Cairo

Articles

Sequential versus combination chemotherapy with capecitabine, irinotecan, and oxaliplatin in advanced colorectal cancer (CAIRO): a phase III randomised controlled trial Miriam Koopman, Ninja F Antonini, Joep Douma, Jaap Wals, Aafke H Honkoop, Frans L G Erdkamp, Robert S de Jong, Cees J Rodenburg, Gerard Vreugdenhil, Olaf J L Loosveld, Aart van Bochove, Harm A M Sinnige, Geert-Jan M Creemers, Margot E T Tesselaar, Peter H Th J Slee, Marjon J B P Werter, Linda Mol, Otilia Dalesio, Cornelis J A Punt

Summary Background The optimum use of cytotoxic drugs for advanced colorectal cancer has not been defined. Our aim was to investigate whether combination treatment is better than sequential administration of the same drugs in patients with advanced colorectal cancer. Methods We randomly assigned 820 patients with advanced colorectal cancer to receive either first-line treatment with capecitabine, second-line irinotecan, and third-line capecitabine plus oxaliplatin (sequential treatment; n=410) or first-line treatment capecitabine plus irinotecan and second-line capecitabine plus oxaliplatin (combination treatment; n=410). The primary endpoint was overall survival. Analyses were done by intention to treat. This trial is registered with ClinicalTrials.gov with the number NCT00312000. Findings 17 patients (nine in the sequential treatment group, eight in the combination group) were found to be ineligible and were excluded from the analysis. 675 (84%) patients died during the study: 336 in the sequential group and 339 in the combination group. Median overall survival was 16·3 (95% CI 14·3–18·1) months for sequential treatment and 17·4 (15·2–19·2) months for combination treatment (p=0·3281). The hazard ratio for combination versus sequential treatment was 0·92 (95% CI 0·79–1·08; p=0·3281). The frequency of grade 3–4 toxicity over all lines of treatment did not differ significantly between the two groups, except for grade 3 hand-foot syndrome, which occurred more often with sequential treatment than with combination treatment (13% vs 7%; p=0·004). Interpretation Combination treatment does not significantly improve overall survival compared with the sequential use of cytotoxic drugs in advanced colorectal cancer. Thus sequential treatment remains a valid option for patients with advanced colorectal cancer.

Introduction Fluorouracil has been the standard of care for patients with advanced colorectal cancer for decades. The availability of irinotecan and oxaliplatin has increased the median overall survival for these patients.1 Several studies have shown a clinical benefit of adding irinotecan or oxaliplatin to fluorouracil in first-line treatment.2–4 Each drug can also confer a benefit in patients refractory to first-line fluorouracil-based chemotherapy.5,6 In terms of survival, there does not seem to be an optimum sequence of administration of irinotecan and oxaliplatin,7 and the optimum timing of delivery of these agents has not been clearly defined. The survival results of randomised studies in which combination treatments were investigated should be interpreted with caution, since salvage treatments were not a prospective part of the study designs, which might have affected the results. Furthermore, the availability of irinotecan and especially oxaliplatin was variable during the period these studies were under way.8 Capecitabine, an oral fluoropyrimidine, has at least equivalent efficacy and a clinically meaningful safety www.thelancet.com Vol 370 July 14, 2007

advantage compared with fluorouracil treatment.9,10 Phase II studies of the combination of capecitabine with either irinotecan or oxaliplatin show efficacy but also toxicity comparable with combination schedules with fluorouracil.11,12 As a result, capecitabine is now often used as an alternative to fluorouracil in the treatment of advanced colorectal cancer, either as monotherapy or in combination therapy. In the CApecitabine, IRinotecan, Oxaliplatin (CAIRO) trial, our aim was to determine whether first-line combination treatment is better than sequential administration of the same drugs in terms of overall survival in patients with advanced colorectal cancer.

Methods Patients Patients aged over 18 years were eligible if they had histologically proven colorectal cancer that was in an advanced stage not amenable to curative surgery, together with measurable or assessable disease parameters (serum carcino embryonic antigen as the only parameter for disease activity was not allowed), and no previous systemic

Lancet 2007; 370: 135–42 See Comment page 105 See Articles page 143 Radboud University Nijmegen Medical Centre, Nijmegen, Netherlands (M Koopman MD, Prof C J A Punt MD); Netherlands Cancer Institute (NKI), Biometrics Department, Amsterdam, Netherlands (N F Antonini MSc, O Dalesio MSc); Rijnstate Hospital, Arnhem, Netherlands (J Douma MD); Atrium Medical Centre, Heerlen, Netherlands (J Wals MD); Isala Hospital, Zwolle, Netherlands (A H Honkoop MD); Maasland Hospital, Sittard, Netherlands (F L G Erdkamp MD); Martini Hospital, Groningen, Netherlands (R S de Jong MD); Meander Medical Centre, Amersfoort, Netherlands (C J Rodenburg MD); Maxima Medical Centre, Veldhoven, Netherlands (G Vreugdenhil MD); Amphia Hospital, Breda, Netherlands (O J L Loosveld MD); Zaans Medical Centre, Zaandam, Netherlands (A v Bochove MD); Jeroen Bosch Hospital, Den Bosch, Netherlands (H A M Sinnige MD); Catharina Hospital, Eindhoven, Netherlands (G-J M Creemers MD); Leiden University Medical Centre, Leiden, Netherlands (M E T Tesselaar MD); St Antonius Hospital, Nieuwegein, Netherlands (P H Th J Slee MD); Viecuri Hospital, Venlo, Netherlands (M J B P Werter MD); and Comprehensive Cancer Centre East (IKO), Nijmegen (L Mol Msc) Correspondence to: Prof C J A Punt, Department of Medical Oncology, Radboud University Nijmegen Medical Centre, P O Box 9101, 6500 HB Nijmegen, Netherlands [email protected]

