Cushing’s syndrome in early infancy due to isolated sporadic bilateral micronodular adrenocortical disease associated with myosin heavy chain 8 mutation: diagnostic challenges, too many!

Abstract

Endogenous Cushing’s syndrome (CS) is rare in infancy. Bilateral micronodular adrenocortical disease (BMAD), either primary pigmented nodular adrenocortical disease or the non-pigmented isolated micronodular adrenocortical disease is an important aetiology of CS in this age group, which requires bilateral adrenalectomy for cure. BMAD may be isolated, or a component of Carney complex. Isolated sporadic BMAD without other systemic manifestations poses a diagnostic challenge. Paradoxical cortisol response to dexamethasone suggests, while adrenal histopathology and mutational analysis of the culprit genes confirm BMAD. BMAD was suspected in 6-year-old infant with midnormal adrenocorticotrophic hormone, inconclusive adrenal and pituitary imaging and paradoxical increase in cortisol following high dose of dexamethasone. Exome sequencing revealed heterozygous c.354+1G>C (5′ splice site) variant in the myosin heavy chain gene (MYH8), located in chromosome 17. This particular variant has not been reported in the literature. In view of suspected phenotype and its absence in the population databases, the variant was classified as pathogenic.

Background

Bilateral micronodular adrenocortical disease (BMAD), a rare form of bilateral adrenal dysplasia, is phenotypically characterised by multiple small adrenocortical nodules with internodular cortical atrophy. If nodules are pigmented due to lipofuscin deposition, the disorder is termed as primary pigmented nodular adrenocortical disease (PPNAD). PPNAD is often associated with adrenocorticotrophic hormone (ACTH)-independent Cushing’s syndrome (CS), which may either be isolated, or associated with Carney complex (CNC), a multiple neoplasia syndrome, with spotty skin pigmentation, myxomas in different organs and endocrine overactivity.¹ PPNAD may be sporadic or inherited in an autosomal dominant manner. Germline mutations in one of the several genes like protein kinase A regulatory 1-alpha subunit (PRKAR1A), phosphodiesterase (PDE) 11A, PDE8B or rarely germline duplications of the catalytic subunit of protein kinase A (PRKACA) gene have been identified in patients, both with sporadic and familial PPNAD with/without CNC.^{2 3} A non-pigmented form of BMAD, in which CS usually manifests at an early age, and, is not associated with other endocrine or non-endocrine manifestations, is known as isolated micronodular adrenocortical disease (IMAD). This disorder is often sporadic and is related to PDE mutations only. Myosin heavy chain gene (MYH8) mutation is associated with isolated arthrogryposis (multiple joint contractures) or a particular variant of the CNC characterised by distal arthrogryposis with atrial myxomas. Such patients have arthrogryposis, manifested primarily as trismus and pseudocamptodactyly (contracture of the fourth and fifth fingers), but no adrenal involvement.⁴

Case presentation

A 6-month-old-baby, born of non-consanguineous union, was brought with recurrent lower respiratory tract infections and rapid weight gain for the preceding 4 months. The parents informed that the child had not established a sleep pattern and had also been suffering from recent onset irritability and disturbed sleep. She had delayed developmental milestones and there was no history of prolonged steroid use in any form. Her perinatal period was uneventful and family history was unremarkable.

The baby had generalised obesity with rounded face, facial hirsutism and solitary pigmented nevus on upper back with widespread Mongolian spots and Tinea corporis (figures 1 and 2). Her weight was 8 kg (50–85th centile), length was 68 cm (50th centile), blood pressure was 140/90 mm of Hg (>97th centile). There were no café-au-lait macules, skeletal deformities, contractures, clitoromegaly or abdominal lump. Systemic examinations were also normal.

Figure 1.

Profile of the child. Note the generalised obesity (A), Mongoloid spots over the back and Tinea corporis over the buttocks (B).

Figure 2.

Rounded face with fat deposition in cheeks (A) and a solitary blue coloured nevus on upper back (black arrow in B).

