This is a “scribe notes” post. I am working on implementing odds ratio as a forecaster for ME/CFS and encountered some issues. In my work experience, this means taking a significant step backwards to look at the data better. I will look at the bacteria with a high frequency of being reported in tests first. The genus of greatest interest from preliminary work are:
Blautia
Faecalibacterium
Clostridium
Lachnospira
Streptococcus
Collinsella
Anaerostipes
Parabacteroides
Anaerotruncus
Bifidobacterium
Pseudobutyrivibrio
The data is from BiomeSight samples. The red line is what would be expected with no influence. The population percentile included those with ME/CFS. Removing them would increase the differences more (I went lazy).
Charts
Blautia
Lower levels are clearly significant.
Faecalibacterium
Lower levels are some significance.
Clostridium
Lower levels are clearly significant.
Lachnospira
Higher levels are some significance for low values.
Streptococcus
Does not appear to have much significance
Collinsella
Higher levels are some significance for below average values
Anaerostipes
Lower levels are significance for all values.
Parabacteroides
Does not appear to have much significance
Anaerotruncus Lower levels are significance for all values.
Bifidobacterium
Higher levels are significance for all values.
Pseudobutyrivibrio
Higher levels are significance for all values.
Next Steps
We can visually see how bacteria is shifted for these genus. The challenge is converting it to a formula to forecast that cross validates. Stay tune.
A reader wrote me today with the following question
I read your article on microbial involvement, can you explain in more detail why you recommended cutting out Lactobacillus? If I interpreted your analysis, you said lactobacillus was rare in ME/CFS, doesn’t that mean increasing it could be beneficial? “..lactobacillus shows up barely in only one result..”.Meaning lactobacillus is rare for ME/CFS patients?
I meant that there is no clear evidence of a lactobacillus deficiency with ME/CFS patients microbiomes. There are reasons to believe that it may be harmful and helps maintain ME/CFS state.
History
I am a facts/study based individual that have been reading studies, conference reports since 1990’s. Back in 1999, on eGroups CFSFM-Experimental, taking probiotics were often suggested. Why? Because probiotics has been promoted as a cure-all for all conditions. A influencer snake oil. Reported results on CFSFM-Experimental were disappointing.
My mind proceeded logically. So ask the US National Library of Medicine (PubMed), “Which probiotics have been helpful for ME/CFS? Given that there were 2400 studies on ME/CFS then, I expected to find a few dozen by then — after all, it would likely be one of the first choices by naturopaths who would rush to publish their results!! There were none that used lactobacillus probiotics. Even today, we have just 32 studies mentioning “chronic fatigue syndrome” probiotics in the 8500 studies posted. [Sarcasm] “Surely, there would have been a rush with all of the ME/CFS specialists to use lactobacillus probiotics given all of this evidence”.
Being a scientist, I know that what gets published are positive results — not no result nor negative results.
Reading conference papers presented by specialist on ME Research UK, I came across a report of a conference panel by active practitioners where the consensus was no benefit. I have worked as a professional technical writer and very “phrasing aware”, I read the wording to indicate that probiotics likely did harm in some of their patients. Slamming probiotics tend to be view as a heresy with many health influencers.
There Appears No Significant Objective Evidence that lactobacillus helps!
Yes, you will find testimonials — but that is not objective evidence. They may have helped because the person did not have ME/CFS (self diagnosis) or a different condition. It is incomprehensible that there have not been dozens (or hundreds) of studies trying lactobacillus — studies that are unpublished because of unfavorable results.
Why may it be EVIL?
Again, conference papers from Australia’s Alison Hunter Memorial Foundation play an important role here. From the Way-Back machine I retrieved items no longer on their site and pasted it into 1998 Was a very good year…. The key finding was “The mean distribution of E.coli as percentage of the total aerobic microbial flora for the control subjects and CFS patients was 92.3% and 49% ” Not a little drop, but almost half the level!
NOTA BENA: The typical (cheap) 16s tests used for most modern microbiome studies effectively ignore E.Coli. Shotgun testing (much more expensive) finds E.Coli in almost every sample. Some 16s finds it in 1 in a thousand samples as shown by the table below. Modern studies not repeating these results is a direct consequence of their methodologies!!
This was the motivation for my trying Mutaflor Probiotics (E.Coli Nissle 1917) which I happen to have in the house because my wife has Crohn’s and it made a huge difference for her (with lots of studies reporting it too!!!). I had a severe Jarisch–Herxheimer reaction for two weeks and a rapid recovery from ME/CFS afterwards.
If you look at Odds Ratios for Metabolites and ME/CFS, you will see that E.Coli probiotics has the biggest impact on the metabolite imbalance with ME/CFS
Going over to the E.Coli page on Microbiome Prescription we see that Lactobacillus constantly reduces E.Coli. So we are moving from levels that are 50% of normal levels to even lower levels.
IMHO, for ME/CFS, Lactobacillus probiotics are EVIL
Yes there are a few lactobacillus that will help some symptoms (and likely make other symptoms worse). Unless you are very sure that it has the actual probiotic strain used in studies, don’t do it. See Probiotics — what is advertised may not be what you get.