135

Articles

820 patients with advanced colorectal cancer

17 patients not meeting eligibility criteria 4 histological diagnosis of colorectal cancer not confirmed 4 abnormal liver or renal function 3 prior chemotherapy for advanced cancer or adjuvant chemotherapy <6 months 2 serious concomitant disease 2 second primary before randomisation 1 no assessable lesion 1 brain metastases

401 patients randomly allocated to sequential chemotherapy

402 patients randomly allocated to combination chemotherapy

397 (99%) patients received first-line capecitabine

398 (99%) patients received first-line capecitabine+irinotecan

4 patients did not receive allocated intervention 1 major protocol violation 3 other reasons 10 patients: first-line treatment still ongoing

4 patients did not receive allocated intervention 2 patient refusal 1 major protocol violation 1 other reason 10 patients: first-line treatment still ongoing

251 (62%) patients received second-line irinotecan 213 (53%) patients received second-line capecitabine+oxaliplatin 136 patients did not receive second-line irinotecan 17 unacceptable toxicity 16 patient refusal 67 no further benefit/poor PS 25 death 2 surgery (hemihepatectomy) 9 other reasons 16 patients: second-line treatment still ongoing

175 patients did not receive second-line capecitabine+oxaliplatin 44 unacceptable toxicity 31 patient refusal 58 no further benefit/poor PS 26 death 7 surgery (hemihepatectomy) 9 other reasons 10 patients: second-line treatment still ongoing

143 (36%) patients received third-line capecitabine+oxaliplatin 92 patients did not receive third-line capecitabine+oxaliplatin 45 no further benefit/poor PS 6 toxicity 24 patient refusal 10 death 7 other reasons 4 patients: third-line treatment still ongoing

401 patients analysed for primary outcome (overall survival) 0 patients excluded

402 patients analysed for primary outcome (overall survival) 0 patients excluded

Figure 1: Trial profile

treatment for advanced disease. Previous adjuvant chemotherapy was allowed provided that the last administration was given at least 6 months before randomisation. Eligible patients had to have a WHO performance score 0–2 and adequate hepatic, bone marrow, and renal function. Exclusion criteria were: serious concomitant disease preventing the safe administration of chemotherapy or likely to interfere with the study assessments; other malignancies in the past 5 years with the exception of adequately treated carcinoma in situ of the cervix and squamous or basal cell carcinoma of the skin; pregnancy or lactation; patients (both male and female) with reproductive potential not implementing adequate contraceptive measures; central nervous system metastases; serious active infections; inflammatory bowel 136

disease or other diseases associated with chronic diarrhoea; previous extensive irradiation of the pelvis or abdomen (excluding 5×5 Gy irradiation for rectal carcinoma); concomitant (or within 4 weeks before randomisation) administration of any other experimental drug; concurrent treatment with any other anti-cancer therapy. The study was approved by the Central Committee of Human-related Research and by the local ethics committees of all participating centres. Written informed consent was required from all patients before study entry.

Procedures This open-label, randomised trial was done by the Dutch Colorectal Cancer Group (DCCG) in 74 centres in the Netherlands. Patients were enrolled between January, www.thelancet.com Vol 370 July 14, 2007

Articles

2003, and December, 2004. Eligible patients were randomly assigned to either sequential or combination treatment in a 1:1 ratio. Randomisation was done centrally by a minimisation technique with stratification according to WHO performance status (0–1 vs 2), serum lactate dehydrogenase (LDH) concentration (normal vs abnormal), previous adjuvant treatment (yes vs no), predominant localisation of metastases (liver vs extrahepatic), and treatment centre. All treatment cycles were administered at intervals of 3 weeks. In the sequential treatment group, first-line treatment consisted of capecitabine (1250 mg/m² twice daily) for 14 days, second-line treatment of irinotecan (350 mg/m²) on day 1, and third-line treatment of capecitabine (1000 mg/m² twice daily) for 14 days plus oxaliplatin (130 mg/m²) on day 1. Patients assigned to combination treatment received capecitabine (1000 mg/m² twice daily) for 14 days plus irinotecan (250 mg/m²) on day 1 as first-line treatment, and capecitabine (1000 mg/m²) twice daily for 14 days plus oxaliplatin (130 mg/m²) on day 1 as second-line treatment. It was recommended to start irinotecan at 75% of the dose in patients over 70 years of age; escalation to full dose when no serious toxicity occurred was recommended. Tumour response was assessed every 9 weeks with CT scans. Assessment was done by the investigators, who used the Response Evaluation Criteria for Solid Tumours (RECIST).13 Toxic effects were assessed according to the US National Cancer Institute Common Toxicity Criteria, version 2.0. At each visit, a history was taken and physical examination, blood haematology, and biochemistry tests were done. Modification of the dose of all drugs in cases of haematological or non-haematological toxic effects were done as specified per protocol. An independent data monitoring committee reviewed safety data on a regular basis. Before a patient entered the next line of treatment, the initial eligibility criteria had to be met. Treatment in both study arms was to be continued for at least 6 months or until disease progression or unacceptable toxicity, whichever came first. After 6 months, it was recommended to continue treatment until disease progression or unacceptable toxicity. However, continuation of any treatment schedule after 6 months in the absence of disease progression or unacceptable toxicity was at the discretion of the investigator. Patients who discontinued any treatment for reasons other than disease progression or unacceptable toxicity had to be assessed for disease status every 3 months. If disease progression occurred after a treatment-free interval of 3 months or less, the next line of treatment was to be initiated. If this interval was longer than 3 months, the previous line of treatment was to be resumed. The QLQ-C30 questionnaire of the European Organisation for Research and Treatment of Cancer (EORTC) was used to assess the patients’ wellbeing during the study. Participation in this part of the study was www.thelancet.com Vol 370 July 14, 2007

