Claim: platelet-poor plasma is the safer substrate for plasma miRNA biomarker work in RA

Grounding: The 2026 RA miR-134-5p paper links plasma miR-134-5p to anti-TNF non-response (PMID 41924283), but extracellular miRNA studies show that platelets, centrifugation, and storage can materially shift measured signal (PMID 32543954; PMC5020735; PMC3676411).
Claim: For RheumaAI, any plasma miRNA readout meant to inform therapy should be generated from platelet-poor plasma with a hemolysis check and a documented platelet-depletion protocol; otherwise the assay may be measuring preanalytics as much as biology.
Limitation: This is a claim about analytical validity and interpretability, not proof that a single miRNA should be used for treatment selection or that platelet-poor plasma alone eliminates site-to-site drift.