Factitious Cushing's syndrome masquerading as Cushing's disease

Authors

  • Tilenka Thynne,

    Corresponding author
    1. Southern Adelaide Diabetes and Endocrine Services, Repatriation General Hospital and Flinders Medical Centre, Adelaide, SA, Australia
    • Correspondence: Dr Tilenka Thynne, Southern Adelaide Diabetes and Endocrine Services, Daws Rd, Daw Park, SA 5041, Australia. Tel.: +61 8 8275 1094; Fax: +61 8 8275 1215; E-mail: tilenka.thynne@health.sa.gov.au

    Search for more papers by this author
  • Graham H. White,

    1. Southern Adelaide Pathology, Flinders Medical Centre, Adelaide, SA, Australia
    Search for more papers by this author
  • Morton G. Burt

    1. Southern Adelaide Diabetes and Endocrine Services, Repatriation General Hospital and Flinders Medical Centre, Adelaide, SA, Australia
    2. Flinders University, Adelaide, SA, Australia
    Search for more papers by this author

Summary

Context

Factitious Cushing's syndrome is extremely rare. The diagnosis is challenging as cross-reactivity of synthetic corticosteroids or their metabolites in immunoassay measurements of plasma or urinary cortisol can make distinguishing between true and factitious Cushing's syndrome difficult. Adrenocorticotropin (ACTH) is usually suppressed in factitious Cushing's syndrome.

Patient

A 54-year-old woman presented with clinical and biochemical features of Cushing's syndrome and an unsuppressed ACTH concentration. She denied recent exogenous corticosteroid use.

Investigations and results

Initial investigations revealed a markedly elevated urinary free cortisol, mildly elevated midnight salivary cortisol and normal morning cortisol concentration. Plasma ACTH was not suppressed at 13 ng/l (RR 10–60 ng/l). A pituitary MRI was normal, but inferior petrosal sinus sampling (IPSS) revealed a post corticotrophin releasing hormone ACTH ratio >20:1 in the left petrosal sinus. Ketoconazole therapy amplified discordance between the urinary free and morning plasma cortisol concentrations. Further investigation of this discordance using high-pressure liquid chromatography tandem mass spectrometry (HPLC-MS/MS) revealed a urinary free cortisol excretion of only 20 nmol/24 h, but prednisolone excretion of 16 200 nmol/24 h.

Conclusions

Factitious Cushing's syndrome can mimic endogenous ACTH-dependent hypercortisolism during initial investigations and IPSS. This case highlights the importance of (i) recognizing the significance of discordant results; (ii) using an ACTH assay capable of reliably differentiating ACTH-dependent from ACTH-independent Cushing's syndrome; and (iii) appreciating that IPSS is only useful to localize the source of ACTH in confirmed ACTH-dependent Cushing's syndrome. In this case, measurement of corticosteroids by HPLC-MS/MS was essential in reaching the correct diagnosis.

Introduction

Approximately, 0·5–1% of the community are prescribed long-term corticosteroids to treat inflammatory or autoimmune diseases.[1] Although treatment with exogenous corticosteroids is the commonest cause of Cushing's syndrome,[2] factitious Cushing's syndrome due to surreptitious corticosteroid ingestion has rarely been reported.

Immunoassays for plasma and urinary cortisol have variable cross-reactivity with structurally closely related steroids, particularly synthetic corticosteroids and their metabolites. In such assays, apparently high cortisol concentrations in plasma and urine arising from exogenous corticosteroid ingestion could be misinterpreted as biochemical confirmation of a clinical diagnosis of Cushing's syndrome. In most cases of Cushing's syndrome due to excess exogenous corticosteroid intake, the typical presentation is one of progressive clinical signs of hypercortisolism in conjunction with a suppressed plasma adrenocorticotropin (ACTH) concentration.[3] This finding, in association with normal or atrophied adrenal glands on imaging, provides a clue to consider surreptitious corticosteroid ingestion. We present a case of factitious Cushing's where ACTH was not suppressed and inferior petrosal sinus sampling (IPSS) was interpreted to be consistent with a pituitary source.