Sequential treatment (N=401)

Combination Total treatment (N=803) (N=402)

Age Age at randomisation (years)

64·0 (27–84)

63·0 (31–81)

63·0 (27–84)

>70 years

93 (23%)

81 (20%)

Male

252 (63%)

255 (63%)

507 (63%)

Female

149 (37%)

147 (37%)

296 (37%)

174 (22%)

Sex

Performance status PS0

257 (64%)

244 (61%)

501 (62%)

PS1

126 (31%)

142 (35%)

268 (33%)

PS2

18 (5%)

16 (4%)

34 (4%)

277 (69%)

285 (71%)

562 (70%)

75 (19%)

92 (23%)

167 (21%)

118 (29%)

115 (29%)

233 (29%)

2 (<1%)

8 (<1%)

Predominant localisation of metastases Liver Liver-only Extrahepatic Unknown

6 (2%)

LDH at randomisation Normal

256 (64%)

257 (64%)

513 (64%)

Abnormal

145 (36%)

145 (36%)

290 (36%)

Previous adjuvant therapy Yes

55 (14%)

56 (14%)

111 (14%)

No

346 (86%)

346 (86%)

692 (86%)

251 (63%)

227 (57%)

478 (60%)

28 (7%)

32 (8%)

60 (8%)

119 (30%)

141 (35%)

260 (32%)

Site of primary tumour Colon Rectosigmoid Rectum Multiple tumours

2 (<1%)

2 (<1%)

4 (<1%)

Missing

1 (<1%)

0

1 (<1%)

Data are median (range) or n (%).

Table 1: Baseline characteristics of randomised and eligible patients

proposed to the first 620 patients entered in the study. Questionnaires were to be completed within 1 week before randomisation and at every 9 weeks thereafter until the end of study treatment. Death of any patient within 30 days of the last administration of study drugs had to be reported as a serious adverse event, and the relation between treatment and death was assessed as unrelated, remote, possible, or probable. Central review was done by the study coordinators (CJAP, MK) of all patient files when death occurred within 30 days of the last administration of study drugs and when accompanied by any other event than disease progression, irrespective of the cause reported for this event. Their assessment, together with that made by the investigator, was then submitted to the independent data monitoring committee, whose assessment on the relation with treatment was deemed final.

Statistical analysis With an anticipated median overall survival for sequential treatment of 14 months, and assuming a median overall 137

Articles

1·0

Combination chemotherapy Sequential chemotherapy Log-rank p=0·3281

Survival probability

0·8

0·6

0·4

0·2

0 0

10

Total deaths Combination chemotherapy Sequential chemotherapy

20 30 Survival (months)

110 228 122 245 Number of patients Event Combination chemotherapy 402 84% (336) Sequential chemotherapy 401 85% (339)

40

302 317 Censored 16% (66) 15% (62)

50

333 336 335 339 Median survival (95% CI) 17·36 (15·32–19·24) 16·30 (14·28–18·15)

Figure 2: Overall survival by treatment arm

survival for combination treatment of 17·5 months, we calculated that, to have 80% power to detect a 20% reduction in the hazard of death (hazard ratio 0·80) at a significance level of 5%, we would need to observe 630 events (deaths), which would require a sample size of 800 patients. The primary endpoint was overall survival, and was calculated as the interval from the date of randomisation until death from any cause or until the date of last follow-up. Secondary objectives were to assess progression-free survival, tumour response, toxicity profile, and quality Sequential treatment (N=401)

Combination treatment (N=402)

Overall survival (months)

16·3 (14·3–18·1)

17·4 (15·2–19·2)

1-year survival rate (%)

64% (59–69)

67% (62–72)

Progression-free survival first-line (months)

5·8 (5·1–6·2)

p value

0·3281 0·38

7·8 (7·0–8·3)

0·0002

PFS2 (months)

8·7 (8·2–9·6)

10·3 (9·3–10·8)

0·15

PFS3 (months)

