Weekly 5-fluorouracil and leucovorin: achieving lower toxicity with higher dose-intensity in adjuvant chemotherapy after colorectal cancer resection

Transcription

1 Original article Annals of Oncology 15: , 2004 DOI: /annonc/mdh134 Weekly 5-fluorouracil and leucovorin: achieving lower toxicity with higher dose-intensity in adjuvant chemotherapy after colorectal cancer resection K. Patel, D. A. Anthoney, A. M. Crellin, D. Sebag-Montefiore, J. Messruther & M. T. Seymour* Cancer Research UK Clinical Centre, Cookridge Hospital Leeds, Leeds LS16 6QB, UK Received 18 September 2003; revised 18 November 2003; accepted 22 December 2003 Background: Current standard therapy following resection of high-risk colon cancer is intravenous bolus 5-fluorouracil (5-FU) with leucovorin (LV), but there is no consensus on the optimum regimen of these drugs: practice ranges from the high toxicity Mayo Clinic schedule to the very low toxicity weekly QUASAR schedule. We present data for a weekly schedule that aims to provide moderately dose-intense treatment with low toxicity. Patients and methods: One hundred and sixty-two patients were studied: 60% male; median age 65 years (36% over 70 years); 94% colorectal primary. Treatment was intravenous bolus (5 min) 5-FU 425 mg/m 2 plus D,L-LV 45 mg flat rate once weekly, for a planned 24 weeks. Data for toxicity, dose-reductions, delays and stoppages were collected. Results: Overall, 20% of patients experienced any grade 3 toxicity, most commonly diarrhoea (14% patients). Dose reductions were made in 35% of patients (although only 21% required 20% or more reduction); toxicity contributed to a decision to stop treatment before 24 weeks in 16% of patients. Median delivered dose intensity (DI) was 96% of planned (407 mg/m 2 /week) during treatment, and 91% of planned (385 mg/m 2 /week) over the full 24 week treatment plan. Female sex and age >70 years were significantly associated with higher rates of toxicity and dose adjustment, and lower delivered DI. Conclusions: Weekly treatment at these doses is convenient and well-tolerated for the large majority of patients, and achieves DI comparable with the 5 days a month QUASAR schedule and other more toxic standard regimens. Key words: adjuvant, colorectal, dose intensity, fluorouracil, leucovorin, toxicity Introduction For over 10 years, systemic adjuvant 5-fluorouracil (5-FU)-based chemotherapy has been standard after resection of Dukes stage C colon cancer in fit patients, with controlled trials continuing in Dukes B disease, where its role is less well defined. Current and planned trials are evaluating new chemotherapy drugs such as oxaliplatin, irinotecan and capecitabine, and other novel agents including anti-angiogenic agents, tumour vaccines and COX2 inhibitors, which have the potential to substantially improve results in the future. One of the first of these, a trial adding oxaliplatin to a 5-FU regimen, reported early results in June 2003 showing a significant improvement in disease-free survival [1], and the results of several other trials of novel adjuvant therapy are expected within the next 2 years. It is to be hoped that these studies will change the outlook for fit, high-risk colon cancer patients. However, many thousands of patients meanwhile, and perhaps low-risk or less fit patients in the future, will continue to receive *Correspondence to: Dr M. T. Seymour, Cancer Research UK Clinical Centre, Cookridge Hospital, Leeds LS16 6QB, UK. Tel: ; Fax: ; Matt.Seymour@leedsth.nhs.uk adjuvant 5-FU therapy, so even now it remains important and relevant to optimise our delivery of this drug. In late 1999, in the light of data recently reported from the large UK randomised trial, QUASAR [2, 3], we reviewed our institution s policy for non-trial adjuvant chemotherapy and decided to adopt a weekly schedule of 5-FU with low-dose folinic acid. However, rather than using the low 5-FU dose (370 mg/m 2 ) and protracted schedule (30 weeks) that had been used in QUASAR, we opted instead to use a higher dose of 425 mg/m 2 5-FU, and to reduce the planned treatment duration to 24 weeks. We anticipated that this modification would allow us to deliver treatment at a higher dose intensity (DI), more comparable to other standard regimens, but to retain the low toxicity and good patient acceptability that had characterised the QUASAR trial. We now present toxicity and tolerability data for the first 162 patients treated with this regimen. Patients and methods Patients Since November 1999, patients at our institution requiring adjuvant 5-FU/ leucovorin (LV), if not entering a clinical trial and considered fit by their 2004 European Society for Medical Oncology

