Free cortisol and free 21-deoxycortisol in the clinical evaluation of congenital 1 adrenal hyperplasia

8 Context: S ome patients with classic congenital adrenal hyperplasia (CAH) survive without glucocorticoid


Introduction
Glucocorticoids, such as cortisol, are crucial for regulating blood pressure, glucose metabolism, and the immune response (1,2).Cortisol is produced by the adrenal cortex and has a diurnal rhythm with the highest concentration in the early morning (3).This rhythm is regulated by the hypothalamus through stimulation of the pituitary gland, which releases adrenocorticotropic hormone (ACTH).Negative feedback by cortisol balances the hypothalamus-pituitary-adrenal (HPA) axis (3).During stress, such as infection or surgery, the cortisol demand increases, which is mediated by an increased ACTH release (4).
Primary adrenal insufficiency (PAI) results in inadequate cortisol production and an inability to increase cortisol concentrations during stress, leading to symptoms like extreme fatigue, weight loss, and abdominal pain (5).In adults, PAI is most often caused by acquired conditions such as auto -immune adrenalitis (Addison's disease, AD).In children, inherited conditions are most common (6), such as classic congenital adrenal hyperplasia (CAH), primarily due to 21-hydroxylase deficiency (21OHD) (90 -95%) or 11-hydroxylase deficiency (11OHD) (∼5%) (7).Patients with these conditions produce insufficient cortisol concentrations with consequently reduced negative feedback to the pituitary gland, causing chronically elevated ACTH levels.This leads to continuous activation of the adrenal cortex and accumulation of precursor steroids before the enzymatic block (7), that are partly shunted into the unaffected adrenal androgen pathway, leading to hyperandrogenism (8,9).So, in contrast to other forms of PAI, the hallmark of classic CAH is not only a decreased concentration of cortisol but also strongly elevated concentrations of precursor steroids and adrenal androgens.Patients with non-classic CAH (NCCAH), a less severe subgroup, generally produce normal cortisol concentrations but still have elevated concentrations of precursor steroids and androgens (4).
Diagnosis of PAI involves measuring morning total cortisol concentrations in blood.To assess HPA axis responsiveness, stimulation tests, such as the insulin tolerance test (ITT) (10) or Synacthen® test (11), are used.In these tests, a suboptimal cortisol response is defined as a total cortisol concentration <500 nmol/L (18 μg/dL) in the PAI guideline (5), but there is a debate on this threshold (12).The CAH guideline specifies a cut-off of <400 -500 nmol/L (14 -18 μg/dL), depending on the measuring method (13).
Cortisol in blood is mostly bound to proteins (±95%), such as corticosteroid-binding globulin (CBG) and albumin (14,15).The measured total cortisol is comprised of the bound and free (i.e., unbound) cortisol, while only the latter determines the biological glucocorticoid activity according to the free hormone hypothesis (16).In patients with 21OHD, elevated levels of 17-hydroxyprogesterone (17OHP) can be metabolized by 11-hydroxylase into 21-deoxycortisol (21DF) (7).Interestingly, previous studies have described untreated classic CAH patients with insufficient cortisol production without overt complaints of cortisol deficiency (17)(18)(19)(20).Our group has demonstrated that 21DF can transactivate the glucocorticoid receptor with 49% of cortisol's potency (17).Hence, 21DF contributes to the glucocorticoid activity, particularly at elevated concentrations as observed in patients with 21OHD (17).Furthermore, precursor steroids can bind to CBG (15,21,22), thereby competing with cortisol for bindings spots and preventing cortisol from binding, therewith increasing the free fraction of cortisol.These mechanisms might contribute to the attenuation of symptoms in patients with untreated classic CAH.So far, no studies have been conducted on free cortisol and free 21DF measurements in patients with CAH.
Our aim is to quantify total and free concentrations of cortisol and 21DF in the morning and 60 minutes after administration of Synacthen® in patients with untreated classic CAH (both 11OHD and 21OHD).We compare these results with patients with NCCAH, patients with other forms of adrenal insufficiency (AI), and controls to investigate the situation of CAH patients in more detail.We hypothesize that free cortisol and (free) 21DF concentrations might give a better reflection of the glucocorticoid activity in patients with 21OHD compared to solely measuring total cortisol concentrations.

