“The pathological nature of the fatigue experienced by ME/CFS sufferers is its inexplicable persistence, severity and its inability to be sufficiently relieved by rest.” Armstrong et. al.

The Naviaux metabolomics study with its findings suggesting that a hypometabolic state is present in chronic fatigue syndrome was thrilling, but it wasn’t the first or even nearly the first ME/CFS metabolomics study. The Aussies (McGregor, Gooley, Butt, and more recently Armstrong) have been plugging away at metabolism and metabolic work for years, and their 2015 study – ignored by most – was as exciting as the Naviaux paper. Ron Davis glommed onto it early and  praised it. Looked at in light of Bob Naviaux’s work, the paper, with its similar core findings and somewhat different interpretations, is exciting indeed.


From the Amstrong SMCI webinar

Christopher Armstrong, the lead author, is an example of the kind of young researcher this field needs so much. An Australian researcher working with Neil McGregor – a metabolomics pioneer in the ME/CFS field  – Armstrong represents hope for the future.

Armstrong was not at IACFS/ME conference but Dr. McGregor was. McGregor’s mostly been working in the shadows for years but he’s out of the shadows now and appears to be in considerable demand. He was not being totally ignored before; a year or so ago the Australian group’s work caught the eye of an ME/CFS researcher – Dr. Fluge – who doesn’t miss much.  Dr. Fluge’s subsequent metabolomics work was one of the highlights of the IACFS/ME Conference.

When I asked Armstrong how he got involved in chronic fatigue syndrome (ME/CFS), he noted that Dr. McGregor and Dr. Henry Butt have studied ME/CFS since the early 1990s. After Butt and Gooley published a paper on gut microbes in ME/CFS. Armstrong’s Ph.D. study examined D-lactate in blood, fecal and urine samples. He then got funded for the larger study that was published in 2015.

The Metabolic Profiling Paper

The idea that energy production is impaired in chronic fatigue syndrome is  inherently appealing. As Armstrong and McGregor noted in the 2015 paper, both fatigue and post-exertional malaise are often the result of impaired energy metabolism.

Tossing the discussion of what initiated this illness aside, they suggested a focus on the “maintaining factors” of the illness and suggested that problems with oxidative stress could be whacking the mitochondria.

Virtually every study that has looked for oxidative stress in chronic fatigue syndrome (ME/CFS) has found it. Perhaps the most exciting oxidative stress studies in ME/CFS have been done by Dikomo Shungu whose brain imaging studies have found increased lactate levels and decreased antioxidant levels. Shungu believes oxidative stress could be causing the cognitive problems in ME/CFS.

Two Metabolomics Studies – Two Approaches

Armstrong has access to both of the machines used in metabolomics; nuclear magnetic resonance (NMR) and mass spectrometry (MS).  While the Aussies have been using nuclear magnetic resonance to study the metabolome, Naviaux has been using mass spectroscopy. I asked Armstrong what the difference was between the two.

Armstrong suggested that both should be used to study metabolites. NMR is a much more reliable but MS is more sensitive (can pick up more metabolites). Because mass spectrometers pick up more metabolites you can do statistical studies on them such as multivariate analyses to ferret out subsets.

With NMR’s you can analyze the same samples on different machines and get the same result. You can’t do that with MS. In fact, MS’s produce variable enough results that you can’t even measure samples on different days with the MS and get the same results; it’s best to do them all in one shot.

The reliability issues are overcome by using large sample sizes, by analyzing the same sample repeatedly and by using quality controls. Armstrong suggested using both machines in the same study but acknowledged that it was expensive. He noted that running his NMR samples on a mass spectrometer resulted in generally the same results.

metabolic pathways

A typically complicated chart of metabolic pathways

Instead of the mass spectrometry used in the Naviaux metabolomics study, the Australian group used a process called nuclear magnetic resonance (NMR). They did an untargeted search; i.e. they looked at all the major biochemical pathways in the body in an attempt to find metabolites they could use as diagnostic biomarkers.

Metabolites are very small molecules left over when larger molecules or compounds are broken down.  Our genes produce proteins which do the work of the cell. In the process of doing that, they interact with and get broken down into metabolites.

Unexpected metabolites that show up, or metabolites that show up in higher or lower concentrations than expected, indicate that some sort of alteration in metabolism has occurred. It’s these breakdowns or signs of disease or disturbance that the metabolomics studies are looking for.

