• Welcome to Phoenix Rising!

    Created in 2008, Phoenix Rising is the largest and oldest forum dedicated to furthering the understanding of and finding treatments for complex chronic illnesses such as chronic fatigue syndrome (ME/CFS), fibromyalgia (FM), long COVID, postural orthostatic tachycardia syndrome (POTS), mast cell activation syndrome (MCAS), and allied diseases.

    To become a member, simply click the Register button at the top right.

Reduced diversity and altered composition of the gut microbiome in individuals with ME/CFS

Kyla

ᴀɴɴɪᴇ ɢꜱᴀᴍᴩᴇʟ
Messages
721
Location
Canada
https://microbiomejournal.biomedcentral.com/articles/10.1186/s40168-016-0171-4
(Open access)

Reduced diversity and altered composition of the gut microbiome in individuals with myalgic encephalomyelitis/chronic fatigue syndrome
  • Ludovic Giloteaux,
  • Julia K. Goodrich,
  • William A. Walters,
  • Susan M. Levine,
  • Ruth E. Ley and
  • Maureen R. HansonEmail author
Microbiome20164:30
DOI: 10.1186/s40168-016-0171-4

© The Author(s). 2016

Received: 29 February 2016

Accepted: 11 May 2016

Published: 23 June 2016

Abstract
Background
Gastrointestinal disturbances are among symptoms commonly reported by individuals diagnosed with myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS). However, whether ME/CFS is associated with an altered microbiome has remained uncertain. Here, we profiled gut microbial diversity by sequencing 16S ribosomal ribonucleic acid (rRNA) genes from stool as well as inflammatory markers from serum for cases (n = 48) and controls (n = 39). We also examined a set of inflammatory markers in blood: C-reactive protein (CRP), intestinal fatty acid-binding protein (I-FABP), lipopolysaccharide (LPS), LPS-binding protein (LBP), and soluble CD14 (sCD14).

Results
We observed elevated levels of some blood markers for microbial translocation in ME/CFS patients; levels of LPS, LBP, and sCD14 were elevated in ME/CFS subjects. Levels of LBP correlated with LPS and sCD14 and LPS levels correlated with sCD14. Through deep sequencing of bacterial rRNA markers, we identified differences between the gut microbiomes of healthy individuals and patients with ME/CFS. We observed that bacterial diversity was decreased in the ME/CFS specimens compared to controls, in particular, a reduction in the relative abundance and diversity of members belonging to the Firmicutes phylum. In the patient cohort, we find less diversity as well as increases in specific species often reported to be pro-inflammatory species and reduction in species frequently described as anti-inflammatory. Using a machine learning approach trained on the data obtained from 16S rRNA and inflammatory markers, individuals were classified correctly as ME/CFS with a cross-validation accuracy of 82.93 %.

Conclusions
Our results indicate dysbiosis of the gut microbiota in this disease and further suggest an increased incidence of microbial translocation, which may play a role in inflammatory symptoms in ME/CFS.
...
 

msf

Senior Member
Messages
3,650
Ah, as I expected, they only used the Fukuda criteria - does this make KDM´s study (see thread on IiME conference) more relevant for people with ME then?
 

msf

Senior Member
Messages
3,650
Ah, as I expected, they only used the Fukuda criteria - does this make KDM´s study (see thread on IiME conference) more relevant for people with ME then?

´ME/CFS patients had significantly higher plasma LPS levels than healthy individuals (median ME/CFS—119.43 pg/mL vs. controls—74.74 pg/mL, P < 0.0005´ - please tell me someone isn´t going to say this isn´t valid because they didn´t correct for multiple comparisons.
 

msf

Senior Member
Messages
3,650
Looks like technology may have advanced since the KDM study, or the KDM group didn´t choose the best tool for the job:

´Recently, a study used high-throughput 16S rRNA gene sequencing to investigate the presence of specific alterations in the gut microbiota of ME/CFS patients from Belgium and Norway [12]. The authors amplified the V5 and V6 hypervariable 16S rRNA regions and sequenced the amplicons using a Roche FLX 454 sequencer, which resulted in an average of only 6000–7000 reads/sample. In contrast, we amplified the V4 hypervariable region of the 16S rRNA gene, sequenced amplicons using the MiSeq Illumina platform, obtained an average of many more reads/sample (98,000), and compared the resulting sequences to a different database, the Greengenes non-redundant reference database [31]. Our analysis showed that within-sample diversity is lower in the ME/CFS specimens compared to controls.The same indices in the Fremont et al. [12] study did not differ between ME/CFS and healthy subjects [12], likely due to the lower read number they obtained. Lower richness has also been observed in unhealthy or inflammatory states [32, 33] and has been associated with IBD, necrotizing enterocolitis [34], and greater abdominal discomfort levels in patients with food intolerances [35, 36].´
 

A.B.

Senior Member
Messages
3,780
Ah, as I expected, they only used the Fukuda criteria - does this make KDM´s study (see thread on IiME conference) more relevant for people with ME then?

´ME/CFS patients had significantly higher plasma LPS levels than healthy individuals (median ME/CFS—119.43 pg/mL vs. controls—74.74 pg/mL, P < 0.0005´ - please tell me someone isn´t going to say this isn´t valid because they didn´t correct for multiple comparisons.

I'm not qualified but I believe the very low p value makes it unlikely for this to be due to chance. Maybe someone who knows more about correcting for multiple comparisons can weigh in.
 

msf

Senior Member
Messages
3,650
I was (hopefully) being sarcastic. They seem to have corrected for numerous comparisons later in the paper, but whether it is relevant to the LPS measurement I cannot say.
 

Bob

Senior Member
Messages
16,455
Location
England (south coast)
Abstract said:
Using a machine learning approach trained on the data obtained from 16S rRNA and inflammatory markers, individuals were classified correctly as ME/CFS with a cross-validation accuracy of 82.93 %.
Mark's blog says that Maureen Hanson mentioned this at the Invest in ME conference. And she says they are aiming to get 100% accuracy.
They also found an association of specific bacterial groups with ME/CFS, as distinct from controls: they found that anti-inflammatory bacterial species (ruminococcae, which produce butyrate, an anti-inflammatory fatty acid) and species of bifidobacterium (which produce lactic acid) were reduced in ME/CFS patients. And with a combination of blood and gut assays they could classify 83% of samples correctly as coming from patients or controls.

...

Although her study was small, she does believe that they could use these metabolites to distinguish 100% of patients from controls, and she is hopeful that they will be able to develop this research into a real biomarker test for ME/CFS.

I haven't read the paper yet.
 
Last edited:

msf

Senior Member
Messages
3,650
I remember Fluge and Mella referenced Maes in one of their papers, but I can´t remember which paper they referenced. I would be interested to hear their take on this.
 

A.B.

Senior Member
Messages
3,780
Anyone want to show their results? My metagenomics stool analysis (through Red Labs):

Fimicutes were 49%, normal would be 50-85% in Europeans. The test indicated that I have low diversity. My sCD1 was also high, but just barely. Seems loosely in line with the findings here.
 

msf

Senior Member
Messages
3,650
Yeah, almost all of the KDM patient results I have seen are in line with the findings in this study.

I´ll find and post my results later.
 
Last edited:

user9876

Senior Member
Messages
4,556
Ah, as I expected, they only used the Fukuda criteria - does this make KDM´s study (see thread on IiME conference) more relevant for people with ME then?

´ME/CFS patients had significantly higher plasma LPS levels than healthy individuals (median ME/CFS—119.43 pg/mL vs. controls—74.74 pg/mL, P < 0.0005´ - please tell me someone isn´t going to say this isn´t valid because they didn´t correct for multiple comparisons.


I think we need to be careful with P values. What they basically say is that the chance that the measures come from the same distribution are small (in this case). However, the distributions can be overlapping which means that the measure is not a good bio marker and I would go further and say that this measure is clearly not the cause. It could be part of the cause but since the distributions overlap there are healthy people with the same levels as people with ME. Hence we need to do more to understand what the role of this and the other measures could be within the overall system. It could for example be an issue of this and genes or multiple different measures together.