Case report

A 54-year-old woman with type 2 diabetes of 4 years duration was referred for management of deteriorating glycaemic control over the previous 12 months. Her other medical problems included a recent low trauma distal fibula fracture, poorly controlled hypertension and asthma. The patient reported receiving a 7 day course of prednisolone 2 years earlier for an exacerbation of asthma, but no other corticosteroid use. Physical examination revealed a moon-shaped facies, buffalo hump, central obesity, marked hypertension and a proximal myopathy, suggesting possible Cushing's syndrome.

Two midnight salivary free cortisol samples measured with an electro-chemiluminescence immunoassay (E170, Roche Diagnostics, Sydney, Australia) were mildly elevated at 13 nmol/l and 15 nmol/l (RR <13 nmol/l). Morning (8:00 am) plasma total cortisol concentrations measured with the same immunoassay ranged between 410-480 nmol/l (RR 200–700 nmol/l). However, a 24-h urinary free cortisol measured by the Roche E170 immunoassay after dichloromethane extraction was markedly elevated at 3281 nmol/24 h (RR 50–350 nmol/24 h), confirmed by repeat measurement. The patient's ACTH concentration was 13 ng/l on two occasions (RR 10–60 ng/l, Immulite 2000 XP, Siemens, Sydney, Australia.). No pituitary abnormality was detected by MRI. Adrenal and chest CT scans were normal. IPSS demonstrated low-normal ACTH concentrations and no significant central to peripheral gradient at baseline. However, 5 min after 100 μg corticotrophin releasing hormone administration (CRH, Ferring Pharmaceuticals, Kiel, Germany), there was a left petrosal to peripheral ACTH ratio greater than 20:1 (Ratio 438:14, Table 1). A diagnosis of ACTH-dependent pituitary Cushing's syndrome was made and the patient referred to the neurosurgical unit for pituitary surgery assessment.

Table 1. Inferior petrosal sinus sampling results
TimePeripheral ACTH (ng/l)Right petrosal ACTH (ng/l)Left petrosal ACTH (ng/l)
−2 min111116
Basal10<1014
Post CRH
2 min<101025
5 min1532438
10 min1820173
15 min1620136

Because early neurosurgery was unavailable and the patient's hypertension and glycaemic control were deteriorating, medical therapy with ketoconazole was commenced. Follow-up measurements of plasma and urine cortisol demonstrated the discordance between the markedly elevated urinary free cortisol and morning plasma cortisol, previously in the mid reference interval, had increased (Fig. 1). Although the morning plasma cortisol concentration decreased to 88 nmol/l, the 24-h urinary free cortisol excretion remained elevated at 3682 nmol/24 h. Despite the low plasma cortisol concentration the patient did not develop clinical signs of adrenal insufficiency and there was no improvement in her blood pressure or glycaemic control. Neurosurgery was therefore deferred pending investigation of the discordant biochemistry results, and ketoconazole was discontinued.

Figure 1.

Urinary free (solid line) and plasma total (dashed line) cortisol concentrations in a patient with factitious Cushing's syndrome. (1) Initial diagnosis (2) Ketoconazole commenced (3) Ketoconazole ceased and patient informed of HPLC-MS/MS results.

Investigation of the apparently low plasma cortisol relative to that in urine included dilutional testing that excluded a hook effect. Repeat measurement of plasma cortisol after pre-treatment with heterophilic blocking reagents (Scantibodies Laboratory, California, USA) did not detect heterophile antibody interference. A plasma cortisol binding globulin (CBG) concentration of 496 nmol/l, measured using a two-site non-competitive monoclonal antibody based enzyme-linked immunosorbent (ELISA) assay, was within the reference interval of 312–1324 nmol/l.[4] However, measurement of repeat plasma and urine samples by high-pressure liquid chromatography-tandem mass spectrometry (HPLC-MS/MS) revealed a urinary free cortisol excretion of 20 nmol/24 h, and a prednisolone excretion of 16 200 nmol/24 h (equivalent to an oral prednisolone dose of 40–80 mg). Total plasma cortisol concentration was <5 nmol/l, with a plasma prednisolone concentration of 240 nmol/l.

