Cushing Syndrome in the McCune-Albright Syndrome

Abstract

Context: Cushing syndrome (CS) is a rare but potentially fatal feature of McCune-Albright syndrome (MAS). Optimal management, prognostic features, and long-term follow-up of this disorder have not been described.

Setting: The study was conducted at an academic tertiary care center.

Patients: A total of 112 patients participating in a natural history study at the National Institutes of Health (NIH) were evaluated, and 21 published cases were reviewed.

Interventions: Subjects received observation, medical management, or bilateral adrenalectomy.

Main Outcome Measures: We measured prevalence, prognostic factors, and natural history.

Results: The prevalence of CS among NIH patients was 7.1%. The median age at diagnosis was 3 months. Clinical features included “Cushingoid facies” (66.7%), failure to thrive (60.0%), low birth weight (50.0%), liver disease (36.7%), and heart disease (26.7%). Six patients (20.0%) died, four after adrenalectomy. Death was more likely in patients with comorbid heart disease (odds ratio, 13.3; P < 0.05). Of 23 survivors, 13 underwent adrenalectomy, and 10 exhibited spontaneous resolution. Two patients with spontaneous resolution who were tested later in life (3 and 15 yr after resolution) continued to have low-level, autonomous adrenal function with biochemical adrenal insufficiency. Compared to MAS patients without CS, patients with CS were more likely to have a cognitive/developmental disorder (44.4 vs. 4.8%; P < 0.001; odds ratio, 8.8).

Conclusions: Comorbid heart and liver disease were poor prognostic markers and may indicate the need for prompt adrenalectomy. The high incidence of cognitive disorders indicates a need for close developmental follow-up and parental counseling. Patients with spontaneous resolution of CS may develop adrenal insufficiency, and they require long-term monitoring.

Cushing’s syndrome is present in up to 7.1% of patients with McCune-Albright syndrome; it may remit spontaneously and is associated with later cognitive dysfunction.

McCune-Albright syndrome (MAS) is caused by postzygotic, activating mutations of the α-subunit of the stimulatory G protein (G_sα) that is coupled to many cell surface hormone receptors (1,2,3). The clinical manifestations of the disorder are variable due to the somatic nature of the mutations and the mosaic distribution of affected tissues. These are now recognized to extend well beyond the classic triad of precocious puberty, fibrous dysplasia of bone, and café-au-lait skin pigmentation (4,5,6), and include nonendocrine manifestations such as renal phosphate wasting (7), hepatobiliary dysfunction (8,9), and heart disease (9,10). Hyperfunctioning endocrinopathies include gonadotropin-independent precocious puberty, hyperthyroidism, GH excess, hyperprolactinemia, and hypercortisolism (9).

Hypercortisolism, or Cushing syndrome, affects a minority of patients with MAS and has a quite heterogeneous natural history, ranging from spontaneous resolution to need for adrenalectomy or even death (11). Due to the small number of reported patients, it may be difficult for the clinical practitioner to recognize which patients can be safely monitored or treated medically vs. those patients at high risk of morbidity and mortality who may benefit from prompt adrenalectomy. In addition, long-term outcomes of these patients have not been reported. To address questions regarding diagnosis, prognosis, management, and long-term outcome of this disorder, we analyzed our eight cases of MAS and cortisol excess and reviewed the 21 cases from the literature and one additional case from personal communication.

Patients and Methods

A database of patients with MAS enrolled in a natural history study at the National Institutes of Health (NIH) was queried to identify patients with Cushing syndrome, cortisol excess, or adrenal abnormalities. The study was approved by the institutional review board of the National Institute of Dental and Craniofacial Research, and subjects and/or their guardians provided informed consent/assent. A PubMed search identified an additional 21 patients with the diagnoses of MAS and Cushing syndromes (9,11,12,13,14,15,16,17,18,19,20,21,22,23,24). One additional patient was identified by personal communication (Stratakis, C.; Table 1, patient 9). A case-control analysis was performed comparing patients with MAS and Cushing syndrome (NIH + literature, n = 30) to NIH patients with MAS without Cushing syndrome (n = 104).

Table 1.

Patients with Cushing syndrome and MAS

(Continued) 

Table 1A.

Continued

F/U, Follow-up; PFD, polyostotic fibrous dysplasia; PP, precocious puberty; LVH, left ventricular hypertrophy; FTT, failure to thrive; SGA, small for gestational age; ARDS, acute respiratory distress syndrome; OCD, obsessive compulsive disorder; NR, not reported; Dev., developmental; PC, personal communication; AST, aspartate aminotransferase; CS, Cushing syndrome; ADHD, attention deficit and hyperactivity disorder; Dex, dexamethasone; Phos, phosphate; GI, gastrointestinal; HTN, hypertension; Resp, respiratory; UFC, urinary free cortisol. 

