Symptom-based clusters in people with ME/CFS: an illustration of clinical variety in a cross-sectional cohort 2023,Vaes,Jason et al

Abstract
Background

Myalgic encephalomyelitis (ME)/chronic fatigue syndrome (CFS) is a complex, heterogenous disease. It has been suggested that subgroups of people with ME/CFS exist, displaying a specific cluster of symptoms. Investigating symptom-based clusters may provide a better understanding of ME/CFS. Therefore, this study aimed to identify clusters in people with ME/CFS based on the frequency and severity of symptoms.

Methods
Members of the Dutch ME/CFS Foundation completed an online version of the DePaul Symptom Questionnaire version 2. Self-organizing maps (SOM) were used to generate symptom-based clusters using severity and frequency scores of the 79 measured symptoms. An extra dataset (n = 252) was used to assess the reproducibility of the symptom-based clusters.

Results
Data of 337 participants were analyzed (82% female; median (IQR) age: 55 (44–63) years). 45 clusters were identified, of which 13 clusters included ≥ 10 patients. Fatigue and PEM were reported across all of the symptom-based clusters, but the clusters were defined by a distinct pattern of symptom severity and frequency, as well as differences in clinical characteristics. 11% of the patients could not be classified into one of the 13 largest clusters. Applying the trained SOM to validation sample, resulted in a similar symptom pattern compared the Dutch dataset.

Conclusion
This study demonstrated that in ME/CFS there are subgroups of patients displaying a similar pattern of symptoms. These symptom-based clusters were confirmed in an independent ME/CFS sample. Classification of ME/CFS patients according to severity and symptom patterns might be useful to develop tailored treatment options.

https://translational-medicine.biomedcentral.com/articles/10.1186/s12967-023-03946-6

 

Perhaps this mainly shows how difficult it is to create clusters based on symptom data.

 

Especially when you know how much the symptoms fluctuate between and among people, making most of those attempts completely pointless until medicine actually starts respecting time as important, that those are dynamic systems that can't be pinned down with single points in time.

The very first Long Covid paper showed how staggered symptoms are. It should have set the stage for everything that followed, but of course everyone is still trying to develop that perfect snapshot that captures the whole illness, which is clearly not a thing. It's the lack of adaptation that gets me. Adaptation is the real test of intelligence, having smart answers means nothing when you keep applying them to the wrong problem. But there's zero adaptation, and clearly learning from experience doesn't work.

 

Have only skimmed the study but could find no mention of the key limitation of the DSQ-2 which is that counting the frequency and severity of symptoms without also accounting for the effect of pacing is going to mess up your data. PwME who can pace well enough to avoid PEM most of the time have less frequent and less severe symptoms than pwME who can't pace to the same degree. Additionally many pwME have some symptoms only during PEM so the type of symptoms people report also depends on their pacing ability. All of which means pwME could be assigned to a specific cluster because of their pacing status rather than their underlying ME phenotype.

 

The ME/cvs Stichting who's donators were part of the study wrote a summary of the study. The Stichting is a bit controversial as they've worked with CBT/MUPS proponents in the past and probably will do so again as they're quite keen on working with professor Judith Rosmalen for the biomedical ZonMw program. She now also studies long-COVID at the moment. Anyways, good on them for working together with the likes of Jason this time.

https://mecvs.nl/nieuws/classification-of-me-cfs-patients-based-on-their-common-symptoms/