Possible pitfalls in the workup of ectopic ACTH secretion illustrated by four rare cases

Abstract

In this case report, we highlight four different cases of ectopic adrenocorticotropic hormone (ACTH) secretion with different pitfalls in the diagnostic workup. Ectopic ACTH secretion (EAS) is an uncommon cause of Cushing’s syndrome that accompanies a variety of tumours. It is associated with significant morbidity and mortality. This underlines the importance of early and adequate diagnosis. We will review the causes of EAS and their presentation to increase awareness of this rare and progressive disease.

Background

Cushing’s syndrome (CS) caused by ectopic adrenocorticotropic hormone (ACTH) production is a very rare disease. CS, containing both ACTH-dependent and ACTH-independent cortisol excess, is a rare condition with an incidence of 0.2–5 cases per million people.^1–3 A minority of patients with CS (20%) have ACTH-independent CS which originate from the adrenal glands (adenoma or carcinoma). The majority of CS cases are ACTH-dependent (80%), subdivided by Cushing’s disease (CD; around 70%), ectopic ACTH secretion (EAS; around 10%) and very rarely caused by ectopic corticotropin releasing hormone (CRH) production (<1%).^{1 2} Discrimination between these forms of CS is crucial since EAS is associated with highly malignant tumours, such as small cell lung cancer (SCLC).

Localisation of the source of EAS is therefore crucial, but can be difficult to find and may be delayed for months, with subsequent increased morbidity and mortality.⁴ In 9%–19% of all EAS patients, the source of ACTH production remain occult.^4–6 Classical features of CS, such as central obesity, are not always seen in EAS because of the brief duration due to the rapid onset and progression of disease. Additionally, some symptoms of CS may be covered by manifestations of the associated tumour.^{6 7} Establishment of the underlying cause is therefore challenging.

The heterogeneity in tumours in EAS, the delayed diagnosis and the increased mortality in this specific subgroup of CS underlines the importance of highlighting current pitfalls in the presentation and diagnostic process of CS caused by an ectopic source. Current literature emphasises these difficulties insufficiently. We aim to expose possible pitfalls in the diagnostic workup of EAS by addressing four rare, but illustrative cases. We will review the causes of EAS and their presentation to increase awareness of this rare and progressive disease.

Case presentation

Case 1

A 76-year-old woman was referred by another hospital with an ACTH-dependent CS for bilateral inferior petrosal sinus sampling (BIPSS). Her medical history stated a refractory osteoporosis and rheumatoid arthritis. The biochemical workup was initiated after an incidentaloma was found on abdominal imaging. Endogenous hypercortisolism was demonstrated without suppressed ACTH levels (table 1) and an MRI scan of the pituitary region showed no adenoma. Cortisol was insufficiently suppressed at a high-dose dexamethasone suppression test (HDDST), suggestive for ectopic CS, but a ¹⁸F-fluorodeoxyglucose positron emission tomography with contrast-enhanced CT (FDG-PET/CT) showed no pathological uptake besides activity in the right adrenal gland. She stated significant weight loss, profound muscle weakness and malaise since 6 months before presentation. Multiple hematomas and atrophic proximal muscle groups were found on physical examination. She had no other stigmata. Because of her ongoing clinical deterioration, we decided to repeat imaging to exclude a progressive underlying malignancy first. ⁶⁸Gallium DOTA-(Tyr3)-octreotate (⁶⁸Ga-DOTATATE) PET/CT scan showed only pathological uptake in her right adrenal gland, with nodular changes similar to the previous FDG-PET/CT scan. Since enlarged or nodular adrenal glands are common in CS,⁸ we considered BIPSS after the exclusion of a pheochromocytoma. Serum metanephrines proved to be elevated (three times the upper limit of normal) in absence of classical symptoms. Patient’s condition declined fast and she developed a hypomanic disruption with hallucinations. She was admitted to the internal ward and treated with metyrapone. An ACTH producing pheochromocytoma was considered the most likely diagnosis and an adrenalectomy of her right adrenal gland was performed after alpha-receptor blockade. Histopathology confirmed the diagnosis of an ACTH producing pheochromocytoma with positive ACTH immunostaining (monoclonal mouse anti-ACTH Clone 02A3; Dako).

