Aim: To investigate the possible effects of pentoxifylline metabolites on retinal blood flow in humans.
Methods: A randomized, placebo-controlled, four-period cross-over study that was observer blinded and partly blinded for the eight participants. On one occasion a placebo was given as an intravenous (i.v.) infusion over 100 min. On the other three occasions pentoxifylline was administered as i.v. infusions over 100 min at a rate of 3 mg min(-1). Before two of the pentoxifylline infusions the subjects were pretreated with either ciprofloxacin or rifampicin. Retinal blood flow was measured by scanning laser doppler flowmetry (SLDF) in a selected area of the central temporal retina before, during and until 5 h after the end of infusion. Blood samples for concentration analyses of pentoxifyllin, R-M1, S-M1, M4 and M5 were taken serially and areas under the curves (AUCs) were calculated. Linear mixed models were used for the statistical analyses.
Results: Mean AUCs (ng h ml(-1)) were significantly increased for pentoxifylline (1964 vs. 1453) and S-M1 (5804 vs. 4227), but not R-M1 when pentoxifylline was co-administered with ciprofloxacin. The mean AUC for M5 was significantly reduced when subjects were pretreated with rifampicin (2041 vs. 3080). Pentoxifylline with and without pretreatment with rifampicin significantly increased retinal blood flow assessed as mean flow, pulsation (i.e. 1-systole/diastole), and diastolic flow (but not during systole), compared with placebo. The increases over placebo were more pronounced on diastolic flow, 9.7% (95% confidence interval 4.2, 15.5) than on mean flow, 4.6% (1.1, 8.3) after pentoxifylline administration. With pentoxifylline after rifampicin pretreatment the corresponding differences were 11.7% (5.8, 17.9) and 5.1% (1.4, 7.8) over placebo, respectively. After co-administration of pentoxifylline and ciprofloxacin we saw only a nonsignificant trend towards increased flow during diastole, but a significant decrease in pulsation. When AUCs for pentoxifylline and its metabolites were used as regressor variables to retinal mean flow we found that pentoxifylline, R-M1 and M5 had coefficients with a positive sign indicating that they enhanced the retinal blood flow. In contrast, S-M1 and M4 had coefficients with negative sign and thus appeared to decrease the blood flow in subjects treated with pentoxifylline.
Conclusion: The R-M1 and M5 metabolites of pentoxifylline contributed significantly to the effects of pentoxifylline on retinal blood flow.