Understanding importance of clinical biomarkers for diagnosis of anxiety disorders using machine learning models.

Anxiety disorders are a group of mental illnesses that cause constant and overwhelming feelings of anxiety and fear. Excessive anxiety can make an individual avoid work, school, family get-togethers, and other social situations that in turn might amplify these symptoms. According to the World Health Organization (WHO), one in thirteen persons globally suffers from anxiety. It is high time to understand the roles of various clinical biomarker measures that can diagnose the types of anxiety disorders. In this study, we apply machine learning (ML) techniques to understand the importance of a set of biomarkers with four types of anxiety disorders-Generalized Anxiety Disorder (GAD), Agoraphobia (AP), Social Anxiety Disorder (SAD) and Panic Disorder (PD). We used several machine learning models and extracted the variable importance contributing to a type of anxiety disorder. The study uses a sample of 11,081 Dutch citizens' data collected by the Lifelines, Netherlands. The results show that there are significant and low correlations among GAD, AP, PD and SAD and we extracted the variable importance hierarchy of biomarkers with respect to each type of anxiety disorder which will be helpful in designing the experimental setup for clinical trials related to influence of biomarkers on type of anxiety disorder.

Stress hormone response to the DEX-CRH test and its relation to psychotherapy outcome in panic disorder patients with and without agoraphobia.

This study tested whether the hormonal stress response to the DEX-CRH test may be predictive of the psychotherapy success for panic disorder (PD). Thirty-four patients diagnosed either with agoraphobia with PD or PD without agoraphobia were subjected to cognitive behavioural therapy (CBT). Patients (pre-therapy) and healthy volunteers were exposed to the DEX-CRH test. Blood samples were taken for cortisol and adrenocorticotropic hormone (ACTH) assessment. Established panic-specific questionnaires were handed out for the pre-therapy and post-therapy evaluation of disease severity (with reference to panic beliefs and agoraphobic cognitions, fear of bodily sensations, agoraphobic avoidance behaviour). Repeated measures ANCOVA were conducted for the analysis of the pre-therapy hormonal response, and Pearson's correlation analysis to test for associations with the psychotherapy outcome. Data analyses revealed large effect sizes for CBT in the clinical measures (η2 ≥ 0.321), main effects of time for cortisol and ACTH with no differences between both groups, and significant associations between cortisol release and agoraphobic cognitions for the patients. PD diagnosis had no impact on the hormonal response. However, those patients with higher cortisol release showed less improvement after CBT (significantly for agoraphobic cognitions). Clinical implications of these findings are the prediction of the therapy success from a potential endocrine correlate whose persistency (if assessed repeatedly) during the treatment may predict (non-)response to the current treatment, possibly representing a decision support for a change in treatment to avoid the continuation of an inefficient treatment.

Anxiety disorders and CRP in a population cohort study with 54,326 participants: The LifeLines study.

OBJECTIVES: Growing evidence indicates that inflammatory processes may play a role in the pathogenesis of anxiety disorders. Nevertheless, much remains to be learned about the involvement of inflammation, including C-reactive protein (CRP), in specific anxiety disorders. This study examines the relation between anxiety disorders and CRP. METHODS: Associations of serum CRP with anxiety disorders were determined in a large population study (n = 54,326 participants, mean age = 47 years; 59% female), the LifeLines cohort. Depressive and anxiety disorders (generalized anxiety disorder, social anxiety phobia, panic disorder with or without agoraphobia and agoraphobia without panic disorder) were assessed using the Mini-International Neuropsychiatric Interview. RESULTS: Anxiety disorders, with the exception of social anxiety disorder, were significantly associated with increased CRP. After adjusting for demographics, life style factors, health factors, medication use, depression, and psychological stressors, CRP remained significantly associated with panic disorder with agoraphobia (ß = 0.01, P = .013). Moreover, CRP levels were significantly higher in people with panic disorder with agoraphobia compared to other anxiety disorders, independent of all covariates (F = 3.00, df = 4, P = .021). CONCLUSIONS: Panic disorder with agoraphobia is associated with increased CRP, although the effect size of this association is small. This indicates that neuroinflammatory mechanisms may play a potential role in its pathophysiology.

The bidirectional relationship between anxiety disorders and circulating levels of inflammatory markers: Results from a large longitudinal population-based study.

