Head and neck cancer is a major cause of cancer-related deaths. In general, early stage head and neck cancers are effectively treated with either radiation or surgery. More advanced tumors often require combined-modality therapy with both radiation therapy and surgery. Recent investigations indicate that the addition of chemotherapy may be helpful. One of the newer chemotherapy agents that appears to have significant activity against head and neck cancer is paclitaxel (Taxol; Bristol-Myers Squibb Company, Princeton, NJ). Paclitaxel, originally derived from the western yew Taxus brevifolia, acts by increasing the stability of microtubules and preventing mitosis. Recent evidence indicates that the microtubule system is vital to the release of various cytokines and that modulation of cytokine release may play a major role in the drug's antitumor activity. We report a phase II trial of paclitaxel in patients with head and neck cancer, not only to evaluate its clinical effects, but also to study its effect on cytokine release. We assessed interleukin-1 beta (IL-1 beta) and tumor necrosis factor-alpha production by using a sensitive enzyme-linked immunosorbent assay to assess the serum of patients receiving paclitaxel and to detect cytokine release in vitro. The objective response rate was 36%, with 12% complete responses and 24% partial responses. No IL-1 beta or tumor necrosis factor-alpha was detected in patient serum at any time during the infusion of paclitaxel or after overnight incubation with patient monocytes. No proIL-1 beta was detected in in vitro cultures of paclitaxel-treated patient monocytes. When monocytes were stimulated with endotoxin, IL-1 beta production was greatest at 48 hours, suggesting that paclitaxel can prime cells to produce greater quantities of cytokines after a second stimulus.