The Reification of the Clinical Diagnosis of Myalgic Encephalomyelitis/ Chronic Fatigue Syndrome (ME/CFS) as an Immune and Oxidative Stress Disorder: Construction of a Data-driven Nomothethic Network and Expo-sure of ME/CFS Subgroups

The approach towards myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) remains in a permanent state of crisis with fierce competition between the psychosocial school, which attributes ME/CFS to the perception of effort, and the medical approach (Maes and Twisk, BMC Med, 2010, 8, 35). The aim of this paper is to review how to construct a nomothetic model of ME/CFS using Partial Least Squares (PLS) path analysis and ensembling causome (bacterial translocation as assessed with IgM/IgA responses to LPS), protectome (lowered coenzyme Q10), adverse outcome pathways (AOP) including increased lysozyme, CD38+ T cell activation, cell-mediated immune activation (CMI), and IgM responses to oxidative specific epitopes and NO-adducts (IgM OSENO).Using PLS, we trained, tested and validated this knowledge- and data-driven causal ME/CFS model, which showed adequate convergence, construct and replicability validity.This bottom-up explicit data model of ME/CFS objectivates the descriptive narratives of the ME/CFS phenome, using causome-protectome-AOP data, whereby the abstract concept ME/CFS is translated into pathways, thereby securing the reification of the ME/CFS phenome. We found that 31.6% of the variance in the physiosomatic symptom dimension of ME/CFS was explained by the cumulative effects of CMI and CD38+ activation, IgM OSENO, IgA LPS, lysozyme (all positive) and coenzyme Q10 (inversely). Cluster analysis performed on the PLS-generated latent vector scores of all feature sets exposed three distinct immune groups of ME/CFS, namely one with increased lysozyme, one with increased CMI + CD38 activation + depressive symptoms, and another with increased bacterial translocation + autoimmune responses to OSENO.

Automated summary

This study aimed to construct a model of ME/CFS using PLS path analysis and data on causome, protectome, and AOP. The results showed that 31.6% of the variance in physiosomatic symptoms of ME/CFS could be explained by various factors. Cluster analysis revealed three distinct immune groups within ME/CFS, each characterized by different immune responses.