A systematic review and meta-analysis of urinary biomarkers in myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS)

BACKGROUND

Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS) is a multifactorial illness that affects many body systems including the immune, nervous, endocrine, cardiovascular, and urinary systems. There is currently no universal diagnostic marker or targeted treatment for ME/CFS. Urine is a non-invasive sample that provides biomarkers that may have the potential to be used in a diagnostic capacity for ME/CFS. While there are several studies investigating urine-based biomarkers for ME/CFS, there are no published systematic reviews to summarise existing evidence of these markers. The aim of this systematic review was to compile and appraise literature on urinary-based biomarkers in ME/CFS patients compared with healthy controls.

METHODS

Three databases: Embase, PubMed, and Scopus were searched for articles pertaining to urinary biomarkers for ME/CFS compared with healthy controls published between December 1994 to December 2022. The final articles included in this review were determined through application of specific inclusion and exclusion criteria. Quality and bias was assessed using the Joanna Briggs Institute Critical Appraisal Checklist for Case Control Studies. A meta-analysis according to Cochrane guidelines was conducted on select studies, in particular, those that investigate urinary free cortisol levels in ME/CFS patients compared to healthy controls using the program STATA 17.

RESULTS

Twenty-one studies were included in this review. All of the studies investigated urinary-based markers in ME/CFS patients compared with healthy controls. The reported changes in urinary outputs include urinary free cortisol (38.10%), carnitine (28.6%), iodine (4.76%), and the metabolome (42.86%). In most cases, there was minimal overlap in the main outcomes measured across the studies, however, differences in urinary free cortisol between ME/CFS patients and healthy controls were commonly reported. Seven studies investigating urinary free cortisol were included in the meta-analysis. While there were significant differences found in urinary free cortisol levels in ME/CFS patients, there was also substantial heterogeneity across the included studies that makes drawing conclusions difficult.

CONCLUSIONS

There is limited evidence suggesting a consistent and specific potential urinary-based biomarker for ME/CFS. Further investigations using more standardised methodologies and more stringent case criteria may be able to identify pathophysiological differences with diagnostic potential in ME/CFS patients compared with healthy controls.

A laboratory approach for characterizing chronic fatigue: what does metabolomics tell us?

INTRODUCTION

Manifestations of fatigue range from chronic fatigue up to a severe syndrome and myalgic encephalomyelitis. Fatigue grossly affects the functional status and quality of life of affected individuals, prompting the World Health Organization to recognize it as a chronic non-communicable condition.

OBJECTIVES

Here, we explore the potential of urinary metabolite information to complement clinical criteria of fatigue, providing an avenue towards an objective measure of fatigue in patients presenting with the full spectrum of fatigue levels.

METHODS

The experimental group consisted of 578 chronic fatigue female patients. The measurement design was composed of (1) existing clinical fatigue scales, (2) a hepatic detoxification challenge test, and (3) untargeted proton nuclear magnetic resonance (¹H-NMR) procedure to generate metabolomics data. Data analysed via an in-house Matlab script that combines functions from a Statistics and a PLS Toolbox.

RESULTS

Multivariate analysis of the original 459 profiled ¹H-NMR bins for the low (control) and high (patient) fatigue groups indicated complete separation following the detoxification experimental challenge. Important bins identified from the ¹H-NMR spectra provided quantitative metabolite information on the detoxification challenge for the fatigue groups.

CONCLUSIONS

Untargeted ¹H-NMR metabolomics proved its applicability as a global profiling tool to reveal the impact of toxicological interventions in chronic fatigue patients. No clear potential biomarker emerged from this study, but the quantitative profile of the phase II biotransformation products provide a practical visible effect directing to up-regulation of crucial phase II enzyme systems in the high fatigue group in response to a high xenobiotic-load.

The assessment of the energy metabolism in patients with chronic fatigue syndrome by serum fluorescence emission

CONTEXT

Chronic fatigue syndrome (CFS) is a debilitating fatigue illness that has unknown etiology and lacks an objective diagnostic marker.

OBJECTIVE

To examine the metabolic component of CFS to determine if practitioners can use serum NAD(P)H concentration measurements to monitor metabolism and fatigue status in patients with CFS.

DESIGN

The research team conducted a case-control study, comparing a group of patients who were diagnosed with CFS with a control group of healthy subjects. The team obtained venous blood samples from fasting patients to examine the serum NAD(P)H concentrations.

