Adrenomedullary Function in Patients with Nonclassic Congenital Adrenal Hyperplasia

Abstract

Congenital adrenal hyperplasia (CAH) due to 21-hydroxylase deficiency is classifi ed into three types based on disease severity: classic saltwasting, classic simple virilizing, and nonclassic. Adrenomedullary dysplasia and epinephrine deficiency have been described in classic CAH, resulting in glucose dysregulation. Our objective was to investigate adrenomedullary function in nonclassic CAH and to evaluate adrenomedullary function according to disease severity. Adrenomedullary function was evaluated in response to a standardized cycle ergonometer test in 23 CAH patients (14 females, age 9–38 years; 6 salt-wasting, 7 simple virilizing, 5 nonclassic receiving glucocorticoid treatment, 5 nonclassic not receiving glucocorticoid), and 14 controls (7 females, age 12–38 years). Epinephrine, glucose, and cortisol were measured at baseline and peak exercise. CAH patients and controls were similar in age and anthropometric measures. Patients with nonclassic CAH who were not receiving glucocorticoid and controls experienced the expected stress-induced rise in epinephrine, glucose, and cortisol. Compared to controls, patients with all types of CAH receiving glucocorticoid had impaired exercise-induced changes in epinephrine (salt-wasting: p = 0.01; simple virilizing: p = 0.01; nonclassic: p = 0.03), and cortisol (saltwasting: p = 0.004; simple virilizing: p = 0.006; nonclassic: p = 0.03). Salt-wasting patients displayed the most significant impairment, including impairment in glucose response relative to controls (p = 0.03). Hydrocortisone dose was negatively correlated with epinephrine response (r = − 0.58; p = 0.007) and glucose response (r = − 0.60; p = 0.002). The present study demonstrates that untreated patients with nonclassic CAH have normal adrenomedullary function. The degree of epinephrine deficiency in patients with CAH is associated with the severity of adrenocortical dysfunction, as well as glucocorticoid therapy.

Introduction

Congenital adrenal hyperplasia (CAH) due to 21-hydroxylase (21-OH) deficiency is an autosomal recessive disorder characterized by a defect in cortisol and aldosterone biosynthesis and androgen excess. There is a range of clinical severity and three types of CAH can be distinguished by clinical and molecular genetic criteria: classic salt-wasting (SW), classic nonsalt-wasting or simple virilizing (SV), and nonclassic (NC) forms [1]. In general, the severity of CAH depends on the degree of 21-OH deficiency caused by CYP21A2 mutations: the classic SW form is generally associated with large deletions or mutations that result in no enzyme activity; the SV form is generally associated with CYP21A2 mutations resulting in one to two percent of 21-OH activity; the NC form is generally associated with mutations resulting in 20 – 60 percent of normal 21-OH activity [2]. There are marked clinical differences between the three forms of CAH. The classic form can be life-threatening and girls are born with ambiguous genitalia; while the NC form can be asymptomatic.

Although multiple genes are important in the development of the adrenal cortex and the adrenal medulla [3 – 5], the aberrant adrenomedullary development found in patients with salt-wasting CAH is due to lack of intra-adrenal cellular interactions and impaired cortical-chromaffi n crosstalk. Patients with classic CAH have been found to have adrenomedullary hypoplasia and im paired epinephrine secretion [6]. Glucocorticoid secretion is necessary for the induction of phenylethanolamine N-methyltransferase (PNMT), which converts norepinephrine to epinephrine. Normal glucocorticoid secretion by the zona fasciculata of the adrenal cortex is also necessary for adrenomedullary organogenesis, and a developmental defect in the formation of the adrenal medulla has been shown in patients with SW CAH [6]. Adrenomedullary function has not been evaluated in patients with NC CAH. During development, intra-adrenal cortisol secretion is expected to be normal or near normal in patients with NC CAH. Thus, normal adrenomedullary formation is expected; however, mild impairment is possible. Moreover, suppression of intra-adrenal cortisol secretion might occur due to treatment, possibly affecting PNMT activity.

Exercise is a potent stimulator of epinephrine from the adrenal medulla. As exercise is natural and quantifi able in terms of work load duration, it has been effectively used to measure epinephrine release and sympathetic nervous system function. Standardized exercise testing has been used in the past to evaluate adrenomedullary reserve in patients with classic CAH [7 – 10]. In a short-term high intensity exercise study, patients with classic CAH had significantly lower epinephrine levels at baseline and in response to exercise when compared to age- and sex-matched controls [8]. Patients also did not experience the normal exercise-induced rise in glucose. This impairment was attributed mostly to epinephrine deficiency and could not be corrected by doubling the replacement dose of hydrocortisone prior to exercise [7].