Investigations

Baseline investigations revealed polymorphonuclear leucocytosis (total count: 18100/µL; P₇₂L₂₂M₄B₁E₁), elevated hepatic transaminases (aspartate aminotransferase: 433 U/L; alanine aminotransferase: 154 U/L), high serum cholesterol (322 mg/dL), hypertriglyceridemia (1007 mg/dL) and elevated C-reactive protein (26.1 mg/L). She also had hypokalemia (3.3 mmol/L) and metabolic alkalosis (pH: 7.5, bicarbonate: 38.2 mmol/L). Renal functions and rest of electrolytes were normal. Cortisol and ACTH measurements were performed by Siemens IMMULITE 1000 and Roche cobas e 411 analysers respectively. Pooled samples (3 samples at 15 min interval) were taken for ACTH assay. Hormonal evaluations have been summarised in table 1.

Table 1.

Summary of hormonal evaluation

Date	Parameters	Patient’s value	Age and sex specific reference range
21/03	Free thyroxine (FT4)	1.38	0.89–1.76 ng/dL
21/03	Thyroid stimulating hormone (TSH)	3.13	0.70–6.4 µIU/mL
21/03	8:00 AM cortisol	53.6	4.3–22.4 µg/dL
21/03	8:00 AM plasma ACTH (iced sample)	31.2	7.2–63.3 pg/mL
22/03	Midnight sleeping cortisol	43.6	>4.4 µg/dL in children confirms CS
08/04	8:00 AM cortisol (repeat)	48.2	4.3–22.4 µg/dL
08/04	8:00 AM plasma ACTH (repeat) (iced sample)	29.8	7.2–63.3 pg/mL
08/04	Dehydroepiandrosterone sulfate (DHEAS)	34.9	16–96 µg/dL
09/04	Overnight dexamethasone (15µg/Kg that is,125 µg) suppression test (ONDST)-cortisol	24.6	>1.8 µg/dL is abnormal
11/04	48 hours Low dose DST-cortisol (30µg/Kg/day that is, 60 µg per dose)	18.7	>1.8 µg/dL is abnormal
13/04	High dose DST-cortisol (120µg/Kg/day that is, 240 µg per dose)	10.3	>68% suppression from baseline suggests CD
15/04	Liddle’s test cortisol	34.3	Serum cortisol is 90% of baseline value in 90% of CD. More than 50% rise from baseline is seen upto 75% of PPNAD

ACTH, adrenocorticotrophic hormone; CD, Cushing’s disease; CS, Cushing’s syndrome.

CT scan of the abdomen revealed mildly bulky adrenals without nodularity or mass (figure 3). Contrast enhanced CT scan of the neck and chest was also performed for ectopic ACTH sources, which could document right lower lobe consolidation along with patchy consolidation in superior segment of lower lobe of left lung.

Figure 3.

CT scan (A: non-contrast; B: contrast enhanced) showing mildly bulky adrenals (white arrow: lateral limb of right adrenal, red arrow: medial limb of right adrenal, yellow arrow: left adrenal).

Dynamic MRI of hypothalamopituitary area was unremarkable. However, marked cortical atrophy, loss of hippocampal volume and leptomeningeal enhancement with few possible granulomas were seen in the brain parenchyma (figure 4). Subsequent cerebrospinal fluid analysis confirmed central nervous system tuberculosis.

Figure 4.

T1 weighted MRI images of brain. (A) (coronal non-contrast): note the cerebral atrophy, (B) (coronal postcontrast): note the leptomeningeal enhancement, (C) (sagittal post-contrast): normal appearing pituitary gland.

Transthoracic echocardiography and whole body skeletal survey were normal.

Clinical exome assay was performed with 100% coverage of coding regions of all critical genes (PRKAR1A, PDE11A, PDE8B, MYH8 and PRKACA). A heterozygous 5′ splice site variation in intron 4 of the MYH8 gene (chr17:g.10418886C>G; depth: 305x) that affects the invariant GT donor splice site of exon 4 (c.354+1G>C; ENST00000403437.2) was detected. The variant has not been reported in the 1000 genomes database. The in silico prediction of the variant is damaging by MutationTaster2. Parental testing came negative for the reported mutation.

