 11OHA4 and 11KA4 might be explained by the higher androgenic activity of the former [21, 27]. Our findings indicate a distinct role of 11OHT in distinguishing poor versus good control which is not mirrored as clearly by 11KT. The reason for this is not quite clear; however, 11OHT appears to be directly produced by the adrenal cortex, whereas 11KT is produced in the periphery from excess adrenal metabolites, predominantly from 11OHA4, via 11KA4 [28, 29]. Inadequate sample size limiting the ability for similar observation in 11KT and lower circulating concentrations of 11OHT compared with 11KT in controls and patients with CAH in good control may offer additional explanation for this finding. ROC analysis demonstrates that fold elevation of 11OHT between 3.48 and 3.88 relative to controls had good accuracy in discriminating patients with good versus poor control with 97% to 100% sensitivity although the specificity was limited (40-47%). A model for 11KT did not show good accuracy; however, our ability to assess the discriminatory ability of these noncanonical androgens could be limited by the retrospective nature of the study and the small sample size. In addition, knowledge of 11-oxyandrogen bioavailability and ability to be aromatized is currently lacking and would enhance our understanding of the clinical implications of these novel androgens.

We demonstrate that patients with 21OHD have unmeasured circulating androgens even when apparently in good control based on clinical evaluation. Despite being categorized under “good” clinical control (n = 30), these patients were noted to have as much as 4-fold elevation of 11KT and 2-fold elevation of 11OHT relative to age and sex-matched controls. In vitro studies have shown that the maximum androgenic potency of 11KT approaches that of testosterone while 11OHT has modest androgenic potency [27, 30]. We have previously shown that in patients with classic 21OHD, all 4 11-oxyandrogens are elevated about 3- to 4-fold relative to controls [24] with higher levels associated with presence of testicular adrenal rest tumor in men and higher total adrenal volume [25].While the metabolic implications of elevated 11-oxyandrogens in 21OHD patients in apparent “good” clinical control remains to be elucidated, 11OHA4 and 11KA4 correlated with markers of insulin resistance in patients with polycystic ovarian syndrome, indicating their value as surrogate markers of metabolic risk [31].

The greater biological variability of 17OHP provides unique challenges in identifying patients in good versus poor control while monitoring therapy in 21OHD. However, among the 41 patients with elevated 17OHP but A4 within reference range, we found a median fold elevation of 11KT as high as 2.5 times relative to control suggesting that abnormally high levels of 17OHP in the absence of concomitant high levels of A4 should not be dismissed. Not surprisingly, 13 of these 41 patients (28%) were in poor clinical control.

Excess 17OHP in 21OHD is preferentially converted to 21dF by 11β-hydroxylase, encoded by CYP11B1, which can contribute to the pool of potent androgens via the backdoor pathway [32]. Its use has been proposed for newborn screening in lieu of 17OHP to help reduce rate of false positives as 21dF is not elevated in premature newborns [33, 34]. Superiority of adrenocorticotropin-stimulated 21dF versus 17OHP in the identification of heterozygote carriers has also been demonstrated [35]. However, 21dF was not useful in discriminating patients in good versus poor control in this cohort of patients with discrepancy between conventional biomarkers of 21OHD.

Strengths of our study include the high number of laboratory assessments for a rare disease, confirmation of discordance by repeat LC-MS/MS, use of age- and sex-matched controls, inclusion of both adult and pediatric patients with discrepant 17OHP and A4, and inclusion of both groups of patients—those with elevated A4 but 17OHP within acceptable range and vice versa. Nevertheless, our study has limitations. First, we have used clinical assessment as the gold standard for medical decision-making regarding good versus poor control. The development of clinical signs and symptoms can lag behind biochemical control. For example, it may take 6 to 12 months after an intervention to see an improvement in hirsutism [36]. In addition, clinical assessments can be subjective and observer dependent. Second, our sample size of 46 patients is relatively small particularly in the group with elevated A4 with only 5 patients. Third, we had fewer adults (n = 18) in the study. Fourth, controls were not matched for BMI, although they were age and sex matched. Finally, our study is limited by its retrospective nature and hence, definitive conclusions cannot be drawn.

In conclusion, existing conventional biomarkers of 21OHD often suffer from biological variability and discordance while monitoring therapy. Ultimately, it is the summation of clinical presentation and biomarkers that help with medical decision-making, and serum levels of 11-oxyandrogens offer an exciting possibility of improving the biochemical assessment of these patients with discrepancy between conventional biomarkers (Fig. 3). Prospective studies will allow us to delineate the potential of these novel androgens as a biomarker for disease control in 21OHD.

Figure 3.

Graphical abstract.

Glossary

Abbreviations

11KA4

11-ketoandrostenedione

11KT

11-ketotestosterone

11OHA4

11-hydroxyandrostenedione

11OHT

11-hydroxytestosterone

11-oxyandrogen

11-oxygenated androgen

16OHP

16α-hydroxyprogesterone

17OHP

17-hydroxyprogesterone

21dF

21-deoxycortisol

21OHD

21-hydroxylase deficiency

A4

androstenedione

CAH

congenital adrenal hyperplasia

LC-MS/MS

liquid chromatography/tandem mass spectrometry

ROC

receiver operating characteristic

Additional Information

Disclosures: D.P.M. received unrelated research funds from Diurnal Limited through the National Institutes of Health Cooperative Research and Development Agreement.

Data Availability

The data sets generated during the current study are not publicly available but are available from the corresponding author on reasonable request.