Gut Microbiome and Plasma Proteomics Yield an 85% Diagnostic Model for Fibromyalgia

Fibromyalgia remains the diagnosis you make by exclusion. No blood test. No imaging finding. No biopsy. The 2010/2011 ACR criteria rely entirely on symptom self-report — the Widespread Pain Index and the Symptom Severity Scale. This is clinically functional, but it leaves FM uniquely vulnerable to the credibility problem: patients are told their suffering is not real because nothing shows up on a lab report. Duran-Gonzalez and colleagues at Pronacera (Sevilla) decided to look where most FM research has not — simultaneously in the gut and the blood.

What the data architecture looks like

The study began with 892 women and narrowed to 242 after exclusion criteria — a rigorous filtering that strengthens the final cohort. Each participant contributed both blood (for nano-LC-MS/MS plasma proteomics) and feces (for 16S rRNA V3-V4 amplicon sequencing). This dual-sample design is the study's core strength: instead of looking at one biological layer and speculating about the other, they measured both.

The proteomics identified 30 differentially expressed proteins. The most dramatically overexpressed was zinc-alpha-2-glycoprotein (ZA2G, ratio 2.05, p=0.013) — a fat-mobilizing protein linked to cachexia and energy metabolism. Dopamine beta-hydroxylase (DOPO, ratio 1.78, p=0.019) converts dopamine to norepinephrine; its elevation in FM plasma fits the known catecholaminergic dysregulation in chronic pain. The underexpressed proteins include glyceraldehyde-3-phosphate dehydrogenase (GAPDH) — a glycolytic enzyme whose depletion signals mitochondrial metabolic blockade.

The metagenomics found 19 differentially abundant bacterial taxa across 6,030 zero-radius OTUs. Alpha diversity was significantly lower in FM patients on both Shannon and Simpson indices — a finding consistent with intestinal dysbiosis seen in other chronic pain conditions. The functional prediction via PICRUSt revealed that glycolysis/gluconeogenesis and pyruvate metabolism were the most affected pathways. The enzymatic activities driving this — hydroxymethylglutaryl-CoA reductase, GAPDH (NADP+), glutathione reductase — all point toward compromised mitochondrial function and antioxidant capacity.

The lead actor in the microbial cast is Streptococcus salivarius. This oral-origin bacterium was the primary contributor to all five significantly enriched enzymatic activities in FM patients. Its overgrowth in the gut suggests oral-gut translocation — a phenomenon increasingly recognized in systemic inflammatory conditions.

The diagnostic model

When the researchers built separate machine learning classifiers (deep neural networks) for each omics layer, metagenomics achieved AUC=0.78 and proteomics AUC=0.77. Neither alone is clinically sufficient. But the integrated multi-omics model — combining significant microbial taxa and proteins into a single classifier — reached AUC=0.85 with standard deviation of 0.07 across cross-validation folds. This is not perfect, but it is the strongest objective discriminator for FM published to date.

The complementarity matters. The two data layers carry non-redundant information. Plasma proteins reflect systemic metabolic and inflammatory state; gut microbiome composition reflects the intestinal environment that modulates neural signaling via the vagus nerve and HPA axis. Together they capture a biological signature that neither alone can.

What this means at the bedside

This is not yet a clinical test. The model needs external validation in independent cohorts, prospective design, and standardization of the multi-omics pipeline. The sample is all-female, reflecting FM epidemiology (89% female) but limiting generalizability. The control group (N=43) is small relative to the patient group (N=199), which can inflate discriminative performance.

But the conceptual advance is significant. FM has been stuck in a diagnostic paradigm that relies on the patient's ability to articulate diffuse symptoms to a clinician who believes them. A lab-based biomarker panel — even an imperfect one — shifts that dynamic. It moves FM toward the same diagnostic legitimacy that rheumatoid arthritis gained when rheumatoid factor became a standard test, even though RF is neither perfectly sensitive nor specific.

For practitioners working with FM patients now: the gut-brain-mitochondrial axis framework proposed here aligns with emerging evidence that probiotics, dietary interventions targeting gut permeability, and mitochondrial support (CoQ10, magnesium) show modest benefit in FM. This study does not prove causation, but it provides a mechanistic scaffold for interventions you may already be considering.