The aim of our study was to evaluate adrenomedullary function in patients with NC CAH. We evaluated adrenomedullary function using a standardized exercise test. Patients with NC CAH were evaluated in relation to patients with classic CAH and healthy controls. We hypothesized that a trend in epinephrine secretion and glucose response would be observed corresponding to phenotype or the degree of 21-OH deficiency with SW patients having the most severe adrenomedullary impairment. We also hypothesized that NC patients would not differ from controls. Because of possible confounding effects of glucocorticoid therapy on adrenomedullary function, we evaluated 2 groups of patients with NC CAH, those receiving glucocorticoid therapy and those not receiving treatment. To our knowledge, this is the fi rst study to examine adrenomedullary function and reserve in patients with NC CAH.

Subjects and Methods

Subjects

Twenty three otherwise healthy patients with CAH (9 males and 14 females, age range 9 – 38 years) and 14 healthy controls (7 males and 7 females, age range 12 – 38 years) participated in the standardized exercise study. Controls were recruited from the NIH Clinical Center Healthy Volunteer Program, a program that maintains a database of healthy volunteers recruited from the surrounding community. The patients with CAH were classified as classic SW (4 males, 2 females), classic SV (2 males, 5 females), NC-treated (2 males, 3 females), or NC-untreated (1 male, 4 females) based on clinical presentation and genotype (Table 1).

Table 1.

Clinical presentation, genotype, and cortisol response to a cosyntropin stimulation test (nonclassic patients) of patients with CAH

Patient #	Age (years)	Sex	Pheno-type	Genotype	Presentation	Cosyntropin 250 |jg test
						Cortisol 0 min	i Cortisol + 60 min
						(nmol/l)	(nmol/l)
1	18	F	SW	In2G/R356W	Ambiguous genitalia
2	17	M	SW	deletion/deletion	Adrenal crisis at 3 weeks
3	17	M	SW	In2G/In2G	Adrenal crisis at 5 weeks
4	14	F	SW	deletion/deletion	Ambiguous genitalia
5	13	M	SW	deletion/deletion	Prenatal diagnosis (family history)
6	13	M	SW	In2G/In2G	Adrenal crisis at 3 weeks
7	27	F	SV	I172N/In2G	Ambiguous genitalia
8	24	F	SV	Il72N/deletion	Ambiguous genitalia
9	17	F	SV	I172N/I172N	Ambiguous genitalia
10	15	M	SV	Il72N/deletion	Growth acceleration, pubarche at 4 years
11	13	F	SV	I172N/In2G	Ambiguous genitalia
12	10	F	SV	Il72N/deletion	Ambiguous genitalia
13	9	M	SV	I172N/Q318X	Newborn screen
14	37	F	NC	P30L/deletion	Infertility at 25 years
15	36	F	NC	V281L/V281L	Hirsutism at 15 years		579
16	14	F	NC	V281 L/deletion	Family history and acne	331	441*
17	11	M	NC	V281L/deletion	Growth acceleration, pubarche at 6 years	235	414*
18	10	M	NC	V281 L/deletion	Growth acceleration, pubarche at 7 years	218	488*
19	18	F	NC-no Rx	V281L/V281L or deletion	Irregular menses with acne, hirsutism at 17 years
20	38	F	NC-no Rx	V281L/V281L or deletion	Family history, genetic testing at33 years	246	574
21	38	F	NC-no Rx	V281L/V281L	Irregular menses/hirsutism at 15 years		579
22	13	M	NC-no Rx	V281 L/deletion	Accelerated growth, advanced bone age at 7 years	218	488 *
23	11	F	NC-no Rx	V281 L/deletion	Family history, genetic testing at 5 years	159	419*

^*Stimulated value below normal based on the criteria of a minimum serum cortisol value of 500 to 550 nmol/l (18–20 μg/dl) 60 min following cosyntropin injection. Conversion factor for calculation of conventional units: cortisol, nmol/l divided by 27.586 = μg/dl

All patients had a genotype that corresponded to the expected clinical phenotype. All SW patients had a history of an adrenal crisis with documented hyperkalemia and hyponatremia during their lifetime. Eight NC patients had cortisol measured during an ACTH stimulation test at the time of diagnosis or in the untreated state, and cortisol levels were slightly below normal in the majority of patients (Table 1).