10·3 (9·0–11·1)

NA

0·19*

First-line

77 (20%, 17–26%)

139 (41%, 36–46%)

<0·0001

Second-line

23 (10%, 6–15%)

24 (12%, 7–17%)

0·46

Overall response rate (CR+PR)†

Third-line

5 (4%, 1–9%)

··

··

of life. Progression-free survival for first-line treatment was calculated from the date of randomisation to the first observation of disease progression or death from any cause. Progression-free survival was also calculated for first-line and second-line treatment (PFS2) and for first-line, second-line, and third-line treatment (PFS3). PFS2 was defined as the time from randomisation until first progression reported after the start of second-line treatment and, if a patient did not start second-line treatment, the date of first progression reported after randomisation or death or time at which the patient was last known to be alive. Likewise, PFS3 was defined as the time from randomisation until first progression, death, or date at which the patient was last known to be alive after the start of third-line treatment, and PFS2 was used in cases where the patient did not start third-line treatment.14 Overall and progression-free survival curves were estimated with the Kaplan-Meier method and compared with the log-rank test. Patients were deemed to be assessable for toxicity if they had started treatment, and for response if they had completed at least three cycles of treatment. The occurrence of grade 3–4 adverse events, hypersensitivity reaction, cardiac ischaemia or infarction, and thrombosis or embolism were compared by χ² test. Patients who completed the quality-of-life questionnaire at baseline and at least once during treatment were included in the analysis of quality of life. Functional and symptomatic scales, overall quality of life, and financial problems were assessed. The largest decrease in quality of life with respect to baseline was calculated. The Wilcoxon rank sum test was used to detect statistical differences between the treatment arms. All analyses were done according to the intention-to-treat principle. All tests were two-sided and p values of less than 0·05 were deemed to be significant. All data received before May, 2007, are included in this report. All analyses were done with SAS version 9.1 and S-plus version 6.2. This trial is registered with ClinicalTrials.gov with the number NCT00312000.

Role of the funding source The funding source had no role in the design, conduct, data collection, data analysis, or interpretation of the study or the results. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.

Disease control rate (CR+PR+SD)† First-line

280 (74%, 69–79%)

297 (87%, 82–90%)

<0·0001

Second-line

162 (71%, 65–77%)

121 (63%, 56–70%)

0·06

Third-line

72 (57%, 48–66%)

··

··

CR=complete tumour response. PR=partial tumour response. SD=stable disease. Data are median (95% CI), %, or n (%, 95% CI). *PFS3 in sequential group vs PFS2 in combination group. †Percentages are based on patients assessable for response.

Table 2: Efficacy results

138

Results 820 patients were randomised, 410 in each arm of the study (figure 1). 17 patients were found to be ineligible and were excluded from the analysis. Reasons for ineligibility were: histological diagnosis of colorectal cancer not confirmed (n=4); abnormal liver or renal function at baseline (4); previous chemotherapy for advanced www.thelancet.com Vol 370 July 14, 2007

Articles

colorectal cancer or adjuvant chemotherapy completed less than 6 months before randomisation (3); serious concomitant disease (2); recent history of second malignancy (2); no evaluable lesion (1), and presence of brain metastases (1). Of the 803 eligible patients, 795 patients received at least one cycle of treatment. The patients’ characteristics were well balanced between the two arms (table 1). In the sequential group, the median number of cycles for patients was six (range 1–45) in first-line; six (1–35) in second-line, and four (1–14) in third-line treatment. In the combination group, the median number of cycles for patients was seven (1–42) in first-line treatment and five (1–23) in second-line treatment. The (arithmetic) median time on treatment (ie, interval between the start of protocol treatment and a patient being off study) was 10·7 (range 0·1–45·1) months in the sequential treatment group and 7·4 (0·1–43·2) months in the combination group (p=0·002). At the time of analysis, 675 (84%) of the 803 eligible patients had died: 336 in the sequential group and 339 in the combination group. The median follow-up for the 128 patients still alive is 31·5 (range 14–49) months. Median overall survival was 16·3 (95% CI 14·3–18·1) months for the sequential treatment group and 17·4 (15·2–19·2) months for the combination group. Figure 2 shows the survival curves by treatment group. The hazard ratio for combination versus sequential was 0·92 (95% CI 0·79–1·08). Comparison of the curves by the log-rank test showed that the difference was not significant (p=0·3281, table 2). In a multivariable analysis that took into account the stratification factors and age 70 years or over, only performance status 2 (hazard ratio 1·44, 95% CI 1·02–2·06; p=0·04) and abnormal serum LDH (hazard ratio 1·9, 1·68–2·33; p<0·0001) were associated with worse overall survival. In first-line treatment, progression-free survival was significantly longer in the combination treatment group than it was in the sequential treatment group (p=0·0002); the hazard ratio for progression for combination versus sequential treatment was 0·77 (95% CI 0·67–0·89; p=0·0002). Progression-free survival was not affected when calculated to disease progression upon which the previous line of treatment was definitely discontinued and treatment-free intervals, after which the previous treatment was resumed, were ignored (6·0 [95% CI 5·4–6·5] months in the sequential group vs 8·0 [7·3–8·4] months in the combination group). PFS2 was not significantly different between the two groups (p=0·15); likewise, the difference between PFS3 in sequential treatment and PFS2 in combination treatment was not significant (p=0·19; table 2). 719 patients were assessable for response in first-line treatment: 379 patients in the sequential group and 340 in the combination group. The overall response rate (ie, complete plus partial tumour response) in first-line was significantly better in the combination group than in the sequential group (p<0·0001; table 2). The disease control www.thelancet.com Vol 370 July 14, 2007