2 569 clinician for full-dose chemotherapy, have been prescribed once-weekly bolus 5-FU plus LV, for 24 weeks. All chemotherapy prescribing is performed using a computerised system (Chemocare version 3.1L; Clinisis Integrated Solutions, Belfast, UK), and a complete list of patients who had started weekly 5-FU/LV chemotherapy between November 1999 and February 2003 was obtained from the database for this study. These patients clinical notes were then examined. The most common indication for adjuvant therapy with this regimen was resected colon cancer, but patients receiving it after resection of other gastrointestinal primary cancers, or after resection of liver metastases, were also included in the analysis. Treatment All treatment was administered by an appropriately trained nurse in a nurse-led chemotherapy clinic. Racaemic calcium D,L-LV was given at a flat-rate dose of 45 mg, regardless of patient surface area, as a slow intravenous injection over 5 min. This was followed by 5-FU at 425 mg/m 2, also given over 5 min. Routine prophylactic antiemetics were not used, but all patients received supplies of p.r.n. oral metoclopramide, loperamide and a chlorhexidine mouthwash (Corsadyl ), with verbal and written instructions on their use. The total planned treatment duration was 24 weeks; doses omitted for toxicity or patient choice (e.g. holidays) during this period were not routinely added on at the end of the planned treatment period. Evaluation of toxicity and treatment modification Monitoring of toxicity was performed by the trained nursing staff, who also administered chemotherapy. Toxicity was routinely evaluated at each visit using the National Cancer Institute common toxicity criteria (version 2.0) grading system, and was summarised in the clinical notes on a pre-printed form every 6 weeks. For grade 2 toxicity, treatment was omitted until recovery and then restarted, either at the same dose or with a dose-reduction of 10 25% at the discretion of the clinician responsible. In some cases, rather than continuing treatment at a reduced dose, a decision was made to stop adjuvant treatment altogether; this was more likely if the indication for treatment was uncertain (i.e. Dukes B cancer or non-colon primary), or if the patient was elderly. Data collection Data for treatment prescribed were collected from the Chemocare database, and for toxicity and adverse events from the clinical notes. The number of weeks of treatment received overall and the percentage dose reduction of 5-FU was recorded, so that the proportion of planned dose administered could be calculated. Reasons for dose reductions, delays and premature termination of treatment were noted. Statistics Toxicity and dose reduction evaluation in subgroups was performed using the χ 2 statistic. Previous studies have used different methods of reporting delivered DI. Therefore, for comparison, we calculated delivered DI in two ways: (i) in mg/m 2 /week, taken from the date of the first chemotherapy injection until 1 week after the final dose administered; and (ii) in mg/m 2 /week, taken from the date of first chemotherapy injection until 24 weeks later, i.e. the end of the planned course. The latter method gives reduced DI in patients stopping treatment earlier than planned. Comparison of DI in subgroups was assessed using the non-parametric Mann Whitney test. Results Patient characteristics Between November 1999 and February 2003, 230 patients were identified as having started adjuvant 5-FU-based chemotherapy. Table 1. Patient characteristics Characteristic n (%) Total number 162 Sex [% (male:female)] 60:40 Age Median [years (range)] 65 (32 79) >70 years (%) 36 Preceding surgery Colon primary resection 109 (67) Dukes C 89 Dukes B 20 Rectal primary resection 28 (17) Dukes C 25 Dukes B 3 Pre-op RT 13 Pre-op CRT 7 Colorectal liver metastases resection 15 (9) Other GI primary sites 10 (6) Pancreas 5 Small intestine 5 RT, radiotherapy; CRT, chemoradiotherapy; GI, gastrointestinal. One hundred and sixty-two of these received weekly 5-FU at 425 mg/m 2 with LV (D,L; 45 mg), and have adequate toxicity records for inclusion in this study; characteristics of these patients are summarised in Table 1. Of the 68 patients not included in the analysis, 48 received different 5-FU dosing schedules, usually because of participation in a clinical trial; in 10 patients the recording of toxicity data was inadequate; and the other 10 patients started adjuvant chemotherapy but stopped early because of overt disease progression, and are therefore not assessable for toxicity or treatment delivery. Toxicity in all patients Treatment was generally well tolerated, with only 20% patients experiencing any grade 3 toxicity, and only 2% patients experiencing any grade 4 toxicity, at any time during