Capecitabine for the treatment of advanced gastric cancer

Article history

The research reported in this article of the journal supplement was commissioned and funded by the HTA programme on behalf of NICE as project number 08/95/01. The assessment report began editorial review in June 2010 and was accepted for publication in July 2010. See the HTA programme website for further project information (www.hta.ac.uk). This summary of the ERG report was compiled after the Appraisal Committee’s review. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HTA editors and publisher have tried to ensure the accuracy of the authors’ report and would like to thank the referees for their constructive comments on the draft document. However, they do not accept liability for damages or losses arising from material published in this report. The views expressed in this publication are those of the authors and not necessarily those of the HTA programme or the Department of Health.

Declared competing interests of authors

M Sculpher has a minority shareholding in a consulting company that has undertaken work for Roche during the last 3 years, but not in the clinical area of gastric cancer. He did not participate personally in this consultancy. M Seymour is a co-investigator in a trial (‘321GO’), which includes capecitabine as treatment for patients with advanced gastroesophageal cancer. The trial is peer reviewed and funded by Cancer Research UK, but also receives some supplementary financial support from Roche (£50k over 2 years). M Seymour also attended the ASCO Oncology Conference last year as a guest of Roche. D Suh is also a co-investigator on the ‘321GO’ study. Roche have also offered to sponsor his trip to ASCO this year. Stephen Kelly accepted financial support from Roche in 2008 and 2009 to attend the British Oncology Pharmacists Association Annual Symposium on behalf of the Leeds Teaching Hospitals NHS Trust Pharmacy Department.

Copyright statement

© 2010 Queen’s Printer and Controller of HMSO. This journal is a member of and subscribes to the principles of the Committee on Publication Ethics (COPE) (http://www.publicationethics.org/). This journal may be freely reproduced for the purposes of private research and study and may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NETSCC, Health Technology Assessment, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.

2010 Queen’s Printer and Controller of HMSO

Description of the underlying health problem

Gastric cancer is the 10th most commonly diagnosed cancer in the UK, with approximately 7000 new cases diagnosed in England and Wales every year. Of these, some 80% of patients are unsuitable for curative treatment and are candidates for palliative chemotherapy. It is estimated that just over one-half (around 2900) of these patients with advanced gastric cancer (aGC) receive such treatment.

The standard UK practice for patients with aGC, who are considered fit enough, has consisted of a triplet regimen comprising a fluoropyrimidine, intravenous 5-fluorouracil (5-FU) in combination with a platinum agent (cisplatin or oxaliplatin) and an anthracycline (epirubicin).

Scope of the ERG report

The decision problem addressed was the use of capecitabine (Xeloda) in combination with platinum-based chemotherapy regimens (cisplatin or oxaliplatin) with or without epirubicin, compared with 5-FU in combination with such regimens, in patients with inoperable aGC. Capecitabine is an oral prodrug of 5-FU, and is licensed for the first-line treatment of aGC in combination with a platinum-based regimen. Oral chemotherapies are usually considered to be preferred by patients and may have fewer associated costs and/or adverse events.

The outcome measures considered were overall survival (OS), progression-free survival (PFS), response rates, adverse effects of treatment and health-related quality of life (QoL). These were assessed in direct comparisons by two open-label non-inferiority randomised controlled trials (RCTs), assessing doublet or triplet regimens. 3,4

Economic outcomes included cost per life-year gained (LYG) and cost per quality-adjusted life-year (QALY) gained. The manufacturer proposed evaluating the cost-effectiveness of capecitabine-based regimens compared with 5-FU-based regimens by using cost-minimisation analyses (CMAs).

Methods

The ERG report comprised a critical review of the evidence for the clinical effectiveness and cost-effectiveness of the technology, based upon the manufacturer’s/sponsor’s submission (MS) to NICE as part of the STA process. The ERG checked the literature searches and carried out a search for ongoing trials. The review methodology including inclusion criteria was appraised. The validity assessments of the included RCTs were critiqued and the ERG carried out its own assessment using the CRD guidelines for the critical appraisal of RCTs.