Sequential treatment (N=397)

Combination treatment (N=398)

Total (N=795)

p value*

Non-haematological adverse events Overall grade 3–4 toxicity

271 (68%)

265 (67%)

536 (67%)

0·61

Hypersensitivity reaction (total)

25 (6%)

18 (5%)

43 (5%)

0·27

Cardiac ischaemia/infarction (total)

14 (4%)

14 (4%)

28 (4%)

0·99

Thrombosis/embolism

35 (9%)

41 (10%)

76 (10%)

0·48

Grade 3 hand-foot skin reaction

50 (13%)

26 (7%)

76 (10%)

0·004

Grade 3

83 (21%)

90 (23%)

173 (22%)

0·23

Grade 4

9 (2%)

17 (4%)

26 (3%)

31 (8%)

36 (9%)

67 (8%)

Diarrhoea

Nausea Grade 3 Grade 4

0

1 (<1%)

0·45

1 (<1%)

Stomatitis Grade 3 Grade 4

12 (3%) 0

5 (1%) 1 (<1%)

17 (2%)

0·15

1 (<1%)

Vomiting Grade 3

24 (6%)

Grade 4

4 (1%)

37 (9%) 2 (<1%)

61 (8%)

0·16

6 (<1%)

Neuropathy sensory Grade 3

8 (2%)

Grade 4

1 (<1%)

12 (3%)

20 (3%)

0

1 (<1%)

0·18

Haematological adverse events Anaemia Grade 3

2 (<1%)

1 (<1%)

3 (<1%)

Grade 4

1 (<1%)

2 (<1%)

3 (<1%)

0·18

Neutropenia Grade 3

17 (4%)

25 (6%)

42 (5%)

Grade 4

3 (1%)

4 (1%)

7 (1%)

Grade 3

16 (4%)

24 (6%)

40 (5%)

Grade 4

3 (1%)

4 (1%)

7 (1%)

Grade 3

3 (1%)

3 (1%)

6 (1%)

Grade 4

1 (<1%)

1 (<1%)

2 (<1%)

0·19

Febrile neutropenia 0·18

Thrombocytopenia 0·99

Data are n (%). *p values are presented for grade 3 and 4 toxicities combined.

Table 3: Adverse effects associated with sequential versus combination treatment over all lines

rate (ie, complete response plus partial response plus stable disease) was also significantly better in the combination treatment group than in the sequential treatment group (p<0·0001; table 2). In second-line treatment, the response rate and disease control rates were not significantly different between the two groups (table 2). The results of an interim safety analysis in the first 400 patients that were enrolled have been published.15 In the total patient cohort, there was no significant difference in the frequency of grade 3–4 toxicity over all lines of treatment in either group (p=0·61; table 3). The only significant difference concerned the occurrence of grade 3 hand-foot syndrome, which occurred more frequently with sequential treatment than with 139

Articles

combination treatment (p=0·004; table 3). The frequency of thrombosis or embolism, and of cardiac ischaemia, did not differ significantly between the two treatment groups (table 3). Table 4 shows the adverse events that occurred during first-line treatment. Grade 3–4 diarrhoea occurred significantly more frequently in the combination group than in the sequential group (p<0·0001), as did grade 3–4 nausea (p=0·004), grade 3–4 vomiting (p=0·0002), febrile neutropenia (p<0·0001), and grade 3–4 neutropenia including febrile neutropenia (p<0·0001). Grade 3 handfoot syndrome occurred significantly more frequently in the sequential treatment group than in the combination treatment group (p=0·002). Death with a probable relation to treatment occurred in 11 patients: eight after sequential and three after combination treatment (p=0·13). Causes of death were sepsis (eight patients, of whom seven had neutropenia), diarrhoea (two patients, of whom one had neutropenia), Sequential treatment (N=397)

Combination treatment (N=398)

Total (N=795)

p value*

0·99

Non-haematological adverse events Hypersensitivity reaction (total)

7 (2%)

7 (2%)

14 (2%)

Cardiac ischaemia/infarction (total)

11 (3%)

13 (3%)

24 (3%)

0·68

Thrombosis/embolism (total)

28 (7%)

38 (10%)

66 (8%)

0·20

Grade 3 hand-foot skin reaction

48 (12%)

23 (6%)

71 (9%)

0·002

Grade 3

38 (10%)

87 (22%)

125 (16%)

<0·0001

Grade 4

5 (1%)

15 (4%)

20 (3%)

14 (4%)

33 (8%)

47 (6%)

Diarrhoea

Discussion

Nausea Grade 3 Grade 4

2 (<1%)

5 (1%)

7 (<1%)

Grade 3

2 (<1%)

5 (1%)

7 (<1%)

Grade 4

0

1 (<1%)