Metabolomics studies are tricky, though; your metabolites can alter in response to the slightest changes in the body. Because they’re so sensitive Armstrong believes metabolomics studies could be an ideal way to delineate the subsets found in ME/CFS.  On the other hand, that sensitivity comes at a cost; metabolomics is best assessed using longitudinal or a series studies that can sample the ME/CFS population again and again – arriving at a set of core metabolites.

The group’s 2014 review paper found that ME/CFS metabolism studies found evidence of decreases in mitochondrial output, decreased amino acid production, increased levels of oxidative stress and problems with nitrogen.

The 2015 study examined the metabolites in both the blood and the urine. Increases in the concentration of a metabolite in the blood suggest that problems with a biological process is present.

Urine, on the other hand, is the body’s way of excreting waste and/or toxic products. Urinalysis is important in metabolomics because when faced with unhealthy concentrations of metabolites, our bodies will attempt to eliminate them through our urine.

If a metabolite is too low, then another metabolite may be dumped in order to maintain balance. If a metabolite is too high, then it may be eliminated through the urine.

Unusual levels of metabolites in the blood generally reflect a process that is happening at the moment the blood is drawn. Unusual metabolites in the urine, on the other hand, generally reflect a chronic or ongoing process of metabolite disruption which has become significant enough for metabolite dumping to begin.

Blood and urine metabolite levels are generally interconnected. Metabolic problems that are found in the blood but not in the urine could either reflect a temporary metabolic fluctuation or problems with excreting the metabolites via the urine.

The study included 34 females with ME/CFS who met the Canadian Consensus Criteria and 25 female healthy controls.

Study Results


Six blood metabolites were significantly altered in the ME/CFS group. Glucose levels were increased whereas acetate, glutamate, hypoxanthine, lactate, and phenylalanine were decreased.

A PCA analysis found a clear separation between the two groups using relative abundance data.

Another analysis indicated, interestingly enough, two metabolites whose abundance was not significantly altered, formate and acetate, played a significant role in separating the ME/CFS patients from the healthy controls.


Significantly decreased concentrations of five metabolites (acetate; alanine; formate; pyruvate; and serine) were found in ME/CFS. Comparing the abundance of metabolites relative to each other they found eight altered metabolites; the five decreased metabolites found above plus decreased valine and increased allantoin and creatinine.

A PCA analysis of the relative abundance data for urine demonstrated a moderate separation; i.e. the ME/CFS and healthy controls were moderately different.

Emphasis on Anaerobic Energy Production

At this point, an unusual problem that had surfaced earlier in the study came to the fore. Creatinine levels are often used as a kind of set point to standardize the concentrations of other urinary metabolites, but when the group used creatinine in this fashion all the metabolites registered as decreased. Considering that a questionable result the researchers normalized each sample to the total metabolite concentration.

The Gist

  • Amino acid concentrations suggest that the clean, powerful aerobic energy producing system is being inhibited, while the dirtier and less effective anaerobic energy production system is being used more in ME/CFS
  • The anaerobic energy pathway – glycolysis – which provides pyruvate for the mitochondria and small amounts of ATP is operating strangely
  • Instead of glucose, amino acids such as glutamate are being used to provide fuel for this pathway.
  • Increased usage of glutamate, however, could be contributing to reduced levels of the main antioxidant in the body (glutathione)
  • A similar pattern seen in sepsis and starvation suggests ME/CFS may, in some ways, be similar to those diseases/conditions
  • If ME/CFS patient’s cells exist in a chronic state of mild starvation, a specialized feeding program may be needed to help them recover
  • Both the Naviaux and this paper agree that a hypometabolic state characterized by amino acid depletion is present in ME/CFS
  • Several further Australian studies are underway including a longitudinal study examining treatment options
Creatinine, which is partially derived from creatine, is produced by the liver in order to fast-forward a phosphate or P to the muscles to help them produce ATP during anaerobic exercise.

Confirmation of that hypothesis came when creatinine was negatively correlated with glycolytic processes in the healthy controls. This suggested that as the aerobic energy production pooped out, creatinine levels were increased in order to rush phosphates to the muscles to produce more ATP.