In evaluating the cost effectiveness of capecitabine, the manufacturer used a cost-minimisation approach. The ERG has thus first commented on the appropriateness of using such methodology, within this specific decision problem, taking into consideration NICE’s reference case methods. 5 Next, the ERG assessed the manufacturer’s de novo economic evaluation using Drummond et al. ’s checklist. 6 In response to the ERG’s points of clarification regarding the initial submission, the manufacturer provided additional evidence on the costs of adverse events, drug acquisition inputs and costs of additional survival. The ERG considered this evidence throughout. Based on the identified limitations in the MS, the ERG revisited the base case according to drug use, unit costs of treatments and pharmacy drug preparation costs. The ERG also undertook additional sensitivity analyses based on the revised base case, and conducted a threshold analysis, evaluating the maximum costs that the NHS would be willing to pay for the extension of survival time implied by prespecified cost-effectiveness thresholds.

Summary of submitted clinical evidence

The MS focused on direct evidence from two phase III non-inferiority RCTs. 3,4

Efficacy outcomes from these trials were pooled in an individual patient data (IPD) meta-analysis. 7 REAL-2 (Randomized ECF for Advanced and Locally advanced oesophagogastric cancer-2) was a 2 × 2 factorial trial that compared 5-FU with capecitabine and cisplatin with oxaliplatin. 3 The following regimens were used: epirubicin  +  cisplatin  +  5-FU (ECF); epirubicin  +  cisplatin  +  capecitabine (ECX); epirubicin  +  oxaliplatin  +  5-FU (EOF); and epirubicin  +  oxaliplatin  +  capecitabine (EOX). A second trial, ML17032, compared cisplatin  +  capecitabine (CX) with cisplatin  +  5-FU (CF). 4

REAL-2 found statistically significant non-inferiority of capecitabine on the primary outcome of OS assessed in the per-protocol population adjusted [hazard ratio (HR) 0.89, 95% confidence interval (CI) 0.77 to 1.02]. ML17032 found statistically significant non-inferiority of capecitabine on the primary outcome of PFS in the per-protocol population (adjusted HR 0.85, 95% CI 0.65 to 1.11). Statistically significant non-inferiority on OS (unadjusted HR 0.85, 95% CI 0.64 to 1.13) was also demonstrated.

The IPD meta-analysis of the intention-to-treat (ITT) populations of the REAL-2 and ML17032 trials found a statistically significant benefit in OS for capecitabine compared with 5-FU (unadjusted HR 0.87, 95% CI 0.77 to 0.98, p = 0.027). 7

There was minimal QoL data reported in the MS. The REAL-2 trial was reported as assessing QoL using the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire C30 (EORTC-30), version 3,8 administered at baseline, and after 3, 6, 9 and 12 months. The manufacturer subsequently provided the levels of compliance, data on the baseline scores for all subscales, and changes from baseline at 12 weeks and 24 weeks for the REAL-2 trial. These showed few statistically significant differences between the individual trial arms.

Safety analyses showed some significant differences in adverse events profiles between capecitabine and 5-FU regimens. However, in the REAL-2 trial, all statistical analyses were pairwise comparisons with the ECF arm, which the trial was not powered to assess. Of particular note was grade 3 or 4 neutropenia, which occurred significantly more often in the ECX arm (p < 0.05) and significantly less often in the EOX and EOF arms (p < 0.01) compared with the ECF arm; grade 3 or 4 diarrhoea, which occurred significantly more often in the EOX and EOF arms compared with the ECF arm (p < 0.05); and grade 3 or 4 hand–foot syndrome, which occurred significantly more often in the ECX arm compared with the ECF arm (p < 0.05). In the ML17032 trial, stomatitis occurred more often, and with greater severity, in the CF arm, while hand–foot syndrome was more common in the CX arm.

Summary of submitted cost-effectiveness evidence

The manufacturer’s literature search identified one economic evaluation relevant to this decision problem. This was Roche’s 2007 submission to the Scottish Medicines Consortium for capecitabine in this indication. The methods and results reported are consistent with the current submission.

The MS included a de novo economic evaluation based on a CMA, where the costs of capecitabine-based regimens were compared with their equivalent 5-FU-based regimens (ECX vs ECF, EOX vs EOF, CX vs CF) in the treatment of advanced gastric cancer. A time horizon of 5.5 cycles (each lasting for 21 days) was used in the base-case analysis, representing the duration of treatment of the alternative regimens.