Collapsing taxonomy to the genus level, individuals with ME/CFS were classified correctly and separately from the healthy group with an average success rate of 0.82 ± 0.12.

This suggests something interesting particularly if there are subsets. However, in a very quick skim through their paper I'm not sure what methodology they used. Really they should have three data sets the first to train several classifiers, the second to choose what appears to be the best classifier (test set) and a third to test the performance of the overall classifier (verification set). Without this the classification results could be over stated.

What I would like to see is a clustering over all the variables they collect to see if there are clear clusters or ones representing different subgroups.

I say all this with no understanding of the anything about the biology of the situation and as such these are general comments about data and overlapping groups.
 

A.B.

Senior Member
Messages
3,780
My bacteroides were low, but my bifidobacteria were very high. sIgA (not measured in this study) was extremely high and the most unusual of all tested things. So the bacteroides/bifido findings don't match my results, but the firmicutes/sCD14 do.
 

msf

Senior Member
Messages
3,650
I t
I think we need to be careful with P values. What they basically say is that the chance that the measures come from the same distribution are small (in this case). However, the distributions can be overlapping which means that the measure is not a good bio marker and I would go further and say that this measure is clearly not the cause. It could be part of the cause but since the distributions overlap there are healthy people with the same levels as people with ME. Hence we need to do more to understand what the role of this and the other measures could be within the overall system. It could for example be an issue of this and genes or multiple different measures together.



This suggests something interesting particularly if there are subsets. However, in a very quick skim through their paper I'm not sure what methodology they used. Really they should have three data sets the first to train several classifiers, the second to choose what appears to be the best classifier (test set) and a third to test the performance of the overall classifier (verification set). Without this the classification results could be over stated.

What I would like to see is a clustering over all the variables they collect to see if there are clear clusters or ones representing different subgroups.

I say all this with no understanding of the anything about the biology of the situation and as such these are general comments about data and overlapping groups.

I think you will find that it is the cause of most of the general ME symptoms. What you are going to see next is the same study done in people with ME who have just done exercise, and there is going to be even clearer difference between cases and controls.

Also, you need to remember that they used Fukuda, so they will have included patients who do not have ME but some other condition, most likely one that doesn´t show increased levels of plasma LPS.
 
Last edited:

Simon

Senior Member
Messages
3,789
Location
Monmouth, UK
I'm not qualified but I believe the very low p value makes [P < 0.0005] it unlikely for this to be due to chance. Maybe someone who knows more about correcting for multiple comparisons can weigh in.
Yes, while they don't seem to have corrected for multiple comparisons, this p value would survive it, especially as most of the measures are, unsurprisingly, highly correlated (LPS, LBP ie LPS-binding protein, and sCCD14, which is activated by LPB). This means using eg Bonferroni correction would be too strict ie even easier for that p value to survive correction for multiple comparisons. However, as @user9876 says, the levels overlap substantially between patients and controls, making it a quesitonable biomarker. Looking at the data, there seem to be signs of subgroups - maybe a biomarker for a subgroup?

See my edit of Figure 1 (nb order in my edit is d, b, e)
hansonlps.png


This suggests something interesting particularly if there are subsets. However, in a very quick skim through their paper I'm not sure what methodology they used. Really they should have three data sets the first to train several classifiers, the second to choose what appears to be the best classifier (test set) and a third to test the performance of the overall classifier (verification set). Without this the classification results could be over stated.
Agreed, though they used random forest for classification and also used cross--validation, both of which should reduce overfitting (and false positives). Even so, needs replication.
 
Last edited:

A.B.

Senior Member
Messages
3,780
Looking at the data, there seem to be signs of subgroups - maybe a biomarker for a subgroup?

Do you mean the way LPS clusters in Figure 1? In patients, it's either below average or above average but few have "average" LPS.