SETTING

The study occurred at the Riordan Clinic in Witchita, Kansas.

PARTICIPANTS

The study included 44 CFS patients at the Riordan Clinic and 30 healthy control participants. The CFS patients presented a spectrum of symptoms that had existed for at least 6 months: new, unexplained, persistent, or relapsing chronic fatigue that bed rest did not resolve and that was severe enough to reduce daily activity significantly by 50% in conjunction with headache, muscle pain, pain in multiple joints, and unrefreshing sleep. In the control group, the research team enrolled subjects without diagnosis of disease or injury.

OUTCOME MEASURES

The research team determined levels of serum reduced nicotinamide adenine dinucleotides (NADH and NAD[P]H) by measuring serum fluorescence emission at 450 nm. The team then conducted sensitivity and specificity analyses. Results NAD(P)H concentrations in serum of CFS participants averaged 8.0 ± 1.4 (standard deviation [SD]) nmol/mL, while those in the healthy controls averaged 10.8 ± 0.8 (SD) nmol/mL, a statistically significant difference. Using a cut-off concentration of 9.5 nmol/mL, the research team attained a sensitivity of 0.73 and a specificity of 1.0. An analysis of receiver-operator characteristics yielded an area under the curve of 0.9. The research team compared serum NAD(P)H to several endocrine and metabolic lab parameters. Serum NAD(P)H was directly correlated with serum CoQ10 levels and inversely correlated with urine hydroxyhemopyrrolin-2-one levels.

CONCLUSIONS

Based on these findings, the research team proposed using serum NAD(P)H, measured as an intrinsic serum-fluorescence emission, to monitor metabolism and fatigue status in patients with CFS. Following patients NAD(P)H levels over time may aid in selecting therapeutic strategies and monitoring treatment outcomes.

Does hypocortisolism predict a poor response to cognitive behavioural therapy in chronic fatigue syndrome?

BACKGROUND

There is evidence that patients with chronic fatigue syndrome (CFS) have mild hypocortisolism. The clinical significance of this is unclear. We aimed to determine whether hypocortisolism exerted any effect on the response of CFS to cognitive behavioural therapy (CBT).

METHOD

We measured 24-h urinary free cortisol (UFC) in 84 patients with Centers for Disease Control and Prevention (CDC)-defined CFS (of whom 64 were free from psychotropic medication) who then received CBT in a specialist, tertiary out-patient clinic as part of their usual clinical care. We also measured salivary cortisol output from 0800 to 2000 h in a subsample of 56 psychotropic medication-free patients.

RESULTS

Overall, 39% of patients responded to CBT after 6 months of treatment. Lower 24-h UFC output was associated with a poorer response to CBT but only in psychotropic medication-free patients. A flattened diurnal profile of salivary cortisol was also associated with a poor response to CBT.

CONCLUSIONS

Low cortisol is of clinical relevance in CFS, as it is associated with a poorer response to CBT. Hypocortisolism could be one of several maintaining factors that interact in the persistence of CFS.

Hematologic and urinary excretion anomalies in patients with chronic fatigue syndrome

Patients with chronic fatigue syndrome (CFS) have a broad and variable spectrum of signs and symptoms with variable onsets. This report outlines the results of a single-blind, cross-sectional research project that extensively investigated a large cohort of 100 CFS patients and 82 non fatigued control subjects with the aim of performing a case-control evaluation of alterations in standard blood parameters and urinary amino and organic acid excretion profiles. Blood biochemistry and full blood counts were unremarkable and fell within normal laboratory ranges. However, the case-control comparison of the blood cell data revealed that CFS patients had a significant decrease in red cell distribution width and increases in mean platelet volume, neutrophil counts, and the neutrophil-lymphocyte ratio. Evaluation of the urine excretion parameters also revealed a number of anomalies. The overnight urine output and rate of amino acid excretion were both reduced in the CFS group (P < 0.01). Significant decreases in the urinary excretion of asparagine (P < 0.0001), phenylalanine (P < 0.003), the branch chain amino acids (P < 0.005), and succinic acid (P < 0.0001), as well as increases in 3-methylhistidine (P < 0.05) and tyrosine (P < 0.05) were observed. It was concluded that the urinary excretion and blood parameters data supported the hypothesis that alterations in physiologic homeostasis exist in CFS patients.