BACKGROUND: Although there has been abundant research on chronic low-grade inflammation as a potential mechanism underlying the link between mood disorders and cardiovascular risk, less is known about the role of inflammatory factors and anxiety disorders. The aim of this paper is to evaluate the bi-directional associations between inflammatory markers including interleukin (IL)-6, tumor necrosis factor (TNF)-α, and high sensitivity C-reactive protein (hsCRP) with anxiety disorders and its subgroups. METHODS: The sample consisted of 3,113 participants (53.7% women; mean age: 51.0, S.D. 8.8 years), randomly selected from the general population, who underwent comprehensive somatic and psychiatric evaluations at baseline and follow-up (mean follow-up duration = 5.5 years, S.D. 0.6). Anxiety disorders were assessed with semistructured diagnostic interviews. Inflammatory biomarkers were analyzed in fasting blood samples. RESULTS: After adjustment for potential confounders, current anxiety disorders (ß = 0.09, 95% CI 0.00-0.17) and agoraphobia (ß = 0.25, 95% CI: 0.07-0.43) at baseline were associated with a steeper increase of hsCRP levels over the follow-up period. Current posttraumatic stress disorder (PTSD) was associated with a lower increase of IL-6 levels over the follow-up period (ß = -0.52, 95% CI: -1.00/-0.04). There was no evidence for an association between inflammation markers at baseline and anxiety disorders at follow-up. CONCLUSIONS: The prospective association between agoraphobia at baseline and hsCRP levels over the follow-up period suggests that chronic low-grade inflammation may be a consequence of this condition. The decrease in IL-6 in PTSD also requires further investigation. No evidence was found for chronic low-grade inflammation as a predictor of future anxiety disorders.

Phosphate levels as a possible state marker in panic disorder: preliminary study of a feasible laboratory measure for routine clinical practice.

OBJECTIVE: Low serum phosphate level is considered one of the metabolic adaptations to the respiratory alkalosis induced by hyperventilation associated with panic disorder. The aim of this study was to assess phosphatemia as a possible state marker for panic disorder. METHODS: Sixteen panic disorder patients underwent clinical assessment with a semi-structured interview, a set of rating scales and the self-rated State and Trait Anxiety Inventory (STAI), as well as extraction of venous blood samples at baseline and after 12 weeks of pharmacological treatment. Ten healthy volunteers of similar sex, age and educational level filled out the STAI and gave blood samples at baseline and 12 weeks later. RESULTS: The median (25th-75th percentiles) of phosphate levels (mg/dl) was 2.68 (2.22-3.18) among patients and 4.13 (3.74-4.70) among healthy volunteers respectively (P < 0.001). Seven (44%) patients and no healthy volunteers presented low serum phosphate (<2.50 mg/dl) at baseline; this patient abnormality was corrected in all cases after successful treatment. At baseline, the age-adjusted correlation between phosphate levels and state-anxiety was -0.66 (P < 0.001) among all 26 participants and -0.51 (P = 0.05) among the 16 panic disorder patients. CONCLUSIONS: Measurement of phosphate levels could be easily introduced into clinical practice as a possible marker for chronic hyperventilation in panic disorder, although further investigations with larger sample sizes are necessary to characterize panic disorder patients with low versus normal phosphate levels.

Brain-derived neurotrophic factor in generalized anxiety disorder: results from a duloxetine clinical trial.

BACKGROUND: Brain-derived neurotrophic factor (BDNF) has been implicated in the pathophysiology of depression and anxiety, but has not been examined systematically in generalized anxiety disorder (GAD). The objective of this study was to examine the relationship between baseline BDNF level and treatment response in patients with GAD. METHODS: Patients (N=168) were from China, met criteria for DSM-IV GAD, had a Hospital Anxiety and Depression Rating Anxiety (HADS-A) subscale score ≥10, and a Sheehan Disability Scale (SDS) global functioning total score ≥12 at baseline. Study design was double-blind therapy for 15 weeks with duloxetine 60-120 mg or placebo. Efficacy measures included the HADS-A and Hamilton Anxiety Rating Scale (HAMA) total score. Change from baseline to endpoint for BDNF by treatment group was analyzed using ANCOVA models with baseline BDNF level as a covariate. RESULTS: No significant association was found between baseline plasma BDNF levels and anxiety illness severity. Patients who received duloxetine (n=88) had a significantly greater mean increase in plasma BDNF level (957.80 picograms/ml) compared with patients who received placebo (n=80; 469.93 pg/mL) (P=.007). Patients who met response and remission criteria (with either treatment) had greater mean increases in BDNF at endpoint from baseline (P≤.05) but when compared with nonresponders and nonremitters, respectively, the differences in mean increase were not statistically significant between groups. CONCLUSIONS: BDNF levels significantly increased with duloxetine treatment for GAD, but response and remission outcomes were not clearly related to an increase in plasma BDNF level.

