Diagnosis of CS is usually prompted by clinical suspicion in patients with multiple and progressive signs and symptoms. Screening is also often considered in patients with other features of CS, for example poorly controlled diabetes, hypertension, or unexplained osteoporosis. Patients with an incidentally discovered adrenal mass should be evaluated (Newell-Price et al., 2006; Sharma et al., 2015).
Diagnosis guidelines recommend a choice of four tests for the initial investigation of CS (Nieman et al., 2008; Gilbert and Lim, 2008).
The 24-hour UFC test gives an indication of the free (unbound) cortisol circulating in the blood over a 24-hour period. If the patient is suspected of having CS and the UFC results are normal, they should be repeated two to three times over a 24-hour period to confirm the diagnosis. If all three results are normal, then a diagnosis of CS is highly unlikely (Arnaldi et al., 2003; Nieman et al., 2008).
UFC levels can be highly variable; a value 4-fold greater than normal can be considered as diagnostic of CS. Milder elevations in cortisol levels may be associated with conditions such as chronic anxiety, depression, alcoholism and are also seen in pregnancy (Arnaldi et al., 2003).
However, UFC cannot be considered a single universal screening test as it may not detect preclinical or mild CS (Arnaldi et al., 2003). In addition, the sensitivity and specificity of UFC can be less than ideal when compared with other diagnostic options and so it is not generally recommended when other first-line diagnostic tests are available (Alexandraki and Grossman, 2011).
Late-night salivary cortisol testing involves the measurement of salivary cortisol levels, which correlate with free plasma cortisol. This has high diagnostic sensitivity and specificity and is particularly useful in patients with cyclical changes to cortisol levels, as tests can be easily repeated over time (Arnaldi et al., 2003). It has become increasingly popular as it can be used in out-patients, but the assay has to be validated and normative values agreed. It appears to be less useful in older patients, those with diabetes mellitus and shift workers or patients with variable sleep patterns (Gilbert and Lim, 2008).
The overnight DST is used to differentiate patients without CS from those with milder forms of the condition. Patients ingest an overnight dose of 1 mg dexamethasone between 23:00 and 24:00, and then measure fasting plasma cortisol levels between 08:00 and 09:00 the following morning. A serum cortisol reading <1.8 µg/dL (50 nmol/L) virtually excludes the diagnosis of CS, but false positives are common, particularly in women on oestrogen therapy or patients taking anticonvulsants or medications which induce hepatic enzyme metabolism through cytochrome P450 3A4 (Gilbert and Lim, 2008). The accuracy of this test is limited as non-CS patients can occasionally fail to suppress cortisol to this level due to other acute or chronic conditions. However, the test is low-cost, and easy to administer, and can be used to essentially exclude the diagnosis, especially when it is thought unlikely (Arnaldi et al., 2003).
In this test (usually reserved for consideration in special circumstances), serum cortisol is measured at 09.00 and then 48 hours later at 09.00 after 0.5 mg dexamethasone has been administered at a dose of 0.5 mg 6-hourly. A final level >1.81 µg/dL (50 nmol/L) is around 95% sensitive and specific for Cushing’s Syndrome (Arnaldi et al., 2003; Gilbert and Lim, 2008).
The interpretation of results from any of the initial tests are influenced by clinical pre-test probability so patients who test positively upon initial screening should be referred to an endocrinologist as well as patients who had a normal result but high pre-test probability (Nieman et al., 2008; Gilbert and Lim, 2008). Confirmatory tests may be another of the initial tests or guidelines suggest either the dexamethasone-CRH test or midnight serum cortisol test (Gilbert and Kim, 2008).
The patient must ingest 0.5 mg dexamethasone orally every 6 hours for 48 hours. CRH (1 µg/kg) intravenously is administered two hours after the last dose of dexamethasone and if the plasma cortisol value 15 minutes after CRH administration is greater than 1.4 µg/dL then the patient is likely to have CS (Arnaldi et al., 2003). This is a cumbersome test and there is some evidence that it is no more accurate than the standard 48-hour DST (Martin et al., 2006).
Patients with CS often have early morning serum cortisol levels that are within normal parameters but lack a normal circadian rhythm. Measurements of plasma cortisol levels at midnight have a high degree of sensitivity; measurements of cortisol levels at other times are less helpful for diagnosis of CS (Arnaldi et al., 2003). However, the patient must be hospitalised for at least 48 hours, and guidelines describe different diagnostic cut-offs for both ‘sleeping’ and ‘awake’ serum cortisol levels (Nieman et al., 2008). Despite inconsistent findings in the literature, serum midnight cortisol levels of <207 nmol/L would exclude the diagnosis of CS (Gilbert and Lim, 2008).
There are also other diagnostic tests which can be used and these include the CRH test and the desmopressin test.
In the CRH test, 100 µg of human- or ovine-sequence CRH (actually, CRH-41) is administered intravenously, and cortisol and ACTH are measured over two hours. A rise from basal to peak plasma ACTH above 50% and serum cortisol above 20% is highly suggestive of Cushing’s Disease. Patients with ectopic ACTH syndrome rarely respond while a failure to respond is seen in some 20% of patients with Cushing’s Disease. However, a failure to respond shifts the probability much more towards ectopic ACTH. In addition, a rise in ACTH and/or cortisol above the normal range is highly suggestive of Cushing’s Disease as opposed to a normal patient (Arnaldi et al., 2003; Peters et al., 2006).
The desmopressin test can also be a useful adjunctive tool for the evaluation of hypercortisolaemic patients because it is able to differentiate between mild CS and pseudo-CS. Pseudo-CS can be observed in patients with major depression, alcoholism and alcohol-withdrawal syndrome, severe truncular obesity, poorly controlled non-insulin dependent diabetes and occasionally polycystic ovary syndrome (Moro et al., 2000).
After an overnight fast, patients are injected with 10 µg of the vasopressin analogue desmopressin and blood sampling for plasma cortisol and ACTH measurements are taken 30 minutes before and at 0, 15, 30, 45, 60, 90 and 120 minutes after injection. Plasma cortisol and ACTH levels are significantly higher in patients with CS than those of normal, obese and pseudo-CS patients (Moro et al., 2000).