Generalized anxiety disorder (GAD) is a prevalent and highly disabling mental health condition; however, there is still much to learn with regard to pertinent biomarkers, as well as diagnosis, made more difficult by the marked and common overlap of GAD with affective and anxiety disorders. Recently, intensive research efforts have focused on GAD, applying neuroimaging, genetic, and blood-based approaches toward discovery of pathogenetic and treatment-related biomarkers [1]. Currently studies on biological markers in the human blood such as G protein-coupled estrogen receptor-1, oxidative stress, testosterone, dehydroepiandrosterone sulfate and cortisol performed to determine peripheral markers [2-5]. Plasma appears to be a rational source for proteomic and metabolomic measurements because it is easily accessible and because several molecules from the brain are transported across the blood-brain barrier and reach the circulation. However, drawing inferences from the neurochemical composition of plasma on the processes in the brain is not straightforward [6]. Moreover, only a few studies have been conducted on plasma-based pathogenetic and/or treatment predictors in GAD, indicating the further need to explore such potentially valuable approaches.

Erythropoietin (Epo) has been originally known for and named after its potent stimulation of erythropoiesis [7]. Apart from their hematopoietic actions, Epo is directly neuroprotective in cell culture models and after application in the brain [8-12]. Expression of Epo and its receptor (EpoR) in the nervous system as well as its multifaceted protective actions in cell culture and animal models (e.g., induction of anti-apoptotic, antioxidant and anti-inflam-matory signaling in neurons, glial and cerebrovascular endothelial cells, stimulation of angiogenesis, and neurogenesis), have been reported extensively [13-19]. Epo and EpoR have interesting properties, which make it a candidate for investigation as a novel therapeutic agent in neuropsychiatric diseases. Epo has shown promising procognitive effects in psychiatric disorders, providing support for a neurotrophic drug development approach. This trophic cytokine has not been assessed in the context of anxiety, but recent studies have implicated Epo as having clinical potential. Promising results from clinical testing of Epo in multiple psychiatric diseases strengthen the hypothesis that it produces behavioral effects via central mechanisms associated demonstrating efficient transport of Epo across the blood brain barrier in humans set the stage for further clinical testing in different disease settings [20]. As an add-on drug in chronic schizophrenia, Epo significantly improved cognitive function over placebo and reduced levels of glial damage marker S100B [21]. Interestingly, Epo appears to be the only drug capable of delaying the progressive thinning of the cortex that occurs in schizophrenia, correlating with an improvement in attention and memory [22]. Furthermore, Epo appears to module cognitive and emotional processes in major depressive disorder (MDD) patients [23]. The administration of Epo in MDD patients with treatment-resistant enhanced verbal memory and word recognition, and reduced depression symptoms [24]. In these patients Epo treatment was associated with changes in the right superior frontal gyrus and left hippocampal activity and working memory improvement [25]. In animal models for depression, the expression of Epo mRNA in hippocampus decreases by the effect of chronic unpredictable stress. Especially, this Epo mRNA reduction is efficiently reversed by antide-pressant treatment [26]. In rodents, peripheral administration of Epo produces robust anti-depressant like effects in the forced swimming test and the novelty-induced hypophagia test. In depressed patients, Epo administra-tion for 3 days reduces the neural activity in amygdala, hippocampus and frontoparietal regions, as well as improves mood and hippocampus-related memory [23]. Taken together, this evidence suggests that the Epo serum level is useful to evaluate the cognitive and emotional progression in MDD patients. Furthermore, Epo is a candidate treatment for cognitive impairment in unipolar and bipolar disorders and modulates cognition related neural activity across a fronto-temporo-parietal network [27]. Epo treatment for 8 weeks improves the cognitive processing and seems to prevent neuronal in the hippocampal regions CA1, CA3, and subiculum [27,28]. Taken together, this evidence suggests that Epo is relevant for clinical uses and can be useful as biomarkers to evaluate the mood state and disease progression in bipolar patients.

Although there are information about Epo as a novel therapeutic agent in patients with schizophrenia, multiple sclerosis, depression and bipolar disorders, to the best of our knowledge, there are no study investigating serum the levels of Epo and EpoR and their’s diagnostic values in GAD patients. Therefore, the present study aimed to assess serum Epo and EpoR levels in drug-naive patients who were diagnosed with GAD.

