We evaluated the effects of weekly short infusions of paclitaxel (PAC) on the development of a peripheral neuropathy (PNP) as primary endpoint. Patients with advanced cancer were randomized to a weekly regimen of PAC (100 mg/m2) infused over 1 versus 3 h. PNP was evaluated by a clinical score including sensory symptoms, strength, tendon reflexes and vibratory sense (range 0-12; PNP >3 points). Kaplan-Meier-type curves were calculated. In total, 22 study centers enrolled 121 patients, 92 assessable for analysis. The probability to exceed a PNP score of 3 increased from 0.20 versus 0.30 after six to 0.68 versus 0.47 after 12 administrations (1 versus 3 h: p = 0.66). After 12 weeks of therapy only a quarter of assessable patients were free of PNP. Cox analysis yielded a relative risk of 1.10 for 1-h infusions (p = 0.80). We observed a rapid increasing risk of PNP manifestation in the course of weekly PAC administrations without significant differences between 1- and 3-h infusions. This is in contrast to pharmacokinetic observations indicating that a shortening of infusion time might enhance neurotoxicity by increasing the AUC of Cremophor. A majority of patients experiencing neurotoxic effects require the investigation of potential nerve protectors in future clinical trials accompanying PAC therapy.