Gender effect on the relationship between stress hormones and panic-agoraphobic spectrum dimensions in healthy subjects.

INTRODUCTION: Alterations of the hypothalamic-pituitary-adrenal (HPA) axis and of its peripheral indices have been reported in both normal and pathological anxiety with controversial findings. The aim of the present study was to investigate the possible correlations between serum cortisol and dehydroepiandrosterone-sulfate (DHEA-S) levels and DHEA-S/cortisol ratio, and panic-agoraphobic spectrum dimensions in a sample of healthy subjects. METHODS: Forty-two healthy subjects of both sexes, with no current or lifetime psychiatric disorders, were assessed by means of the Structured Clinical Interview for DSM-IV (SCID-I/P) and the so-called Panic Agoraphobic Spectrum-Self Report lifetime version (PAS-SR). RESULTS: Significant, negative correlations were found between cortisol levels and the total score of the separation sensitivity, panic-like symptoms, and medication/substance sensitivity PAS-SR domains. The PAS-SR total and the panic-like symptoms domain scores were positively related to the DHEAS/cortisol ratio. When the sample was divided in women and men, these correlations were present in women only. DISCUSSION: These findings, while indicating the presence of significant relationships between panic-agoraphobic traits and some indices of HPA axis functioning in healthy women, would suggest this as one of the factors explaining the greater vulnerability of women to cross the line between normal and pathological anxiety. CONCLUSIONS: Further studies are needed to explore gender differences in the relationships between HPA axis alterations and the panic-agoraphobic spectrum dimensions.

Decreased IFN-γ and IL-12 levels in panic disorder.

OBJECTIVE: The aim of this study is to assess the measures of proinflammatory cytokines in patients with panic disorder in comparison with the healthy subjects. METHODS: Twenty three patients with panic disorder with or without agoraphobia and twenty three controls were recruited for the study. Plasma samples of all subjects were analyzed for TNF-α, IFN-γ, IL-1ß, IL-2, IL-6, and IL-12 concentrations and NK-cell activity is measured in the peripheral blood samples of the subjects. RESULTS: We found significant differences on the mean values of IL-12 (p=0.01) and IFN-γ (p=0.02) between the panic disorder and control groups. In a logistic regression analysis, IFN-γ values were significant statistical predictors of the presence of panic disorder (B=-0.07, SE=0.03, p=0.04). CONCLUSION: The most important implication of our results is to suggest a relation between panic disorder and low levels of IFN-γ, compatible with the results of the animal studies showing that IFN-γ plays a role by acting to regulate the development of anxiety-like behaviors.

Stress hormones during flooding therapy and their relationship to therapy outcome in patients with panic disorder and agoraphobia.

In spite of excessive fear during a panic attack, studies have found no or little evidence for an activation of cortisol during natural panic attacks. Whether this phenomenon is related to psychopathology or outcome of psychotherapy is unknown. In this study, 10 patients with panic disorder and agoraphobia were treated with cognitive behavioural therapy including 3 in-vivo exposures (flooding) to individual phobic situations. Before, during and after exposure, the level of subjective fear was assessed and blood was collected simultaneously. Cortisol and ACTH were analysed from plasma. Ten matched healthy control subjects went through the same procedure. Fear and stress hormones during exposure were compared in patients and controls as well as related to therapy outcome at the end of therapy and 2 follow-ups in patients. Results showed that the concentrations of cortisol and ACTH did not significantly increase during exposure. Patients' cortisol concentrations were higher than those of controls at baseline and during exposure, while ACTH concentrations were comparable before and during exposure, and even lower than those of controls at recovery. Cortisol concentrations were moderately but consistently correlated to therapy outcome, i.e. patients with least cortisol release during exposure profited least from therapy. The study showed that a lack of stimulation of the HPA system at repeated confrontation with the phobic situation was related to therapeutic outcome. Mechanisms of action via the influence of cortisol on extinction learning or the inhibition of central excitatory neurotransmission are conceivable.

Respiratory and cognitive mediators of treatment change in panic disorder: evidence for intervention specificity.

