A panel of non-steroidal anti-inflammatory drugs commonly used for therapeutic purposes was assessed for their effects on the respiratory burst of isolated human polymorphonuclear neutrophils. Cells were stimulated with opsonised yeast and the production of reactive oxygen species was measured by amplified chemiluminescence with luminol and lucigenin which are two luminogenic agents measuring different cellular events. A special attention was devoted to the establishment of dose-effect curves and calculation of ED50. Some of the drugs tested (acemetacine, diclofenac, flufenamic acid and niflumic acid) were able to decrease both luminol and lucigenin chemiluminescence in a dose-dependent manner reflecting an inhibitory effect on the respiratory burst. The most potent derivative was flufenamic acid (ED50 8 and 78 microM, respectively, with luminol and lucigenin), followed by diclofenac (21 and 98 microM), niflumic acid (97 and 227 microM) and acemetacine (585 and 427 microM). In contrast, several other drugs (flurbiprofen, ibuprofen, ketoprofen, piroxicam) stimulated both luminol and lucigenin chemiluminescence, suggesting a pro-oxidant activity. Acetylsalicylic acid (up to 1250 microM) was a modest inhibitor (maximum 25% inhibition) showing no dose-dependent effect and tolmetin (up to 125 microM) had no significant effect in both systems. The results were in agreement using both luminogenic agents, except for indomethacin, naproxen and tenoxicam which showed different kinds of effects. The unspecific and complex nature of the measurement systems used did not allow to give a complete mechanistic interpretation of the results, but the comparison with literature data gave some pertinent explanations for both anti- and pro-oxidant effects.