1 (<1%)

0·004

Stomatitis 0·16

Vomiting Grade 3

9 (2%)

Grade 4

1 (<1%)

33 (8%) 1 (<1%)

42 (5%)

Grade 3

1 (<1%)

0

1 (<1%)

Grade 4

1 (<1%)

1 (<1%)

2 (<1%)

0·0002

2 (<1%)

Haematological adverse events Anaemia 0·56

Neutropenia Grade 3

2 (<1%)

Grade 4

0

23 (6%)

25 (3%)

4 (1%)

4 (1%)

<0·0001

Febrile neutropenia Grade 3

2 (<1%)

22 (6%)

24 (3%)

Grade 4

0

4 (1%)

4 (1%)

Grade 3

0

0

0

Grade 4

0

1 (<1%)

1 (<1%)

<0·0001

Thrombocytopenia 0·32

Data are n (%). *p values are presented for grade 3 and 4 toxicities combined.

Table 4: Adverse effects associated with sequential versus combination treatment first-line treatment

140

and neutropenic fever (one patient). Protocol violations were identified in nine of these patients; some of these only being detected after central review. Violations included administration of irinotecan in patients with hyperbilirubinaemia, non-adherence to guidelines for dose reductions, or delays of chemotherapy in case of diarrhoea. Six patients died suddenly, one during sequential and five during combination treatment (p=0·10). Cardiopulmonary risk factors were present in four of these patients. All-cause 60-day mortality did not differ significantly between the two groups (3·0% [12 patients] in the sequential group vs 4·5% [18 patients] in the combination group; p=0·27). 403 patients were available for assessment for quality of life: 203 in the sequential treatment group and 200 in the combination treatment group. Change in financial problems and global health status were similar between the two groups. The decrease in functioning was on average higher for combination treatment on all scales (cognitive, emotional, physical, role, and social; data not shown). The largest decrease was seen for role functioning: a decrease of 20 points for sequential treatment versus 24 points for combination treatment. Also for the symptomatic scales, changes were on average greater for combination treatment, except for pain and dyspnoea (data not shown). The only significant difference in change was seen for diarrhoea: 20 points for sequential versus 28 points for combination treatment (p=0·002).

Our results show that, for patients with advanced colorectal cancer, combination treatment with all effective cytotoxic drugs was no better than their sequential use. Progression-free survival over all subsequent treatment lines was not significantly different between the study groups. Additionally, sequential treatment was associated with less toxicity during first-line treatment than was combination therapy. The high median overall survival seen in patients receiving sequential therapy in which a fluoropyrimidine was administered as the sole agent in first-line treatment supports the validity of the strategy to make all effective drugs available during the course of disease.1 Furthermore, our results for progression-free survival and response rates in first-line treatment were in the range of previously published results with fluorouracil monotherapy.3,4 These data also support the hypothesis that the survival benefit seen for combination chemotherapy in previous phase III studies might have been biased by the insufficient use of salvage treatments in the control arm.8 The Fluorouracil, Oxaliplatin, and CPT11 (irinotecan)— Use and Sequencing (FOCUS) study investigated prospectively the sequential versus combined use of either irinotecan or oxaliplatin in combination with fluorouracil plus levofolinate in advanced colorectal cancer.16 The www.thelancet.com Vol 370 July 14, 2007

Articles

inclusion of all three drugs in the treatment of individual patients was introduced at a later stage, and was done in a small subset of patients only. Overall survival in this study was 15·0 months with sequential treatment with fluorouracil and irinotecan (compared with 16·7 months for combination treatment) and 15·2 months with sequential treatment with fluorouracil and oxaliplatin (compared with 15·4 months for combination treatment). Our results suggest that survival is further improved when all drugs are made available during the course of disease. The proportion of patients who received all three active drugs in our trial is somewhat lower than that in other recent trials.1 Two factors might be relevant in this respect. First, the protocol recommended to treat patients until disease progression. However, if treatment was discontinued for any other reason than progression, the protocol prescribed to resume the initial treatment when the treatment-free interval was longer than 3 months. By contrast with several other studies, the reinstitution of previous therapy in this situation was not defined as subsequent therapy. Second, the proportion of patients with liver-only metastases is lower than usually reported in other phase III trials in this setting, which could have resulted in a study population with worse prognosis on average, and thus in a lower proportion of patients that were able to enter subsequent treatment lines. As first-line treatment, combination therapy was associated with a higher response rate and prolonged PFS compared with monotherapy, as expected. That this did not result in a significant survival benefit cautions against the use of PFS as a surrogate marker for overall survival when several lines of treatment are available. The increased overall frequency of grade 3 hand-foot syndrome is probably related to the higher dose of capecitabine when given as a single agent compared with its use in combination treatment. Even so, the frequency seen here is lower than that reported in previous studies with capecitabine monotherapy (12% vs 16–18%).9,10 The increased frequency of grade 3–4 diarrhoea in first-line treatment in the combination treatment group seen here is comparable with that from three phase II studies in which the same dose and schedule of capecitabine and irinotecan was used.12,17,18 However, two studies reported a higher frequency of toxicity with this regimen than was seen here: one phase III study was discontinued prematurely because of an excess of toxicity (39% of patients experienced grade 3–4 diarrhoea) and mortality (11% treatment related),19 and 46% of patients in a second study experienced grade 3–4 diarrhoea.20 Both of these studies also included the use of celecoxib in the randomisation procedure, and the contribution of this drug to the observed toxicity is uncertain. Geographical differences in terms of folate supplementation or polymorphisms in metabolism have been implicated in the tolerability of fluoropyrimidines,21 and these factors could have a role in the differences in toxicity among these studies. www.thelancet.com Vol 370 July 14, 2007