OBJECTIVE: There are numerous theories of panic disorder, each proposing a unique pathway of change leading to treatment success. However, little is known about whether improvements in proposed mediators are indeed associated with treatment outcomes and whether these mediators are specific to particular treatment modalities. Our purpose in this study was to analyze pathways of change in theoretically distinct interventions using longitudinal, moderated mediation analyses. METHOD: Forty-one patients with panic disorder and agoraphobia were randomly assigned to receive 4 weeks of training aimed at altering either respiration (capnometry-assisted respiratory training) or panic-related cognitions (cognitive training). Changes in respiration (PCO2, respiration rate), symptom appraisal, and a modality-nonspecific mediator (perceived control) were considered as possible mediators. RESULTS: The reductions in panic symptom severity and panic-related cognitions and the improvements in perceived control were significant and comparable in both treatment groups. Capnometry-assisted respiratory training, but not cognitive training, led to corrections from initially hypocapnic to normocapnic levels. Moderated mediation and temporal analyses suggested that in capnometry-assisted respiratory training, PCO2 unidirectionally mediated and preceded changes in symptom appraisal and perceived control and was unidirectionally associated with changes in panic symptom severity. In cognitive training, reductions in symptom appraisal were bidirectionally associated with perceived control and panic symptom severity. In addition, perceived control was bidirectionally related to panic symptom severity in both treatment conditions. CONCLUSION: The findings suggest that reductions in panic symptom severity can be achieved through different pathways, consistent with the underlying models.

Relative hyperhomocysteinemia in patients with panic disorder: a case-control study.

BACKGROUND: The aim of this work was to examine a possible association between a clinically relevant panic disorder and plasma total homocysteine concentration. METHODS: 23 patients with panic disorder with or without agoraphobia confirmed by a standardized clinical interview (Structural Clinical Interview for DSM-IV-German version) and 23 healthy controls matched for gender and age completed questionnaires (SCL-K9, STAI, ADS, STAXI) and had blood drawn after a 15 min rest. Plasma total homocysteine concentrations were measured by competitive enzyme immunoassay. Interfering variables such as age, gender, smoking status, comorbid depression and medication were controlled for. RESULTS: Patients with panic disorder had higher plasma homocysteine concentrations in comparison to the control group (mean value 11.00 vs. 9.14 mumol/l, p = 0.04 with age, gender, smoking status, comorbid depression and antidepressant medication being controlled for). Furthermore, homocysteine plasma concentration was positively correlated with Global Severity of Symptoms (SCL-K9, r(Pearson) = 0.41, p < 0.01). CONCLUSION: The findings of this study suggest a link between elevated plasma homocysteine levels and panic disorder. This raises a new hypothesis of another pathway to an increased risk of cardiovascular events in anxious individuals.

Elevated alpha-amylase but not cortisol in generalized social anxiety disorder.

Stress-system dysregulation is thought to increase the risk for anxiety disorders. Here we describe both hypothalamic pituitary adrenal (HPA) axis and autonomic nervous system (ANS) activity in basal non-challenging conditions and after 0.5mg dexamethasone in generalized social anxiety disorder (gSAD) patients. To ensure stress-free sampling we collected saliva and determined cortisol and alpha-amylase (sAA), the latter a relative new marker of autonomic activity. Forty-three untreated gSAD patients without comorbidity were compared with 43 age and gender matched controls in non-stressed conditions on sAA and cortisol after awakening, during the day (including late evening), and after a low dose (0.5mg) of dexamethasone. Cortisol and sAA were analyzed with mixed models. Additional analyses were done with paired t-tests. Apart from the assessments in the morning, gSAD patients had significantly higher diurnal and post-dexamethasone 1600h sAA levels. No differences between gSAD and controls in any cortisol measurements were found. In conclusion, in gSAD in basal, non-stimulated conditions and after dexamethasone, we found hyperactivity of the ANS, as measured with sAA, but not of the HPA-axis. This suggests a relative increased activity of the ANS as compared to the HPA-axis, in line with the observed hyperarousal in gSAD.

Association between anxiety and factors of coagulation and fibrinolysis.

