Circadian (about 24 h) changes modulate cellular metabolism and proliferation in healthy tissues. As a result, the extent of toxicity of cancer chemotherapy varies by 50% or more according to dosing time in mice or rats. These experimental data led to test the principle of adapting chemotherapy delivery to circadian rhythms--so-called "chronotherapy"--in cancer patients. Clinical trials first aimed to assess toxicities and maximal tolerated dose of chronotherapy schedules. Programmable-in-time pumps became the indispensable tool for reliably testing this principle in a large enough patient population and without any need for hospitalization. Results from randomized clinical trials have indicated that chronotherapy allowed to increase dose intensity (DI, a secondary criterion in these trials) by 13 to 45% as compared to injecting the drug at another time (doxorubicin, cisplatin) or according to a flat rate (5-fluorouracil, floxuridine, oxaliplatin). A non-randomized trial further suggested this principle would also apply to cytokine delivery (interferon alpha). Two consecutive randomized trials have revealed that chronotherapy-related increase in 5-FU DI was further associated with a 20% increase in response rate of patients with metastatic colorectal cancer to the 3-drug association that was used (5-FU, folinic acid and oxaliplatin). In a subsequent phase II trial 5-FU daily dose and 3-drug treatment frequency were increased in 50 patients with this disease. Such intrapatient dose escalation protocol also disclosed a positive correlation between DI of 5-FU and response. New anticancer drugs should benefit from this approach so that both their clinical safety and patient's quality of life are improved, through ambulatory administration.