By contrast with our results, the frequency of grade 3–4 diarrhoea in studies with regimens of infusional fluorouracil plus folinic acid in combination with irinotecan every 2 weeks is usually lower (10–15%).2,14 However, this lower toxicity should be balanced against the disadvantages presented by use of central venous access devices and ambulatory pumps.14,22 The finding that most treatment-related deaths reported here involved protocol violations underscores the importance of education of patients, as well as medical and nursing staff, to prevent severe toxicity, and that strict quality control results in a better assessment of safety results. With the availability of targeted agents such as bevacizumab and cetuximab, the treatment options and outcome for patients with advanced colorectal cancer have changed considerably. However, chemotherapy remains the backbone of systemic treatment in this disease, and our results indicate that sequential treatment remains a valid treatment option for these patients. Contributors The study was designed by the Medical Oncology Task Force of the DCCG, which included RSdJ and CJAP. NFA, OD, MK, LM, and CJAP contributed to the data analysis and drafting of the manuscript. All authors took part in data collection. All authors reviewed and helped revise the manuscript, and approved the final version of the manuscript. CAIRO team members Medical Oncology Task Force: E Balk, A Cats, C J van Groeningen, R L H Jansen, J J M van der Hoeven, R S de Jong, N H Mulder, J W R Nortier, C J A Punt, D J Richel, H P Sleeboom, A S Th Planting, E E Voest. The following investigators all participated in undertaking the study: C Rodenburg-Amersfoort; J van der Hoeven-Amstelveen; D Richel, M Schweitzer, B de Valk, M Soesan-Amsterdam; J Douma-Arnhem; P Nieboer-Assen; F Valster-Bergen op Zoom; R Rietbroek-Beverwijk; G Ras, O Loosveld-Breda; D Kehrer-Capelle a/d IJssel; M Bos-Delft; Z Erjavec-Delfzijl; H Sinnige, C Knibbeler-Den Bosch; W Van Deijk, F Jeurissen, H Sleeboom-Den Haag; H de Jong-Den Helder; A Imholz-Deventer; E Muller-Doetinchem; H van Kamp-Drachten; E Balk-Ede; G Creemers, M Dercksen-Eindhoven; M Legdeur-Enschede; H van Halteren-Goes; A van der Torren-Gouda; G Hospers, R de Jong-Groningen; P Zoon-Harderwijk; J Wals-Heerlen; V Derleyn-Helmond; C Gerrits-Hengelo; C de Swart-Hoofddorp; J Haasjes-Hoogeveen; W Meijer-Hoorn; C de Swart-Haarlem; M Polee-Leeuwarden; M Tesselaar-Leiden; G Jonkers-Leiderdorp; R Brouwer-Leidschendam; R Jansen-Maastricht; P de Jong, P Slee-Nieuwegein; C Punt, H Oosten-Nijmegen; J van Wissen-Oosterhout; M Kuper-Oss; M den Boer-Roermond; A Planting, A vander Gaast, J Stouthard, F de Jongh, T Kok-Rotterdam; J Braun-Schiedam; F Erdkamp-Sittard; G Veldhuis-Sneek; C Geers-Spijkenisse; P van der Werf-Stadskanaal; A van Reisen-Terneuzen; H Roerdink-Tilburg; D ten Bokkel Huinink, R Oltmans, S van der Vegt, E Voest-Utrecht; L van Hulsteijn-Veghel; G Vreugdenhil-Veldhoven; M Werter-Venlo; J Ruit-Vlaardingen; L Kerkhofs-Vlissingen; W Jaspers-Winschoten; P Schiphorst-Winterswijk; J Holleman-Woerden; A van Bochove-Zaandam; O van Dobbenburgh-Zutphen; J de Graaf, A Honkoop-Zwolle. Independent Data Monitoring Committee: P De Mulder (deceased), D Sleijfer, G Stoter. Conflict of interest statement CJAP has received research grants and honoraria from Sanofi-Aventis, Roche, and Pfizer. All other authors declare that they have no conflict of interest.

