Pentoxifylline has been claimed to work a beneficial effect in arterial insufficiency by improving erythrocyte deformability and thus improving blood flow. A number of observations, including the drug concentrations required to work the red cell effect, suggested that this was not likely to be a complete explanation. We therefore examined the effect of pentoxifylline on several granulocyte and platelet functions. Pentoxifylline inhibited platelet aggregation in response to 4 mumol/L adenosine diphosphate; although statistically significant inhibition was seen at 1 mumol/L pentoxifylline, over 200 mumol/L was required for 50% inhibition. The adherence of unstimulated platelets to cultured endothelial cells was not strongly inhibited by pentoxifylline; however, the additional increment in adherence seen in the presence of thrombin was strongly inhibited (50% attenuative dose [AD50] = 18 mumol/L). Granulocyte aggregation in response to C5a was modestly inhibited (AD30 approximately equal to 8 mumol/L; AD50 greater than 1 mmol/L), and the adherence of unstimulated polymorphonuclear neutrophils (PMNs) to endothelium was uninhibited. The C5a-mediated augmentation of PMN adherence to endothelium was mildly inhibited (AD50 = 240 mumol/L). Inhibition of PMN chemotaxis to N-Formyl-methionyl-leucyl-phenylalanine (FMLP) or C5a (AD50 = 12 mumol/L) and inhibition of superoxide production in response to FMLP-cytochalasin B (AD50 = 24 mumol/L) were seen at more clinically credible concentrations. Perhaps most important, pentoxifylline blocked the ability of platelet activation factor to prime neutrophils for enhanced response to subsequent stimuli (AD50 approximately equal to 8 mumol/L; AD60 = 10 mumol/L when production was the indicator system); in vivo, this could broaden the drug's effect to include functions that it does not inhibit potently in a primary fashion. Although pentoxifylline is known to be a phosphodiesterase inhibitor, and we found it to elevate intracellular cyclic adenosine monophosphate in stimulated PMNs, we found it to be only marginally more potent than theophylline in this regard; therefore, the failure of theophylline to inhibit PMN priming suggests that this enzyme inhibition is not a complete explanation of the pharmacologic action of pentoxifylline. We suggest that the effects of pentoxifylline on platelet and granulocyte function are likely to contribute to the drug's clinical efficacy.