Subjects
This study was approved by the local medical ethics committee (Radboudumc, case number 2021-12944).
Patients were included via three routes (see Figure S1  11OHD, n=5) were obtained from an Indonesian cohort as previously described (17 for 17OHP, 0 -191 nmol/L for 11DF, 0 -167 nmol/L for 21DF, and 0 -2,299 nmol/L for cortisol.In-house quality controls of human serum were measured in duplicate in each run.

Measurement of free steroid concentrations with LC-MS/MS
Measurement of free 21DF was incorporated in a clinically validated measuring method for free cortisol and free testosterone.Equilibrium dialysis was performed with samples buffered to pH 7.4 (± 0.03) at 37 °C (± 0.5) by a previously described HEPES buffer (4-(2-Hydroxyethyl)piperazine-1-ethanesulfonic acid; Merck, Rahway, NJ, USA) that resembled the ionic environment of serum (27).Two identical volume compartments (180 μL) were separated by a semipermeable regenerated cellulose membrane (Serva Electrophoresis GmbH, Heidelberg, Germany) with a diameter of 28 mm and maximum permeability of 5 kDA.During dialysis, cells were placed in a water bath at 37 °C (± 0.5) for 5h while continuously rotating.
After dialysis, dialysate was emptied from the cell and SPE was performed with 100 μL preceded by the  S3 (23).A nine-point calibration curve with 21DB (Steraloids, Newport, RI, USA), 21DF, and cortisol (both from Sigma-Aldrich) was used for quantification.The calibration ranges were 0 -115 nmol/L for all three steroids.

Method validation
All measurements were performed in an ISO15189 accredited laboratory.All clinically validated measuring methods are monitored for quality in each run using in duplicate measurements of in-house quality controls of human serum.For validation, a CLSI EP6 protocol (28) in serum was used to assess linearity.
We measured steroid concentrations in a blank sample that was injected after the highest standard to assess carry-over.Imprecision was assessed by a modified CLSI EP5 protocol (30) with pooled human serum samples at a low (n=7 for 21DF; n=20 for cortisol) and high concentration (n=8 for 21DF, n=20 for cortisol).The lower limit of quantification (LLOQ) was defined as the lowest concentration with an inter-assay coefficient of variation (CV) <20% and was determined with a modified EP evaluator protocol from pooled serum by between-day assay repeated measurements (n=10).For 21DF, recovery was calculated in five samples to which 29.94 nmol/L was added.For cortisol, a National Institute of Standards and Technology (NIST) standard (LOT number: 271020) diluted to 1.496 nmol/L was measured each run (n=29).A correction factor of 1.03 was applied to adjust free cortisol levels according to the NIST standard.Possible interfering substances were identified by measuring commercially available steroids and anti-androgens with a molecular mass ≤2 g/mol lower than the analyte or the internal standard.First, retention time interference was checked, and second, MRM transition interference was evaluated.

Data and statistical analyses
Free cortisol percentage was calculated by dividing the free cortisol concentration by the total cortisol concentration from the same sample multiplied by 100%.Since cortisol has a diurnal rhythm, only morning samples (between 08.00 -11.00 AM) were used for the analyses of absolute free and total cortisol concentrations.Results <LLOQ with a signal-to-noise ratio <5 were assigned a value of half the LLOQ.These results were excluded from the calculation of percentages.A Kruskal-Wallis test with post-hoc Dunn's test was used to compare means between different subgroups (classic CAH vs NCCAH vs control).For comparison of related samples (T 0 vs T 60 ), a related-samples Wilcoxon signed rank test was used.Patients from the AI group were excluded for statistical comparison due to the low number of inclusions.An LC-MS/MS method for measuring free cortisol and free 21DF was established and validated.The method was linear between 0.13 and 99.2 nmol/L for 21DF and between 0.08 and 66 nmol/L for cortisol.