IMHO, for brain fog, Lactobacillus probiotics are EVIL
“Of particular concern is the association between certain probiotic strains, such as Lactobacillus species, and the production of bacterial metabolic byproducts (e.g., D-lactate and histamine) implicated in the pathogenesis of brain fogginess, a term describing a state of cognitive dysfunction characterized by symptoms including confusion, impaired judgment, and lack of focus ” [Aug 2024] with many links to further studies.
Bottom line, checking for clinical studies if a probiotics clearly helps is recommended. This search engine may help.
Bifidobacterium also?
In Visual Exploration of Odds Ratios, we see that ME/CFS people have higher then general population amounts of Bifidobacterium. On the flip side, the average amount is reported lower on several studies. This compounds issues with several things that needs to be investigated.
Did the lower bifidobacterium count not found in their average as zero? We use the values only when detected. Looking at the dots, we see that the dots are sparse/rare for lower values suggesting a lower detection rate. This suggests a threshold behavior of bifidobacterium.
Looking at impact on E.Coli, we see most studies say that it decreases E.Coli
There is not enough data to come to a safe conclusion.
Bacterial Metabolic Activity: Bacteria produce and consume various metabolites, which can significantly impact the host’s metabolic environment13.Metabolic Imbalances: Different bacterial compositions can lead to similar metabolite imbalances, making metabolite profiles potentially more informative than bacterial species profiles alone7 8.
Advantages of This Approach
Net Effect: By examining metabolites, we can assess the overall impact of the microbiome on the host, regardless of the specific bacterial species present5.
Consistency: Metabolite imbalances may be more consistent across patients than bacterial species composition, which can vary widely7.
Functional Insight: This approach provides insight into the functional consequences of microbiome dysbiosis in ME/CFS3 8.
The analysis of metabolites across multiple microbiome testing platforms (Ombre, Biomesight, and uBiome) reveals a more consistent pattern of metabolite imbalances compared to bacterial species identification.
Potential Consequences of Low GDP-L-fucose (the top one)
The deficiency in GDP-L-fucose could have several implications:
Altered Immune Response: It may affect the proper functioning of the immune system, potentially impacting inflammatory processes12.
Cancer-Related Changes: Low levels might influence tumor progression or immune evasion mechanisms, as fucosylation is often altered in cancer24.
First-degree relatives: A clinic-based study reported that first-degree relatives of ME/CFS patients had a significantly higher(four times) prevalence of any cancer compared to controls (OR 4.06) [2022]
Cellular Communication: It could disrupt normal cell-cell interactions and signaling pathways dependent on fucosylated glycans3.
Metabolite
Threshold
Low
High
GDP-L-fucose
16.5
12.1
21.2
Holo-[citrate (pro-3S)-lyase]
12.7
1.8
23.4
N,N’-Diacetyllegionaminate
12.1
5.4
22
alpha-Oxo-benzeneacetic acid
11.8
1.9
23.4
Oxalyl-CoA
11.3
2
22.4
S-Methyl-5-thio-D-ribose 1-phosphate
11.2
1.3
30.6
Malonyl-CoA
10.6
2.5
26.3
1,2-Diacyl-3-alpha-D-glucosyl-sn-glycerol
10.5
5.4
20
Translate Low Metabolites to Probiotics
Many of these metabolites are produced by probiotics, so in terms of highest importance and reasonably available probiotics, I produced the list below.
The top one is one is one that had very dramatic positive effect for me when I relapsed into ME/CFS (after the worse herxheimer reaction that I have ever experienced): Mutaflor (E.Coli Nissle 1917). I took it as a result of a 1999 study in Australia reporting very low levels of E.Coli in CFS patients [As a FYI, 16s tests do a very poor reporting on E.Coli].
I asked perplexity, which foods may increase any of the above metabolites. The following were reported:
Spinach
Rhubarb
Beets
Nuts
Chocolate
Tea
Wheat bran
Strawberries
Bottom Line
While the metabolite-focused approach provides valuable insights into the biochemical imbalances associated with ME/CFS, its immediate clinical applications are somewhat limited. The probiotics and the food suggestions are reasonable and I see several of the items appearing on suggestions from the expert system for ME/CFS patients.
This is the title of a new publication in Microbes & Immunity, available here.
Microbiome fluctuations or metabolic endotoxemia are proposed as possible disorder biomarkers. Based on the fact that gut microbiota dysbiosis reverts to a state of eubiosis in long-term patients with this condition, it may be hypothesized that disease progression begins with the loss of beneficial gut microorganisms, particularly short-chain fatty acid producers, leading to more widespread gastrointestinal phenotypes that are subsequently reflected in plasma metabolite levels. These alterations, specific of each individual, thereby result in metabolic and phenotypic shifts and in ME/CFS.