This is a cross-sectional study that was conducted at Kahramanmaras Sutcu Imam University (KSU), between 2019 and 2020. The procedures of the study complied with the ethical standards of the relevant national and institutional committees for human experimentation and with the Declaration of Helsinki of 1975, revised in 2008. The study protocol was approved by University Hospital (KSU) Human Research Ethics Committee (date: October 7, 2019, decision number: 2019/266-4). All participants provided written informed consent and participated in the research voluntarily. The sample size of original study was determined by G*Power analysis (version 3.1.9.4) [29]. A total of 75 subjects were included in the study, 45 newly diagnosed drug-naive patients with GAD (age rages: 18−50 years; male/female: 17/28) and 30 healthy individuals (age ranges: 20−52 years; male/female: 10/20) as control group. The inclusion criteria for patients were as follows: (1) First-episode GAD patients according to Diagnostic and Statistical Manual of Mental Disorders 5th edition (DSM-V) criteria with a structured clinical interview (SCID) by an experienced psychiatrist; (2) have not been previously or currently treated with psychiatric medications. Exclusion criteria for patients were as follows: (1) Presence of comorbid psychiatric disease or personality disorder according to DSM-5 diagnostic criteria (a total of thirteen patients were excluded from the study due to comorbidity, eight of them with comorbid major depressive disorder and five of them with comorbid panic disorder); (2) comorbid substance or alcohol abuse; (3) smoker; (4) have a serious medical condition, such as a head injury, ongoing neurological or physical illness; (5) lactation or pregnancy; (6) mental retardation; (7) treatment of antioxidant or anti-inflammatory agents; (8) severe systemic diseases (epilepsy, diabetes mellitus, hypertension); (9) inadequate blood sampling; (10) Those who do not fulfill the requirements of the research and fill in the forms incompletely (twenty participants were excluded from the study).

The control individuals were consisted of volunteers who were admitted to other clinics of the hospital and did not have psychopathology which was assessed with SCID-1 and SCID-II. All participants were assessed by a trained clinical psychiatrist. Control group was evaluated the same exclusion criteria as patients such as medical comorbidity, malnutrition, drug abuse, smoke and antioxi-dant medication etc. Those who had one of these conditions were excluded from the study. The ages, sexes and body mass indexes (BMIs) of the groups were matched. We used strict exclusion criteria to obtain homogenous groups. Demographic data and complete medical histories were obtained, and physical examinations were performed for all the participants. Socio-demographic form was developed by the researchers to collect information regarding socio-demographic data including age, sex, BMI, marital status, duration of education, psychiatric and medical history, alcohol and substance use. The Hamilton Anxiety Rating Scale (HAM-A) was applied to the patient and control groups to evaluate the severity of symptoms [30,31].

Biochemical Analysis

All blood samples were taken from both groups between 7:30 AM to 17:00 PM, because diurnal variation of Epo or EpoR have been reported in literature [32]. Then, the serum was promptly separated, in a refrigerated centrifuge, and stored at −20°C until analysis. The serum levels of Epo and EpoR were measured by a quantitative sandwich enzyme immunoassay technique (ELISA) using a commercial kit (MyBioSource Company) according to the manufacturer’s instruction. Absorbances of each parameter were measured at 450 nm using a microplate reader (Model 680; BioRad), and the results were calculated using GraphPad PRISM 5.0 (GraphPad Software Inc.). Long transformation was performed for both analyses. All samples were tested in duplicate. In this study, we determined that the intra and inter-assay coefficients of variations for Epo were 6.56% and 10.35%, and 7.03% and 13.98% for EpoR, respectively.

Statistical Analysis

Statistical analysis was performed using the Statistical Package for Social Sciences, ver. 22.0 (IBM Co.). A pvalue of less than 0.05 was considered statistically sig-nificant. The normality of continuous variables was assessed using Shapiro−Wilk’s Wtest. Relation-ships between the categorical variables were evaluated using the chi-square test. To compare of mean differences for normally distributed continuous variables between the two groups, a Student’s ttest was used. While investigating associations of data, correlation coefficients and their significance were calculated with Spearman’s test (for non-normally distributed variables) and Pearson’s test (for normally distributed variables). A receiver operator characteristics (ROC) curve was plotted in order to find the cut-off point.