Subjects
Table 2 shows the maximum deviation from linearity, imprecision, LLOQ, recovery, and ion suppression.
Carry-over was negligible for both components.For cortisol and its internal standard, we evaluated 18hydroxycorticosterone as possible interfering substance and it showed no interference.For 21DF and its internal standard, corticosterone, 11-deoxycortisol, and 18-hydroxy-11-deoxycorticosterone were evaluated and showed no interference.

Classic CAH patients and controls have similar free morning cortisol concentrations
In the morning samples (T 0 ), the median total cortisol concentration in untreated classic CAH patients was 106 nmol/L, significantly lower than NCCAH patients (249 nmol/L, p=0.010) and controls (202 nmol/L, p=0.016) (Figure 1A).
Interestingly, the median free morning cortisol concentration in untreated classic CAH patients (8.62 nmol/L) was not significantly different from NCCAH patients (9.84 nmol/L, p=0.313) and controls (6.94 nmol/L, p=0.573) (Figure 1B).So, in contrary to the total cortisol concentrations, this suggested a comparable cortisol availability.

Concentrations of free 21DF increased after Synacthen® administration in classic 21OHD patients
After Synacthen® administration, total 21DF concentrations significantly increased in NCCAH patients (median at T 60 26.4 nmol/L, p=0.043), but not in classic 21OHD patients (median at T 60 94.6 nmol/L, p=0.121) (Figure 2C).In AI patients, total 21DF concentrations remained <LLOQ at T 60 and in controls, the maximum was twice the LLOQ (0.7 nmol/L) for two patients.Classic 21OHD patients had significantly higher total 21DF concentrations at T 60 compared to controls (p<0.001)(data not shown).
Free 21DF concentrations significantly increased in both classic CAH (p=0.023) and NCCAH patients (p=0.043)(Figure 2D).At T 60 , the median free 21DF concentration in classic CAH patients was 7.34 nmol/L, which was significantly higher compared to NCCAH patients (median 2.22 nmol/L, p=0.041).In AI patients and controls, median free 21DF concentrations were <LLOQ at T 60 (data not shown).In all 11OHD patients, both total and free 21DF concentrations were <LLOQ (data not shown).
In classic 21OHD patients, the percentage of free 21DF increased from a median of 8.4% at T 0 to 10.3% at T 60 (p=0.035).There was no change in the median free 21DF percentage in NCCAH patients (p=0.893),remaining 7.9% both at T 0 and T 60 (data not shown).

