Twenty-nine CFS patients who were diagnosed according to Centers for Disease Control (CDC)¹ criteria (20 females and 9 males; mean age: 41.89±8.53 years; mean disease duration: 80.13±57.16 months) and 20 healthy controls (16 females and 4 males; mean age: 42.50±9.42 years) were included in the study. The patients were selected consecutively from the outpatient population of the Psychiatry Clinic of Erciyes University Medical School. None of the patients was or had ever been diagnosed with any psychiatric disease according to Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, Text Revision (DSM-IV TR) criteria.²⁹ The control group included healthy hospital personnel who did not have any psychiatric disease history or diagnosis. Medical disorders (e.g. neurological, endocrinological, or metabolic) were excluded through the history, clinical examinations, and the evaluation of the results of the laboratory tests in all subjects. None of the subjects was using any medication or oral contraceptive drug that could affect the functions of the HPA axis. The female subjects were all pre-menopausal.

This study was approved by the local Ethics Committee. Written informed consent was obtained from each subject after the study was explained to them.

In the patients the first stage involved a standard clinical evaluation and the administration of the following clinical scales: the Hamilton Depression Rating Scale (HAM-D);³⁰ Hamilton Anxiety Rating Scale (HAM-A);³¹ and Newcastle Research Group ME/CFS Score Card (NRG),³² which consists of 10 items and measures the physical and cognitive functions, and on which a score of ≥15 is highly suggestive of CFS. After this, catheters were placed in the forearm veins of the subjects at 8.00 am after a starvation period of 12 hours and, following a resting period of 30 minutes, blood samples were collected into tubes. The samples were centrifuged directly Separated serum samples were stored at -70℃. Basal blood samples of the female patients were obtained between the 4th and 7th days of their menstrual cycles. Then, all patients were started on fixed dose 8 mg/d oral galantamine hydrobromide. After four weeks of treatment, the same psychiatric measures and blood sampling procedure were repeated in the patients. However, seven patients were later excluded as they were unable to tolerate the treatment because of nausea or would not agree to cooperate, and the study continued with 22 patients. The same psychiatric measures and blood taking procedure for hormonal values were carried out only once in healthy controls.

At the end of the treatment period we separated our patients into two subgroups: responders (n=10) and non-responders (n=12). Twenty percent and more reductions in NRG scores were accepted as a response to treatment arbitrarily.³²

Serum cortisol levels [{Diagnostic System Laboratories (DSL, Webster, TX, USA) DSL-2100 reagents}, (sensitivity: 0.3 µg/dL, coefficient of variation: 8.4)] and DHEAS levels [{DSL (Webster, TX, USA) DSL-3500 reagents}, (sensitivity: 17.0 ng/mL, coefficient of variation: 6.3)] were measured by the radioimmunoassay (RIA) method.

Data distributions for each hormonal evaluation were normal according to the Kolmogorov-Smirnov test. The patients and controls were compared with respect to sex using the χ² test. The significance of differences between age, body mass index (BMI), and clinical scale scores of the patients and controls was assessed with a t-test for independent groups. To compare hormone levels of the patients and controls a two-way analysis of variance (ANOVA) test was performed by taking the presence of disorder and gender as fixed factors, and age and BMI as covariates. A repeated measure ANOVA test was carried out for the comparison of the hormone levels of the patients before and after the treatment. The same tests were used for analyses of subgroups according to the response to the treatment (responders and non-responders). Pearson's correlation test was performed to investigate the relationships between hormone levels and clinical and demographical variables in the patients and controls. The changes in hormonal values and those in clinical scales were calculated as delta values by subtracting the post-treatment values from the pre-treatment values.

Many studies of patients with CFS have found HPA axis dysregulation and hypocortisolism. In these studies cortisol levels were decreased in 24-hour urine,⁷ serial blood samples,¹⁰ or saliva,¹¹ as seen in our study, but it is not clear whether low cortisol levels and HPA hypofunction in CFS are a cause or result of the illness. Taking into account the finding of this study of the reduced cortisol levels of the pre-and post-treatments, we can infer that HPA alterations in CFS are the persisting features of the disease, although we do not make any suggestions as to whether it is a cause or a consequence of the illness. Reduced HPA axis functioning⁶ has also been suggested to cause some of the symptoms, like reduced energy, post-exertional malaise, and headaches¹ seen in CFS patients. In a previous study 5-10 mg/d hydrocortisone therapy resulted in a fall in fatigue scores of 34%.³³ One explanation for hypocortisolism in CFS may be increased glucocorticosteroid receptor sensitivity.³⁴

The finding of unchanged cortisol level after galantamine treatment of this study is consistent with a recent study by Blacker et al.²⁶

How can we explain the finding of the study that while cortisol levels remained lower throughout the galantamine treatment some patients showed symptomatic improvement at partial level? It has been suggested that a functional deficiency of cholinergic neurotransmission may result in hypoactive HPA axis under stress conditions³⁵ and may have a role at pathophysiology of CFS. Cholinergic neurotransmission stimulates growth hormone and insulin-like growth factor I secretions, which plays an important role in the regenerative functions of the peripheral nerves and skeletal muscles.³⁵ Taking together these findings, increased cholinergic activity by galantamine treatment might trigger well-being independently of cortisol levels in some patients. This means, at least in some patients, that it may not be necessary to normalize cortisol levels for an improvement to be achieved. However, if we followed the patients in the long-term we could have a better idea about cortisol levels. As a conclusion we can suggest that in some patients with CFS there may be dysregulation in other systems like the cholinergic system along with the HPA axis.