141

Articles

Acknowledgments The DCCG received grant support from the Commissie Klinisch Toegepast Onderzoek (CKTO) of the Dutch Cancer Foundation (KWF), and unrestricted scientific grants from Roche, Aventis, Sanofi, and Pfizer. Logistic support was provided by Schmidt Consultancy. References 1 Grothey A, Sargent D, Goldberg RM, Schmoll HJ. Survival of patients with advanced colorectal cancer improves with the availability of fluorouracil-leucovorin, irinotecan, and oxaliplatin in the course of treatment. J Clin Oncol 2004; 22: 1209–14. 2 de Gramont A, Figer A, Seymour M, et al. Leucovorin and fluorouracil with or without oxaliplatin as first-line treatment in advanced colorectal cancer. J Clin Oncol 2000; 18: 2938–47. 3 Douillard JY, Cunningham D, Roth AD, et al. Irinotecan combined with fluorouracil compared with fluorouracil alone as first-line treatment for metastatic colorectal cancer: a multicentre randomised trial. Lancet 2000; 355: 1041–47. 4 Saltz LB, Cox JV, Blanke C, et al. Irinotecan plus fluorouracil and leucovorin for metastatic colorectal cancer. Irinotecan Study Group. N Engl J Med 2000; 343: 905–14. 5 Cunningham D, Pyrhonen S, James RD, et al. Randomised trial of irinotecan plus supportive care versus supportive care alone after fluorouracil failure for patients with metastatic colorectal cancer. Lancet 1998; 352: 1413–18. 6 Rothenberg ML, Oza AM, Bigelow RH, et al. Superiority of oxaliplatin and fluorouracil-leucovorin compared with either therapy alone in patients with progressive colorectal cancer after irinotecan and fluorouracil-leucovorin: interim results of a phase III trial. J Clin Oncol 2003; 21: 2059–69. 7 Punt CJ. Irinotecan or oxaliplatin for first-line treatment of advanced colorectal cancer? Ann Oncol 2005; 16: 845–46. 8 Punt CJ. New options and old dilemmas in the treatment of patients with advanced colorectal cancer. Ann Oncol 2004; 15: 1453–59. 9 Hoff PM, Ansari R, Batist G, et al. Comparison of oral capecitabine versus intravenous fluorouracil plus leucovorin as first-line treatment in 605 patients with metastatic colorectal cancer: results of a randomized phase III study. J Clin Oncol 2001; 19: 2282–92. 10 van Cutsem E, Twelves C, Cassidy J, et al. Oral capecitabine compared with intravenous fluorouracil plus leucovorin in patients with metastatic colorectal cancer: results of a large phase III study. J Clin Oncol 2001; 19: 4097–106. 11 Borner MM, Dietrich D, Stupp R, et al. Phase II study of capecitabine and oxaliplatin in first- and second-line treatment of advanced or metastatic colorectal cancer. J Clin Oncol 2002; 20: 1759–66.

142

12

13

14

15

16

17

18

19

20

21

22

Rea DW, Nortier JW, ten Bokkel Huinink WW, et al. A phase I/II and pharmacokinetic study of irinotecan in combination with capecitabine as first-line therapy for advanced colorectal cancer. Ann Oncol 2005; 16: 1123–32. Therasse P, Arbuck SG, Eisenhauer EA, et al. New guidelines to evaluate the response to treatment in solid tumors. European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada. J Natl Cancer Inst 2000; 92: 205–16. Tournigand C, Andre T, Achille E, et al. FOLFIRI followed by FOLFOX6 or the reverse sequence in advanced colorectal cancer: a randomized GERCOR study. J Clin Oncol 2004; 22: 229–37. Koopman M, Antonini NF, Douma J, et al. Randomised study of sequential versus combination chemotherapy with capecitabine, irinotecan and oxaliplatin in advanced colorectal cancer, an interim safety analysis. A Dutch Colorectal Cancer Group (DCCG) phase III study. Ann Oncol 2006; 17: 1523–28. Seymour MT, Maughan TS, Ledermann JA, et al. Different strategies of sequential and combination chemotherapy for patients with poor prognosis advanced colorectal cancer (MRC FOCUS): a randomised controlled trial. Lancet 2007; 370: 143–52. Bajetta E, Di BM, Mariani L, et al. Randomized multicenter phase II trial of two different schedules of irinotecan combined with capecitabine as first-line treatment in metastatic colorectal carcinoma. Cancer 2004; 100: 279–87. Borner MM, Bernhard J, Dietrich D, et al. A randomized phase II trial of capecitabine and two different schedules of irinotecan in first-line treatment of metastatic colorectal cancer: efficacy, quality-of-life and toxicity. Ann Oncol 2005; 16: 282–88. Kohne C, de Greve J, Bokemeyer C, et al. Capecitabine plus irinotecan versus 5-FU/FA/irinotecan +/- celecoxib in first line treatment of metastatic colorectal cancer. Safety results of the prospective multicenter EORTC phase III study 40015. J Clin Oncol 2005; 23 (suppl 16s): 252s. Fuchs CS, Marshall J, Mitchell E, et al. A randomized trial of first-line irinotecan/fluoropymidine combinations with or without celecoxib in metastatic colorectal cancer (BICC-C). J Clin Oncol 2006; 24 (suppl 18s): 147s. Haller DG, Cassidy J, Clarke S, et al. Tolerability of fluoropyrimidines appears to differ by region. J Clin Oncol 2006; 24 (suppl 18s): 149. Colucci G, Gebbia V, Paoletti G, et al. Phase III randomized trial of FOLFIRI versus FOLFOX4 in the treatment of advanced colorectal cancer: a multicenter study of the Gruppo Oncologico Dell’Italia Meridionale. J Clin Oncol 2005; 23: 4866–75.

www.thelancet.com Vol 370 July 14, 2007