^a Age (in years) at diagnosis of Cushing syndrome (line 1), at diagnosis of MAS (line 2), and at last reported follow-up (line 3). ^b Cushing syndrome was first feature of MAS. ^c Fig. 1F; ^d Fig. 1, A–C; ^e Fig. 1, D and E. 

Statistical analyses were performed using GraphPad InStat Version 3 (GraphPad Software, Inc., San Diego, CA).

Results

Natural history of Cushing syndrome in MAS

Eight patients with hypercortisolism were identified in the NIH cohort of 112 patients (7.1%). The characteristics of MAS patients from NIH, personal communication, and the medical literature are summarized in Tables 1 and 2. The median age at diagnosis was 3.1 months (range, birth to 44 months). A history of low birth weight for gestational age (<10th centile) was noted in 15 patients (50%). The most frequent clinical features of cortisol excess at the time of diagnosis were “Cushingoid facies” (66.7%), failure to thrive (60.0%), hypertension (33.3%), nephrocalcinosis (30.0%), hirsutism (26.6%), hyperglycemia (20.0%), and linear growth failure with normal or excess weight gain (10%). Additional clinical features included liver disease such as cholestasis or elevated transaminases (36.7%) and heart disease such as cardiomyopathy (26.70%).

Table 2.

Clinical features of patients with Cushing syndrome

Age at diagnosis (yr)
Mean	0.5
Median	0.3
Range	0–3.7
Small for gestational age	15 (50.0)
Clinical features at diagnosis
Round face	20 (66.7)
Failure to thrive	18 (60.0)
Hypertension	10 (33.3)
Hirsutism	8 (26.6)
Nephrocalcinosis	9 (30.0)
Hyperglycemia	6 (20.0)
Linear growth arrest	3 (10.0)
Method of diagnosis
Laboratory (elevated cortisols with no suppression to dexamethasone)	25 (83.3)
Clinical	2 (6.7)
Not reported	3 (10.0)
Outcome
Death	6 (20.0)
Early death (before adrenalectomy)	2 (33.3)
Late death (after adrenalectomy)	4 (66.7)
Medical therapy	4 (13.3)
Successful (resolution)	2 (50.0)
Unsuccessful (death)	2 (50.0)
Adrenalectomy	17 (56.7)
Successful (resolution)	13 (76.5)
Unsuccessful (death)	4 (23.6)
Spontaneous resolution	10 (33.3)
Without medical therapy	8 (80)
After medical therapy	2 (20)
Not reported	1 (3.3)
MAS features
Fibrous dysplasia	28 (93.3)
Café-au-lait spots	28 (93.3)
Precocious puberty	22 (73.3)
Hyperthyroid	17 (56.7)
Renal phosphate wasting	5 (16.7)
GH excess	3 (10.0)
Other clinical feature
Liver disease	11 (36.7)
Heart disease	8 (26.7)

Data are expressed as number of patients (percentage), except for age. 

Diagnosis of Cushing syndrome was based on elevated serum or urine cortisol with failure to suppress to dexamethasone in 25 patients (83.3%). Detailed information on laboratory evaluation of Cushing syndrome was available in six NIH patients, all of whom had lack of cortisol suppression to high-dose dexamethasone and/or ACTH levels below the lower limit of normal, consistent with cortisol excess of adrenal origin. In two patients evaluated at NIH, the diagnosis of Cushing syndrome was made retrospectively based on clinical criteria. The method of diagnosis was not reported in three cases from the literature (25,26).

Six patients (20.0%) died, four of whom had undergone prior bilateral adrenalectomy. In one patient from the dermatology literature (17), management and outcome were not reported. Among the 23 known survivors (76.7%), 13 underwent adrenalectomy, and 10 exhibited resolution of Cushing syndrome without adrenalectomy. Two of the patients who spontaneously improved had previously been treated medically with cortisol-lowering drugs (ketoconazole and metyrapone). The most common pathological finding in patients who underwent adrenalectomy was nodular hyperplasia (11 of 14).

