The Fibromyalgia Overlap
Fibromyalgia is ME/CFSβs closest nosological relative β the two conditions co-occur at rates of 22β47% depending on criteria and sample, and their clinical overlap has driven decades of debate about whether they are distinct diseases or points on a severity spectrum.
Mass spectrometry of 2,083 cerebrospinal fluid proteins found no significant differences between ME/CFS patients with and without comorbid fibromyalgia (AUC 0.67 β barely above chance). The authors concluded both conditions βfall along a common illness spectrum.β
This is the strongest lumping evidence available: the central nervous system milieu β sampled directly via CSF rather than inferred from peripheral blood β is indistinguishable between ME/CFS and ME/CFS+FM. If the CSF protein composition is identical, the argument that they are distinct brain diseases weakens substantially.
Consequence: If CSF proteome identity holds in larger replication, the clinical distinction between ME/CFS and fibromyalgia is a distinction of symptom thresholds, not of underlying biology β a severity-level split rather than a disease-level split.
Severity applicability: Unknown β CSF studies recruit ambulatory patients; lumbar puncture in severe/bedbound patients is rare.
(Certainty: 0.50.) ME/CFS, fibromyalgia, and IBS share genetic architecture β rg=0.75 between ME/CFS and IBS β but Hirsch 2025 found different specific DNA signals at shared loci. This is βallelic heterogeneity within shared biological pathwaysβ β the same genes are involved, but different causal variants and likely different tissue-specific regulatory mechanisms.
Three competing mechanistic models explain the shared heritability: (1) glutamatergic (shared synaptic dysfunction), (2) serotonergic (shared neurotransmitter dysregulation), and (3) autonomic (shared cardiovascular/sympathetic dysfunction). Local genetic correlation analysis can distinguish these β if the glutamatergic signal persists after controlling for autonomic, the shared biology is synaptic; if the autonomic signal persists after controlling for glutamatergic, the shared biology is cardiovascular.
Falsifiable prediction: Local genetic correlation (LAVA or Ο-HESS) at glutamatergic gene loci will exceed the genome-wide average for ME/CFS-FM-IBS, confirming neuronal sharing; at autonomic receptor loci, it will exceed for ME/CFS-POTS but not for IBS, confirming condition-specific regulatory divergence.
Consequence: βSame genes, different regulationβ is a sophisticated middle ground between lumping and splitting β it acknowledges shared biology while explaining why different conditions manifest differently. It also predicts that drugs targeting shared pathways (glutamatergic modulators) may benefit multiple conditions, while drugs targeting condition-specific regulatory nodes (autonomic modulators) will show condition-specific efficacy.
(Certainty: 0.50.) PEM status, not the diagnostic label (ME/CFS vs fibromyalgia), is the critical variable determining exercise response.
Depression (exercise beneficial) β PEM-negative FM (graded exercise neutral to mild benefit) β PEM-positive FM and ME/CFS (exercise harmful) β severe ME/CFS (exercise harmful at any dose). This spectrum is supported by the BRANDO framework: unblinded exercise trials (regardless of condition) consistently show subjective improvement, while objective outcomes (CPET, actigraphy) show no change or decline β suggesting that the treatment effect in positive trials is unblinding bias, not biological benefit.
Gulf War illness, Long COVID, and fibromyalgia all share the exercise intolerance profile with ME/CFS, each with identical blinding vulnerabilities and psychosomatic framing. The nosological implication: exercise response is not a disease-specific feature β it is a cross-condition dimension that is currently binned into different disease labels. Reclassifying patients by PEM status rather than by diagnosed condition would produce more homogeneous treatment-relevant subgroups than the current disease categories.
Falsifiable prediction: A trial stratifying patients by PEM status (2-day CPET deterioration) rather than by diagnosis will show a larger treatment Γ subgroup interaction for exercise-based interventions than a trial stratifying by diagnosis. If the diagnosis-stratified trial shows equal or larger interaction, the PEM-status hypothesis is weakened.
Consequence: If PEM status is the clinically important variable, treatment trials for ME/CFS, fibromyalgia, Long COVID, and GWI should all stratify by PEM rather than by diagnosis β a cross-condition reorganization of research methodology that would shift funding from disease-specific to mechanism-specific clinical trials.