Epilepsy is a common affliction that involves inflammatory processes. There are currently no definitive chemical diagnostic biomarkers in the blood, so diagnosis is based on a sometimes expensive synthesis of clinical observation, radiology, neuro-psychological testing, and interictal and ictal EEG studies. Soluble ICAM5 (sICAM5), also known as telencephalin, is an anti-inflammatory protein of strictly central nervous system tissue origin that is also found in blood. Here we have tested the hypothesis that plasma concentrations of select inflammatory cytokines, including sICAM5, might serve as biomarkers for epilepsy diagnosis. To test this hypothesis, we developed a highly sensitive and accurate electrochemiluminescent ELISA assay to measure sICAM5 levels, and measured levels of sICAM5 and 18 other inflammatory mediators in epilepsy patient plasma and controls. Patient samples were drawn from in-patients undergoing video-EEG monitoring, without regard to timing of seizures. Differences were defined by t-test, and Receiver Operating Condition (ROC) curves determined the ability of these tests to distinguish between the two populations. In epilepsy patient plasmas, we found that concentrations of anti-inflammatory sICAM5 are reduced (p = 0.002) and pro-inflammatory IL-1β, IL-2, and IL-8 are elevated. TARC (thymus and activation regulated chemokine, CCL17) concentrations trend high. In contrast, levels of BDNF and a variety of other pro-inflammatory mediators are not altered. Based on p-value and ROC analysis, we find that the ratio of TARC/sICAM5 discriminates accurately between patients and controls, with an ROC Area Under the Curve (AUC) of 1.0 (p = 0.034). In conclusion, we find that the ratio of TARC to sICAM5 accurately distinguishes between the two populations and provides a statistically and mechanistically compelling candidate blood biomarker for drug resistant epilepsy.