Treated patients with CAH who underwent the exercise study were on conventional therapy with hydrocortisone and fludrocortisone (latter in classic patients only) and were in good clinical control based on physical examination and laboratory evaluation done within 2 months prior to the exercise testing. All patients had 17-hydroxyprogesterone levels between 3 and 45 nmol/l (100 – 1500 ng/dl); plasma renin activity was within the normal reference range; and growth rates were within 2 SD for age in children. Nonclassic patients who were not on medications had either never received glucocorticoid or had discontinued glucocorticoid treatment at least 18 months prior to the study. Female patients did not have new signs of virilization, and all subjects were nonsmokers. Healthy volunteers were age-, sex- and BMI-group-matched.

At an initial screening visit, all subjects had a thorough history and physical examination. Subjects also underwent baseline electrocardiogram to ensure their eligibility for high-intensity exercise. Female subjects had negative pregnancy tests. The study was approved by the Eunice Kennedy Shriver National Institute of Child Health and Human Development Institutional Review Board. All adult participants and parents of participating children gave their written consent. All children gave their assent.

Exercise protocol

Each subject performed two exercise tests on two consecutive days. All tests were monitored by a physician and performed on a cycle ergonometer (Sensor medics Ergoline 800, Sensor medics Corp., Yorba Linda, CA, USA) in the morning after an overnight fast. Subjects were instructed not to eat or drink anything (except water or noncaloric, noncaffeinated beverages) after midnight. Forty-five to 60 min prior to each exercise test, participants drank water (5 ml/kg body weight) [current weight (kg) × 5 ml water = 350 ml for a 70 kg person] to provide adequate hydration. One hour prior to each test, treated patients were given their morning doses of hydrocortisone and fludrocortisone. Patients were instructed to abstain from caffeine, alcohol, and strenuous exercise at least 24 h prior to the testing.

The first day, participants underwent maximal incremental exercise tests to determine their maximal aerobic capacity (VO_2max), as previously described [7, 8]. Maximal heart rate, respiratory exchange ratio (RER), and rating of perceived exertion (RPE) were recorded at peak exercise. RPE was measured using the Borg scale [11]. VO_2max was reached at the point when at least two of the following criteria were met: 1) plateau in oxygen uptake of O₂ at 2.0 ml O₂/kg∙min or less; 2) heart rate within 5% of 195 beats/minute; 3) RPE of 17 or greater, or 4) respiratory exchange ratio 1.10 or greater.

The following day, subjects underwent the standardized 20-minute exercise test. An indwelling line was placed in the forearm 45 min or more before the exercise test for blood drawn before and during exercise. Exercise began with a 3 min warmup with no workload followed by 5 min of exercise at an intensity of 50% VO_2max, 10 min at 70% VO_2max, and 5 min at 90% VO_2max [8].

Assays

Assays were done at the Clinical Center laboratories at the National Institutes of Health, Bethesda, Maryland, USA. Plasma epinephrine was determined by liquid chromatography with electrochemical detection [12]. Detection limits were 5 – 10 pmol/l. Heparinized whole blood was used for glucose. Cortisol was measured by chemiluminescence immunoassay.

Statistical analysis

Height and body mass index standard deviation (SD) scores were determined using U.S. anthropometric reference data [13]. Group differences in clinical characteristics were analyzed using Analysis of Variance. When significant (p < 0.05) group differences were found, post hoc t-test was performed to establish which groups differed. Exercise-induced changes in glucose, epinephrine, and cortisol were assessed by calculating the change between baseline and peak exercise, the expected time of maximal stimulation of the adrenal medulla, and were compared to healthy controls using a 2-tailed t-test with step-down Bonferroni correction. Epinephrine has been found to lack a normal distribution in the general population and thus was logtransformed for analyses. Pearson’s correlation coefficient was calculated for the dose response impact of hydrocortisone on exercise-induced changes in epinephrine and glucose.

Results

All subjects undergoing the standardized exercise study were similar in age and anthropometric measures (Table 2). All patients receiving treatment were on hydrocortisone and overall, NC patients were receiving lower doses of hydrocortisone than classic patients (p < 0.01) [SW vs. SV vs. NC (mean ± SEM): 14. 6 ± 0.6 vs. 16.6 ± 1.2 vs. 9.6 ± 1.2mg/m²/day] (Table 2). Prior cortisol response to ACTH stimulation was similar between NC patients receiving glucocorticoid treatment and NC patients not receiving treatment.

Table 2.