β-Alanine and γ-aminobutyric acid in chronic fatigue syndrome

BACKGROUND

Due to the occurrence of sleep disturbances and fatigue in chronic fatigue syndrome (CFS), an investigation was performed to examine if there is an abnormal excretion of gamma-aminobutyric acid (GABA) and/or its structural analogue beta-alanine in the urine from CFS patients. Both GABA and beta-alanine are inhibitory neurotransmitters in the mammalian central nervous system.

METHODS

The 24 h urine excretion of GABA and beta-alanine was determined by isotope dilution gas chromatography mass spectrometry in 33 CFS patients and 43 healthy controls. The degree of symptoms in both patients and controls was measured by grading of three typical CFS symptoms using a Visual Analogue Scale.

RESULTS

Men had a significantly higher excretion of both beta-alanine and GABA than women. Comparing CFS patients with healthy controls showed no significant difference in excretion of neither beta-alanine nor GABA. No correlation was found between the excretion of beta-alanine or GABA and any of the three characteristic CFS symptoms measured. However, two female and two male CFS patients excreted considerably higher amounts of beta-alanine in their 24 h urine samples than control subjects.

CONCLUSIONS

Increased excretion of beta-alanine was found in a subgroup of CFS patients, indicating that there may be a link between CFS and beta-alanine in some CFS patients.

Urinary cortisol and cortisol metabolite excretion in chronic fatigue syndrome

OBJECTIVES

Reduced basal hypothalamic-pituitary-adrenal (HPA) axis output in chronic fatigue syndrome (CFS) has been inferred from low cortisol levels in blood, saliva, and urine in some studies. Because > 95% of cortisol is metabolized before excretion, we assessed cortisol output by assay of both cortisol metabolites and free cortisol in 24-hour urine collections and also investigated sex differences in these between CFS and control groups.

METHOD

We calculated total urinary cortisol metabolites (TCM) and cortisol metabolite ratios from individual steroid data in 40 patients (20 males and 20 females) with CFS who were free of medication or comorbid psychiatric disorder likely to influence the HPA axis. Results were compared with those of 40 healthy volunteers (20 males and 20 females) well matched for age and body mass index. Data for free cortisol was obtained on 28 of the patients and 27 of the controls.

RESULTS

The mean of TCM and cortisol metabolite ratios was not significantly different between patients and controls for either sex (p > .05 for all parameters). Previously established sex differences were confirmed in our controls and were found to be similar in CFS for TCM and the ratios 11OH/11OXO, 5alpha/5beta THF, and 20OH/20OXO (see text) (p < .005, p < .05, p < .05, and p < .005, respectively). Urinary free cortisol values were numerically (but not statistically) lower in patients with CFS than controls, and correlated inversely with fatigue levels in patients.

CONCLUSION

The finding of normal urinary cortisol metabolite excretion in patients with CFS is at variance with earlier reports that CFS is a hypocortisolemic state. If serum and saliva cortisol levels are lower in CFS, this would suggest that metabolic clearance of cortisol is faster in patients with CFS than controls. This study also demonstrates that sex differences must be taken into account when interpreting results in patients with CFS.

CFSUM1 and CFSUM2 in urine from patients with chronic fatigue syndrome are methodological artefacts

McGregor et al. reported increased levels of an unidentified urinary compound (CFSUM1) in patients with chronic fatigue syndrome (CFS), with reduced excretion of another unidentified compound (CFSUM2), and suggested the possibility of chemical or metabolic 'markers' for CFS. The identity of CFSUM1 as reported was erroneous and the identities of these compounds have remained unknown until now. Urine samples were obtained from 30 patients with ME/CFS, 30 age- and sex-matched healthy controls, 20 control patients with depression and 22 control patients with rheumatoid arthritis. Samples were prepared using the published methods of McGregor et al. to produce heptafluorobutyryl-isobutyl derivatives of urinary metabolites. Alternative preparations utilised isopropyl, n-butyl and trifluoroacetyl derivatives. These were separated and identified using gas chromatography-mass spectrometry. CFSUM2 was identified as being partially derivatised [isobutyl ester-mono-heptafluorobutyryl (HFB)] serine. CFSUM1 was identified as partially derivatised pyroglutamic acid, being the isobutyl ester without formation of a HFB derivative. Both CFSUM1 and CFSUM2 are artefacts of the sample preparation procedure and previously reported quantitative abnormalities of CFSUM1 and CFSUM2 in urine from patients with ME/CFS are also artefactual. Pyroglutamic acid may be of primarily dietary origin. The methods used cannot provide reliable qualitative or quantitative data on urinary metabolites. No clinical or biochemical significance can be drawn between these compounds in ME/CFS or any other clinical conditions.