Seventy-five individuals were included in the study. The mean age of the GAD group (n = 45) was 36.25 ± 6.02 years, and 28 (62.22%) were female. The mean age of the control group (n = 30) was 35.06 ± 5.56 years, and 20 (66.66%) were female. No significant differences were found between the groups in terms of age, sex, BMI, marital status and duration of education (p > 0.05). HAM-A scores were significantly higher in patients than controls (medians were 14 and 8 respectively, p < 0.001). The average episode duration was 31.2 weeks for patients. These results were shown in Table 1.

Epo levels were significantly lower in patients than healthy controls (medians were 6.78 mIU/ml and 10.75 mIU/ml, respectively, p < 0.001). The highest and lowest bounds for Epo in GAD patients were 7.64 mIU/ml and 5.26 mIU/ml, respectively. The highest and lowest bounds for Epo in the control group were 14.76 mIU/ml and 9.13 mIU/ml, respectively (Fig. 1). EpoR levels were significantly higher in patients than in controls (medians were 1.42 ng/ml and 0.69 ng/ml respectively, p = 0.038). The highest and lowest bounds for EpoR in GAD patients were 1.88 ng/ml and 1.18 ng/ml, respectively. The highest and lowest bounds for EpoR in the control group were 0.97 ng/ml and 0.52 ng/ml, respectively (Fig. 2). More-over, the ratio of Epo/EpoR was significantly lower in GAD patients than controls (p < 0.05) (Fig. 3).

While analyzing correlations, we found significant and high correlations between HAM-A scores and Epo or between HAM-A scores and EpoR levels. As seen in Table 2, according to Pearson correlation analysis, there was a negative significant correlation between Epo levels and HAM-A score (r = −0.749, p < 0.001). However, there was a positive significant correlation between EpoR levels and HAM-A score (r = 0.755, p < 0.001). Additionally, the results of correlation analysis showed that there is a negative correlation between Epo and EpoR levels (r = −0.840, p < 0.001) as shown in Table 2.

A ROC curve was plotted for Epo and EpoR levels. Areas under the curve were 0.901 for Epo (p < 0.001), and 0.912 for EpoR (p < 0.001). These findings indicate that Epo and EpoR levels are diagnostic. The cut-off point was 6.11 mIU/ml for Epo, and all of the patient group Epo levels were under the cut-off point. The sensitivity and specificity of Epo were 100% and 85.9%, respectively. For EpoR, the cut-off point was 0.62 ng/ml, and all of the patient group EpoR levels were above the cut-off point. The sensitivity and specificity of EpoR were 98.2% and 86.3% (Figs. 4, 5).

To our knowledge this is the first study investigation of Epo and EpoR in patients with GAD. We found that the levels of Epo in patients with GAD were lower than the control patients. However, EpoR levels were increased in these patients. This situation may be due to increased of neuroinflammation in patients with GAD. Camacho recommended that anxious-depression should be considered as a chronic inflammatory phenomenon but the only longitudinal study found the association between GAD and increased C-reactive protein level to be attributable to body mass index and medication use [33,34]. Some authors reported C-reactive protein levels elevated in male patients with current anxiety disorders and immune dysregulation in patients with a late-onset anxiety disorder [35]. Many studies showed that an integrated specificity model express specific patterns of biological responses to specific psychological states and an anxiety-specific effect on inflam-matory activity in clinically anxious individuals [36-38]. Recent studies have characterized Epo as a potent anti-inflammatory cytokine in chronic inflammatory disorders and infectious diseases [39]. Also, recently, multiple lines of evidence have shown that both endogenous and exogenous Epo has protective roles in central nervous system injury processes, such as ischemia-reperfusion injury. Although the presence of functional EpoRs in neurons has been challenged, Epo selectively reduced inflammatory and oxidative stress processes associated with brain ischemia, and prevented neuronal apoptosis [40-45].

Moreover, in our study, decreased levels of Epo/EpoR ratio may due to low Epo levels. Until now, Epo/EpoR ratio has not been reported in patients with GAD. So, we did not compared to our results. This study suggests that efforts aiming to increase either Epo expression or the activation of EpoR in the GAD may be a promising target for GAD treatment, especially in stopping the progression, and potentially reversing the well known behavioral morbidities.