Clinical characteristics of subjects

	Classic SW CAH	Classic SV CAH	Nonclassic CAH	Nonclassic CAH-no Rx	Controls
No.	6	7	5	5	14
Age (years)	15.3±0.9	16.6±2.5	24.3 ± 5.5	23.6 ± 6.0	19.1 ±2.1
BMI SD score	0.9±0.3	1.2±0.2	1.4±0.2	0.9 ± 0.3	1.5±0.3
Weight SD score	0.6±0.4	0.1 ± 0.6	0.5±0.9	1.3±0.4	0.4 ± 0.1
Height SD score	0.5±0.7	1.0±0.3	1.4±0.3	1.4±0.7	1.4±0.4
Hydrocortisone dose (mg/m2/day)	14.6±0.6	16.6± 1.2	9.6 ± 1.2

Values are the mean ± SEM

The increase in epinephrine in response to exercise was significantly different between all CAH groups receiving treatment and the control group (SW: p = 0.01; SV: p = 0.01; NC-treated: p = 0.03; Fig. 1a). The SW patients had the lowest epinephrine levels followed by the SV and NC patients. The epinephrine response to exercise was not significantly different between NCuntreated patients and controls. Although the normal exerciseinduced rise in glucose was blunted in all CAH patients receiving glucocorticoid, the difference in the glucose levels between baseline and peak exercise was significantly different from controls for the SW group only (p = 0.03; Fig. 1b).

Fig. 1.

Epinephrine (a), glucose (b), and cortisol (c) concentrations during a standardized high-intensity exercise test in patients with saltwasting (SW), simple virilizing (SV) and nonclassic (NC) CAH, and healthy controls (CTRL). A subset of NC patients were not receiving glucocorticoid therapy (No Rx). Baseline refers to time zero of the exercise test and peak indicates the time of maximal exertion at 20 min. The difference between baseline and peak levels were compared to healthy controls, * * p ≤ 0.01; * p < 0.05. Conversion factors for calculation of conventional units: epinephrine, pmol/l divide by 5.461 = pg/ml; glucose, mmol/l divide by 0.055 = mg/dl; cortisol, nmol/l divided by 27.586 = μg/dl.

The change in cortisol level between baseline and peak exercise was significantly different between all CAH groups receiving treatment and the control group (SW: p < 0.004; SV: p = 0.006; NC-treated: p = 0.03; Fig. 1c). Cortisol levels in CAH patients receiving hydrocortisone refl ected medication (hydrocortisone) levels. As expected, the control group did have an overall rise in cortisol level in response to exercise, while this was not observed in any of the CAH patients on medication, including the NC patients. The NC patients not receiving glucocorticoid treatment did show a rise in cortisol level that was somewhat less than in the healthy controls, but this difference failed to reach statistical significance (p = 0.08).

For all CAH patients, hydrocortisone dose was inversely correlated with the exercise-induced change in epinephrine (r = − 0.58; p = 0.007) and glucose (r = − 0.60; p = 0.002).

Discussion

Our data demonstrate that patients with the mild or nonclassic form of CAH who are not receiving glucocorticoid therapy have normal adrenomedullary function. However, CAH patients in good clinical control on glucocorticoid therapy, including those with NC CAH, have decreased epinephrine reserve and reduced epinephrine response to high-intensity, short-term exercise. Thus, chronic glucocorticoid therapy appears to contribute to the decreased epinephrine secretion seen in patients with CAH. Our fi nding that patients with classic SW CAH display the most severe impairment of adrenomedullary function supports previous studies of impaired adrenomedullary function in classic CAH [7, 8, 10]. Proper adrenomedullary formation depends on intra-adrenal interactions between adrenomedullary and adrenocortical cells with adequate cortisol output [14]. Differences in adrenomedullary structure and development have been previously described in classic SW CAH patients and this abnormal development is likely the major cause of the markedly decreased adrenomedullary response observed in our SW patients. Abnormal adrenomedullary development may also be present to some degree in less severely affected CAH patients. This developmental defect may account, in part, for the subnormal epinephrine responses in the SV group. However, intra-adrenal cortisol secretion is expected to be normal or near normal during development in patients with NC CAH.