We advised the patient of the results of the HPLC-MS/MS analysis. The patient strongly denied taking oral prednisolone within the last 2 years. She also denied topical, inhaled or parenteral corticosteroid use, or taking any other herbal medications or supplements. Her general practitioner had not prescribed prednisolone for over 2 years and her local pharmacy had not dispensed it. However, 4 weeks later, her urinary free cortisol concentration by immunoassay had normalized (Fig. 1) and prednisolone was not detected in her plasma or urine by HPLC-MS/MS on repeat analysis. The patient continues to attend clinic and has not demonstrated any clinical signs of adrenal insufficiency. Twelve months later, the patient's weight has fallen by 22 kg and she has stopped her diabetes therapy, but remains moderately hypertensive.

Discussion

Factitious disorders are characterized by physical or psychological symptoms that are voluntarily initiated by the patient without a recognizable goal or gain, other than assuming the role of a patient.[5] Reports of the surreptitious use of insulin or sulphonylureas to produce factitious hypoglycaemia are well recognized [6, 7] and screening for sulphonylureas is an established part of the investigation of hypoglycaemia in the nondiabetic patient.[8]

In contrast, despite the widespread use of corticosteroid medications, there are few reports of factitious Cushing's syndrome,[9-16] accounting for less than 1% of patients with Cushing's syndrome. The largest series comprised six of 860 patients evaluated for hypercortisolism at the National Institute of Health Clinical Centre.[9] Typical patient characteristics of factitious Cushing's syndrome included a history of depression or anxiety, multiple surgeries for unrelated conditions in the past, and a medical-related occupation or close contact with a medical worker. Our patient differed as she had no known history of depression or anxiety, and did not work, or have close contact with healthcare professionals, but did subsequently report that she had a friend who had recently been investigated for Cushing's syndrome.

Case reports and series suggest the typical laboratory features of factitious Cushing's syndrome are erratically fluctuating urinary free cortisol concentrations and suppression of ACTH.[9-16] Our patient, however, had persistently elevated urinary free cortisol concentrations by immunoassay and an unsuppressed ACTH concentration, although at the lower end of the reference interval. Measurement of plasma ACTH has traditionally played a central role in localizing the cause of glucocorticoid excess and determining the pathway of further investigation.[2, 17] Most cases of ACTH-independent Cushing's syndrome are correctly classified by demonstrating a suppressed plasma ACTH,[18] though cases of ACTH-independent Cushing's syndrome with a plasma ACTH at the lower end of the reference interval have been reported.[19, 20] This has usually been in patients with mild glucocorticoid excess,[21] not the apparently ten-fold elevation in urinary free cortisol found in our patient.

A reliable ACTH assay is crucial to differentiate ACTH-dependent from ACTH-independent Cushing's syndrome. However, the performance of ACTH immunoassays at low concentrations was questioned by a recent study.[22] Seven patients expected to have a suppressed ACTH (three patients on high dose exogenous corticosteroids, three patients with adrenal Cushing's syndrome and one patient with isolated ACTH deficiency) had their plasma ACTH measured by 7 different immunoassays across 35 laboratories. The ACTH concentrations were frequently reported to be within the normal reference interval.[22] In particular, 19% of ACTH measurements performed in these patients with the Immulite 2000 immunoassay, used to assess our patient, were within the normal reference interval (i.e. unsuppressed).[22] With a lower limit of detection of only 9 ng/l with the Immulite 2000 immunoassay ACTH assay, our patient's plasma ACTH may have been misclassified as “normal” when in fact was suppressed.

Intermittent corticosteroid ingestion by our patient could also explain the lack of ACTH suppression and peripheral response to CRH. A recent study reported that prednisolone 60 mg/day for 7 days only partially suppressed serum ACTH, but induced adverse changes in bone and carbohydrate metabolism.[23] This suggests repeated intermittent high dose prednisolone may manifest as a clinical picture of Cushing's syndrome without complete suppression of the hypothalamic–pituitary–adrenal axis.

A peripheral CRH test has been recommended for patients with Cushing's syndrome and a low-normal plasma ACTH.[24] Although not performed in our patient, the 60% increase in peripheral ACTH following CRH administration during IPSS was consistent with Cushing's disease.[20, 25] Measurement of dehydroepiandrosterone sulphate, which is suppressed by exogenous glucocorticoid administration,[26] could have provided a clue to the correct diagnosis.