Predicting clinical outcome in Cushing syndrome

We sought correlations between the clinical features of MAS (e.g. hyperthyroidism) and outcome (death, adrenalectomy, or spontaneous resolution). The presence of heart disease (including all cardiac disorders listed in Table 1) was a poor prognostic marker for survival [Fischer’s exact test, P = 0.018; odds ratio (OR), 13.3; 95% confidence interval (CI), 1.6–107]. There was a trend toward lower likelihood of spontaneous resolution in patients with liver disease (P = 0.098; OR, 0.12; 95% CI, 0.01–1.2). No other significant associations between clinical features of MAS and outcome were found.

The mean age at diagnosis of patients who died was 2.4 ± 1.6 months (range, birth to 4 months) vs. 7.1 ± 10.1 months (range, 1 wk to 3.7 yr) for survivors; however, this difference was not statistically significant.

Comparison of MAS patients with and without Cushing syndrome

A case-control analysis of MAS patients without Cushing syndrome (Table 3) revealed that the presence of Cushing syndrome was correlated with an increased number of common clinical features of MAS, including café-au-lait skin pigmentation (Fig. 1, D and E), fibrous dysplasia, precocious puberty (Fig. 1E), hyperthyroidism, renal phosphate wasting, and GH excess. Patients with Cushing syndrome had a mean of 3.4 of these features, vs. a mean of 2.9 in patients without Cushing syndrome (P = 0.041, unpaired t test). Specific features of MAS that were associated with Cushing syndrome were café-au-lait spots (P = 0.0013; OR, 7.8; 95% CI, 1.8–34.8), precocious puberty (P = 0.036; OR, 2.7; 95% CI, 1.1–6.7), and hyperthyroidism (P = 0.0035; OR, 3.7; 95% CI, 1.6–8.7). In addition, MAS patients with Cushing syndrome who survived beyond 1 yr of age were more likely to have a cognitive disorder, including specific learning or speech disorders, or global developmental delay (P = 0.0006; OR, 8.8; 95% CI, 2.6–30.0). Interestingly, three of the NIH patients with Cushing syndrome had speech apraxia, vs. 0 of 104 patients without Cushing syndrome (P = 0.01; OR, 26.6; 95% CI, 1.3–530.8).

Table 3.

Comparison of MAS patients with vs. without Cushing syndrome

	Cortisol excess	No cortisol excess	P value	OR (95% CI)
Male	10 (33.3)	40 (38.5)	0.67	0.8 (0.3–1.9)
Female	20 (66.7)	64 (61.5)
Café-au-lait	28 (93.3)	66 (63.5)	0.0013	7.8 (1.8–34.8)
Fibrous dysplasia	28 (93.3)	103 (99.0)	0.13	0.1 (0.01–1.6)
Precocious puberty	22 (73.3)	52 (50.0)	0.036	2.75 (1.1–6.7)
Hyperthyroid	17 (56.7)	27 (26.0)	0.0035	3.7 (1.6–8.7)
Phosphate wasting	5 (16.7)	34 (32.7)	0.11	0.41 (0.1–1.2)
GH excess	3 (10.0)	19 (18.3)	0.40	0.5 (0.1–1.8)
Cognitive disordera	8 (44.4)	5 (4.8)	0.0006	8.8 (2.6–30.0)
Total no. of MAS manifestationsb	3.4 ± 1.1c	2.9 ± 1.4c	0.041
Age (yr)	28.8 ± 16.9c	10.2 ± 6.3c	<0.0001

Data are expressed as number of patients (percentage) unless explained otherwise. Significant differences shown in bold. 

^aAmong children who survived to at least 1 yr of age. 

^bNot including Cushing syndrome. 

^cMean ± sd. 

Figure 1.

A–C, Patient 6 as neonate with no clinical signs of cortisol excess (A); at age 3 months, with frequent infections and hirsutism of forehead (B); and at age 6 months, with resolved hirsutism (C). D and E, Patient 7 at age 5 yr demonstrating typical café-au-lait skin pigmentation (D) and at age 15 months demonstrating breast development from precocious puberty (E). F, Computed tomography of the abdomen in patient 5 demonstrating a large adrenal cyst (white arrow). This patient has persistent autonomous adrenal function and poor adrenal reserve.

Persistent autonomous adrenal function after clinical resolution of Cushing syndrome

Patients 4 and 5 exhibited spontaneous resolution of hypercortisolism followed by residual autonomous adrenal function with low adrenal reserve, a phenomenon not previously reported in the literature.

