Cost-effectiveness of second-line treatment with irinotecan or infusional 5-fluorouracil in metastatic colorectal cancer

Annals of Oncology 11: 157-161, 2000. © 2000 Khmer Academic Publishers Printed in the Netherlands.

Original article Cost-effectiveness of second-line treatment with irinotecan or infusional 5-fluorouracil in metastatic colorectal cancer C. Levy-Piedbois,1 I. Durand-Zaleski, 2 H. Juhel,4 C. Schmitt,5 A. Bellanger4 & P. Piedbois 3 'institut Gustave Roussy, Department of Public Health and Informatics, Villejuif, Departments of 2 Public health, 3Oncology, Hopital Henri Mondor, AP-HP, Paris; 4 Pharmacy, Hopital Pitte-Salpetriere, AP-HP, Parts; iARCOS, Issy-les-Moulineaux, France

(continuous infusion, AIO, or LV5-FU2 regimens) to irinotecan alone. Costs were derived from the accounting system in Background: It has been shown that irinotecan is superior to two university hospitals in Paris, France. infusional 5-fluorouracil (5-FU) in patients with advanced Results: The range in total healthcare costs was 14,135 to colorectal cancer after 5-FU failure. In a recent trial, median 12,192 US$ patient between management approaches, with survival was 10.8 months for patients treated with irinotecan, irinotecan chemotherapy costing most and 5-FU-continuous compared to 8.5 months in patients receiving infusional 5-FU. infusion least. If survival was included as a treatment benefit, Considering the statistically significant but clinically relatively the cost-effectiveness ratio of irinotecan over 5-FU ranged small advantage of irinotecan over 5-FU, cost effectiveness from 9,344 to 10,137 USS per year of added survival. should also be part of treatment decision. Conclusions: The least expensive management for metaPurpose: To relate the costs of each management approach static colorectal was 5-FU infusion but the additional cost of to overall survival in patients with metastatic colorectal irinotecan was balanced by the added months of survival, with a cost-effectiveness ratio close to that of other cancer treatcancer. Patients and methods: The healthcare costs and medical ments. benefits (treatment-added survival) of second-line chemotherapy in patients (infusional 5-FU: 129, irinotecan: 127) were compared. Data on overall survival were drawn from a Key words: 5-fluorouracil, chemotherapy, colorectal cancer, multicenter randomised trial that compared infusional 5-FU cost /effectiveness analysis, irinotecan Summary

Introduction

Colorectal cancer is a major health care problem in developed countries [1], and management of the large number of patients with metastases places a heavy load on both oncology services and healthcare resources. In first-line treatment, four meta-analyses conducted by our group confirmed that 5-FU plus leucovorin [2], 5-FU plus methotrexate [3], 5-FU continuous infusion [4], or hepatic arterial infusion FUDR [5] were superior to intravenous 5-FU bolus. Until recently, most patients with metastatic colorectal cancer received low-dose infusional 5-FU when they experienced a tumour progression under bolus 5-FU regimens [6]. It has been showed that irinotecan yields a prolongation of overall survival in these patients compared to infusional 5-FU, with a similar quality of life in both groups. In a recent study, median survival was 10.8 months in patients receiving irinotecan, and 8.5 months in the 5-FU group (/* < 0.035) [7]. The acquisition costs of the drug irinotecan are much higher than those of 5-FU. Although the decision about which treatment is most appropriate for an individual patient should be based on patient and disease-related factors, budgetary constraints also re-

quire that treatment is allocated as cost-effectively as possible. As M. O'Connell [8] pointed out in the Commentary which accompanied the paper cited previously, "In view of the small benefit, the side-effects and cost of irinotecan must be judiciously weighted against the benefit." This requires comparable information about the costs and the effectiveness (prolonged survival) of the various treatments [9,10]. In the present study, we have related the costs of each management approach to its overall survival. Patients and methods Patients Irinotecan and infusional 5-fluorouracil (5-FU) patients were treated in a previously reported trial [7]. Infusional 5-FU patients were recruited into a randomised trial in the same centres and with the same entry criteria and follow-up procedure as irinotecan patients. Three 5-FU regimens were allowed: 5-FU continuous infusion with Lokich regimen [11], AIO [12], or LV5-FU2 [13]. 'Lokich' regimen consisted of fluorouracil 250-300 mg/m 2 /day. "AIO regimen'consisted of fluorouracil 2,600-3,000 mg/m 2 /day over 24 hours with or without folinic acid 20-500 mg/m 2 /day intravenously weekly for six weeks, with a two-week rest period between cycles. 'LV5-FU2' regimen con-

158 sisted of folinic acid 200 mg/m2 intravenously over two hours followed by fluorouracil 400 mg/m2 intravenous bolus, then fluorouracil 600 mg/m2 continuous infusion over 22 hours/day on the first two days of every two-week period. The choice of the regimen was left to each investigator, based on the patient previousfluorouracilregimen. Treatment was continued until tumour progression, unacceptable toxicity, patient refusal to continue chemotherapy, or death.

