Measuring cortisol levels in hair: potential clinical applications in Cushing’s syndrome

Keywords::

• cortisol
• Cushing’s syndrome
• cyclic
• hair
• hypercortisolism
• saliva
• serum

Biochemical diagnosis of Cushing’s syndrome

Cushing’s syndrome (CS) is a medical condition characterized by chronically increased cortisol exposure due to either endogenous or exogenous corticosteroids leading to a typical combination of signs and symptoms, such as abdominal obesity with purple striae, moon face, proximal myopathy, depressive symptoms, dyslipidemia, hypertension, diabetes and osteoporosis. Causes of endogenous CS include adrenocorticotropic hormone-producing pituitary adenoma, ectopic adrenocorticotropic hormone production and adrenal neoplasia. Diagnosis is confirmed by biochemical testing. The following first-line screenings tests are recommended in clinical practice: 24-h urine cortisol excretion (at least two measurements, since in CS there is variability in daily cortisol production), late night salivary cortisol (also at least two measurements), and 1 mg overnight or 2 mg 48-h dexamethasone suppression test [1]. If one test is abnormal and exogenous corticosteroid use and physiological causes of hypercortisolism are excluded, additional testing with one of the aforementioned tests is recommended.

In some cases diagnosis of CS is a straightforward diagnosis. However, a substantial number of patients with clinical suspicion of CS but initially normal test results develop new or progressive signs or symptoms of CS and should be retested. Similarly, patients with suspected cyclic CS may have normal test results when the disorder is quiescent and should be retested when the clinical picture of CS is present again [2]. In addition, some patients with mild CS may have discordant test results, whereas some patients without CS may have only minimally abnormal but discordant test results. At present there is no uniform recommendation for the diagnostic process to discriminate these specific groups of patients. Thus, establishing a correct diagnosis of CS can be difficult and sometimes delayed or even missed. An adequate test to evaluate chronic exposure to cortisol excess would therefore be worthwhile.

Measurement of cortisol in hair

In recent years a promising new tool to determine long-term cortisol levels has been developed and well validated in various studies [3–7]. Animal studies, as well as human studies, demonstrated that cortisol can reliably be measured in extracts obtained from scalp hair using ELISA (for review see [4]). In contrast to serum, urine and saliva cortisol measurements, hair provides the opportunity to determine cortisol exposure of the past. Since the growth rate of hair is approximately 1 cm per month [8], cortisol levels of each month can be determined by cutting the hair sample in fragments of 1 cm. The most proximal hair segments contain the most recent cortisol levels, and the distal segments show previous cortisol levels. In this way, timelines of historical cortisol levels have been documented [3,4,9,10]. Recently, Stalder et al. reported that hair cortisol concentrations show an important intraindividual stability over time in healthy individuals, suggesting that these measurements contain a strong trait component [11]. However, there is extensive evidence that hair cortisol is a dynamic marker of cortisol status, as shown by studies inducing various psychological and/or physical stress [12–15].

Potential applications for diagnosis of Cushing’s syndrome

As hair cortisol seems to reflect chronic systemic cortisol exposure it could be a potential diagnostic tool for CS. In fact, recent pilot studies indeed showed that cortisol levels in scalp hair of patients with CS are significantly higher compared with healthy controls [3,16]. These studies used a proximal hair segment of the posterior vertex of the head and compared cortisol levels to hair cortisol levels of normal individuals. Nearly all patients with CS had convincingly higher hair cortisol levels than controls, indicating that hair cortisol analysis could be a test with high sensitivity. The specificity is not clear at present. No persons with elevated cortisol levels due to other causes, such as psychiatric diseases, alcohol dependence, infections, or morbid obesity have been included in these studies comprising CS patients.

Hair cortisol measurements can be used in two ways. First, a short hair fragment of a period of interest such as 3 cm, cut as close to the scalp as possible, can be used. This would roughly correspond to average cortisol levels of the past 3 months before obtaining the hair sample. It is a robust parameter, also feasible in most males with short hair, and is not dependent on the time of sample collection, acute stress, diet, sleeping conditions and use of oral contraceptives.