Diurnal excretion of urinary cortisol, cortisone, and cortisol metabolites in chronic fatigue syndrome

OBJECTIVE

The aim of this study was to obtain comprehensive information on basal hypothalamic-pituitary-adrenal (HPA) axis activity in chronic fatigue syndrome (CFS) patients who were not affected by medication or comorbid psychiatric disorder likely to influence the HPA axis.

METHOD

Steroid analysis of urine collections from 0600 to 2100 h at 3-h intervals in CFS patients and in controls.

RESULTS

Urinary free cortisol and cortisone concentrations showed a significant normal diurnal rhythm, but levels were lower across the cycle in CFS. In contrast, while urinary cortisol metabolites also showed a normal diurnal rhythm, levels were not significantly different between the CFS and controls at any time. Derived metabolite ratios were similar in both groups.

CONCLUSION

This study provides further evidence for reduced basal HPA axis function in patients with CFS, based on lower free cortisol and cortisone levels, but this is not corroborated by cortisol metabolite data. The difference between these measures cannot be explained by an altered timing of the diurnal rhythm.

Plasma and urinary carnitine and acylcarnitines in chronic fatigue syndrome

Contradictory reports have suggested that serum free carnitine and acylcarnitine concentrations are decreased in patients with chronic fatigue syndrome (CFS) and that this is a cause of the muscle fatigue observed in these patients. Others have shown normal serum free carnitine and acylcarnitines in similar patients. We report here studies on free, total and esterified (acyl) carnitines in urine and blood plasma from UK patients with CFS and three control groups. Plasma and timed urine samples were obtained from 31 patients with CFS, 31 healthy controls, 15 patients with depression and 22 patients with rheumatoid arthritis. Samples were analysed using an established radioenzymatic procedure for total, free and esterified (acyl) carnitine. There were no significant differences in plasma or urinary total, free or esterified (acyl) carnitine between UK patients with CFS and the control groups or in renal excretion rates of these compounds. The data presented here show that, in the CFS patients studied, there are no significant abnormalities of free or esterified (acyl) carnitine. It is thus unlikely that abnormalities in carnitine homeostasis have any significant role in the aetiology of their chronic fatigue.

Urinary and plasma organic acids and amino acids in chronic fatigue syndrome

Previous work by others have suggested the occurrence of one or more chemical or metabolic 'markers' for ME/CFS including specific amino acids and organic acids and a number of unidentified compounds (CFSUM1, CFSUM2). We have shown elsewhere that CFSUM1 is partially derivatised pyroglutamic acid and CFSUM2 partially derivatised serine and have suggested and demonstrated that the analytical methods used were unsuitable to identify or to accurately quantify urinary metabolites. We have now made a detailed analysis of plasma and urinary amino acids and of urinary organic acids from patients with ME/CFS and from three control groups. Fasting blood plasma and timed urine samples were obtained from 31 patients with CFS, 31 age and sex-matched healthy controls, 15 patients with depression and 22 patients with rheumatoid arthritis. Plasma and urinary amino acids and urinary organic acids were determined using established and validated methods and data compared by statistical analysis. None of the previously reported abnormalities in urinary amino acids or of organic acids could be confirmed. Results however provide some evidence in patients with ME/CFS for underlying inflammatory disease and for reduced intramuscular collagen with a lowered threshold for muscle micro-injury. These factors in combination may provide a basis for the fatigue and muscle pain that are the major symptoms in these patients.

Urinary electrophoretic profiles from chronic fatigue syndrome and chronic fatigue syndrome/fibromyalgia patients: a pilot study for achieving their normalization

Aim of our study was to determine if there were distinct, disease-related patterns of urinary analytes in chronic fatigue syndrome (CFS) and chronic fatigue syndrome/fibromyalgia (CFS/FM) compared to normal controls (NC). Urine was collected from these subjects for two consecutive 24 h periods and aliquots were submitted to micellar electrokinetic chromatography (MEKC). To compensate for the differences in peak migration times, these were normalized from the 35 min duration of run to a 100-point scale, and each peak was assigned its normalized time measure. Peak heights were also normalized by dividing the mAU by that of the internal standard (creatinine) and multiplying by 100. MEKC with normalization for peak height and migration time generated comparable results within each of the patient groups. CFS/FM and CFS had significant differences in peaks compared to NC that may be of significance as biomarkers of illnesses.