Health benefit measures Survival was estimated by a Kaplan-Meier method, from date of randomisation to death, whatever its cause, or last visit. Duration of survival was not discounted, to be consistent with the cost computations. Costs for cost-effectiveness analysis

Cost derivation Costs were computed from the viewpoint of the hospital, over the total duration of patient survival, or three-year follow-up. We did not include non hospital healthcare costs and non healthcare costs because data on these were not available from the initial trial and because, in France, there is little healthcare delivery for oncology patients outside the hospital. The hospital perspective does not misrepresent total healthcare costs in countries where healthcare is essentially delivered in the hospital even after failure of chemotherapy. Due to the fact that median survival was less than one year and the largest part of the costs were to be incurred during the initial chemotherapy cycles we did not introduce any discounting. Healthcare costs included: chemotherapy, hospital admissions for administration of chemotherapy and for subsequent complications, hospital outpatient clinic visits. Other drug treatments for symptom palliation were excluded under the assumption that they would be equivalent in each treatment arm. The costs of hospital admissions and tests, were drawn from the 1998 Henri Mondor and Pitie Salpetriere (Paris, France) cost accounting system and represent actual costs and not charges to the payer. Hospital costs comprised the cost of chemotherapy, follow-up outpatient clinic visits, chemotherapy and management of complications. Overall survival and costs were estimated from the time of randomisation until the death of the patient, or last visit. French francs were converted into USSusing the OECD purchasing power parity index (6/1).

The costs calculated were those borne by the hospital. The analysis assessed the extra cost/year of added survival (cost-effectiveness), comparing irinotecan to each of the three 5-FU regimens used in the control group [14, 15]. We computed three cost-effectiveness ratios, which correspond to regimens in different countries. We did not compute incremental cost effectiveness ratios comparing the three 5-FU regimen among themselves1 these regimens correspond to practice patterns and habits in different European countries which are unlikely to be changed by the result of a cost analysis. Moreover, the study was not designed to compare the costs of 5-FU infusion. The only effect of choosing a variety of comparators for irinotecan was that the costeffectiveness ratios would be relevant in a given country countries.

Sensitivity analysis The sensitivity analysis explored a range of 0.5-3.5 months for the gain in survival: this range was considered reasonable in the absence of data on the 95% confidence interval of the gain in survival in the initial paper [7, 8]. Another analysis explored the possibilities of suppressing severe diarrhoea (i.e., which would result in additional hospital costs) from the complications of irinotecan by using irinotecan associated with LV5-FU2 chemotherapy, which has been recently showed to be particularly interesting in terms of efficiency and tolerance [16]. We estimated this other cost-effectiveness ratio by adding the cost of LV5FU2 chemotherapy and subtracting the costs of diarrhoea form the irinotecan arm.

Treatment regimens We considered that all chemotherapy was administered by oncology nurses on an out-patient basis. Patients attended an out-patient consultation with the supervising oncologist every four weeks for examination, and review of blood tests and CT scans, which were performed at different intervals depending on regimen. When those tests were performed outside the hospital, their costs were not included, consistently with the viewpoint chosen. We performed a time-motion study to estimate the time spent in each protocol by hospital staff: oncologist, oncology nurses and head nurse, and secretary. Time spent was estimated from actual surveys and varied, mostly for MDs and nurses, with the type of chemotherapy administered. Time was valued using hospital salaries plus employees benefits for each category of personnel. The time motion-study used actual observation of the time spent for preparing and delivering care for each type of visit and thus the personnel costs take into account the differing intensity of each visit. We collected detailed information on the supplies used in each protocol. Those supplies were valued at their purchase price by the hospital. Drugs were valued at their purchase price. Total drug usage was the product of the actual daily prescription per patient by duration of treatment. The cost of complications occurring during the chemotherapy was estimated from the Diagnosis Related Groups (DRG) cost schedule currently used in France, by relating the cause of admission and the number of patients concerned to the cost of the corresponding DRG. This cost is computed yearly for a sample of public hospitals over the country and is not a reimbursement price. Second-line chemotherapy was stopped when disease progression occurred. From this time to the death of the patients or last visit, the costs of post-progression therapy were estimated from the number of admissions and the cause of hospitalisation as recorded in the initial trial. Patients were managed in oncology (80"'u 90'!') or in surgery (10%-20%) wards, and the costs were computed for each department.