Differential diagnosis

In this child, aged 6 months, pretest probability of adrenal CS was very high. Hence, adrenal CT scan was performed pretty early in the diagnostic algorithm, once we came across unsuppressed morning cortisol and equivocal ACTH. CT scan effectively ruled out adrenal tumour and bilateral macronodular adrenal hyperplasia (BMAH). ACTH values fell within the ‘grey zone’, so ACTH-dependent CS was also considered. Rapid onset severe hypercortisolism, hypokalemia and metabolic alkalosis pointed towards possible ectopic ACTH syndrome. However, ACTH values were not unequivocally high and CT of neck, thymus, chest and abdomen were negative for ectopic ACTH secreting tumour. Appearance of the adrenals on CT, plasma ACTH concentrations and the age of the patient were consistent with BMAD, either PPNAD or IMAD. Liddle’s test cortisol value was corroborative with our diagnosis. Negative family history, normal echocardiography and absence of convincing skin lesions pointed towards sporadic isolated form of the disease.

Treatment

The child was treated with broad spectrum antibiotics and antitubercular agents according to the institutional protocol. Ketoconazole was withheld due to elevated liver enzymes. Emergency bilateral adrenalectomy was planned.

Outcome and follow-up

Clinical condition of the child deteriorated. She was transferred to intensive care unit for assisted ventilation and pressor support. She succumbed to sepsis with multiorgan failure. Her parents decided against medical autopsy. The skin lesion was also not biopsied. The child did not fulfil the criteria for CNC. In absence of adrenal histopathology, the diagnosis of PPNAD or IMAD could not be established with confidence. However, age at presentation, CT appearance of adrenals, ACTH value, cortisol trends during DST and the MYH8 variant collectively pointed towards BMAD. The other aetiologies of CS in infancy were ruled out and BMAD remains the only diagnosis in this child by the principle of Occam’s razor.

Discussion

Screening tests, those have been recommended in patients with suspected endogenous CS, are overnight dexamethasone suppression test (ONDST), at least two measurements for urinary free cortisol (UFC), at least two measurements of late night salivary cortisol (LNSC), 48 hours-low dose dexamethasone suppression test (LDDST) and in selected cases, midnight serum cortisol (MNSC). However, in early infancy, a 24-hour urine collection is almost impossible in the absence of a catheter in situ, particularly in the outpatient setting. The diurnal rhythms of ACTH and cortisol begin to set in at around 6–12 months and often are not well established until 3 years of age, making LNSC or MNSC less reliable as screening tests in early infancy.⁵ This is partly related to establishment of normal sleep pattern, which is evident at around 6 months of age. So, dexamethasone suppression test (DST) remains the only dependable screening test in early infancy. This child had a disturbed sleep pattern; so, we had to measure post-DST cortisol instead of relying on MNSC as the sole screening test. The next step in the algorithm is to look for ACTH dependency of cortisol production. 8:00 hour’s plasma ACTH higher than 20 pg/mL is suggestive of ACTH-dependent disease while in most patients with adrenal tumours, the value is less than 5 pg/mL. However, there is considerable overlap in the ACTH levels, as upto 25% patients with adrenal CS may have ACTH values within normal reference range, while some 10% of patients with Cushing’s disease (CD) (CS due to pituitary corticotroph adenoma) display ACTH concentrations below 10 pg/mL.⁶ Low-normal or intermittently detectable ACTH, encountered in adrenal CS like BMAH or PPNAD, likely is a consequence of mild hypercortisolism during the early phase of the disease or cyclic hormonogenesis.⁶ In a series of nine cases of PPNAD (six index cases and three family members of one case) with overt or subclinical CS, the ACTH value ranges from 1.2 to 13.9 pg/mL and in another cohort of 11 patients with overt disease, the range is between 2 and 12 pg/mL.^{7 8} An intermediate value of plasma ACTH (5–20 pg/mL), thus poses significant challenge to the treating endocrinologist. Corticotrophin-releasing hormone (CRH) stimulation test and high dose DST (HDDST) (either overnight single dose or eight doses over 48 hours) have been suggested when ACTH concentration falls between 5 and 29 pg/mL in a child with CS. If HDDST cortisol is not suppressed, the classic Liddle’s test needs to be performed. The other indications of Liddle’s test are suspected adrenal disease and inconclusive/negative imaging studies.⁹