Table 1.

Overview of diagnostic results in all cases

	1 mg DMX suppression test	24-hour urinary free cortisol	Midnight salivary cortisol	Midnight serum cortisol	ACTH	Potassium nadir
Reference range	<0.05 μmol/L	90–450 nmol/24 hours	1–6 nmol/L	<0.2 μmol/L	5–70 ng/L	3.8–5.0 mmol/L
Case 1	1.23	586	42	0.78	59	3.7
Case 2	1.70	–	65	1.56	234	2.7
Case 3	1.67	860	17	0.34	74	1.7
Case 4	1.08	–	–	–	629	2.7

Midnight serum cortisol measured as part of a 24-hour serum cortisol profile.

ACTH, adrenocorticotropic hormone; DMX, dexamethasone.

Case 2

A 64-year-old man was presented at our outpatient clinic with rapid deterioration of his type 2 diabetes mellitus, malaise and severe back pain. Physical examination was unremarkable besides a mild hepatomegaly. Biochemical analysis showed a mild hypokalaemia (3.2 mmol/L) and increased liver function tests. Imaging revealed a diffuse metastasised carcinoma, with a primary tumour most likely located in the right lung. Patient was admitted to the hospital for diagnostic workup of the metastasised carcinoma and improvement of his serum glucose level. Histopathology of a liver metastasis confirmed the diagnosis SCLC. One week later, patient was readmitted with severe hypokalaemia with ECG deviations (2.7 mmol/L). Both hypertension and diabetes were difficult to control. Therefore, despite the lack of any Cushing stigmata, patient was tested for CS. These tests (table 1) were suggestive for an ACTH-dependent CS. We asked the pathologist for an ACTH immunostaining of the liver metastasis. Surprisingly, this was negative (figure 1). Because of the rapid deterioration of patient’s condition, palliative chemotherapy and high-dose dexamethasone (DMX) were started and further biochemical analysis was therefore impossible. We requested an immunostaining of CRH on the liver metastasis (polyclonal rabbit anticorticotropin releasing factor Sigma, product number C5348) (figure 2), which was strongly positive and concluded a CRH producing small cell neuroendocrine carcinoma of the lung. Unfortunately, it was not possible to measure serum CRH in our hospital.

Figure 1.

Negative adrenocorticotropic hormone immunostaining of a liver metastasis (case 2).

Figure 2.

Positive immunohistochemical staining for corticotropin releasing hormone in a liver metastasis (case 2).

Case 3

A 59-year-old woman was admitted to the hospital with a second cerebral stroke. Her medical history was marked by cardiovascular disease, pulmonary embolism and the first stroke in the previous year besides de novo type 2 diabetes mellitus. At admission, hirsutism, hypertension, profound hypokalaemia (1.7 mmol/L) and mild hypernatriaemia (150 mmol/L) were observed. Patient was tested for CS and ACTH-dependent hypercortisolism was demonstrated with increased 24-hour urinary free cortisol (UFC), impaired DMX suppression test and absent circadian rhythm of serum cortisol. The HDDST showed sufficient suppression, indicating CD but a CRH test was inconclusive. MRI of the pituitary region showed no adenoma and FDG-PET/CT scanning was unremarkable. Patient was referred to our hospital for BIPSS, however, at follow-up, the cortisol day profile was normal as well as salivary cortisol and UFC. Her diabetes medication could also be discontinued. The suspicion of a cyclic CS rose and at multiple occasion, biochemical evaluation was repeated and cyclic CS demonstrated (figure 3).

Figure 3.

Morning cortisol (nmol/L) and potassium levels (mmol/L) over 6 years in a patient with cyclic Cushing’s syndrome (case 3). The cyclic pattern shows a recurrence of every ~2 years and illustrates cortisol influence on potassium levels.