Results Healthcare costs

Cost data was available for all patients included in the initial trials. The detailed costs of chemotherapy are presented Table 1, the total cost of the drug was 7,290 $/patient for irinotecan and ranged from $93— S608 S/patient for infusional 5-FU. The hospital resources used in treatment are shown in Table 2. This table includes only chemotherapy received on protocol, postprogression therapy was excluded. The costs of complications during the treatment period were higher in the irinotecan arm compared with in the infusional 5-FU arm (Table 3). The costs of management after chemotherapy failure were lower in the irinotecan arm, due to fewer hospital days per patient: although hospitalisations were more frequent in the irnotecan arm, total length of stay was shorter since the majority of admissions were outpatients (Table 4). The total cost of the treatment was S 14,135 for irinotecan patients versus 512,192-512,344 for infusional 5-FU (Table 5). When the added cost of irinotecan was related to the additional 2.3 month benefit, the costeffectiveness ratios ranged from 59,344-510,137 per year of added survival.

159 Table 1. Chemotherapy drug costs, for both cytotoxic drugs and. when applicable, folinic acid. All costs are in 1999 USS. Irinotecan arm (n = 127)

Infusional 5-FU arm (n = 129) LV5-FU2 (n = 35)

Planned dose per treatment session Irinotecan

5-FU

350 m g / n r every three weeks -

Folinic acid

-

Cost per treatment session (S) 1217 Actual number of treatment sessions administered per patient 5.99 Total drug cost per patient (S) 7290

Lokich (n = 39)

AIO (n = 55)

-

-

2000 mg/m 2 every two weeks 400 mg/m 2 every two weeks

-

270 mg/m 2 per day -

2600 mg/m 2 every week + two weeks rest 500 mg/m 2 every week + two weeks rest

Diarrhea Occlusion Diarrhea + infection Diarrhea + other Fever + neutropenia Anemia Pulmonary embolism Other Total

Irinotecan (n = 127)

5-FU (n - 129)

Number of patients

Total cost

Number of patients

7 7 6 9 3 2 1 17 52

20.773 20,020 17.805 15,435 8,903 7,075 6,875 29,315 126,201

1 6 0 6 1 0 0 11 25

Total per patient 1

30

3.6

92 90

12 94

204

93

608

Irinotecan («=127) LV5-FU2 (« = 35)

6.0

13.6

Total cost 2,968 17,160 0

10,290 2,968 0 0 28,000 61,474

476

994

47

Table 4. Cost of patient treatments (best supportive care) after the termination of chemotherapy. Costs are in 1999 USS.

Table 2 Cost of chemotherapy administration, excluding the cost of drugs, for the total duration of chemotherapy cycles. Costs are in 1999 USS.

Number of hospital visits per patient Labour costs per visit Nurses Head nurses Ancillaries Physicians Secretaries Total labour cost per patient Supplies and outpatient costs per patient Port-a-cath insertion procedure Total drug administration costs cost per patient

Table 3. Cost of complications during chemotherapy. Complications were retrieved for patients with irinotecan and for all patients with 5-FU regimens, regardless of the type of regimen used. Costs are in 1999 USS.

5-FU (n = 129)

Total hospital days per patient Number of hospitahsations/patient Total cost per patient

5-FU (n = 129)

8.5 52 4,908

14.4 24 8.743

Table 5 Total hospital costs per patient, in 1999 USS.