ACTH in this child was not clearly high. Moreover, in infants and young toddlers, autonomous cortisol secretion from one or both the adrenals is encountered in overwhelming majority of cases with endogenous CS. In an analysis of 60 infants with CS, 48 had adrenal tumours (adrenocortical carcinoma in 25), nine had adrenal hyperplasia (nodular hyperplasia in five), two had ectopic ACTH syndrome (neuroblastoma and pheochromocytoma in one each) and one had CD.¹⁰ So, we first asked for CT adrenals. Bilateral adrenocortical hyperplasia is classified into two groups by the size of nodules: BMAH (nodule diameter ≥1 cm) and BMAD (nodule diameter less than 1 cm), either PPNAD or IMAD. CS due to ACTH independent BMAH may be encountered in an infant with McCune Albright syndrome (MAS) or Beckwith-Wiedemann syndrome (BWS). Such a clinical presentation of MAS, though rare, usually is seen before 6 months of age. CT scan in this child without other features of MAS or BWS was not suggestive of BMAH. The adrenals on CT scan in patients with BMAD may appear enlarged, may be nodular having the classical ‘string of beads’ appearance, but, often are reported as normal. Appearance of the adrenals on CT scan suggested two possibilities in this child; ACTH-dependent CS or BMAD. The MRI of pituitary and whole body CT scan did not show any eutopic or ectopic ACTH secreting tumour. Then, we performed ONDST (8 April to 9 April), LDDST (9 April to 11 April) and HDDST (11 April to 13 April) back to back (table 1). We also drew a sample for cortisol estimation 48 hours after the last dose of dexamethasone (15 April). This was a deviation from the classic Liddle’s test (6-day DST), where standard 48 hours LDDST is followed by standard 48 hours HDDST and a more than 50% rise of urinary 17-hydroxysteroids and UFC excretion on day 6 from baseline is considered as positive test. We did not estimate urinary steroids and relied on serum cortisol instead. Moreover, we performed ONDST and Liddle’s test in succession for two reasons; first, to confirm autonomous cortisol production and reach the definitive diagnosis as early as possible in this child with very high clinical suspicion for endogenous CS, having altered sleep pattern and significant hypercortisolemia (waiting for clearance of dexamethasone, used in ONDST, from systemic circulation before performing LDDST or Liddle’s test seemed inappropriate) and second, to increase the cumulative dose of dexamethasone and thereby increasing the specificity of Liddle’s test in this child with very high index of suspicion for BMAD. We also checked serum cortisol after completion of ONDST, LDDST and HDDST in addition to day 6 cortisol of the Liddle’s test. More than 50% suppression of serum cortisol following HDDST is seen in 90% of corticotroph microadenomas, 50% of macroadenomas and upto 30% of patients with ectopic ACTH secretion.¹¹ At least 68% suppression from baseline during HDDST has been suggested in children to differentiate CD from adrenal or ectopic CS.¹² Moreover, greater than 30% suppression in the 48 hours-LDDST may be considered as the surrogate marker of suppressed HDDST.¹¹ PPNAD is characterised by a paradoxical increase in cortisol response to dexamethasone during Liddle’s test in up to 50%–75% of cases, that distinguishes it from other forms of adrenal CS.¹³ It has also been suggested that Liddle’s test serum cortisol is 90% of baseline cortisol value in 90% of patients with CD.¹²

Taking pre-ONDST cortisol as baseline, we noticed a decrease of 61.2%, 78.6% and 28.8% following LDDST, HDDST and Liddle’s test cortisol, respectively. However, if the cortisol value, measured immediately before the first dose of dexamethasone for LDDST (measured on 9 April) is taken as baseline, the decrements were 24% and 58% following LDDST and HDDST, respectively. These figures were inconsistent with CD. Liddle’s test cortisol, however, increased by 39.4% in that context. The coefficient of variation (CV) for cortisol estimation in IMMULITE 1000 is 7.6%±2.2% and 7.3%±2.6% for mean cortisol concentration of 29.1 µg/dL and 35.8 µg/dL, respectively. The change in cortisol concentration thus, is almost four times than the CV of the test, performed. More interestingly, we noticed 233% increases in Liddle’s test cortisol from post-HDDST cortisol. Paradoxical cortisol response to dexamethasone may also be seen in about 20% cases of adrenal adenoma and very rarely in adrenal malignancy (angiosarcoma) or even CD, both corticotroph micro and macroadenoma.^{14 15} Eutopic or ectopic ACTH sources and adrenal tumours were effectively ruled out in this child with CS. Though the increment in Liddle’s test cortisol (day 6) from baseline was less than 50%, we believe that the response was consistent with ‘paradoxical response’ due to the following reasons. The time interval between post-HDDST and day 6-cortisol was not enough for complete clearance of a high cumulative dose of dexamethasone from circulation as the biological half-life of the drug is 36–54 hours. Moreover, the degree of incremental response in PPNAD demonstrates interindividual variation, and the in vitro and in vivo responses are also different in a same patient (intra-individual variation) at different point of time.⁸ The paradoxical cortisol response to dexamethasone observed in vivo in PPNAD patients can be reproduced in vitro, confirming a direct stimulatory action of the drug on hyperplastic adrenocortical tissue. Factors responsible for this paradoxical response are increased presence of GR mRNA, prolonged half life of GR mRNA, upregulation of GR transcriptional coregulators and nuclear trafficking. This paradoxical response is encountered in both sporadic and familial forms, either isolated or a part of CNC. It was initially suggested that increased cAMP signalling mediates this response, and PRKAR1A alterations are not an absolute prerequisite for this phenomenon.¹⁶ However, in a more recent in vitro study, GR-mediated effect on PRKACA has been suggested to be the underlying reason of this paradoxical response to dexamethasone.⁸ We have detected MYH8 mutation in this child and it has been suggested that myosin heavy chain and PKA are involved in intersecting metabolic pathways regulating growth of the affected organs.

