Cyclo-oxygenase (COX) inhibitors may have a role in reducing inflammation in asthma and other pulmonary diseases. COX inhibitors have different selectivities for the two COX isoenzymes (COX 1 and COX 2) which vary between purified enzyme and intact cell preparations. The relative selectivity of COX inhibitors has not been studied in human airway cells. A number of COX inhibitors in cultured human airway cells were compared which exclusively express either COX 1 (primary degree cultured human airway smooth muscle (HASM) cells) or COX 2 (A549 pulmonary epithelial cell-line) as measured by Western blotting. COX activity was assayed by prostaglandin (PG)E2 production following 30 min incubation with 5 mM arachidonic acid. COX activity in both cell types was similar; HASM cells 92.2+/-12.1 ng PGE2 x mg-1 protein, A549 cells 87.7+/-24.4 ng PGE2 mg-1 protein. In HASM cells the median inhibitory concentration (IC50) was >10-5 M for nimesulide, 3.2 x 10-6 M for N-(2-cyclohexyloxy-4-nitrophenyl)-methanesulphonamide (NS398), 1.8 x 10-8 M for flurbiprofen, 6.7 x 10-9 M for indomethacin and >10-5 M for aspirin. In A549 cells the IC50 was 1.8 x 10-9M for nimesulide, 4.1 x 10-9 M for NS398,6.2 x 10-10 M for flurbiprofen, 1.3 x 10-8 M for indomethacin and >10-5 M for aspirin. Sodium valerate had no effect in either HASM or A549 cells. The COX 2:COX 1 selectivity ratio (COX 2 IC50/COX I IC50) was <0.0001 for nimesulide, 0.001 for NS398, 0.03 for flurbiprofen and 1.9 for indomethacin. In conclusion the present study has shown that cyclo-oxygenase inhibitors have a range of selectivities for cyclo-oxygenase 1 and cyclo-oxygenase 2 in intact human airway cells. The relative cyclo-oxygenase 2 selectivity of N-(2-cyclohexyloxy-4-nitrophenyl)-methanesulphonamide and nimesulide may have implications for the treatment of asthma and other inflammatory pulmonary diseases.