Impaired Glycolysis

These correlations, paired with a decrease of amino acid concentrations, implicate an increasing utilization of amino acids as a source of energy production through the citric acid cycle, largely via glutamate. Armstrong et. al. 

ATP production

The aerobic energy producing pathway; note that glycolysis provides some ATP but that its main function in aerobic energy production is to provide pyruvate for the mitochondria.

Glycolysis is the metabolic pathway that converts glucose into pyruvate while producing small amounts of ATP. In glycolysis no oxygen has been used and only small amounts of ATP are produced.

Pyruvate then enters the mitochondria and gets oxidized in the aerobic energy production pathways to produce high amounts of ATP.

Glycolysis, then, is a part of both the anaerobic and aerobic energy production systems. When not enough oxygen is present for the aerobic energy production pathway to function properly  – such as during intense exercise – glycolysis becomes a particularly important source of energy. Not much energy is produced, though, and the by-products of anaerobic energy production such as lactate are toxic.

Note then that anaerobic energy production or glycolysis is always occurring at least to some degree, and a healthy body is well adapted to take care of the small amounts of toxic byproducts produced by it during normal functioning. A decline in the aerobic energy production process appears to have left people with ME/CFS more dependent than usual on glycolysis to produce energy.

Glycolysis appears to have problems in ME/CFS as well, though. Glucose is the preferred substrate for glycolysis. The high blood glucose levels in the ME/CFS patients combined with the decreased concentrations of the metabolic endpoints of glycolysis (alanine, pyruvate), plus reduced levels of acetate, suggested instead that glycolysis had taken a hit in ME/CFS.

Reduced blood lactate levels could reflect decreased glycolysis as well. All in all, the ME/CFS patients appeared to be having trouble converting carbohydrates (glucose) into energy in the glycolytic pathway.

If I have this very complex subject right this suggests both energy producing pathways – the aerobic and anaerobic – are not functioning well in chronic fatigue syndrome. The fact that glucose – the best fuel for glycolysis – is not being used in will impact aerobic energy production but it’s not clear to my befuddled brain if that is the main issue or if other problems with aerobic energy production exist.

The Glucose / Amino Acid Switch

What then were the people with ME/CFS using to produce the substrate – pyruvate – used by the aerobic energy production process? The low levels of amino acids found suggested they were breaking down non-essential amino acids (alanine, glutamate and proline) to produce it, instead of carbohydrates.

(Armstrong noted that the high glucose levels could also be the result of increased gluconeogenesis. Gluconeogenesis maintains glucose levels by breaking down non-carbohydrate substrates such as amino acids and lipids.; Note that increased gluconeogenesis would also result in increased glucose levels and reduced amino acid levels).

“…suggests in ME/CFS a depletion of nonessential amino acids in the blood is being used to fuel the citric acid cycle and produce ATP in the absence of sufficient glucose usage via glycolysis.”  Armstrong et. al. 

The amino acid depletions suggested that levels of an enzyme called aspartate transaminase (AST) used to break down amino acids should be dramatically increased. They weren’t able to test that hypothesis themselves but a recent Stanford/Columbia study finding threefold increases in AST levels suggested that the Aussie team’s findings and hypothesis regarding amino acid depletion were correct.  The study also found increases in other enzymes used to support amino acid use in the TCA cycle. Enzymes associated with producing aerobic energy in the electron transport chain were reduced.


The differences between blood and urine glutamate levels between the two groups may tell the tale. Glutamate levels were higher in the blood of the healthy controls and lower in their urine. (This presumably indicated glutamate levels were “in bounds” in the blood of the healthy controls; there was no need to dump glutamate into the urine). Glutamate levels, on the other hand, were essentially the same in both the blood and the urine in ME/CFS patients.

The authors believe this may be because glutamate in the blood of ME/CFS patients is being transformed into glutamine and then into amino acids for use as an energy substrate. (When amino acids are used in this way glutamate is usually the amino acid that is used up.) With the glutamate in the blood being used up as an energy substrate, little glutamate is left to be dumped into the urine; ergo, glutamate is low in both the blood and urine of ME/CFS patients.


Instead of glucose, glutamate, an amino acid, was being broken down to provide energy. (Glutamate enzyme)

The negative correlation of blood glutamate levels and urine creatinine levels also suggested that glutamate being used as a substrate for ATP production in ME/CFS.