We did not perform CRH stimulation test (due to unavailability) or inferior petrosal sinus sampling (due to lack of expertise) to establish pituitary source of ACTH, neither we measured serum chromogranin A or urinary hydroxyindoleacetic acid to confirm possible ectopic ACTH secreting neuroendocrine tumours in this child. Despite that we reached a final diagnosis of BMAD due to the following reasons: CD or ectopic ACTH syndrome are extremely rare causes of CS in this age group, and the changes in cortisol concentrations following DSTs are not concordant with findings commonly observed either in CD with normal pituitary MRI or ectopic ACTH syndrome.^{10 12} Though mutation in MYH8 gene has not identified in a large database of CNC, a recent study has demonstrated a particular MYH8 variant (transition from G to A at nucleotide 2094 (G2094A) in exon 16 that predicted an arginine to glutamine missense mutation at amino acid 674 (R674Q)) in families with typical spotty pigmentation of the skin, cardiac and cutaneous myxomas, distal arthrogryposis (the trismus–pseudocamptodactyly syndrome) and pituitary tumour without adrenal involvement, and, designated the syndrome as a variant form of CNC.^{4 17} This child with CS did not have myxomas or skeletal deformities, but had a mutation in MYH 8, different from that observed from the reported cohort, probably suggesting a novel pathogenic variant.

Learning points.

• Many of the recommended screening tests for endogenous hypercortisolemia have their own limitations, when used during early infancy. Collection of 24-hour urinary sample for free cortisol is a challenge in these children and measurement of late night salivary cortisol or midnight serum cortisol is not useful before establishment of ‘normal’ sleep pattern. Serum cortisol after dexamethasone suppression tests (overnight or 48-hour low dose), thus remains the reliable screening test in early infancy.

• Autonomous secretion of cortisol from one (tumour) or both (bilateral nodular adrenocortical disease) the adrenals is encountered in overwhelming majority of Cushing’s syndrome (CS) in early infancy. Adrenal imaging should be used upfront in the diagnostic algorithm, even if plasma adrenocorticotrophic hormone (ACTH) is not suppressed.

• ACTH concentrations, in the lower half of the normal reference range, are often encountered in bilateral macronodular adrenal hyperplasia (BMAH) or bilateral micronodular adrenocortical disease (BMAD). Liddle’s test reliably differentiates these two entities, as paradoxical cortisol response is seen in about 75% cases of BMAD, and, not in BMAH. Adrenal imaging also play a role.

• More than 50% elevation on day 6-serum cortisol from baseline during Liddle’s test may not be observed in all cases of BMAD. Any elevation from baseline cortisol and/or from posthigh dose dexamethasone cortisol, measured on day 4, may be considered as positive response in appropriate settings.

• The classic Liddle’s test with measurements of serum cortisol on days 0, 2, 4 and 6, thus, should be done routinely in all infants with endogenous CS with equivocal ACTH (5–30 pg/mL) and negative imaging (adrenal and pituitary) to pick up BMAD. The protocol may also be considered even in adults with similar parameters, where facilities for bilateral inferior petrosal sinus samplings are not available.