OBJECTIVE Paclitaxel is most commonly infused over 3 hs rather than the original schedule of 24 h as the briefer infusion duration results in greater convenience, similar efficacy, significantly less myelosuppression, and less cost. While differences in toxicity between 3- and 24-h infusions are well described, there is little information about the effect of modest prolongation of infusion duration, which is often employed in patients who develop hypersensitivity reactions. To assess whether prolonging a 3-h infusion significantly increases the degree of neutropenia, we reviewed our data from a randomized, crossover trial of 3-h versus 6-h versus 24-h regimens of paclitaxel. METHODS Results from 12 patients who were randomized to receive one 3-h, one 6-h and one 24-h infusion of paclitaxel in varied sequences during their first three cycles of treatment were analysed. The blood counts were monitored closely throughout each cycle of treatment. RESULTS Crossover trial methodology was used to assess the differences in the degree of neutropenia caused by the three different infusion regimens. The 24-h infusion regimen resulted in significantly worse neutropenia than the 3- or 6-h infusion regimens. There was no statistically significant difference between the 3- and 6-h infusion regimens with respect to all endpoints. The estimated mean difference in the duration of grade 3 or 4 neutropenia between the 3- and 6-h infusion regimens (6 h - 3-h) was 1.1 day (95% CI: -0.9, 3.2), and for grade 4 neutropenia, the estimated mean difference in the duration was 0.8 day (95% CI: -0.4, 2.0). CONCLUSIONS Increasing the duration of paclitaxel infusions from 3 to 6 h does not result in a statistically significant increase in the degree of neutropenia. Any additional neutropenia is likely to be of brief duration.