BACKGROUND: Psychological stress and anxiety have been shown to produce an activation of coagulation and fibrinolysis. Resulting hypercoagulability is a risk factor for cardiovascular diseases, and could therefore contribute to an increased prevalence of coronary artery disease in anxiety patients. However, hemostasis function has not yet been studied in patients with clinically relevant anxiety disorders. METHODS: A group of anxiety patients (panic disorder with agoraphobia or social phobia) and a healthy control group (each n = 29) completed some questionnaires [SCL-K9 (a short form of the SCL-90-R), State Trait Anxiety Inventory, ADS (general depression scale)], and had blood drawn after a 15-min rest period. To assess the reaction of the hemostatic system by global entities, sum scores were computed from parameters of coagulation and fibrinolysis (fibrinogen, FVII, FVIII, vWF, F1 + 2, TAT, D-dimer, alpha(2)-AP, PAP, tPA, PAI-1). Interfering variables, such as age, gender, alcohol consumption and smoking status, were controlled. RESULTS: Anxiety patients scored higher in a composite hemostatic score and a sum score of fibrinolysis in comparison to the control group, with a predominant activation of inhibitors in fibrinolysis. However, the psychological variable with the closest association to hemostasis was not trait anxiety, but self-perceived worry about blood drawing before blood sampling was performed. CONCLUSIONS: The coagulation and fibrinolysis system is activated in the direction of a hypercoagulable state in patients with severe phobic anxiety, triggered by fear of blood drawing. This could be one mediating factor for the increased risk of cardiovascular diseases in this population. Acute situational phobic anxiety should be monitored closely when studying the association between anxiety and hemostasis.

The speed of onset of action of alprazolam-XR compared to alprazolam-CT in panic disorder.

This study compares the speed of onset of action of the extended release (XR) formulation of alprazolam with that of the compressed tablet (CT) formulation in a sample of outpatients with DSM-IV panic disorder. Diary records of hourly antianxiety benefit from a 9-week open label switch study of 30-patients stabilized on alprazolam- CT for 3 weeks and then switched to an equivalent dose of alprazolam-XR, were used to examine the timing and magnitude of clinical benefit on both formulations. The magnitude of benefit at the first hour after the first morning dose was similar for both formulations. The peak benefit, over the hours after the first morning dose, was also similar and 90% of peak benefit that was achieved in the first hour on both formulations. Mean time to peak benefit was similar (1.5 h for alprazolam-CT vs. 1.6 h for alprazolam-XR) and the percent of patients achieving peak benefit in the first hour was also similar. Compared to the CT formulation, alprazolam-XR had a much longer duration of therapeutic action (11.3 +/- 4.2 h vs. 5.1 +/- 1.7 h). The results, which may be related to the biotechnology (and resultant pharmacokinetic profile) of the XR preparation, suggest that alprazolam-XR has value as a "rescue" as well as a prophylactic or maintenance treatment in panic disorder. These results must be viewed in the context of the study limitations including its small size, the lack of independence of groups in a switch study, and the limitations of the diary records used.

Hyponatremia-induced seizure during carbamazepine treatment.

We report the case of a 54-year-old woman who was admitted for benzodiazepine withdrawal. After 6 weeks of carbamazepine treatment (600, then 200 mg) the patient suddenly suffered from a grand mal seizure. Laboratory findings revealed a clinical significant hyponatremia of Na 125 mmol/l (baseline: 143 mmol/l). CCT and ECG were normal. To our knowledge, this is the first description of a seizure related to hyponatremia in an adult carbamazepine-treated patient.

Nerve growth factor serum concentrations rise after successful cognitive-behavioural therapy of generalized anxiety disorder.

Generalized anxiety disorder (GAD) is a chronic stress disease with permanent physical tension and cognitive strain. Raised nerve growth factor (NGF) serum levels were reported as an acute stress reaction in soldiers before their first parachute jump even before the rise in cortisol. Taking GAD as a clinical model of chronic stress, we measured NGF in the serum of 22 patients with GAD before and after cognitive-behavioural therapy (CBT) and compared them to those of healthy normal controls. Treatment response was tested by the values of the State and Trait of Anxiety Inventory (STAI) and the Hamilton Anxiety Scale (HAM-A) as treatment outcome variables. The NGF values of patients and controls were similar at baseline (p=0.8941); however, with successful treatment, corresponding to a mean reduction in the HAM-A by more than 50% and a reduction in the clinical global impression scale (CGI) median from 4 to 1, the patients' NGF serum concentrations rose significantly (p=0.0006) which might correspond to an altered stress reaction, possibly contributing to good therapeutic response with CBT. There were 3 patients with a HAM-A decrease of less than 15%. In those patients NGF rose only marginally. Hence, the increase in serum NGF seems to indicate good treatment response.

Relationship between anxiety and thyroid function in patients with panic disorder.

