The purpose of the study was to compare the results of sepsis scoring (clinical examination and clinical pathology) to the concentrations of matrix-metalloproteinases (MMPs) -2, -8, and -9; tissue-inhibitor of metalloproteinases (TIMPs) -1 and -2; and inflammatory chemokines interleukin (IL) 1 beta and tumor-necrosis-factor-alpha (TNF-alpha) in plasma and peritoneal fluid of equine colic patients. A modified sepsis scoring including general condition, heart and respiratory rate, rectal temperature, mucous membranes, white blood cell count (WBC), and ionized calcium was applied in 47 horses presented with clinical signs of colic. Using this scoring system, horses were classified as negative (n=32, <= 6/19 points), questionable (n=9, 7-9/19 points), or positive (n=6, >= 10/19 points) for sepsis. MMPs, TIMPs, IL-1 beta, and TNF-alpha concentrations were evaluated in plasma and peritoneal fluid using species-specific sandwich ELISA kits. In a linear discriminant analysis, all parameters of sepsis scoring apart from calcium separated well between sepsis severity groups (P<0.05). MMP-9 was the only biomarker of high diagnostic value, while all others remained insignificant. A significant influence of overall sepsis scoring on MMP-9 was found for peritoneal fluid (P=0.005) with a regression coefficient of 0.092, while no association was found for plasma (P=0.085). Using a MMP-9 concentration of >113 ng/ml in the peritoneal fluid was found to be the ideal cutoff to identify positive sepsis scoring (>= 10/19 points; sensitivity of 83.3% and specificity of 82.9%). In conclusion, MMP-9 was found to be a biomarker of high diagnostic value for sepsis and endotoxemia in equine colic. The evaluation of peritoneal fluid seems preferable in comparison to plasma. As abdominocentesis is commonly performed in the diagnostic work-up of equine colic, a pen-side assay would be useful and easy-to-perform diagnostic support in the decision for therapeutic intervention.