Given the findings of lower pre- and post-treatment cortisol levels, higher pre-treatment DHEAS levels compared to post-treatment, and higher DHEAS/cortisol molar ratios in the patient group, we can suggest that there may be a disturbance in the 'metabolic shift' from androgens to glucocorticoids.

DHEA/cortisol or DHEAS/cortisol molar ratios were used in many studies as variants that reflected the homeostasis between androgens and glucocorticoids at adrenal level.^17,36 Scott et al.³⁷ performed a study in CFS patients without any psychiatric comorbidity. They found that DHEA and DHEAS levels were lower than in healthy controls and explained this finding with mild adrenal cortical atrophy (failure) secondary to ACTH deficiency. Scott et al.³⁷ also did not identify any difference between CFS patients, depressives, and healthy controls with respect to basal cortisol levels. De Becker et al.³⁸ demonstrated that basal DHEA levels were normal, but DHEA responses to ACTH stimulation were blunted in CFS patients. No cortisol levels were reported in that study. Cleare et al.³⁶ found higher basal DHEA levels in CFS patients with no psychiatric comorbidity compared to healthy controls; however, they did not report any difference in DHEAS levels. In the same study, they did not find any difference between patients and controls with respect to cortisol/DHEA molar ratio. In addition, the investigators administered low dose (5-10 mg/d) hydrocortisone to CFS patients and determined significant decreases in DHEA levels in patients who reported declined fatigue scores with treatment. They found significant reductions in DHEAS after hydrocortisone treatment. Similarly, we also found decreased DHEAS levels after galantamine treatment in CFS patients. This may be interpreted as the fact that the acetylcholinergic effect of galantamine facilitates the release of CRH hormone and in turn, causes elevated DHEAS levels to decrease through recovering the existing tentative disturbance in metabolic shift from cortisol to DHEAS.

Some studies have reported low^39,40 or normal^36,41 morning DHEAS levels in CFS patients. An investigator reported a slightly high DHEA levels in CFS patients.¹⁸ Similarly, Goldberg and Lichten⁴² reported DHEAS levels above the reference limit in 20% of their patient group, made up of 140 females with CFS. In our study increased pre-treatment DHEAS/cortisol molar ratio may have been due to impaired HPA axis activity and decreased cortisol levels or a defect in metabolic shift from androgen to glucocorticoid production. It was proposed that the literature findings about higher cortisol/DHEA molar ratios in major depression were the most important indicators of excessive cortisol exposure of the brain.^43,44 On the other hand, our finding of high DHEAS/cortisol molar ratios might lead to further decreases in the low cortisol effects on the brain. However, only basal levels were investigated in this study. Circadian measurements might provide further information on this subject.

The controversies in all these studies might be explained by the differences in clinical characteristics (psychiatric comorbidity, medicines), methodological discrepancies (differences in the timing of blood sampling or laboratory techniques), and other factors that might affect functions of the HPA axis, like inactivity. Our study had the advantages of including pure and untreated CFS patients.

When we separated our patients into two subgroups as responders and non-responders to treatment, the pre-treatment basal DHEAS levels and DHEAS/cortisol molar ratios of the responders were significantly higher than those of the controls; however, it was interesting to find the values of the non-responders similar to those of the controls. To our knowledge, these findings are unique in the literature. It was concluded that high levels of pre-treatment DHEAS activity and its ratio to cortisol might be an indicator of the responsiveness to treatment.

It is possible to explain the high pre-treatment basal DHEAS levels and DHEAS/cortisol molar ratio in responders and the decrease in DHEAS levels with galantamine hydrobromide treatment in all CFS patients and especially in responders in several ways. The first is that high pre-treatment DHEAS levels and DHEAS/cortisol molar ratio in responders may be due to hypocortisolism supposed to be present in CFS. Improvement in relative hypocortisolism with treatment may provide a return to the normal DHEAS pattern. However, we did not observe any significant increase in cortisol levels. A longer period might be needed for such a finding. The second possible explanation is that some abnormalities in the metabolic pathway of DHEAS like 21-hydroxylase enzyme deficit⁴² may lead to higher pre-treatment DHEAS synthesis. The third is that normalization of high DHEAS levels in responders may be secondary to improvement in some clinical features such as sleep disturbances, anxiety or some depression symptoms which may be available in these patients.⁴⁵ Indeed, the finding of the present study of a positive correlation between reduction of the DHEAS levels with galantamine treatment and recovery in clinical depression and anxiety scales in only respondent patients supports this hypothesis strongly. The fourth is that abnormal glucocorticoid receptor sensitivity may improve with treatment and the production may shift from DHEAS to cortisol. Finally, the cholinergic effect of galantamine that facilitates release of CRH may cause reduction in DHEAS levels and the normalization of high DHEAS/cortisol molar ratio. We think that the duration of our experimental treatment may not have been sufficient to show these changes.

There are several limitations of the study. A small number of patients and the absence of placebo control group might have prevented us from observing the therapeutic and hormonal effects of the galantamine more accurately. Additionally, the short duration of the treatment might obscure the long-term effects of the drug on fatigue scores or on the hormones such as cortisol.

In conclusion, decreases in basal DHEAS levels and DHEAS/cortisol molar ratios in CFS patients with galantamine treatment may somewhat support the hypothesis that cholinergic deficit may be present in CFS. On the other hand, DHEAS findings of responders suggest that high pre-treatment values of this hormone may be an indicator of the treatment responsiveness. If DHEA and DHEAS levels play an important role in CFS physiopathology, we think that, in addition to the employment of non-invasive techniques like the measurement of cortisol levels in the saliva, the evaluation of DHEA (-S) levels with response to naturalistic stressors (challengers) like exercise or social stress in future studies will be beneficial.