BIPSS was conducted during an episode of hypercortisolaemia and revealed no ACTH-gradient, ectopic CS was concluded. Block and replacement therapy was started with metyrapone and hydrocortisone. A somatostatin receptor scintigraphy (SSRS) and endoscopy revealed no focus. ⁶⁸Ga DOTATATE PET/CT scan indicated a focus in the pancreas tail, but histopathology was negative for a neuroendocrine tumour.

Case 4

A 59-year-old woman with type 2 diabetes mellitus was admitted to the hospital for analysis of a severe hypertension and hypokalaemia (2.7 mmol/L). She suffered from fatigue and muscular weakness. Physical examination showed hypertension, palsy of the legs and facial hair growth. Laboratory results confirmed hypokalaemia (table 1). First, diagnostic tests were used to screen for CS. In addition, tests for pheochromocytoma were performed. Biochemical testing revealed an ACTH-dependent CS and elevated plasma metanephrines (three times the upper limit of normal). In addition, both CRH stimulation test and HDDST were suggestive for ectopic CS. Imaging studies, such as FDG-PET/CT showed a lesion of the left adrenal gland but no pathological uptake was seen on ¹²³Iodine-labelled metaiodobenzylguanidine scintigraphy. Nevertheless, we suspected an ACTH-producing pheochromocytoma of the left adrenal gland. Surgery was delayed due to severe infection, but finally a laparoscopic left adrenalectomy was performed after preoperative treatment with alpha-adrenergic and beta-adrenergic blocking agents. The clinical diagnosis was histologically confirmed employing immunohistochemical markers (figure 4). Postoperatively, hydrocortisone supplementation was administered and the patient recovered rapidly.

Figure 4.

Histopathologic examination shows a pheochromocytoma with positive adrenocorticotropic hormone staining (case 4).

Outcome and follow-up

Most of our cases are still alive and doing well (cases 1, 3 and 4). Our patient in case 2 showed a rapid deterioration. As a result, palliative chemotherapy was started. We decided not to add adrenal-directed therapy for CS, due to the intense chemotherapy and poor prognosis. After start of chemoradiation, potassium supplementation was reduced from 120 to 0 mmol/day and glucose control improved. Unfortunately, disease progressed after 6 months with concomitant increase of Cushing manifestations. Patient died 7 months after his first presentation.

A focus for ectopic production was found in three out of four cases. In case 3, no focus is found despite regular follow-up with anatomical and functional imaging modalities. Currently, she is still being treated with block and replacement therapy. Surgery was performed successfully in cases 1 and 4. Both patients recovered after surgery and only a mild cognitive impairment persisted in case 1. In case 4, the antidiabetic and antihypertensive treatment could be stopped. In hindsight, the diabetes mellitus was the first manifestation of hypercortisolism 1 year earlier. These patients are still under yearly surveillance.

Discussion

In this case report, we highlighted four different cases of EAS with different pitfalls in the diagnostic workup. The establishment of the diagnosis EAS should be made without any delay and the presenting manifestations can help in the discrimination between CD and EAS. Some clinical features (such as gender, age, presence of hypokalaemia and ACTH levels) possibly have more discriminatory power to differentiate between CD and EAS than biochemical testing.⁹ Most of our cases (three out of four) had a typical presentation with hypertension, diabetes mellitus and hypokalaemia as presenting symptoms. However, our first case is illustrative for the heterogeneity in clinical manifestations within EAS. This patient presented herself with rapid clinical deterioration with profound weight loss instead of weight gain. In literature, weight loss is reported in 21%–25% of patients with EAS.^{6 10} Weight loss should therefore not exclude CS. An explanation for weight loss could be the rapid onset and the underlying neoplastic process.^{6 7} On the other hand, central obesity is less often seen in EAS, compared with CD (14% vs 90%, respectively).⁷ Most reported features of EAS in literature are muscle weakness (63%–95%), hyperpigmentation (31%–90%), hypertension (60%–76%), easy bruising (50%–62%) and peripheral oedema (37%–57%).^{4 7 10}