Patient 4 was diagnosed with Cushing syndrome at age 4 months. At 6 months of age, her serum cortisol levels were markedly elevated without diurnal variation [2400 h cortisol values, 19.7–40 μg/dl (531–1094 nmol/liter); 0800 h cortisol values, 20.3–27.8 μg/dl (548–751 nmol/liter)], ACTH was low [2–3 pg/ml (0.44–0.67 pmol/ liter)], and she had a lack of cortisol suppression to dexamethasone. Clinical features of Cushing syndrome resolved by 1 yr of age, and serum cortisol values gradually decreased, but with persistent lack of diurnal variation. Evaluation at NIH at age 3 5/12 yr revealed serum cortisol values of 1.8–2.1 μg/dl (50–58 nmol/liter) at 2400 h and 2.3–3.3 μg/dl (63–91 nmol/liter) at 0800 h, demonstrating a continued lack of diurnal variation. ACTH remained undetectable [<5 pg/ml (<1.1 pmol/liter)], indicating continued autonomous adrenal function. ACTH stimulation testing (250 μg Cosyntropin) showed a peak cortisol response of 18 μg/dl (496 nmol/liter) at 60 min, indicating borderline adrenal reserve.

Patient 5 was diagnosed with Cushing syndrome at age 18 months, at which time her morning cortisol was 27.4 μg/dl (755 nmol/liter) and nighttime cortisol was 21.1 μg/dl (581 nmol/liter). Evaluation at NIH at age 15 8/12 yr revealed serum cortisol values of 4.7–5 μg/dl (129–138 nmol/liter) at 2400 h and 4.7–5.2 μg/dl (129–143 nmol/liter) at 0800 h, demonstrating persistent lack of diurnal variation. ACTH was undetectable, indicating autonomous adrenal function. Computed tomography imaging of the adrenal glands showed a 5-cm cyst of the right adrenal gland (Fig. 1F). ACTH stimulation testing (250 μg Cosyntropin) showed a peak cortisol response of 15.4 μg/dl (424 nmol/liter) at 60 min, meeting the biochemical criteria for inadequate adrenal reserve. However, the patient did not have any signs or symptoms of adrenal insufficiency and had, in fact, had major orthopedic surgery 2 yr earlier without developing adrenal crisis. Nonetheless, it was felt prudent to educate her parents on the management of adrenal insufficiency and prescribe injectable hydrocortisone for surgery, severe illness, or injury.

Discussion

Cushing syndrome in the context of MAS is a rare disorder, and publications to date have largely consisted of case reports. In this study, we analyzed eight previously unpublished cases with long-term follow-up, as well as the published cases. In addition, we included a large, well-characterized case-control cohort of patients with MAS without cortisol excess. Our goal was to provide useful guidelines on the diagnosis, prognosis, management, and long-term outcome of this disorder.

It is possible that the NIH cohort may be biased toward a more severe spectrum of MAS and that the prevalence of 7.1% reported here may be an overestimate of the true prevalence of Cushing syndrome in MAS. Conversely, lack of access to medical records or follow-up on cases taken from the literature may have led to an underestimation of other clinical features, such as nephrocalcinosis or cognitive delay.

This study showed that hypercortisolism may be the first recognized manifestation of MAS in some cases (often appearing before the development of the café-au-lait skin pigmentation). Therefore, prompt and appropriate care of the infant or young child with MAS and Cushing syndrome depends on having a high index of suspicion for the diagnosis. In the child in whom the diagnosis is suspected, a rapid assessment for significant comorbidities should be performed, including hyperthyroidism and cardiac and liver disease. Evaluation for complications of glucocorticoid excess should be conducted, including hypertension, hyperglycemia, and nephrocalcinosis. If present, nephrocalcinosis is best managed by correcting the glucocorticoid excess, whereas hypertension and hyperglycemia can be medically managed pending definitive therapy. Patients with significant cortisol excess are immunocompromised, and prophylaxis for pneumocystis carinii pneumonia should be considered in all patients before definitive therapy.

The clinical features of cortisol excess in patients with MAS were generally similar to those seen in Cushing syndrome of other etiologies. One notable exception is that the majority of MAS patients with Cushing syndrome demonstrated failure of both linear growth and weight gain (failure to thrive), in contrast to the excessive weight gain typically associated with cortisol excess. It is possible that, in some cases, this failure to gain weight was related to comorbidities such as hyperthyroidism or heart disease.