This suggested that ME/CFS patients were using glutamate instead of glucose as an energy resource.

Glutamate, interestingly, given Shungu’s findings of low glutathione levels in the brain, plays an important role in glutathione synthesis. If glutamate is being used as an energy source it may not be available for the synthesis of glutathione – the body’s chief antioxidant.


As markers of anaerobic metabolism (formate, glycine, hypoxanthine, and lactate) increased in the ME/CFS patients, the levels of another marker of anaerobic metabolic activity, creatinine, decreased in the urine. The urine decreases suggested that creatinine – which is used to increase anaerobic energy production during intense exercise – was being used up quickly in the blood.  That creatinine depletion suggested, as have the other findings, that ME/CFS patients’ bodies were in the kind of energy deficit healthy people only reach when they exert themselves very vigorously.

These findings, of course, fits very well with the exercise studies showing that aerobic energy production has gotten hit hard leaving a significant number of patients relying on the dirtier, less efficient and far less powerful process of anaerobic energy production to produce their energy. At the IACFS/ME Christopher Snell said the metabolomics findings correlated perfectly with their exercise results.

Starving for Energy?

Why is this happening? Armstrong suggested a couple of reasons. He noted that many of the metabolomic anomalies he found in ME/CFS are also found in sepsis and starvation. All show reductions in amino acids and lipids and increased levels of glucose. In both diseases proteins and lipids are used to produce maintain low energy levels while glucose is used for other matters – such as immune cell proliferation in sepsis.

Armstrong speculated that an infection or autoimmune process may have triggered a sepsis-like condition which then lead to a state of chronic starvation. During sepsis immune cells rely entirely upon glycolysis to proliferate wildly. They are so energy hungry during this process that they can deplete the system of essential cofactors perhaps leading to a state of chronic cellular starvation.

In starvation amino acids and fats are preferentially used to feed the TCA or Krebs cycle instead of glucose. Likewise, in anorexia the mitochondria switch to amino acids and lipids to fuel ATP production. The Aussie team believes the inability to use glucose properly may be contributing to a kind of low-level chronic starvation of mitochondria.

This state of low-level starvation is not particularly easy to escape. When the body begins starve it robs the tissues of many of the cofactors (vitamins/minerals) needed to utilize foods. If those cofactors aren’t provided along with the food a problem called refeeding syndrome can result.  That’s an intriguing issue given the problems some severely ill patients have with gaining weight.

Low energy

Is a kind of cellular starvation present in ME/CFS?

Armstrong speculated that the treatment protocols similar to those used to safely bring people out of starvation might be able to help in ME/CFS. Those protocols involve providing nutrients in specific stages based on their metabolic state. Bob Naviaux has also endorsed a stepwise approach to solving ME/CFS patients’ metabolomics issues.

Naviaux has reported that he hopes to start small clinical trials in the upcoming year and the Australian group is doing likewise. They’re providing ME/CFS patients with combinations of metabolites, vitamins, and minerals and then monitoring their metabolomics over time to see the effect they’ve had.

Broadly Similar Findings Thus Far

Armstrong stated that thus far broadly similar findings pervade the ME/CFS metabolomics studies. They suggest that:

  • Increased oxidative stress is present;
  • Lipids (fats) are being used to produce ATP;
  • Issues of purine metabolism and with folate cycle and methionine are present;
  • Reduced glycolysis/ increased glucose and increased use of amino acids for ATP production are present.


Armstrong said it was too early to recommend specific treatments based on metabolomics results. The longitudinal studies underway, do however, include treatments the researchers believe may help to improve ME/CFS patients’ metabolic profiles.

Armstrong suggested that Rituximab could be helping, at least in part, by reducing the energy drain caused by expansion of B-cells.

Armstrong didn’t mention paleo or ketogenic-type diets. Do the problems with glucose metabolism suggest that high fat/ moderate protein and low carbohydrate diets are indicated? I don’t know but I’ve been very happy with my results from a paleo-type diet over the last couple of months.

Next Up

Next up for the Aussie group include studies on

  • Longitudinal metabolomics and genomics while providing interventions;
  • Metabolomics of immune cells;
  • Producing a large-scale symptom database for patient stratification;
  • Large-scale genetic markers and metabolite population studies.

Your Donations Keep Health Rising on the Web

Share This