[Role of pathological delayed-type hypersensitivity in chronic fatigue syndrome: importance of the evaluation of lymphocyte activation by flow cytometry and the measurement of urinary neopterin]

Chronic fatigue syndrome or benign myalgic encephalomyelitis has been extensively described and investigated. Although numerous immunological abnormalities have been linked with the syndrome, none have been found to be specific. This article describes the detection of delayed-type hypersensitive responses to certain common environmental antigens in almost fifty per cent of patients with this syndrome. Such hypersensitivity can be detected by the intradermal administration of antigens derived from commensal organisms like the yeast Candida albicans, and then monitoring for a systemic reaction over the following six to forty eight hours. This approach can be consolidated by performing lymphocyte activation tests in parallel and measuring in vitro T-cell activation by Candida albicans antigens by three-colour flow cytometry based on CD3, CD4 and either CD69 or CD25. Another useful parameter is the kinetics of neopterin excretion in the urine over the course of the skin test. The results showed that the intensity of the DTH response correlated with the number of T-cells activated in vitro. Various factors have been implicated in the fatigue of many patients, notably lack of sleep. However, it remains difficult to establish causality in either one direction or the other. This work is in the spirit of a multifactorial approach to the group of conditions referred to as "chronic fatigue syndrome".

Urinary free cortisol in chronic fatigue syndrome

OBJECTIVE

The authors measured 24-hour urinary free cortisol in a group of well-characterized patients with chronic fatigue syndrome.

METHOD

They obtained 24-hour urine collections from 121 consecutive clinic patients with chronic fatigue syndrome and 64 comparison subjects without the syndrome.

RESULTS

Urinary free cortisol was significantly lower in the subjects with chronic fatigue syndrome regardless of the presence or absence of current or past comorbid psychiatric illness. Lower levels of urinary free cortisol were not related to medication use, sleep disturbance, or disability levels.

CONCLUSIONS

There is mild hypocortisolism in chronic fatigue syndrome. Whether a primary feature or secondary to other factors, hypocortisolism may be one factor contributing to the symptoms of chronic fatigue syndrome.

Delayed-type hypersensitivity and chronic fatigue syndrome: the usefulness of assessing T-cell activation by flow cytometry--preliminary study

Chronic fatigue syndrome or benign myalgic encephalomyelitis has been extensively described and investigated. Although numerous immunological abnormalities have been linked with the syndrome, none have been found to be specific. This article describes the detection of delayed-type hypersensitive responses to certain common environmental antigens in almost fifty per cent of patients with this syndrome. Such hypersensitivity can be detected by the intradermal administration of antigens derived from commensal organisms like the yeast Candida albicans albicans, and then monitoring for a systemic reaction over the following six to forty-eight hours. This approach can be consolidated by performing lymphocyte activation tests in parallel and measuring in vitro T-cell activation by Candida albicans albicans antigens by three-colour flow cytometry based on CD3, CD4 and either CD69 or CD25. Another useful parameter is the kinetics of neopterin excretion in the urine over the course of the skin test. The results showed that the intensity of the DTH response correlated with the number of T-cells activated in vitro. Various factors have been implicated in the fatigue of many patients, notably lack of sleep. However, it remains difficult to establish causality in either one direction or the other. This work is in the spirit of a multifactorial approach to the group of conditions referred to as "chronic fatigue syndrome".

Urinary free cortisol excretion in chronic fatigue syndrome, major depression and in healthy volunteers

Urinary free cortisol excretion (UFC) was compared in 21 patients with chronic fatigue syndrome (CFS), in 10 melancholic depressives and in 15 healthy controls. Patients with depression had UFC values which were significantly higher than healthy comparison subjects, whereas UFC excretion of CFS patients was significantly lower than the comparison group. These findings are in keeping with currently held hypotheses of hyperactivity and hypoactivity of the hypothalamic-pituitary-adrenal (HPA) axis in depression and chronic fatigue syndrome respectively. Five of the 21 CFS patients had a co-morbid depressive illness. This sub-group retained the profile of UFC excretion of those with CFS alone, suggesting a different pathophysiological basis for depressive symptoms in CFS.