Lokich AIO (n = 39) (« = 55)

13.3

Irinotecan (n = 127)

Irinotecan (n = 127) LV5-FU2 (n = 35)

129

18 10 7 13 5 320

33 15 12 7 3 952

25 10 7 5 2 643

537

623

1960

1347

1090

0

890

890

890

943

3802

2880

2517

22 10 7 2 2

5-FU (A; = 129)

Chemotherapy: drug + administration Complications during chemotherapy Follow-up Total cost per patient

Lokich AIO (« = 39) (n = 55)

8,233

3,006

2,973

3,125

994 4,908 14,135

476 8,743 12,225

476 8,743 12,192

476 8,743 12,344

during the treatment period, the total cost of the irinotecan arm became $17,708 and the cost-effectiveness ratio $29,373 per year of added survival compared to the Lokich regimen.

Sensitivity analysis Discussion When the benefit in survival varied between 0.5 and 3.5 months, the cost-effectiveness ratio ranged from $3,000$45,000 per year of added survival (Figure 1). When we considered combining LV5-FU2 to irinotecan, the total cost of chemotherapy (including the drug, the supplies and the staff) became $12,243 per patient. Assuming that this combination of chemotherapy suppressed the cost of admissions for diarrhoea

The patients included in this study were drawn from a randomised study to allow comparison of irinotecan and infusional 5-FU in the treatment of metastatic colorectal cancer. There was a 14% variation in total hospital cost between the various management approaches studied (Table 5), with the Lokich regimen costing least. The additional healthcare cost of irinotecan over infu-

160 90000

80000 -

0,5

1,0

1,5

2,0

+3,0

3,5

months of added survival —•—irinotecan vs LV5FU2 —m- irinotecan vs Lokich irinotecan vs AIO Figure 1. Cost-effectiveness (USS per year of added survival) of irinotecan versus infusional 5-FU. Costs are healthcare costs only; nonhealthcare costs were excluded. Costs were those associated with each treatment; effectiveness ranged from 0.5 to 3.5 months of added survival. The arrow indicates the base case.

sional 5-FU was close to $10,000 per life year gained depending on the infusional 5-FU regimen chosen. A cost issue clearly related to the organisation of healthcare delivery is the magnitude of follow-up procedures costs. These costs were incurred by the hospital because few alternative facility is available for delivering supportive care after the failure of chemotherapy. In countries where such alternatives to hospital care exist, the added cost of treating patients with irinotecan would be close to $9,000 compared with infusional 5-FU regimens which would result in a cost-effectiveness ratio of $50,000 per added life-year. The sensitivity analysis suggested that combining LV5-FU2 to irinotecan increases the costeffectiveness ratio compared with the Lokich regimen. Table 1 shows that the unit cost of irinotecan drug is from 9-80 times more expensive than infusional 5-FU regimens. However, the higher cost of irinotecan is counterbalanced by the lower costs of its administration (reduced personnel costs) and the lower disease costs after progression. Irinotecan has received approval in most countries for the treatment of patients with metastatic colorectal cancer after failure of 5-FU. The 5-FU regimens considered in this study encompass practice patterns in most European countries and some centres in the USA. Finally, the hospitals chosen in the present study to value

resources used are tertiary care university centres, which do not differ from centres used to managed such patients in other countries. However, supportive care was a major cost driver, if only treatment and complication costs were considered the cost-effectiveness ratio differ by an order of magnitude: thus, decisions regarding the external validity of our results should consider whether hospice and skilled nursing facilities available in the country would reduce the cost of supportive care and drive the cost-effectiveness ratio up to $50,000 per life year. Our objective was to assist the decision making process within the hospital. The question answered was whether the high acquisition costs of the drug were possibly offset by savings on other line items. This cost analysis did not include ambulatory costs and non healthcare costs which are not borne by the hospital. Further studies are required to document prospectively the ambulatory costs (such as patient transportation or nurses visits) and non healthcare costs. In summary our results suggested that the least expensive management for metastatic colorectal was 5-FU infusion, but that the additional cost of irinotecan was balanced by the added months of survival, with a cost-effectiveness ratio close to that of other cancer treatments and acceptable with regards to the criteria

161 proposed by Laupacis et al. [17, 18]. The added 2.3 months of survival might be considered, in the view of some, as a small benefit, although, in relative terms, it represents a 25% improvement for patients with a 8.5 months life expectancy.

Acknowledgements The authors are indebted to AERO (European Association for Research in Oncology) and to ARCOS (Analyses et Recherches en Economie de la Sante) for their support, to Rhone-Poulenc Rorer Laboratories for providing individual patient data, and to Dr E. Antoine for his assistance.