treatment. Most common was diarrhoea, which occurred in 14% of patients. Grade 3 lethargy was reported in 7% of patients (Table 2). There were seven toxicity-related hospital admissions (4% of patients), five with diarrhoea and one with uncomplicated neutropenic fever. One elderly patient (age 74 years) was admitted with diarrhoea and then developed neutropenic fever. In this latter case, the patient also developed a deep venous thrombosis and, although recovering fully from the acute chemotherapy toxicity, subsequently died of a pulmonary embolism. Grade 1 and 2 toxicities were common, with mild or moderate lethargy reported in 77% of patients, nausea in 56% and diarrhoea in 48%. Skin and mucosal toxicity was also common, with hand

3 570 Table 2. Maximum toxicities per patient (National Cancer Institute common toxicity criteria version 2.0) (n = 162 patients) Grade 1 (%) Grade 2 (%) Grade 3 (%) Grade 4 (%) Diarrhoea Mucositis 17 1 Lethargy Nausea Hand foot Watery eyes Leucocytes Granulocytes Platelets 10 1 Figure 1. Timing of dose reductions. foot dermatitis in 39%, sore eyes in 35% and stomatitis in 18% of patients (Table 2). Angina occurred in one patient. Rectal cancer patients who had received pre-operative radiotherapy or chemoradiotherapy appeared to tolerate post-operative adjuvant chemotherapy just as well as those who had not received pre-operative treatment. However, patients aged 70 years experienced significantly more grade 3 diarrhoea than younger patients (P = 0.05), and significantly more grade 2 or 3 watering of the eyes (P = 0.001). Female patients experienced toxicity more frequently than male patients, with 21% of women experiencing grade 2 or 3 nausea compared with 8% of males (P = 0.02). Female patients were also more likely to experience grade 2 or 3 lethargy and hand foot dermatitis (P = 0.06 and 0.06, respectively). Dose reductions, stoppages and delays In all, 59 of 162 (36%) patients received the entire 24-week planned course without any toxicity-related dose reductions or delays. The dose of 5-FU was reduced at some point in 57 (35%) patients, although only 34 (21%) patients underwent dose reduction of 20% or more. Pre-operative radiotherapy or chemoradiotherapy did not result in more dose reductions in subsequent adjuvant chemotherapy (P = 0.52). Female patients required more dose reductions than males (P = 0.02). These dose reductions were made at any time from week 4 to 22 (median 10.5 weeks) (Figure 1). Treatment was stopped before the planned end-date in 32 (20%) patients. The reason for stopping was toxicity in 26 (16%) patients. Clinicians were more likely to recommend stopping rather than reducing treatment in elderly patients, hence 29% of patients 70 years old, but only 7% of patients <70 years, stopped treatment before 24 weeks (P = 0.001). The difference in stoppage rate between male and female patients did not reach significance (12% versus 23%; P = 0.08). Administration of chemotherapy was delayed for 1 week at some point in 70 (43%) patients, most commonly because of toxicity (35 patients, 22%) or holidays (18 patients, 11%). It was

4 571 Table 3. Delivered dose and DI Figure 2. Delivered dose intensity (DI). Upper line: delivered DI as percentage of planned DI (425 mg/m 2 /week) calculated from the start of treatment until 1 week after the last dose. Lower line: delivered DI as percentage of planned DI (10.2 g/m 2 /24 weeks) calculated from the start of treatment until 24 weeks later. not unit policy to routinely add missed treatments to the end of the course after the planned 24 weeks treatment duration, although this was done in approximately half of these cases, usually at the patient s request. DI of chemotherapy administered The regimen has a planned 5-FU DI of 425 mg/m 2 /week. Because of the relatively low rate of dose reductions and delays, the delivered DI calculated in the normal way, up to 1 week after stopping treatment, was median 96% of planned DI (407 mg/m 2 /week). The delivered DI exceeded 90% of planned DI in 67% of patients and exceeded 80% of planned DI in 90% of patients (Figure 2, upper line). If instead, DI is calculated over the full 24 weeks planned treatment course (10.2 g/m 2 /24 weeks), giving a reduced DI for patients who stop treatment early for any reason, the median delivered DI was still 91% of planned DI (9.23 g/m 2 /24 weeks, equivalent to 385 mg/m 2 /week) and exceeded 80% of planned DI in 71% patients (Figure 2, lower line). Delivered DI was significantly lower in patients with colon cancer than in those with rectal cancer (P = 0.01). As anticipated from the dose reduction and stoppage data (above), patients over the age of 70 years received less dose-intense treatment than those <70 