Discussion
This is the first study that evaluated free cortisol and free 21DF concentrations in untreated patients with classic CAH (both 21OHD and 11OHD), NCCAH, other forms of AI, and controls.We found that despite lower total cortisol concentrations, free cortisol concentrations were similar in untreated classic CAH patients and controls in the unstimulated state.This could explain the mild or absent symptoms of cortisol deficiency in some patients with classic 21OHD or 11OHD.In addition, classic 21OHD patients had elevated concentrations of total and free 21DF, which is not observed in AI patients and controls, increasing the glucocorticoid activity.After Synacthen®, classic CAH patients showed no increase in total cortisol concentrations and minimal increase in free cortisol concentrations, but classic 21OHD had increased free 21DF concentrations.This might explain why the described patients survived periods of stress without glucocorticoid treatment.
We developed an LC-MS/MS method to measure free cortisol and free 21DF.The mean free cortisol percentage in controls (3.7%) was consistent with previous findings (3.95%) (31,32).The relative increase in free cortisol concentration (610%) after Synacthen® was greater than the relative increase in total cortisol (198%), as observed in other studies (31)(32)(33).Therefore, evaluating free concentrations may provide more informative insights than total concentrations.
Glucocorticoids perform numerous functions, and they are particularly necessary during stress.Thus, glucocorticoid levels should be interpreted in two contexts: the daily baseline state and the state under physical stress.
Cortisol in blood is mostly bound to CBG and albumin, but only its free form is biologically active (34).In the baseline state, we found that while total cortisol concentrations were lower in untreated classic CAH patients compared to controls, free cortisol concentrations were similar.This is supported by data from Androgens may lower CBG levels, potentially increasing the free cortisol concentrations (37,38).
In addition, classic 21OHD patients had increased 21DF concentrations in the basal state, confirming previous findings (39)(40)(41).Our group showed that 21DF has 49% of cortisol's potency in activating the glucocorticoid receptor (17).The median free 21DF concentration in these patients (5.32 nmol/L) is hypothetically equivalent to (0.49 x 5.32 =) 2.61 nmol/L free cortisol, potentially increasing glucocorticoid activity beyond what total cortisol concentrations suggest.This could explain why some classic CAH patients have managed without glucocorticoid medication during daily circumstances.Patients with 11OHD are not capable of producing 21DF, but their free cortisol concentrations were already similar to controls.
During physical stress, such as illness or surgery, the demand of glucocorticoid increases.Classic CAH patients were not able to increase their total cortisol concentrations after ACTH stimulation.In our study, these patients did not reach the 400-500 nmol/L total cortisol threshold after Synacthen® administration (13).They also showed only a minor increase in free cortisol concentrations, which did not match the increase observed in controls.Interestingly, 21OHD patients exhibited a significant rise in free 21DF up to a median of 7.34 nmol/L.Given its ability to activate the glucocorticoid receptor, this increase may significantly enhance the glucocorticoid activity.However, whether this concentration of free glucocorticoids is sufficient to prevent an Addisonian crisis remains uncertain.This raises questions about how previously described untreated classic 21OHD patients have managed physical stress, such as dengue fever and surgery, without glucocorticoid treatment (17)(18)(19)(20).Other protective measures might contribute to their resilience during stress.For instance, androgens may enhance glucocorticoid activity either directly or by facilitating glucocorticoid receptor binding (36).Additionally, synthetic ACTH administration has been shown to increase free cortisol fractions in healthy individuals (34).The continuous ACTH pressure in classic CAH patients might also boost free cortisol fractions and glucocorticoid activity.Further research is needed to explore these possibilities in patients with CAH.
Some classic CAH patients have survived without glucocorticoid treatment despite low total cortisol concentrations and inability to increase cortisol during stress (17)(18)(19)(20).This observation may be explained show minimal symptoms of adrenal insufficiency and may refuse glucocorticoid therapy (17,42).However, untreated patients face significant risks, including an Addisonian crisis, development of testicular adrenal rest tumors, formation of adrenal myelolipomas, and adverse cardiovascular effects (43,44).
Genotype-phenotype correlations in CAH are not always accurate.Our study also revealed discrepancies between genetic predictions and clinical phenotype, especially with the R356W mutation in CYP21A2.
Stoupa et al. (45) reported that 60% of their NCCAH patients had an inadequate cortisol response to Synacthen®, while none of them experienced periods of cortisol deficiency, highlighting that insufficient total cortisol concentrations do not always correlate with clinically relevant adrenal insufficiency.
Our findings suggest that measuring total cortisol concentrations alone may not adequately assess adrenal cortex function in all forms of PAI.Additional measurement of free cortisol may provide a more accurate reflection of glucocorticoid activity.Furthermore, in 21OHD patients, evaluating (free) 21DF concentrations should be considered to comprehensively evaluate the glucocorticoid activity.This approach is particularly useful for patients with milder forms of CAH to indicate the necessity of glucocorticoid treatment.More research is necessary to compute cut-off values for defining adrenal cortex insufficiency and to integrate these measures into clinical practice.Measuring free cortisol concentrations is also relevant for other conditions where discrepancies between free and total cortisol are present, such as hypoproteinemia in critical illness (34), CBG-deficiency (14), and oral contraceptive use (46).
Our study has several strengths and limitations.It is the first to quantify free cortisol and free 21DF concentrations in untreated classic CAH patients and compare these levels with controls, NCCAH patients, and AI patients.We used a state-of-the-art dialysis method to obtain free concentrations followed by a sensitive LC-MS/MS method, which is the gold standard for serum steroid measurements (13).
Furthermore, deuterium or 13 C-isotopes of each steroid were used as internal standards to allow for the correction of matrix effects.Moreover, all included CAH patients were genetically confirmed.Limitations include the low number of patients, especially for AI patients.Therefore, this group was excluded from statistical analyses.Cortisol has a diurnal rhythm, so we included only samples taken before 11.00 am for the analysis of cortisol concentrations to mitigate this issue.The control group had a higher percentage of males than the patient groups.In clinical practice, there is no sex separation for cut-off values (5,13), Downloaded from https://academic.oup.com/jcem/advance-article/doi/10.1210/clinem/dgae591/7740916 by guest on 31 August 2024 Downloaded from https://academic.oup.com/jcem/advance-article/doi/10.1210/clinem/dgae591/7740916 by guest on 31 August 2024 Downloaded from https://academic.oup.com/jcem/advance-article/doi/10.1210/clinem/dgae591/7740916 by guest on 31 August 2024 Serum samples from untreated classic CAH patients (21OHD, n=17;