The aim of this study was to investigate correlations between thyroid function and severity of anxiety or panic attacks in patients with panic disorder. The authors examined 66 out-patients with panic disorder (medicated, n=41; non-medicated, n=25), and measured their free thriiodothyronine (T3), free thyroxine (T4) and thyroid-stimulating hormone (TSH) levels. Significant correlations between the thyroid hormone levels and clinical features were observed in the non-medicated patients. The more severe current panic attacks were, the higher the TSH levels were. In addition, severity of anxiety correlated negatively with free T4 levels. In this study, we discuss relationship between thyroid function and the clinical severity or features of panic disorder.

Serum cholesterol level comparison: control subjects, anxiety disorder patients, and obsessive-compulsive disorder patients.

OBJECTIVE: To determine whether panic disorder is associated with elevated serum cholesterol levels. Serum cholesterol levels of panic disorder patients are reported to be elevated. This could explain the higher-than-expected cardiovascular mortality in this population. Some evidence exists wherein cholesterol levels are also increased in patients with general anxiety disorder and phobias. To date, there are only 2 reports on cholesterol levels of obsessive-compulsive disorder (OCD) patients, giving controversial results. METHOD: We compared serum cholesterol levels of anxiety disorder patients, OCD patients, and normal control subjects with each other (n = 60 in each group). Serum cholesterol was measured in each subject before treatment. Subjects of the 3 groups were matched by age and sex. RESULTS: Patients with anxiety disorders and OCD had elevated cholesterol levels, compared with normal control subjects. Cholesterol levels in OCD patients were comparable with those in patients with phobia. CONCLUSIONS: Our data support the assumption that elevation in cholesterol level is not a specific feature of panic disorder (as most assumed), but more generally associated with anxiety disorders. Increased cholesterol levels in patients with anxiety disorders and OCD may be of clinical relevance.

Memory performance in panic disorder patients after chronic use of clomipramine.

This study investigated the chronic use (6.3 +/- 0.5 years; mean +/- SEM) of therapeutic doses of clomipramine (57.0 +/- 8.0 mg/day) by outpatients with panic disorder/agoraphobia who were currently in remission to assess impairment of memory and psychomotor functions. In addition, the association between test performance and serum levels of clomipramine (CMI) and its active metabolite desmethylclomipramine (DCMI) was also assessed. Patients and healthy volunteers matched for sex, age and educational level were submitted to rating scales and to memory and psychomotor tests. There was no significant difference between groups regarding any variable, except for metamemory. Significant associations were found between (i) longer-term clomipramine treatment and poorer performance in the implicit test and (ii) higher serum levels of clomipramine or desmethylclomipramine, or both (CMI + DCMI) and lower performance in central executive tests and metamemory. The results showed that low doses of CMI chronically administered to panic patients are associated with diminished metamemory and impaired priming and working memory. Further investigations are needed to confirm these results and to determine whether the chronic use of higher therapeutic doses of tricyclic antidepressants is associated with more intense deleterious effects on memory and psychomotor functions.

GABA(A) receptor-modulating neuroactive steroid composition in patients with panic disorder before and during paroxetine treatment.

OBJECTIVE: Previous studies have shown that neuroactive steroids modulate anxiety and stress reactivity. However, no data on the possible role of these gamma-aminobutyric acid(A) (GABA(A)) receptor-modulating neuroactive steroids in patients with anxiety disorders are available. METHOD: The concentrations of 3alpha,5alpha-tetrahydroprogesterone (3alpha,5alpha-THP), 3alpha,5beta-THP, 3beta,5alpha-THP, and their precursors were studied in the plasma of 10 patients with panic disorder and 10 matched healthy comparison subjects. In addition, the effects of paroxetine treatment on neuroactive steroid concentrations were studied in the panic disorder patients over a 24-week period. RESULTS: Unexpectedly, patients with panic disorder had significantly greater concentrations of the positive allosteric modulators 3alpha,5alpha-THP and 3alpha,5beta-THP and significantly lower concentrations of 3beta,5alpha-THP (a functional antagonist for GABA(A) agonistic steroids), which might result in greater GABA(A) receptor-mediated neuronal activity. Paroxetine treatment did not affect neuroactive steroid concentrations, which were highly stable over 24 weeks. CONCLUSIONS: Differences in neuroactive steroid composition in patients with panic disorder were the opposite of those seen in patients with major depression and may reflect counterregulative mechanisms against the occurrence of spontaneous panic attacks.