The aim of our study was to determine the best combination of clinical factors in predicting post-concussion syndrome (PCS) â€“ assessed with the British Columbia Post-Concussion Symptom Inventory (BC-PSI) â€“ at 3 months following mild TBI. Patients with mTBI (16-60 years) were defined as Glasgow Coma Scale (GCS) score between 13-15, loss of consciousness (LOC)<30 min and post-traumatic amnesia (PTA)<24 hours. Clinical/demographic factors were age, sex, GCS, LOC, PTA, traumatic intracranial finding on MRI (within 72 h), other somatic injuries, plasma-derived proteins associated with CNS damage (Tau, GFAP and NFL) and inflammation (IFNï§, IL-8, Eotaxin, MIP-1ï¢, MCP-1, IP-10, IL-17A, IL-9, TNF, FGF-basic PDGF and IL-1ra). Blood was drawn 24-72 hours, 2 weeks, and 3 months after injury. The discriminatory value (AUC analysis) of each individual blood biomarker predictor was at chance level. For each timepoint, combining all blood biomarkers also did not predict PCS at clinically relevant levels (AUCs 0.66-0.71). At 24-72 hours, the blood biomarkers included in the best-subset model predicting PCS were TNF, IL-9, IL-17A and IL-8. At both 2-weeks and 3 weeks, a combination of PDGF, IL-8 and IFNï§, best predicted PCS. The baseline clinical injuries predicting PCS together with 24-72 hours blood were woman, longer PTA, traumatic intracranial findings, lower GCS, and with 2-week and 3-month blood: woman, longer PTA, and traumatic intracranial findings. In conclusion, our study shows that blood biomarker level in both acute, subacute and chronic phase of MTBI were associated with PCS, but these biomarkers did not have clinical specificity/sensitivity to reliably predict PCS.