Figure 2 1 Figure 1 2
Figure 2 Increase of total and free cortisol and 21DF concentrations (nmol/L) before (T0) and after (T60) Synacthen® in patients with classic congenital adrenal hyperplasia (CAH), non-classic congenital adrenal hyperplasia (NCCAH), adrenal insufficiency but not CAH (AI), and controls.(A) Change in total cortisol concentrations; (B) Change in free cortisol concentrations; (C) Change in total 21DF concentrations; (D) Change in free 21DF concentrations.*p≤0.05,** p≤0.01, ***p≤0.001,†excluded from statistical analysis.A C C E P T E D M A N U S C R I P T Figure 2 271x274 mm (DPI) A C C E P T E D M A N U S C R I P T

Table 1 .
Downloaded from https://academic.oup.com/jcem/advance-article/doi/10.1210/clinem/dgae591/7740916 by guest on 31 August 2024 R I P T respectively.No increase was observed in AI patients (Figure2B).At T 60 , the median free cortisol concentration in classic CAH patients was 9.85 nmol/L, which was significantly lower than in NCCAH patients (33.8 nmol/L, p=0.009) and controls (63.6 nmol/L, p<0.001).AI patients had a median of 5.08 nmol/L (data not shown).All controls had free cortisol increases ≥37 nmol/L after stimulation.NCCAH patients had free cortisol concentrations ranging from 22 to 45 nmol/L after Synacthen® stimulation.Classic CAH patients did not reach free cortisol concentrations seen in controls (maximum 21.8 nmol/L).

Table 1
R I P T Downloaded from https://academic.oup.com/jcem/advance-article/doi/10.1210/clinem/dgae591/7740916 by guest on 31 August 2024because there are no clinically relevant differences in concentrations of cortisol and 21DF between males and females (47,48).Therefore, we did not stratify for sex.To conclude, untreated classic CAH patients have basal state free cortisol concentrations in the same range as controls.Moreover, patients with 21OHD produce elevated concentrations of total and free 21DF, a precursor steroid with glucocorticoid activity.Both mechanisms might explain their relatively mild symptoms of adrenal insufficiency.During stress, 21OHD patients exhibit increased (free) 21DF concentrations, offering potential additional protection compared to other forms of PAI who did not produce 21DF.Future research should focus on the role of free cortisol and free precursor steroids, especially in milder forms of CAH to better assess the glucocorticoid activity and determine the need for glucocorticoid treatment.47.Eisenhofer G, Peitzsch M, Kaden D, et al.Reference intervals for plasma concentrations of adrenal steroids measured by LC-MS/MS: Impact of gender, age, oral contraceptives, body mass index and blood pressure status.Clin Chim Acta.Jul 2017;470:115-124.doi:10.1016/j.cca.2017.05.002 48.El-Farhan N, Pickett A, Ducroq D, et al.Method-specific serum cortisol responses to the adrenocorticotrophin test: comparison of gas chromatography-mass spectrometry and five automated immunoassays.Clin Endocrinol (Oxf).May 2013;78(5):673-80.doi:10.1111/cen.12039Demographics of included patients.Continuous data is given as median (IQR).

Table 2
Linearity, Imprecision (within and between day) on low and high level, lower limit of quantification (LLOQ), recovery, and ion suppression for free 21DF and free cortisol measured by LC-MS/MS after dialysis.