years old (P = 0.01), and DI was lower in female patients than in males (P = 0.02) (Table 3). Discussion Currently, the best-recognised reference regimen in adjuvant colon cancer treatment is the Mayo Clinic schedule: 5-FU bolus 425 mg/m 2 plus LV 20 mg/m 2 daily for 5 days, repeated every 4 5 weeks for six cycles. This was superior to control in NCCTG Trial [4], and at least equivalent to three other regimens in the large US Intergroup trial INT-0089 [5]. It was also, when combined with levamisole, superior to 6 months of 5-FU and levamisole alone in NCCTG trial [6]. The Mayo Clinic All patients 91 Percentage of planned DI delivered over 24 weeks (median) Primary site 0.01 Colon 88 Rectum 98 Pre-op treatment (rectum) No pre-op treatment 96 Pre-op RT 100 Pre-op CRT 96 Age years 85 <70 years 95 Sex 0.02 Male 95 Female 85 DI, dose intensity; NS, not significant; RT, radiotherapy; CRT, chemoradiotherapy. schedule is recognised by licensing authorities as a reference regimen in adjuvant studies. The Machover regimen has, likewise, been validated against a non-treatment control [7]; like the Mayo regimen it is given for five consecutive days each month, but uses a lower dose of 5-FU ( mg/m 2 ) with higher dose LV (200 mg/m 2 ). The weekly Roswell Park regimen was superior to a previous regimen, MOF, in NSABP trial C-03; it involves 6 weekly treatments of 5-FU at 500 mg/m 2 with high-dose LV (500 mg/m 2 ), followed by 2 weeks without treatment, for six cycles (48 weeks) [8]. However, these standard regimens, although effective, produce rates of serious toxicity that have sometimes exceeded normal definitions of maximum tolerable dose. The Mayo schedule was established in a trial in which 24% of patients had grade 3 4 diarrhoea and 36% had grade 3 4 mucositis; by the time cycle 6 was administered, fewer than 30% patients were still on treatment and receiving 80% or more of the planned dose [4]. Subsequently, nearly 1000 patients received the same regimen in INT-0089, and the rates of grade 3 and 4 toxicity were again high: mucositis in 18%, diarrhoea in 21% and neutropenia in 24% [9]. A recent retrospective review by Tomiak et al. [10] highlighted the problems of administering the Mayo Clinic schedule in routine clinical practice, with 35% of patients not receiving even the second cycle of treatment as planned because of toxicity. On the other hand, a recent phase III trial reported lower rates of toxicity, although 70% of patients still required dose modification at some point during treatment [11]. The Machover regimen appears less toxic than the Mayo schedule; rates of reported toxicity varied widely in the three trials making up the 1995 IMPACT report [7], but in a recent trial, 26% of 453 patients treated with this regimen experienced grade 3 or 4 toxicity [12]. The Roswell Park schedule, in the P NS

5 572 NSABP C-03 trial, produced grade 3 or 4 diarrhoea in 28% patients [8]. Two recent trials have looked at infusional regimens as potentially less toxic and/or more active than bolus 5-FU/LV in the adjuvant setting. Both showed reduced rates of grade 3 4 toxicity, although against this must be offset the risks, inconveniences and expense of infusional treatment; disappointingly, neither showed evidence of increased efficacy in the adjuvant setting [12, 13]. An ongoing industry trial of oral capecitabine will be of interest, since it presents a potentially more attractive method of protracted fluoropyrimidine dosing [14]. The large UK-based trial, QUASAR, has been important in identifying simple but better-tolerated regimens of bolus 5-FU and LV. It included 4929 patients randomised to receive adjuvant 5-FU at 370 mg/m 2 (as in the Machover regimen [7]), together with either high- or low-dose L-leucovorin (175 versus 25 mg) and with either levamisole or placebo. This trial demonstrated, with high statistical power, that levamisole does not add to the activity of 5-FU/LV, and that low-dose LV is as effective as, and less toxic than, high-dose LV [2]. In QUASAR, participating centres chose to use a 5-day, 4-weekly schedule for six cycles (as in the Mayo and Machover regimens) or to give the same doses once weekly for 30 weeks. Similar numbers of patients received the two schedules (2559 and 2370, respectively), and, although not randomised, these groups were fortuitously very well matched for pre-treatment characteristics, including age and stage. The lack of randomisation means that comparisons must be interpreted with caution, but with this important proviso there was no evidence for any difference in relapse-free survival [hazard ratio (HR) 0.99; 95% confidence interval (CI) ] or overall survival (HR 0.97; 95% CI ) between patients receiving the two administration schedules. On