This has been my primary hypothesis for many years and lead to my writing Microbiome Prescription to normalize the dysbiosis. The bacteria shifts will be reported differently from different labs because of a lack of standardization of microbiome tests (See The taxonomy nightmare before Christmas… ).
Below I have pulled the lab specific dysbiosis shifts for ME/CFS at thhe genus level.
Biomesight Tests
The number are the percentile ranking for the bacteria listed. Each match increases the odds of ME/CFS by 1.5.
Methylonatrum >= 99
Planifilum >= 98.9
Thiohalorhabdus >= 98.9
Granulicatella >= 98.7
Amedibacillus >= 98.6
Enterococcus >= 98
Methylobacillus >= 98
Emticicia >= 97.7
Candidatus Tammella >= 97.7
Dokdonella >= 97.6
Runella >= 97.4
Collinsella <= 18
Pseudobutyrivibrio <= 16.4
Calothrix <= 13.8
Lachnospira <= 11.4
Veillonella <= 10.8
Shuttleworthia <= 8.5
Parabacteroides <= 8.4
Bifidobacterium <= 8.4
Faecalibacterium <= 7.5
Actinobacillus <= 7.5
Pedobacter <= 7.2
Coprococcus <= 4.6
Moorella <= 3.5
Natronincola <= 2.9
Mediterraneibacter <= 2.5
Oscillospira <= 1.8
Anaerovibrio <= 1.7
Sphingobacterium <= 1.6
Thryve
Collinsella <= 44.2
Bifidobacterium <= 41.1
Gemmiger <= 23
Dorea <= 19
Butyrivibrio <= 8
Fusicatenibacter <= 5.7
Lachnoclostridium <= 5.2
Gordonibacter <= 5.1
Eubacterium <= 4.8
Prevotella <= 4
Brassicibacter <= 3.9
Sporobacter <= 3.5
Johnsonella <= 3.5
Phascolarctobacterium <= 3.4
Dialister <= 3.4
Agathobacter <= 3.3
Lactonifactor <= 3.2
Murimonas <= 3.2
Moryella <= 3.1
Niabella <= 3.1
Ruminococcus <= 3
Hungatella <= 3
Anaerotruncus <= 2.8
Odoribacter <= 2.7
Lactobacillus <= 2.4
Eggerthella <= 2.3
Cellulosilyticum <= 2.1
Haemophilus <= 2
Holdemania <= 2
Terrisporobacter <= 2
Anaerobutyricum <= 1.9
Howardella <= 1.8
Hydrogenoanaerobacterium <= 1.8
Lachnobacterium <= 1.8
Hespellia <= 1.7
Agathobaculum <= 1.6
Pseudoflavonifractor <= 1.6
Desulfotomaculum <= 1.6
Intestinibacter <= 1.6
Barnesiella <= 1.5
Eisenbergiella <= 1.5
Facklamia <= 1.5
Acetatifactor <= 1.5
Ruminiclostridium <= 1.5
Natranaerovirga <= 1.5
Enterocloster <= 1.4
Caloramator <= 1.4
Thomasclavelia <= 1.4
Desulfovibrio <= 1.4
Parabacteroides <= 1.4
Shuttleworthia <= 1.3
Peptoniphilus <= 1.3
Akkermansia <= 1.3
Alistipes <= 1.3
Marvinbryantia <= 1.3
Acetivibrio <= 1.2
Paraprevotella <= 1.2
Veillonella <= 1.2
Bacteroides <= 1.2
Butyricimonas <= 1.2
Romboutsia <= 1.2
Ethanoligenens <= 1.2
Phocaeicola <= 1.2
Anaerocolumna <= 1.2
Blautia <= 1.1
Coprococcus <= 1
Lachnospira <= 1
uBiome
Veillonella <= 22.4
Hespellia <= 19
Bifidobacterium <= 15.1
Anaerosporobacter <= 7.2
Streptococcus <= 6.8
Eggerthella <= 6.4
Phascolarctobacterium <= 4.1
Eisenbergiella <= 3.9
Marvinbryantia <= 3.4
Holdemania <= 3.1
Dorea <= 3
Anaerostipes <= 3
Oscillibacter <= 2.7
Dialister <= 2.3
Kluyvera <= 2.1
Peptoclostridium <= 1.7
Erysipelatoclostridium <= 1.4
Intestinimonas <= 1.3
Sarcina <= 1.3
Roseburia <= 1.3
Phocaeicola <= 1.2
Terrisporobacter <= 1.2
Bacteroides <= 1.1
Bottom Line
Bifidobacterium is common across all of these, and lactobacillus shows up barely in only one result. Since there can be some hostility between Bifidobacterium and Lactobacillus, I would suggest cutting out lactobacillus probiotics.
I repeated this for Bifidobacterium species and one species shone.
Bifidobacterium longum <= 9.9
My personal preference for a source is Maple Life Science which sells direct from factory resulting in very fresh and alive probiotics.
Some examples of analysis people with Long COVID and ME/CFS are there. I am working on building an algorithm to build suggestions based on odds ratio which should improve the suggestions more.
CFS Remission 