IPSS is useful only for localizing the source of ACTH in patients with ACTH-dependent Cushing's syndrome. In this setting, it distinguishes Cushing's disease from the syndrome of ectopic ACTH production with 95% sensitivity and 100% specificity.[27] However, IPSS results are potentially misleading in the patient with pseudo-Cushing's, ACTH-independent Cushing's syndrome and eucortisolaemic patients.[28, 29] As normal subjects can produce a similar increase in ACTH following CRH administration to that of patients with Cushing's disease,[29] clear-cut persistent hypercortisolism must be present at the time of undertaking IPSS. The expected IPSS result in factitious Cushing's syndrome is a blunted ACTH increase after CRH stimulation.[16] However, our case demonstrates that false positive IPSS results can occur in patients taking exogenous corticosteroids if the hypothalamic–pituitary–adrenal axis is not completely suppressed, similar to reports in patients with adrenal Cushing's syndrome.[30] The case highlights how critical it is to confirm a diagnosis of ACTH-dependent Cushing's syndrome to correctly interpret IPSS results.

The key to the diagnosis of factitious Cushing's syndrome in our patient was recognition of the discordance between the normal plasma cortisol, mildly elevated salivary cortisol, and marked elevation of urine cortisol concentrations. This discrepancy was present during the initial investigations, but the significance was not fully appreciated until treatment with ketoconazole amplified the discordance. There are several potential causes of error in the measurement of urine and plasma cortisol by immunoassay (Table 2). The correct diagnosis of factitious Cushing's syndrome was established in our patient by measuring cortisol concentrations in plasma and urine by HPLC-MS/MS. Immunoassays measuring cortisol lack analytical specificity due to cross-reactivity with corticosteroids. Exogenous corticosteroid ingestion could cause a high, normal or low plasma cortisol concentration measured with an immunoassay, depending on the degree of cross-reactivity with a particular immunoassay, the timing of prednisolone ingestion relative to plasma sample collection and the patient's own endogenous cortisol production. The fall in morning plasma cortisol concentration in our patient while on ketoconazole could be explained by a reduction in endogenous cortisol production, if the plasma sample was collected prior to her prednisolone ingestion. By discriminating and quantifying analytes by their mass:charge ratio, HPLC-MS/MS has high analytical specificity for cortisol and structurally similar corticosteroids such as prednisolone and 11-deoxycortisol. Our case suggests that there is an important role for cortisol measurement by HPLC-MS/MS in the setting of discordant or clinically unusual results.

Table 2. Potential causes of error in the measurement of urine and plasma cortisol by immunoassay
 Effect on cortisol immunoassayInvestigation
  1. CBG, cortisol binding globulin; HPLC-MS/MS, high pressure liquid chromatography-tandem mass spectrometry. *A number of CBG mutations have been reported that result in a low CBG concentration,[32] but none to date in the setting of concurrent Cushing's syndrome. **Reported with testosterone[33] and aldosterone[34] immunoassays, but not with cortisol immunoassays.

Heterophile/human anti-mouse antibodiesPositive or negative interference can cause falsely high or low serum/plasma concentrations respectively.[31] As antibodies are not filtered into urine, urinary free cortisol concentration is unaffected.Blocking tubes
CBG deficiencyLow plasma total (bound) cortisol relative to measurements of unbound or “free” cortisol in urine.*Measure CBG
Hook effectAnalyte concentrations exceed the antibody-antigen binding capacity of the assay, potentially producing falsely low concentrations in plasma and urine.**Dilutional studies
Cross-reacting corticosteroidImmunoassays measuring cortisol lack analytical specificity due to cross reactivity with corticosteroids (see text for more details).HPLC-MS/MS

In conclusion, factitious Cushing's syndrome is an important diagnosis to consider in the evaluation of a patient with apparent hypercortisolism. Exogenous steroid ingestion can present with biochemical results, including during IPSS, which appear consistent with ACTH-dependent Cushing's disease. The key to the diagnosis of factitious Cushing's syndrome is the recognition of discordant test results and the limitations of immunoassays, and measurement of cortisol and synthetic corticosteroids in plasma and urine using HPLC-MS/MS.

Acknowledgements

We thank Mr. B McWhinney, Chemical Pathology, Central Laboratory, Pathology Queensland, Brisbane for performing the plasma and urinary cortisol and prednisolone HPLC-MS/MS assays.

Conflict of intersest

The authors declare that they have no conflict of interest and nothing to declare. No financial support has been received in relation to this article.

Ancillary