Since the pharmacological profiles of various non-steroidal anti-inflammatory drugs (NSAIDs) might depend on their differing selectivity for cyclooxygenase 1 (COX-1) and 2 (COX-2), we developed a new screening method using human peripheral monocytes. Monocytes from healthy volunteers were separated, and the cells were incubated with or without lipopolysaccharide (LPS). Monocytes without LPS stimulation exclusively expressed COX-1 on Western blotting analysis, whereas LPS stimulation induced COX-2 expression. Unstimulated monocytes (COX-1) and LPS-stimulated monocytes (COX-2) were then used to determinethe COX selectivity of various NSAIDs. The respective mean IC50 values for COX-1 and COX-2 IC50 (microM), and the COX-1/COX-2 ratio of each NSAID were as follows: celecoxib, 82, 6.8, 12; diclofenac, 0.076, 0.026, 2.9; etodolac, > 100, 53, > 1.9; ibuprofen, 12, 80, 0.15; indometacin, 0.0090, 0.31, 0.029; meloxicam, 37, 6.1, 6.1; 6-MNA (the active metabolite of nabumetone), 149, 230, 0.65; NS-398, 125, 5.6, 22; piroxicam, 47, 25, 1.9; rofecoxib, > 100, 25, > 4.0; S-2474, > 100, 8.9, > 11; SC-560, 0.0048, 1.4, 0.0034. The percentage inhibition of COX-1 activity at the IC50 of COX-2 also showed a wide variation among these NSAIDs. The bioassay system using human monocytes to assess the inhibitory effects of various NSAIDs on COX-1 and COX-2 may become a clinically useful screening method.