We report the results of a randomised phase II trial of docetaxel tested as a single agent in patients with recurrent head and neck cancer using methotrexate as a control arm to validate the results. Eligibility criteria included: histologically-confirmed squamous cell carcinoma, measurable disease, adequate haematological, renal and hepatic functions, no prior chemotherapy for recurrent cancer, signed informed consent. 40 mg/m2 methotrexate was given as a short weekly bolus i.v. injection, and 40 mg/m2 docetaxel was administered as a one hour weekly infusion. A total of 57 patients were randomised based on a ratio of 2/1:37 and 20 patients received docetaxel and methotrexate, respectively. Patient characteristics included 49 males and 8 females; the median age was 59 years (range: 43-82 years). Twenty-eight patients had a local-regional relapse and 29 had distant metastasis, the median disease-free interval was 7.9 months (range: 0-165 months). For patients treated with docetaxel, the following grade 3-4 toxicities occurred: neutropenia (12.5%) with febrile neutropenia in one patient (1%), anaemia (19%) mucositis (9%) and ungueal toxicity (9%). In the methotrexate arm, the grade 3-4 toxicities were: anaemia (15%) and mucositis (5%). The response rate was significantly higher in the docetaxel arm with 27% (95% confidence interval (CI): 21.7-32.3%) of objective responses versus 15% (95% CI: 11.2-18.8%) in the methotrexate arm. Overall survival and time to progression were super-imposable between the docetaxel and methotrexate treatments. Docetaxel given as a weekly infusion has a high activity in patients with head and neck cancer. A phase III trial is needed to test if this translates into a survival benefit for docetaxel use.