Another way to use hair cortisol measurements is constructing a timeline of the past by obtaining a long hair sample from the posterior vertex and cutting it into fragments of 1 cm, with each centimeter corresponding to approximately 1 month [3,9,10]. It is a unique aspect of hair cortisol measurements that cortisol levels can be determined over prolonged periods of time. In patients with CS, hair cortisol levels changed in parallel with the clinical course of the disease with, for example, (extremely) elevated cortisol levels before treatment, significant decreases in hair cortisol levels after resection of the adrenocorticotropic hormone-producing pituitary adenoma, and gradually further decreasing postoperative cortisol levels during hydrocortisone tapering [3,16]. The latter has also been demonstrated by Gow et al., who reported a significant correlation between hair cortisol and dose of corticosteroids in patients with adrenal insufficiency using the most proximal 2 cm of hair [17].

The function of hair cortisol measurements as a historical calendar of chronic cortisol levels provides new opportunities for diagnosing cyclic CS. Cyclic CS is a rare condition, characterized by recurrent episodes of hypercortisolism interspersed by episodes of normal cortisol production [2]. The period in between the episodes of cortisol excess varies from several days to years. In cyclic CS, which can be associated with all causes of CS, establishing the diagnosis is difficult due to normal test results during intercyclic phases and the fluctuations in the clinical picture. Recently, we showed several patients with a cyclic pattern of increased cortisol secretion [Manenschijn L, Koper JW, van den Akker ELT, de Heide LH, Feelders RA, van Rossum EFC, Unpublished Data]. Hair samples were obtained during an episode of clinical cortisol excess, which was confirmed by urine cortisol levels and further work-up for CS. By retrospectively determining the hair cortisol levels of the past year we demonstrated that earlier phases of hypercortisolism had occurred. These previous elevations in hair cortisol levels corresponded with episodes of Cushingoid signs and symptoms. Thus, hair cortisol could be a valuable diagnostic tool, in particular during phases between episodes of hypercortisolism, in patients who are clinically suspected of cyclic CS.

Advantages & limitations of hair cortisol measurements

Although several studies show that cortisol levels can be reliably measured over quite an extensive historical time span there is still controversy as to the exact time frame in which cortisol levels can be reliably determined. In the past few years many studies have shown that a period of at least 3–6 months is well measurable. After this term, in some studies, a decline in cortisol levels is detected [10,18]. At present this phenomenon is not well understood, but might be due to external factors affecting the hair integrity or may be related to the techniques of processing hair samples and measuring cortisol levels used. However, other studies using slightly different protocols with respect to the processing of the hair samples and the extraction of cortisol, than those mentioned previously show substantially longer periods, up to several years, in which cortisol can be reliably determined [3,16]. It is still under investigation what the exact differences are between the techniques of hair cortisol measurement, which obtain relatively short-term (a few months) or long-term (up to years, depending on the length of the hair) historical cortisol levels. Either way, in all studies hair cortisol levels provide insight into cortisol exposure of significantly longer periods of time than serum, saliva or urine cortisol. In addition, hair cortisol is independent of many critical factors (e.g., acute stress, time of saliva or blood sample collection, use of oral contraceptives, and so on) which complicate interpretation of regular test results used in the work-up of suspected CS. Other advantages of hair cortisol include a noninvasive procedure, which is also suitable for children, as well as easy storage of hair samples at room temperature. Recent validation studies did not show an effect of hair color, frequency of hair washing and use of hair styling products on hair cortisol levels [3,6], but slightly lower levels in dyed or bleached hair. In most studies in adults no relationships between hair cortisol and gender or age were observed. However, at the extreme ages (young and old) as well as in males, preliminary data suggest that there may be higher cortisol levels [6].

Conclusion

Preliminary data show that hair cortisol measurement appears to be a promising tool for diagnosis of CS and may in particular be useful in the detection of cyclic variants. To evaluate whether other causes of hypercortisolism can be distinguished from CS using hair measurements should be investigated. It could be speculated that hair cortisol levels in individuals with a physiological cause of hypercortisolism are elevated but to a lesser extent than CS patients. Another clinical application may be the monitoring of hydrocortisone replacement therapy, which has been shown to correlate with doses used [3,17]. Thus, hair cortisol measurement could be developed into a convenient and noninvasive test to contribute to the diagnosis of CS and evaluate therapy, with the advantages of providing information about past cortisol levels and easy sample collection, independent of usual methodological issues of current tests. Additional research is needed to determine sensitivity and, importantly, specificity of this test.

Financial & competing interests disclosure

EFC van Rossum received a research grant from The Netherlands Organisation for Scientific Research (grant number 916.96.069). The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.