The cost-minimisation approach was based on evidence from two clinical trials – REAL-2 and ML17032 – reporting that capecitabine is at least as effective as intravenous 5-FU (see above). The calculations considered costs relating to drug acquisition and to drug administration. The drug administration costs comprised the costs of hospital visits (central line insertion, delivery of chemotherapy, and subsequent care by a nurse to flush central line and change the pump), pharmacy drug preparation costs, ambulatory pump costs and NHS transport costs (Table 1). The manufacturer assumed that there were no significant economically important differences in the incidence or severity of adverse events between capecitabine and 5-FU-based regimens, and therefore the costs of treatment-related adverse events were not included in the analysis. After a request for clarifications, the manufacturer presented the expected costs associated with the relevant adverse events.

The results of the manufacturer’s base-case analysis showed that capecitabine regimens are cost saving compared with their equivalent 5-FU-based regimens. The total net cost savings for capecitabine-based regimens were £1620 (ECX vs ECF), £1572 (EOX vs EOF), and £4210 (CX vs CF). Capecitabine remained cost saving in the manufacturer’s one-way sensitivity analysis, scenario analysis and worst-case analysis. The manufacturer did not conduct probabilistic sensitivity analysis or subgroup analysis. The submission also included a threshold analysis that explored the additional effectiveness (in terms of QALYs) needed for 5-FU to be considered cost-effective.

Commentary on the robustness of submitted evidence

The MS appears to include all relevant evidence from completed RCTs with respect to the question of efficacy; the ERG’s search revealed no additional completed RCTs, although one additional ongoing trial was located.

The ERG identified a number of issues and errors in the review process, which had the potential to exclude relevant studies. However, it did not appear that this had impacted on the results of the review.

Two RCTs that directly addressed the comparison between capecitabine and 5-FU in combination with platinum in the licensed population were included.

The REAL-2 trial was large (n = 1002), adequately powered, and closely reflective of UK standard practice. The patient population was also representative of those UK patients who were considered fit enough for standard chemotherapy, although these patients are significantly younger than the UK aGC patient population as a whole. The trial included a majority of patients who are outside the licensed indication, having advanced inoperable cancer of the oesophagus or gastroesophageal junction. The ERG’s clinical experts confirmed that treatment for each of these cancers would follow the same course as that for advanced inoperable gastric cancer. There was also no evidence of a statistically significant difference in prognosis based on primary tumour location. 9

The ML17032 (n = 316) trial assessed doublet therapy, which the ERG’s clinical advisors indicated would be used in patients who were considered unable to tolerate triplet therapy. However, such doublets would be given at a lower dose than was used in the trial. The trial population was also unrepresentative of UK patients, being younger and having a different ethnic composition. When the non-inferiority analyses of efficacy outcomes were performed using a margin of 1.25 relative to the efficacy of 5-FU, rather than 1.40 as the protocol had specified, the trial had only 50% power to detect statistically significant non-inferiority.

Both trials were necessarily open-label, and REAL-2 was unblinded for all outcomes, whereas for ML1703 the MS reported blinded outcome assessment only for the primary outcome of PFS. The ERG requested these independently assessed data for the outcomes of tumour response and adverse events. The manufacturer subsequently supplied these data for response rates and, although differences in the data sets were present, there was no indication of systematic bias.

The primary weakness of the initial MS was the limited QoL data. This is of particular importance where the decision problem centres on an issue of clinical non-inferiority and patient preference.

With respect to economic evaluation, the ERG deems CMA to be an appropriate framework with which to analyse the decision problem. However, it should be noted that the appropriateness of using such an approach is dependent not only on clinical evidence from the REAL-2 and ML17032 trials, but also on evidence relating to QoL and adverse events. The weaknesses identified above regarding the evidence presented by the manufacturer are therefore relevant. The manufacturer has also failed fully to consider uncertainty when justifying the use of CMA.

Overall, cost estimates for the CMA were generated appropriately. The ERG identified a number of shortcomings and potential uncertainties related to resource utilisation, unit costs, utilities and sensitivity analysis in the MS. However, these were considered minor, and additional analysis provided by the manufacturer, and further evaluations by the ERG, showed no impact on the overall conclusions. Results from the MS and ERG’s additional analysis are compared in Table 2.

A full probabilistic analysis was not performed, so the probabilities that capecitabine is less and more costly than its comparators have not been formally quantified. However, the mean estimates, sensitivity analyses and worst-case scenario are consistent in suggesting that capecitabine has a lower mean cost than 5-FU-based regimens. Further analyses exploring the additional costs incurred by the NHS resulting from extending survival duration show that these are unlikely to have a material effect on decision-making regarding capecitabine.

Disclaimers

The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the Department of Health.