the other hand, toxicity was markedly different, with highly significantly less diarrhoea (P <0.001), stomatitis (P <0.001), neutropenia (P <0.001) and dermatitis (P <0.001) in patients on the once-weekly schedule. Consequently, there were lower rates of dose reduction (P <0.001), and a higher percentage of planned dose was delivered (P <0.001) [3]. Reaction to these data has been mixed. Some oncologists now use the 5-day monthly treatment at 370 mg/m 2 5-FU with lowdose LV, on the basis that the QUASAR trial randomly validated this schedule against a standard regimen. Others, reassured by the large size and well-balanced patient characteristics in QUASAR s non-randomised comparison of schedules, have adopted the weekly regimen, which gives the same doses weekly for 30 weeks, so giving the same total planned dose (11.1 g/m 2 ) but with lower planned DI (370 versus 462 mg/m 2 /week). We reasoned that, given the low 5-FU dose used in the weekly QUASAR schedule and the near absence of toxicity for many patients, there was scope to increase the dose while maintaining acceptably low toxicity levels, and that this may allow for more dose-intensive treatment, comparable to other validated treatments. This report of our experience with this regimen must be interpreted with caution; as a non-randomised series it is subject to selection bias, and cannot provide evidence for efficacy. However, within this limitation the premise of the approach has been confirmed: weekly 5-FU at 425 mg/m 2, with low dose D,L-LV 45 mg flat rate, for 24 weeks, is feasible, with minimal toxicity to patients. Consequently, even in this non-trial population where treatments were often omitted for patients holidays and the threshold for reducing or stopping treatment was relatively low, the median DI achieved was 96% of planned DI whilst on treatment (407 mg/m 2 /week), and 91% of planned DI over 24 weeks (385 mg/m 2 /week). Although dose modification or delay was used at some point in the majority of patients, this was usually for mild toxicity, and only 20% of patients experienced any grade 3 toxicity at any stage. Notably, 36% of our patients were over the age of 70 years, and, in common with previous studies [15 17], we found that female sex and advanced age were significantly correlated with higher toxicity and lower delivered DI. How does this compare with other schedules? The Mayo and Roswell Park schedules have planned DIs of 472 and 375 mg/m 2 / week, respectively; it is clear that the deliverable DI, particularly for the Mayo schedule, is well below planned DI, but trial reports have not provided these data. The Machover regimen was reported, in the three trials making up the IMPACT report, at variable median delivered DI from 85% to 98% of planned (393 to 454 mg/m 2 /week) [7]. Delivered DI data were reported for QUASAR: the median delivered DI was 398 mg/m 2 for the 5-day monthly and 345 mg/m 2 for the once weekly schedule [3]. Thus the higher dose weekly schedule described here achieved similar DI to the 5-days QUASAR schedule. Does this represent an acceptable regimen for non-trial delivery of 5-FU? QUASAR firmly established its 5-day monthly schedule, with low-dose LV, to be as effective as the more toxic and expensive Machover schedule [2]; however, the status of QUASAR s weekly schedule as a standard option is more contentious, depending as it does on a very large but non-randomised comparison [3]. The regimen we have described here delivers a similar DI to QUASAR s 5-day monthly schedule. Its toxicity is low and acceptable: 20% of our patients underwent 20% dose reduction, which compares with 17% of patients in the QUASAR weekly and 42% in the 5-day monthly schedules. Comparison with trials using the Mayo, Machover and Roswell Park schedules would suggest that we have also achieved at least as high delivered DI, with lower toxicity, than those schedules. With any adjuvant treatment regimen, certain proof of equivalent efficacy can only be achieved by an adequately powered randomised trial, involving several thousands of patients. Comparisons of different 5-FU schedules are, thankfully, no longer a priority for such large trials programmes, which are now rightly focused on novel therapies. However, we would suggest that the comparative toxicity and DI data presented here would support the use of 5-FU 425 mg/m 2 plus low-dose LV, weekly for 24 weeks, as a simple, convenient and relatively low-toxicity treatment option for non-trial patients requiring 5-FU adjuvant therapy. Our subgroup analysis would, in common with others findings, suggest the need for caution in elderly female patients since, even with this low toxicity schedule, these patients have the highest risk of toxicity.