Cortisol excess is almost certainly present in utero in the majority of MAS patients with Cushing syndrome, as evidenced by low birth weight and early age of presentation. As with other endocrinopathies in MAS, cortisol excess is caused by tonic activation of G_sα in the adrenal cortex, leading to unregulated overproduction of cAMP, and thus cortisol. Cushing syndrome is unique among MAS-related endocrinopathies, however, in its tendency to spontaneously resolve. This phenomenon of spontaneous resolution may reflect the fact that the G_sα mutation is found within cells in the fetal zone of the adrenal cortex, which undergoes rapid apoptosis in the postnatal period. During gestation, the bulk of the adrenal gland consists of the steroidogenic “fetal zone,” which broadly expresses G_sα (27,28). The fetal zone undergoes rapid degeneration after birth and is gone by 1 yr of age (29). In the cases of Cushing syndrome in MAS that resolve spontaneously, this may reflect the appropriate (albeit delayed) involution of fetal adrenal cells.

We report long-term follow-up on two patients with spontaneous resolution of hypercortisolism, in whom biochemical testing revealed continued autonomous adrenal function even after resolution of overt Cushing syndrome. More importantly, ACTH stimulation testing on these patients showed low peak cortisols, suggesting the possibility of adrenal insufficiency under stress. Such patients should be educated on the signs and symptoms of adrenal insufficiency and may require stress-dosing of glucocorticoids for illness or surgery.

The prognosis of patients with Cushing syndrome in this study was most strongly associated with the presence of comorbid heart disease, which was associated with increased odds of death. There was a weaker association of comorbid liver disease with outcomes of death or adrenalectomy. Because both heart and liver disease in MAS have in some cases been shown to be a direct consequence of G_sα mutations in the affected tissues, it is possible that the poor outcome of Cushing syndrome in this group is merely a reflection of a greater total body burden of G_sα mutation and, thus, sicker patients. Similarly, the greater number of MAS-associated endocrinopathies seen in patients with Cushing syndrome may reflect a greater burden of mutation-carrying cells. Nonetheless, recognition of comorbid heart or liver disease in children with MAS and cortisol excess should prompt consideration of early adrenalectomy when medically feasible. Although patients who have had adrenalectomy, if not appropriately treated, are at risk for adrenal crisis and death, in only one subject was this identified as a possible cause of death. Nonetheless, this possibility remains, and appropriate education of families is essential.

The high incidence of developmental problems among survivors of Cushing syndrome has not been previously noted, and again this raises the question of the effects of exposure to excess glucocorticoid in utero or during early postnatal life. In addition, the finding of speech apraxia in three of 10 NIH patients is quite striking, given that the prevalence of this disorder in the general population is estimated at only 0.125–1.3% (30). In children who developed Cushing syndrome later than the neonatal period, minor declines in cognitive function have been observed after treatment, particularly in younger children, but overall IQ scores remained in the normal range (31). Studies of cognitive function in children at risk for congenital adrenal hyperplasia or premature birth who were exposed to dexamethasone in utero have not shown significant differences in cognitive outcome related to glucocorticoid exposure (32,33,34,35). However, affected infants with Cushing syndrome due to MAS may have been exposed to significantly higher levels of glucocorticoid than those used therapeutically for lung maturation and congenital adrenal hyperplasia. An alternative explanation of cognitive impairment in MAS is the possibility of mutated G_sα in the central nervous system as part of the mosaic distribution, with Cushing syndrome survivors being more affected due to greater total body mutation burden. Transgenic mice in which a similarly activating G_sα mutation (Q227L) was selectively expressed in the brain showed a neurobehavioral phenotype including short- and long-term memory deficits, and impaired associative and spatial learning (36,37). Interestingly, these animals actually exhibit decreased cAMP levels in the cortex and hippocampus (as opposed to the increase that might be expected from constitutively activated G_sα). This was found to be due to compensatory increases in enzymes that break down cAMP: phosphodiesterases 1 and 4 (37,38); and the phosphodiesterase 4 inhibitor, rolipram, was found to reverse some aspects of the neurobehavioral phenotype. cAMP levels are not known to be decreased in the brains of patients with MAS, but it is worth considering whether phosphodiesterase inhibitors may be helpful in those patients with significant cognitive or neuropsychiatric dysfunction. This approach should be considered with caution, however, because phosphodiesterase inhibitors could theoretically worsen other manifestations of MAS, such as hyperthyroidism.

In summary, Cushing syndrome occurs in up to 7.1% of patients with MAS and may be characterized by low birth weight or failure to thrive, as well as more classic features of cortisol excess. Comorbid heart and liver disease appear to be poor prognostic markers and may indicate the need for prompt adrenalectomy. The high incidence of cognitive disorders in survivors indicates a need for close developmental follow-up. Patients with spontaneous resolution of Cushing syndrome may develop adrenal insufficiency, and adrenal function should be monitored long term in this group.