Dehydroepiandrosterone sulfate deficiency in chronic fatigue syndrome

The chronic fatigue syndrome (CFS) is a condition of unknown etiology, characterized by a persistent debilitating fatigue, the muscle-related symptoms and the neuropsychiatric symptoms. Recently, it has been reported that the patients with CFS might have impaired activation of the hypothalamic-pituitary-adrenal axis, and suggested that a part of the patho-genesis of CFS might be associated with abnormalities of the endocrine system. Herein, we show that the majority of Japanese patients with CFS had a serum dehydroepiandrosterone sulfate (DHEA-S) deficiency. Serum DHEA-S is one of the most abundantly produced hormones which is secreted from the adrenal glands, and its physiological function is thought to be a precursor of sex steroids. DHEA-S has recently been shown to have physiological properties, such as neurosteroids, which are associated with such psychophysiological phenomena as memory, stress, anxiety, sleep and depression. Therefore, the deficiency of DHEA-S might be related to the neuropsychiatric symptoms in patients with CFS.

Preliminary determination of the association between symptom expression and urinary metabolites in subjects with chronic fatigue syndrome

Chronic fatigue syndrome (CFS) patients have a urinary metabolite labeled CFSUM1 with increased incidence (P < 0.004) and relative abundance (P < 0.00003). The relative abundances of urinary CFSUM1 and beta-alanine were associated with alterations in metabolite excretion and symptom incidence. In 20 CFS patients and 45 non-CFS subjects, symptom/metabolite associations were investigated by assessing symptom sensitivity and specificity, and symptom indices of total symptom incidence, CFS core symptoms, cognitive, neurological, musculoskeletal, gastrointestinal, infection-related and genitourinary symptom indices, as well as a visual analogue pain scale of average pain intensity. Thirty-three symptoms had significant (P < 0.005) sensitivity and specificity in the CFS patients compared to that in the non-CFS controls. Severe fatigue was the only symptom with 100% sensitivity and specificity and CFSUM1 excretion was the primary metabolite for expression of this symptom. All nine symptom indices had elevated responses in the CFS patients (all P < 0.0000001). Multiple regression analyses indicated that all the symptom indices had significant correlations (R) with changes in the urinary excretion of metabolites (P < 0.0001). CFSUM1 and beta-alanine were the first and second metabolites correlated with the CFS core symptom index and CFSUM1 was primarily associated with infection-related and musculoskeletal indices whereas beta-alanine was primarily associated with gastrointestinal and genitourinary indices. The strong associations of CFSUM1 and beta-alanine with CFS symptom expression provide a molecular basis for developing an objective test for CFS.

Preliminary determination of a molecular basis of chronic fatigue syndrome

Chronic fatigue syndrome (CFS/ME) is a debilitating fatigue illness that has an unknown etiology. We studied 20 chronic fatigue syndrome (CFS) patients, who complied with the Oxford and American CDC definitions, and 45 non-CFS subjects. Participants completed questionnaires, were clinically examined, and had first morning urine specimens collected, which were screened by gas chromatography-mass spectrometry for changes in metabolite excretion. Multivariate analysis of the urinary metabolite profiles differed significantly in the CFS patients compared to the non-CFS patients (P < 0.004). The CFS patients had increases in aminohydroxy-N-methylpyrrolidine (P < 0.00003, referred to as chronic fatigue symptom urinary marker 1, or CFSUM1), tyrosine (P < 0.02), beta-alanine (P < 0.02), aconitic acid (P < 0.05), and succinic acid (P < 0.05) and reductions in an unidentified urinary metabolite, CFSUM2 (P < 0.0007), alanine (P < 0.005), and glutamic acid (P < 0.02). CFSUM1, beta-alanine, and CFSUM2 were found by discriminant function analysis to be the first, second, and third most important metabolites, respectively for discriminating between CFS and non-CFS subjects. The abundances of CFSUM1 and beta-alanine were positively correlated with symptom incidence (P < 0.01 and P < 0.001, respectively), symptom severity, core CFS symptoms, and SCL-90-R somatization (P < 0.00001), suggesting a molecular basis for CFS.