References 1. Williams N. Colorectal cancer: Epidemiology, aetiology, pathology, clinical features and diagnosis. In Keighley M. Williams N (eds). Surgery of the Anus, Rectum and Colon. Vol. 1. Oxford: WB Saunders, 1993; 830-3. 2. Advanced Colorectal Cancer Meta-Analysis Project. Modulation of 5-fluorouracil by leucovorin in patients with advanced colorectal cancer: Evidence in terms of response rate. J Clin Oncol 1992; 10: 896-903. 3. Advanced Colorectal Cancer Meta-Analysis Project. Meta-analysis of randomized trials testing the biochemical modulation of 5-fluorouracil by methotrexate in metastatic colorectal cancer. J Clin Oncol 1994; 12:960-9. 4. Meta-Analysis Group In Cancer. Efficacy of intravenous continuous infusion of 5-fluorouracil compared with bolus administration in advanced colorectal cancer. J Clin Oncol 1998; 16: 301-8. 5. Meta-Analysis Group in Cancer. Reappraisal of hepatic arterial infusion in the treatment of non-resectable liver metastases from colorectal cancer. J Natl Cancer Inst 1996; 88: 252-8. 6. Thinon P, Cunningham D, Findlay M et al. Pooled analysis of phase II trials with low-dose 5-fluorouracil continuous infusion (LD 5-FU CI) as a second-line chemotherapy in advanced colorectal cancer. Proc Am Soc Clin Oncol 1998; 17: 1047. 7. Rougier P, Van Cutsem E, Bajetta E et al. Randomised trial of irinotecan versus fluorouracil by continuous infusion after fluorouracil failure in patients with metastatic colorectal cancer. Lancet 1998; 352: 1407-12.

8. O'Connell MJ. Irinotecan for colorectal cancer: A small step forward. Lancet 1998; 352: 1402. 9. Drummond M. Stoddart G. Torrance G. Methods for Economic Evaluation of Health Care Programmes. Oxford: Oxford University Press 1989; 18-34. 74-83. 10. Task Force on Principles for Economic Analysis of Health Care Technology. Economic analysis of health care technology. A report on principles. Ann Intern Med 1995; 122: 61-71. 11. Lokich JJ, Ahlgren JD. Gullo JJ et al A prospective randomized comparison of continuous infusion fluorouracil with a conventional bolus schedule in metastatic colorectal carcinoma: A MidAtlantic Oncology Program study. J Clin Oncol 1989; 7: 425-32. 12. Weh MJ, Wilke HJ. Dierlamm J et al. Weekly therapy with folinic acid (FA) and high-dose fluorouracil (F-FU) 24-hour infusion in pretreated patients with metastatic colorectal carcinoma Ann Oncol 1994; 5: 233-7. 13. De Gramont A, Bosset JF, Milan C et al. Randomized trial comparing monthly low-dose leucovorin and fluorouracil bolus with bimonthly high-dose leucovorin and fluorouracil bolus plus continuous infusion for advanced colorectal cancer: A French intergroup study. J Clin Oncol 1997; 15: 808-15. 14. Canadian Coordinating Office for Health Technology Assessment. Guidelines for the Economic Evaluation of Drugs. Ottawa, Canada 1994. 15. Weinstein MC, Siegel JE, Gold MR et al. Recommendations of the panel on cost-effectiveness and health. JAMA 1996; 276: 1253-8. 16. Douillard JY, Cunningham D. Roth AD et al. A randomized phase III trial comparing irinotecan (IRI) + 5-FU/folinic acid -FA) to the same schedule of 5-FU/FA in patients (pts) with metastatic colorectal cancer (MCRC) as front-line chemotherapy (CT). Proc Am Soc Clin Oncol 1999; 18: 899. 17. Laupacis A, Feeny D, Detsky A et al. How attractive does a new technology have to be to warrant adoption and utilization? Tentative guidelines for using clinical and economic evaluations. Can Med Assoc J 1992; 146: 473-81. 18. Smith TJ, Hillner BE, Desch CE Efficacy and cost-effectiveness of cancer treatment: Rational allocation of resources based on decision analysis. J Natl Cancer Inst 1993; 85: 1460-74. Received 6 December 1999; accepted 14 December 1999. Correspondence to: I. Durand-Zaleski, MD, PhD Sante Publique, Hopital Henri Mondor 51 avenue du Marechal de Lattre deTassigny 94010 Creteil France E-mail: isabelle.durand-zaleski(ahmn.ap-hop-pans.fr