The presence of hypokalaemia is another clinical manifestation indicating EAS. Studies comparing signs and symptoms between EAS and CD showed a prevalence of hypokalaemia between 50% and 100% compared with ~10%, respectively.4 6 7 9–11 The mean serum potassium in patients with EAS was lower as well (2.6 mmol/L in EAS vs 4.3 mmol/L in CD).⁷ This difference is the result of a significant relationship between 24 hours UFC excretion and the presence of hypokalaemia, which we try to illustrate with figure 3. Several mechanisms can lead to hypokalaemia in CS.¹¹ In healthy individuals, mineralocorticoid receptors (MRs) are not activated by cortisol.¹² One important mechanism is the conversion of cortisol into the inactive cortisone by the enzyme 11-beta-hydroxysteroid dehydrogenase type 2 (11β-OHSD2). In EAS, the extremely high cortisol levels can exceed the metabolic capacity of 11β-OHSD2.¹³ This mechanism seems to be the most important contributor to hypokalaemia.¹¹ Other studies suggested the inhibition of 11β-OHSD2 by high levels of ACTH in EAS or the hyperexcretion of non-aldosterone mineralocorticoids like desoxycorticosterone but evidence is limited.^{14 15}

Physicians should also be aware of a possible vascular origin when confronted with neurological complaints in a patient with CS. Our third case presented herself with a second cerebral stroke besides a history of pulmonary embolism. Patients with hypercortisolism have altered hemostatic parameters, which can lead to an increased risk of thromboembolic events.^{16 17} The risk of thrombosis is not specified between the forms of CS. The association between EAS and a neoplastic process could be an argumentation for a higher risk of thromboembolic events.⁶

Our cases not only illustrate multiple pitfalls in presentation but also in the diagnosis of EAS. In the diagnostic workup for ACTH-dependent CS, anatomical imaging studies, like MRI or CT-scan, can establish the source of hypercortisolaemia when a clear pituitary adenoma (>6 mm) or ectopic tumour (in absence of a pituitary adenoma) is visualised. However, MRI of the pituitary has a low sensitivity (46%–67%) due to the small size of adenomas in CD.¹⁸ Specificity is substandard as well due to the prevalence of incidentalomas of the pituitary in 10.7%.¹⁹ As a result, further biochemical testing is often necessary to differentiate between CD and EAS.

Most pituitary adenomas in CD remain sensitive for CRH stimulation or DMX suppression while ectopic tumours are resistant. CRH is regarded the most accurate test of the two with sensitivity varying between 70%–93% for ACTH response to CRH and 61%–91% for cortisol response.^20–22 Specificities vary between 70%–100%. The HDDST alone is a poor diagnostic tool since 20% of patients with CD fail to suppress cortisol as well with more than 50%. The poor negative predictive value of the CRH stimulation test and the high proportion of discordant results between the two tests (up to 65%) stretches out the need for distinctive and non-invasive diagnostics.²³ BIPSS has a sensitivity of 88%–100% and specificity of 67%–100%, for the differentiation between CD and EAS.²⁴ However, it is an invasive procedure and can lead to false-negative results due to venous anomalies of the petrosal sinus or to the inexperience of the personnel performing the procedure. Using prolactin levels can increase accuracy of BIPSS.²⁵ False results can be seen as well in cyclic CS or CRH producing tumours. CRH producing tumours do not suppress pituitary ACTH secretion, in contrast to EAS, but maintain or stimulate pituitary ACTH secretion.²⁶ As a result, CRH producing tumours could behave like pituitary adenomas regarding HDDST, CRH stimulation test and BIPPS.²⁷ However, the majority of case reports on CRH producing tumours reported test results consistent of EAS.²⁸ Although we did not perform BIPSS in our second case due to positive CRH staining on tumour tissue, it is important to take notice of this pitfall.

