BACKGROUND Fluorouracil (5-FU) has remained the mainstay of treatment of advanced colorectal cancer disease for nearly 40 years, and despite the implementation of various strategies to increase response rates no substantial improvement in survival has been achieved. 5-FU efficacy has been enhanced by modulating its cytotoxicity with leucovorin (LV) or by administering it as a continuous intravenous infusion. These regimens resulted in a few-fold improvement of tumour response rate in patients with metastatic disease compared with standard 5-FU treatment. The figures are still low (25%), and survival is affected only modestly [1, 2]. Cisplatin and carboplatin are also able to modulate 5-FU cytotoxicity in experimental systems [3]. Recent results of experimental studies suggested that high doses of carboplatin were necessary to achieve biochemical modulation of 5-FU cytotoxicity in vivo [6]. The use of chronobiologically determined drug over a 24-hour period may allow an increase in antitumour effect and a decrease in side effects. Experiments in animals revealed large changes in the toxic effects of 5-FU and platinum analogues, depending on the circadian rhythm of drug administration [9, 10]. The aim of the study was to investigate the toxicity and efficacy of combination of high dose carboplatin, fluorouracil and leucovorin in patients with advanced colorectal adenocarcinoma. Carboplatin and 5-FU were administered in circadian-dependent rhythm regimen in a 4-h infusion. METHODS An open, prospective study was carried out in 99 patients enrolled in the study. The following treatment schedule was used: carboplatin 150 mg, a 4-hour-infusion, start at 8 a.m., day 1-7; 5-FU 750 mg/m2, a 4-hour-infusion, start at 6 p.m., day 1-5; leucovorin 100 mg/m2, bolus i.v. at 8 p.m., day 1-5. The cycles were repeated every 28 days. Dose modifications were not planned. Treatment response and toxicity were evaluated according to WHO criteria [11]. Time to progression was calculated from the beginning of chemotherapy. Survival was calculated from start of chemotherapy using the Kaplan-Meier method [12]. RESULTS We treated 99 patients (27 colon and 72 rectal cancers) with metastatic diseases. There were 52 males and 47 females, average age 58 years (range 34-72). There were 82 patients (pts) with previous surgical treatment, 30 pts. with previous radiotherapy and 19 pts. with previous adjuvant chemotherapy. Performance status was 0 for 9 pts., 1 for 53 pts., 2 for 33 pts. and 3 for 4 pts. We carried out 469 cycles (mean 4/pts.). Ninety seven pts. were evaluated. The response rate was 28% (95% CI: 19.9-37.9) with 4/97CR, 23/97PR, 48/975D and 22/97PD. Mean time to progression was 5.5 months (range 2-18). The mean survival time was 7 months (range 4-27). The mean survival time was 11 months for responders and 6 months for non-responders. WHO grades 3 and 4 leukopenia were observed in 19 pts (19.2%). WHO grades 3 and 4 thrombocytopenia occurred in 21 (21.2%) pts. The main nonhaematologic toxicity were diarrhea, nausea/vomiting and mucositis. No grades 4 of nonhaematological toxicity were seen, except diarrhea grade 4 was observed in one patient. There was no life-threatening toxicity. CONCLUSIONS Carboplatin is able to modulate 5-FU cytotoxicities [3]. A high dose of carboplatin is necessary to achieve biochemical modulation of 5-FU cytotoxicities in vivo [6]. The studies of colon tumour cell lines demonstrated sensitivities to carboplatin when used at clinically achievable dose level [4]. Clinical studies of carboplatin, as a single agent or in combination, in the treatment of colorectal cancer, reported controversial results [19, 20]. Animal studies indicated large and predictable changes in the toxic effects of carboplatin and 5-FU, depending on the circadian rhythm of drug administration. Mechanisms included 24 h changes in the activities of several enzymes involved in 5-FU and carboplatin catabolism or in the anabolism of its cyt