Preprint Beyond pain: Using Unsupervised Machine Learning to Identify Phenotypic Clusters of Small Fiber Neuropathy 2024 Murin et al

Discussion in 'Other health news and research' started by Hoopoe, Nov 15, 2024.

  1. Hoopoe

    Hoopoe Senior Member (Voting Rights)

    Messages:
    5,483
    Lilas, voner, hotblack and 7 others like this.
  2. wigglethemouse

    wigglethemouse Senior Member (Voting Rights)

    Messages:
    1,128
    Not sure what is meant by lower intensity of neuropathic pain. Cluster 1 looks to have intense pain and cluster 3 less pain with intense fatigue. Cluster 1 and cluster 3 seem to have similar levels of fatigue.
    upload_2024-11-15_16-47-14.png

    Also:
     
    Lilas, Hutan, hotblack and 5 others like this.
  3. Hoopoe

    Hoopoe Senior Member (Voting Rights)

    Messages:
    5,483
    In cluster 3 the fatigue tends to be the worst, or one of the worst, symptoms.
     
  4. wigglethemouse

    wigglethemouse Senior Member (Voting Rights)

    Messages:
    1,128
    Cluster 1 and cluster 3 look very similar for fatigue based on the data. If anything Fig A vs FigC shows a higher intensity of fatigue on average for cluster 1, with what look like similar levels of prevalence of fatigue in Clusters 1 & 3 - Fig D & Fig F.

    upload_2024-11-16_9-12-59.png


    So their wording seems off. Why highlight that cluster 3 had especially severe fatigue (5.6) when cluster 1 had a higher average value of fatigue (7.2)?

    I don't understand their method of "Unsupervised machine learning" but it sure looks like it's based on the surveys. If that is the case, then all this paper is telling us is that MFN vs SFN is indistinguishable from each other by looking at symptoms, and SFN causes widespread pain and fatigue and other symptoms, which many people may not be aware off.

    The clusters seem meaningless to me for interpreting anything.

    This seemed interesting as lipids has been highlighted quite a bit in ME.
     
    Sean, oldtimer, Hutan and 1 other person like this.
  5. Hutan

    Hutan Moderator Staff Member

    Messages:
    32,224
    Location:
    Aotearoa New Zealand
     
    Sean, oldtimer and wigglethemouse like this.
  6. Hoopoe

    Hoopoe Senior Member (Voting Rights)

    Messages:
    5,483
    A cluster 3 patient would be a patient for whom fatigue is the worst symptom. They would have other symptoms but usually of low to moderate intensity. I always thought of sensory nerve abnormalities (pain, abnormal sensations, lack of sensation) as being the main symptoms of SFN but this study shows something different.
     
    Last edited: Nov 16, 2024
    Sean, oldtimer, Hutan and 1 other person like this.
  7. Hutan

    Hutan Moderator Staff Member

    Messages:
    32,224
    Location:
    Aotearoa New Zealand
    It's certainly interesting and raises many questions for me. Was there some biased selection of the participants? How common is a finding of SFN in people who would rate themselves as healthy? How many of the people in the Type 3 group would meet ME/CFS diagnostic criteria?

    SFN has been raised as a comorbidity of ME/CFS for a while now. I'm not sure that we have any really good data on its prevalence though. It's too easy for the selection of participants to be biased.

    Possibly, people who have some pins and needles, the occasional burning pains in their feet but are otherwise healthy might just ignore those symptoms, thinking 'everyone probably gets that'. I think that would have been the case for me. Whereas if someone is debilitated by fatigue/PEM/fatiguability, they might start assessing and reporting every symptom they have, in order to find a clue?
     
    Last edited: Nov 17, 2024
  8. Sean

    Sean Moderator Staff Member

    Messages:
    8,870
    Location:
    Australia
    According to standard annual bloods my lipids have always been within normal, apart from a couple of borderline results over the years.

    Plus a couple of times my fasting sugar levels have been borderline, but the follow up test for diabetes has been negative.

    Only consistently abnormal standard lab result for me is alkaline phosphatase, which is always high (typically around 140-160, IIRC), and has been since at least my mid-20s.
     
  9. Amw66

    Amw66 Senior Member (Voting Rights)

    Messages:
    6,940
    If ALP is high how is your GGT ?
    If it's not high as well then it can be indicative of bone issues / vit D deficiency
    ( Took a look at LFTs when daughter had low vitD)

    GGT being high has been mooted as being potentially indicative of early NAFLD , and perhaps insulin resistance .
     
  10. Sean

    Sean Moderator Staff Member

    Messages:
    8,870
    Location:
    Australia
    Don't know if I have been tested for GGT levels. But I don't think there have been any liver function tests.

    The high ALP was originally noted by the professor of immunology who diagnosed me with ME/CFS. Neither he nor any other clinician since has thought it something worth following up on, and I have asked about it a number of times. Main responses have been that it was too unspecific a finding to tell us much, particularly in the absence of other findings or symptoms, and that is was not so high as be a significant concern. The normal range is not well defined and can very quite a bit, and depending on which authority you read it can overlap with my range, and my understanding is that it is not considered worth following up on until it is above about 200, particularly in the absence of other signs or indicators. All complicated by having osteoarthritis, and also getting old (early 60s), both of which can be associated with raised ALP.

    Will ask my doctor about GGT next visit.
     
  11. Sean

    Sean Moderator Staff Member

    Messages:
    8,870
    Location:
    Australia
    Interestingly, one of the main findings of the DECODE genetic study is liver dysfunction (plus insulin resistance and low level inflammation), and liver problems are one of the main suspects for raised ALP levels.
     

Share This Page