• prevention of HFS
  

• treatment of HFS
  

Although pyridoxine has been used empirically for the prevention of capecitabine associated HFS, its efficacy needs to be demonstrated in prospective controlled trials. We estimated that the HFS rate with placebo and pyridoxine would be 0.35 and 0.18, respectively, and we therefore calculated that a sample size of 345 patients would be necessary to detect these hazard rates with an 80% power (β=0.2) and two-sided significance level of α=0.05. We assumed a follow up loss rate of 10%, thus requiring 380 patients to be randomized. Chemotherapy-naive patients with gastrointestinal tract cancers who were scheduled for capecitabine-containing chemotherapy were assigned to receive oral pyridoxine or placebo in randomized double-blind placebo controlled study. Pyridoxine 100 mg b.i.d was prescribed to the patients in the pyridoxine group, identical placebo 100 mg b.i.d was prescribed in the placebo group by the closed envelop randomization. Patients were stratified by chemotherapy regimen: capecitabine alone (X), capecitabine and cisplatin (XP), or docetaxel, capecitabine, and cisplatin (DXP). Patients were observed until NCI CTC grade 2 or 3 HFS developed or capecitabine-containing chemotherapy ended. Patients in the placebo group who developed grade 2 or 3 HFS were randomized to receive pyridoxine or placebo for the next chemotherapy cycle to determine whether pyridoxine could improve HFS, and the same treatment was continued for 2 chemotherapy cycles.