Proper adrenomedullary function also depends on intra-adrenal interactions between adrenomedullary and adrenocortical cells with adequate cortisol output [15, 16]. Although the dose of hydrocortisone of the nonclassic group was only slightly above what is considered physiological replacement [17], the absent cortisol rise observed for the NC-treated patients indicates that the mean dose of hydrocortisone given (9.6 mg/m²/day) was high enough to suppress endogenous cortisol secretion. Moreover, we found a significant inverse correlation between hydrocortisone dose and exercise-induced epinephrine secretion, suggesting that the glucocorticoid medication played a role in suppressing epinephrine. Interestingly, the NC-untreated patients had somewhat lower exercise-induced cortisol levels than the controls. This failed to reach statistical significance (p = 0.08), but is consistent with our fi nding of suboptimal cortisol response to ACTH stimulation at the time of diagnosis in the majority of the NC CAH patients in our study. This observation is also consistent with the recent fi nding by Bidet et al. [18] of weak (< 15 μg/dl) response to ACTH administration in approximately one-third of patients with NC CAH. Epinephrine secretion was normal in our NC-untreated patients, suggesting that intra-adrenal cortisol secretion was not relevantly reduced. Patients with classic SW CAH have been found to have more severe adrenomedullary impairment than patients with SV CAH. Plasma metanephrine, the O-methylated metabolite of epinephrine, has been found to be an accurate measure of CAH phenotype amongst classic patients [19], supporting a correlation between adrenomedullary impairment and clinical phenotype in CAH. This method of assessing phenotype was found to be as accurate as genotyping in predicting disease severity, but was only performed for classic CAH patients.

In adults, epinephrine counters hypoglycemia, although this compensation is not believed to be critical if glucagon and insulin secretion is intact [20]. Yet, in a fasting state, the adrenal medulla may provide compensation for hypoglycemia [21]. Children are known to be more prone to hypoglycemia than adults [22], and much of their glycemic control appears to depend on epinephrine secretion [23 – 26]. The hypoglycemia-associated automonic failure that occurs in type 1 diabetes is a clinical example of defective adrenomedullary epinephrine response. A recent study by Bao et al. [27] showed that prior exposure to elevated cortisol resulted in reduced epinephrine response to exercise in patients with type 1 diabetes. Thus, the interplay between cortisol and epinephrine regulation likely occurs at several levels and might even involve central mechanisms. This requires further study.

Adrenomedullary impairment has been described in other forms of glucocorticoid deficiency. For example, patients with isolated glucocorticoid deficiency display a low baseline epinephrine and minimal adrenomedullary response to physical stress [28]. This patient population with congenital cortisol deficiency likely has both developmental defects in adrenomedullary formation and adrenomedullary suppression due to glucocorticoid treatment.

This has not been studied. Although the clinical implications of impaired epinephrine secretion in response to increased metabolic stress are not entirely clear, the demonstrated epinephrine deficiency could contribute to or be responsible for the tendency of children with classic CAH to develop hypoglycemia when ill [29 – 32], and possibly play a role in the tendency of children with hypopituitarism or other forms of adrenal insufficiency to develop hypoglycemia. Our data suggest that iatrogenic suppression of endogenous cortisol and epinephrine secretion might commonly coexist. Another possible clinical implication of adrenomedullary hypofunction is hyperleptinemia and insulin resistance. Leptin, secreted by adipose tissue, is regulated by multiple factors including the adrenal medulla and epinephrine deficiency has been associated with elevated insulin and leptin levels in patients with classic CAH [10, 33].

The majority of our NC patients displayed mild decreases in cortisol secretion. As there is a continuum of disease severity, even among subtypes, there is a possibility that some untreated patients with NC CAH may not have a normal cortisol and epinephrine response to exercise or other stressors. Mild epinephrine and cortisol deficiency may or may not be clinically significant. Miscarriage risk was recently found to be significantly lower in glucocorticoid treated NC CAH women than in untreated NC CAH women [34], suggesting that the mild hormone imbalances characteristic of NC CAH may indeed have subtle clinical implications. Our study did not address the question of potential clinical clinical implications. A limitation of our study is the small sample size. However, this sample size is acceptable for an exercise study, and statistical significance was found with limited analyses.

In summary, this fi rst study of adrenomedullary function in patients with NC CAH demonstrates normal stress induced epinephrine response in patients not receiving glucocorticoid therapy. Adrenomedullary impairment was most significant in SW CAH, but was observed across all subtypes of CAH receiving treatment, including those with the NC type. Thus, the degree of epinephrine deficiency in patients with CAH is associated with the severity of adrenocortical dysfunction, as well as glucocorticoid therapy. It appears that patients with classic CAH mostly have an inborn epinephrine deficiency, whereas NC CAH patients mostly develop epinephrine deficiency under glucocorticoid replacement. Future research is needed to clarify the clinical implications of adrenomedullary impairment and epinephrine deficiency.