High correlation values constitute an important part of our findings. The correlation coefficient is shown with the ‘r’ symbol. A “rvalue” ≤ 0.35 represents low or weak correlation, between 0.36 and 0.67 shows moderate correlation, 0.68 to 0.90 shows high correlation, and 0.90 to 1.0 shows very high correlation [46]. We found that a significant negative correlation between HAM-A scores and Epo levels. However, we found that a positive correlation between HAM-A scores and EpoR levels. There are no studies between Epo and HAM-A scores, or between EpoR and HAM-A scores in patients with GAD. So, we did not compared to our results. We believe that a high correlation coefficient forms a basis for the detection of biomarkers.

Identifying diagnostic biomarkers for psychiatric disorders is a rising topic of interest. Anxiety disorders are one of the most commonly reported disorders in psychiatry, causing a high medical and socio-economic burden. Due to the ambiguity of the diagnosis and a large number of underdiagnosed patients, researchers are looking for laboratory tests that could facilitate the diagnosis of anxiety disorders in clinical practice and would allow for the earliest possible implementation of appropriate treatment. Unfortunately, research in the field of biomarkers is hampered by insufficient knowledge about the etiopathogenesis of anxiety disorders, the significant heterogeneity of anxiety disorders, frequent comorbidities, and low specificity of biomarkers. The development of appropriate biomarker panels and their assessment using new approaches may have the prospective to overcome. We think that ROC analysis is important in order to make a disease-specific evaluation and to examine its possible diagnostic value, since there is no previous study on the investigated parameters in our study. The areas under the curve (AUC) obtained on the ROC curve is used to measure the diagnostic value of the marker (0.6 to 0.7, poor; 0.7 to 0.8, fair; 0.8 to 0.9, good; 0.9 to 1, very good) [47]. In previous studies on diagnostic performance in patients with GAD, malondialdehyde, superoxide dismutase and paraoxonase (AUCs: 1.0, 1.0, and 0.980, respectively) were reported to have very good diagnostic value, while catalase (AUC: 0.648) was reported to have poor diagnostic value [3,48]. To the best of our knowledge, our study is the first study investigating the diagnostic value of Epo and EpoR with a ROC curve in GAD. We found satisfactory statistical results in the ROC curve we applied to determine the diagnostic value of Epo and EpoR levels. We calculated the AUCs for Epo and EpoR were 0.901 and 0.912, respectively. This result provides a very good potential ability to distinguish patients with GAD from healthy controls. Although we obtained very good diagnostic value, we do not consider this result as a new biomarker discovery, but we believe that more studies are needed to test the diagnostic value of these parameters. Besides, inflammatory cytokines including Epo could be related to not only GAD, but also other psychiatric disorders. Even though Epo and its receptor showed excellent diagnostic value in this study compared to healthy controls, these biomarkers may not specific to GAD. Furthermore, the low levels of Epo may not indicate the high possibility of GAD. However, our study shows possible implication for hormon-receptor (Epo-EpoR) level as an important research area for detection of peripheral biomarkers.

The major limitations of the present study were the small number of patients, its cross-sectional design, and the lack of pre- and post-treatment Epo-EpoR levels mea-surements. Also, In terms of diagnostic values of Epo and EpoR validity as a biomarkers in GAD needs confirma-tion. Nevertheless, to our knowledge, there is no study in the literature assessing the serum Epo and EpoR levels in GAD patients; accordingly, this study has the feature of being the first study on this issue. Moreover, all patients were drug-naive. The strength of our research is that our patients and healthy controls are the most homogenous groups studied thus far. The outcome of the present study is an important in terms of providing data for a treatment approach through target receptors in GAD.

Our current findings showed that serum Epo/EpoR ratios were significantly lower in the patients with GAD than in the control subjects. Also, Epo/EpoR ratio may be an important biomarker for GAD. We thought that recombinant erythropoietin may be a good choice for GAD treatment. Furthermore, we found high diagnostic values for Epo and EpoR levels for GAD. Although our results show excellent diagnostic value, we do not interpret these data as a discovery of a new biomarker. These findings should be considered preliminary and needing verifica-tion by further studies.

We gratefully to Psychiatrist Dr. Ebru Findikli for key role in physical examination of patients and in the obtain of samples and perform this study’s medical record. Also, we thank to Psychiatrist Dr. Onur Hursitoglu for reviewing the revised form of our article.

No potential conflict of interest relevant to this article was reported.

None.