An improved method is reported for measurement of myeloperoxidase (MPO) activity in tissues. While spectrophotometric methods based on oxidation of O-dianisidine or other dyes have been reported for MPO measurement in pure polymorphonuclear leukocytes (PMNs), these methods often fail to accurately assay MPO activity in tissues. We observe that tissue myoglobin or vascular hemoglobin markedly effects the spectrophotometric assay for MPO. Under optimal conditions of 0.53 mM O-dianisidine, 0.15 mM H2O2, pH 6.0, either myoglobin or hemoglobin produced absorbance at 460 nm in a concentration-dependent manner similar to that of MPO. In perfused heart tissue, myoglobin caused a major problem with the assay resulting in an inability to obtain accurate linear results as a function of MPO concentration and PMN number. To eliminate the effect of tissue myoglobin or vascular hemoglobin on the assay, one-step gel filtration chromatography of tissue extracts was introduced. MPO, myoglobin, and hemoglobin were easily separated using a Sephadex G-75 column according to the difference in their molecular weights. A linear relationship between the MPO activity and PMN number was observed only after processing the tissue extracts through the Sephadex G-75 column. Thus, MPO activity in PMN-containing tissues can be precisely quantitated after one-step purification on a molecular exclusion column. Enzyme purification with removal of myoglobin is essential for obtaining accurate measurement of MPO activity and quantitation of PMNs in muscle tissue.