In 12%–19% of the ectopic production, no focus is found on initial imaging.^29–31 We illustrated this in our third patient. For initial imaging, thin cut multislice anatomical imaging of the chest, abdomen and pelvic is advised. Sensitivities for CT and MRI in 231 patients with EAS were 66.2% and 51.5%, respectively.⁵ Additional functional imaging like FDG-PET/CT and ¹¹¹In-pentetreotide scintigraphy have limited value due to the high proportion of false negatives. Hypercortisolism can suppress tumorous somatostatin subtype 2 receptor concentrations, leading to false-negative results in SSRS. Well-differentiated neuroendocrine tumours as the source of EAS have a low metabolic rate and therefore are often FDG-PET/CT negative. However, when an aggressive tumour is suspected, FDG-PET/CT has additional value and is the modality of choice. We prefer to perform ⁶⁸Ga-labelled somatostatin DOTA PET/CT (⁶⁸Ga-DOTA PET/CT) if conventional imaging is negative since its sensitivity is superior over other imaging modalities.⁵ If ⁶⁸Ga-DOTA PET/CT is negative as well, we perform FDG-PET/CT. This is in line with the proposed integrated approach from a systematic review with 231 EAS patients.⁵ Although ⁶⁸Ga-DOTA PET/CT superiority is well-established in well-differentiated neuroendocrine tumours larger studies are needed to confirm these results in EAS.³² In the rare case of ACTH producing pheochromocytomas ⁶⁸Ga-DOTA PET/CT is recommended as well.³³ Furthermore, adrenal adenomas can create a pitfall in patients with EAS, as was illustrated in our first and fourth cases.⁸ Functional imaging is necessary in such cases to confirm or exclude the adrenal gland as a focus.

If a possible source of EAS is identified, histopathology must be obtained to confirm the diagnosis. Immunostaining for ACTH is negative in very rare cases of ectopic ACTH-dependent CS with approximately 20 cases in literature describing isolated CRH producing tumours without co-production of ACTH.²⁸ It is likely that the actual percentage of CRH production is probably higher because of underreporting and under-ascertainment. In clinical practice, CRH is not routinely measured in the circulation or by immunostaining techniques.²⁶ There are a few pitfalls in the immunohistochemistry of tissue. First, when tissue is obtained from a metastasis, it is possible that the tumour tissue is negative for ACTH due to dedifferentiation. Second, negative ACTH immunoreactivity can be the result of a high-secretory capacity of the ectopic source leaving reduced storage of ACTH in its tissue. Third, inefficient procession of pro-opiomelanocortin (POMC) can lead to false-negative results.³⁴ A polyclonal antibody that also reacts with POMC is advised, so aberrant prohormone processing of POMC to ACTH will not lead to false-negative results.

In summary, EAS is an uncommon cause of CS that accompanies a wide variety of tumours. The clinical features of CS may be masked by the symptoms of the underlying malignancy accompanying EAS. In this case report, we presented rare cases of EAS and possible diagnostic pitfalls and challenges with a source difficult to identify. EAS is associated with significant morbidity and mortality, and this underlines the importance of early and adequate diagnosis.

Patient’s perspective.

Case 3: The highest burden of disease comes from the unpredictability associated with the cyclic pattern of the disease of my wife. We never know when signs and symptoms will flare up again and therefore, we live day-by-day. My wife has a reasonable performance status without the manifestations of Cushing’s syndrome but this decreases in times of disease activity. This uncertainty is increased by the differences in manifestation. Every cycle has presented itself with different symptoms, like nausea or transformation of facial appearance. As a result, minor physical complains increase our anxiety and concern for a flare of her disease.

Case 4: I have been through some tough times, and it took a while before I was fully recovered. I have had muscle and joint problems for a long time. Finally, I am very happy with the result, because my diabetes disappeared and insulin injections were no longer needed. I just started exercising and I really enjoy being active.

Learning points.

• Ectopic adrenocorticotropic hormone secretion (EAS) is an uncommon cause of Cushing’s syndrome (CS) that accompanies a variety of tumours.

• Classical features of CS are not always seen in EAS because of the brief duration due to the rapid onset and progression of disease. Common features of patients with EAS are hypokalaemia, muscle weakness, hyperpigmentation, hypertension, easy bruising and peripheral oedema.

• Localisation of the source of EAS can be difficult, with consequent increased morbidity and mortality. This underlines the importance of early and adequate diagnosis.

• There are multiple pitfalls in the diagnostic process of CS caused by EAS. Biochemical tests, imaging modalities, BIPPS and histopathology can all be false negative by various reasons. Diagnosis and treatment should therefore take place in experienced centres.