6 573 Acknowledgements The authors wish to acknowledge the contributions of C. Johnston, statistician, and the following clinicians, nurse specialists and data managers who were involved in treating patients and recording data: A. Chaudhury, D. M. Bottomly, M. Braun, N. Charnley, D. Dodwell, K. Dyker, L. Garbutt, S. Goulding, P. Hatfield, J. Lowes, A. Melcher, S. Pashley, C. Peedle, D. Jackson, S. Jagdev, F. Roberts, M. P. Snee and K. Whitmarsh. This research was supported by Cancer Research UK and the Cookridge Gastrointestinal Oncology Research Unit Fund. References 1. de Gramont A, Banzi M, Navarro M et al. Oxaliplatin/5-FU/LV in adjuvant colon cancer: results of the international randomized MOSAIC trial. Proc Am Soc Clin Oncol 2003; 22: 1015 (Abstr). 2. QUASAR Collaborative Group. Comparison of fluorouracil with additional levamisole, higher-dose folinic acid, or both, as adjuvant chemotherapy for colorectal cancer: a randomised trial. Lancet 2000; 355: Kerr DJ, Gray R, McConkey C, Barnwell J. Adjuvant chemotherapy with 5-flourouracil, L-folinic acid and levamisole for patients with colorectal cancer: non-randomised comparison of weekly versus four-weekly schedules. Ann Oncol 2000; 11: O Connell MJ, Mailliard JA, Kahn MJ et al. Controlled trial of fluorouracil and low-dose leucovorin given for 6 months as postoperative adjuvant therapy for colon cancer. J Clin Oncol 1997; 15: Haller DG, Catalano PJ, MacDonald JS, Mayer RJ. Fluorouracil, leucovorin and levamisole adjuvant therapy for colon cancer: five-year final report of INT0089. Proc Am Soc Clin Oncol 1998; 17: 982 (Abstr). 6. O Connell MJ, Laurie JA, Kahn M et al. Prospectively randomized trial of postoperative adjuvant chemotherapy in patients with high-risk colon cancer. J Clin Oncol 1998; 16: International Multicentre Pooled Analysis of Colon Cancer Trials (IMPACT) Investigators. Efficacy of adjuvant fluorouracil and folinic acid in colon cancer. Lancet 1995; 345: Wolmark N, Rockette H, Fisher B et al. The benefit of leucovorinmodulated fluorouracil as postoperative adjuvant therapy for primary colon cancer: results from National Surgical Adjuvant Breast and Bowel Project Protocol C-03. J Clin Oncol 1993; 11: Macdonald JS, Astrow AB. Adjuvant therapy of colon cancer. Semin Oncol 2001; 28: Tomiak A, Vincent M, Kocha W et al. Standard dose (Mayo regimen) 5-fluorouracil and low dose folinic acid: prohibitive toxicity? Am J Clin Oncol 2000; 23: Punt CJA, Nagy A, Douillard J-Y et al. Edrecolomab alone or in combination with fluorouracil and folinic acid in the adjuvant treatment of stage III colon cancer: a randomised study. Lancet 2002; 360: André T, Colin P, Louvet C et al. Semimonthly versus monthly regimen of fluorouracil and leucovorin administered for 24 or 36 weeks as adjuvant therapy in stage II and III colon cancer: results of a randomized trial. J Clin Oncol 2003; 21: Saini A, Norman AR, Cunningham D et al. Twelve weeks of protracted venous infusion of fluorouracil is as effective as 6 months of bolus 5FU and folinic acid as adjuvant treatment in colorectal cancer. Br J Cancer 2003; 88: Twelves C, Wong A, Nowacki MP et al. Improved safety results of a phase III trial of capecitabine vs bolus 5-FU/leucovorin as adjuvant therapy for colon cancer (the X-ACT Study). Proc Am Soc Clin Oncol 2003; 22: 1182 (Abstr). 15. Meta-analysis Group in Cancer. Toxicity of fluorouracil in patients with advanced colorectal cancer: effect of administration schedule and prognostic factors. J Clin Oncol 1998; 16: Stein BN, Petrelli NJ, Douglass HO et al. Age and sex are independent predictors of 5-fluorouracil toxicity. Analysis of a large scale phase III trial. Cancer 1995; 75: Sloan JA, Goldberg RM, Sargent DJ et al. Women experience greater toxicity with fluorouracil-based chemotherapy for colorectal cancer. J Clin Oncol 2002; 20: