Bevacizumab in combination with fluoropyrimidine-based chemotherapy for the first-line treatment of metastatic colorectal 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 08/211/01. The assessment report began editorial review in November 2009 and was accepted for publication in March 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.

Declared competing interests of authors

none

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

Colorectal cancer is the third most common cancer in the UK, with 36,766 new cases diagnosed in England and Wales in 2005. 2 Metastatic disease is, in the majority of cases, incurable and treatment is palliative in nature. Although local radiotherapy and, less commonly, surgery both have a role, metastatic disease is essentially a systemic disease requiring systemic treatment. Traditionally this has meant cytotoxic chemotherapy although, in recent years, passive immunotherapy in the form of monoclonal antibody treatment has been added to chemotherapy regimens. Commonly used regimens include oral capecitabine monotherapy, oral capecitabine + intravenous (IV) oxaliplatin (XELOX), IV 5-fluorouracil + folinic acid + oxaliplatin (FOLFOX), IV 5-fluorouracil + folinic acid + irinotecan (FOLFIRI), IV 5-fluorouracil ± folinic acid, and oral capecitabine + IV irinotecan (XELIRI). With current standard first-line chemotherapy, median survival is around 15–20 months. 3–5

Scope of the evidence review group report

The objective of the appraisal was to evaluate the clinical effectiveness and cost-effectiveness of bevacizumab, within its licensed indications, in combination with oxaliplatin and either 5-fluorouracil or capecitabine for the treatment of metastatic colorectal cancer. The comparator was oxaliplatin or irinotecan, including chemotherapy regimens without bevacizumab. Measured outcomes included overall survival, progression-free survival, response rate, adverse effects of treatment and health-related quality of life.

The licensed indication permits the use of bevacizumab in combination with fluoropyrimidine-based chemotherapy for the treatment of patients with metastatic colorectal cancer but does not specify a line of treatment. The NICE scope included the use of bevacizumab in combination with oxaliplatin-based chemotherapy in individuals with histologically confirmed metastatic colorectal cancer as first-line therapy (for patients not previously treated for metastatic disease), and as second-line therapy. The manufacturer’s submission (MS), however, focuses on first-line use only.

The main evidence presented in support of the clinical effectiveness of bevacizumab was based on one phase III, multicentre, multinational, randomised, open-label study (NO16966 trial). 6 This two-arm study was originally designed to demonstrate the non-inferiority of XELOX compared with FOLFOX-4 in adult patients with histologically confirmed metastatic colorectal cancer who had not previously been treated. Following randomisation of 634 patients, the open-label study was amended (additional phase II and III studies that were published demonstrated the benefit of adding bevacizumab to irinotecan, 5-fluorouracil and folinic acid)7,8 to include a 2 × 2 factorial randomised (partially blinded for bevacizumab) phase III trial (n = 1401) with the coprimary objective of demonstrating superiority of bevacizumab in combination with chemotherapy (B-XELOX or B-FOLFOX-4) compared with chemotherapy alone (P-XELOX or P-FOLFOX-4). The dose of bevacizumab was 5 mg/kg every 2 weeks (B-FOLFOX-4) or 7.5 mg/kg every 3 weeks (B-XELOX).

The scope of the manufacturer’s cost-effectiveness submission focused on a comparison with regimens containing oxaliplatin which was considered to be the most relevant comparator. A comparison with irinotecan-based chemotherapy was also included for completeness. The manufacturer submitted additional analyses in response to the ERG clarification questions. Further data and analyses were also submitted following the first committee meeting. These included further data on the patient access scheme’s (PAS’s) operating costs as well as pharmacy and preparation costs for bevacizumab.

Summary of submitted clinical evidence

The manufacturer’s main analysis pooled data from the initial two arms and the 2 × 2 factorial part of the NO16966 trial and compared the addition of bevacizumab to chemotherapy with chemotherapy alone. The manufacturer’s primary pooled analysis of superiority (using the intention-to-treat population) showed that after a median follow-up of 28 months, the addition of bevacizumab to chemotherapy (B-XELOX/B-FOLFOX-4 combined) significantly improved progression-free survival and overall survival compared with chemotherapy alone (P-XELOX/P-FOLFOX-4/XELOX/FOLFOX-4 combined) in adult patients with histologically confirmed metastatic colorectal cancer who were not previously treated [median progression-free survival 9.4 vs 7.7 months (absolute difference 1.7 months); hazard ratio (HR) 0.79, 97.5% confidence interval (CI) 0.72 to 0.87; p = 0.0001; median overall survival 21.2 vs 18.9 months (absolute difference 2.3 months); HR 0.83, 97.5% CI 0.74 to 0.93; p = 0.0019].

A secondary pooled analysis of superiority (requested by the ERG as it was believed to be more appropriate) restricted to patients in the second 2 × 2 factorial part of the NO16966 study as per the original statistical trial plan (B-XELOX/B-FOLFOX-4 combined vs P-XELOX/P-FOLFOX-4 combined) found similar results [median progression-free survival 9.4 vs 8.0 months (absolute difference 1.4 months); HR 0.83, 97.5% CI 0.72 to 0.95; p = 0.0023; median overall survival 21.3 versus 19.9 months (absolute difference 1.4 months); HR 0.89, 97.5% CI 0.76 to 1.03; p = 0.0769].

The manufacturer’s pooled analysis of non-inferiority (using the eligible patient population and the intention-to-treat population) showed that the XELOX (XELOX/P-XELOX/B-XELOX combined) and FOLFOX-4 (FOLFOX-4/P-FOLFOX-4/B-FOLFOX-4 combined) based regimens were equivalent for both progression-free survival and overall survival (p-values were stated as not significant, but these values were not reported). No analysis was undertaken for the factorial design (P-XELOX/B-XELOX combined versus P-FOLFOX-4/B-FOLFOX-4 combined).

A pre-defined subgroup analysis on progression-free survival found that the statistical superiority of bevacizumab plus chemotherapy was evident in the XELOX subgroups (B-XELOX vs P-XELOX; HR 0.80, 97.5% CI 0.66 to 0.96; p-value not reported) but did not reach the significance level in the FOLFOX-4 subgroups (B-FOLFOX-4 vs P-FOLFOX-4; HR 0.89, 97.5% CI 0.74 to 1.06; p-value not reported). Additional post hoc exploratory analyses, following the results from the Adjuvant Colon Cancer End Points (ACCENT) study,9 found that there was a significant and direct correlation between time to recurrence after surgery and survival after recurrence in patients whose disease recurred after surgery and adjuvant treatment. Removing the subgroup of patients that may have slower tumour progression after adjuvant treatment (an imbalance between treatment groups with regard to an important prognostic factor that was not recognised at the start of the NO16966 trial) significantly improved (i.e. lowered) the HRs for adding bevacizumab to chemotherapy compared with chemotherapy alone for both overall survival and progression-free survival. Depending on the analyses conducted (e.g. exclusion of patients with prior adjuvant chemotherapy from all four treatment arms of the factorial study, from FOLFOX groups only or from P-FOLFOX group only) the HRs for overall survival ranged from 0.83 to 0.85 (p < 0.03) and the HRs for progression-free survival ranged from 0.74 to 0.77 (p < 0.0001). Although this may be plausible, the ERG notes that caution should be exercised as this is a post hoc exploratory analysis.

The majority of adverse events were generally associated with cytotoxic chemotherapy. FOLFOX-4-based regimens were generally associated with increased neutropenia/granulocytopenia, and XELOX-based regimens were generally associated with increased diarrhoea and hand and foot syndrome. Adverse events that could be potentially related to bevacizumab included increased frequencies of high blood pressure, proteinuria, bleeding, gastrointestinal perforation, thromboembolic events and wound healing complications. Serious (grade 3) or life threatening (grade 4) adverse events that occurred more commonly in patients receiving bevacizumab plus chemotherapy (B-XELOX/B-FOLFOX-4 combined) than those receiving chemotherapy alone (P-XELOX/P-FOLFOX-4/XELOX/FOLFOX-4 combined) were thromboembolic events (7.8% vs 5.1%, respectively), hypertension (4.0% vs 0.8%, respectively), proteinuria (3.5% vs 0.9%, respectively) and bleeding problems (1.9% vs 1.5%, respectively). Grade 3 and 4 gastrointestinal perforations and wound healing complications were rare (< 1%). Similar results were observed when data were restricted to the factorial analyses.

The rates of discontinuation were higher in the bevacizumab containing groups (B-XELOX/B-FOLFOX-4 combined, 30.8%) than in the no bevacizumab containing groups (P-XELOX/P-FOLFOX-4/XELOX/FOLFOX-4 combined, 25.3%), Corresponding data, restricted to the 2 × 2 factorial analyses, yielded similar results (B-XELOX/B-FOLFOX-4 combined, 30.8% vs P-XELOX/P-FOLFOX-4 combined, 20.8%). The statistical analysis comparing the rates of discontinuation between treatment groups was not reported in the MS or in the manufacturer’s supplementary evidence.

Summary of submitted cost-effectiveness evidence

Cost-effectiveness was estimated using a Microsoft excel model with four states: pre-progression on treatment, pre-progression and post treatment, progressive disease and dead. An area under the curve approach was used to estimate the disease progression of metastatic colorectal cancer patients. The distribution of patients between health states was used to calculate total direct costs and quality-adjusted life-years (QALYs) for each intervention. Costs were considered from an NHS and Personal Social Services perspective. Cost-effectiveness was expressed in terms of incremental cost per QALY with a time horizon of 8 years, which is equivalent to a lifetime horizon in the population of interest. The analysis focused on the interventions B-XELOX and B-FOLFOX. The model was populated with efficacy data from the N016966 trial but as discussed in the clinical effectiveness section these trial data have been analysed in several different ways. Data on treatment duration and dose intensity were also based on the N016966 trial. Survival data were modelled using Kaplan–Meier data up to median survival of 28 months and a Weibull distribution after this point. The ERG requested several changes to the model inputs and modelling assumptions (including additional analyses).

A summary of the key incremental cost-effectiveness ratios (ICERs) included in the submission are presented in Table 1. Of the several analyses presented by the manufacturer, the ERG considered the analysis using the 2 × 2 part of the N016966 trial, with the XELOX and FOLFOX arms pooled, with patients with prior adjuvant treatment excluded to be the most appropriate. This analysis produced ICERs of £36,006 and £31,174 for B-XELOX versus XELOX and B-FOLFOX versus FOLFOX, respectively. The inclusion of patients with prior adjuvant chemotherapy resulted in higher ICERs. Unpooling the XELOX and FOLFOX arms affected the individual XELOX and FOLFOX ICERs in different directions. While no systematic review was undertaken with irinotecan as a comparator, a cost-effectiveness analysis was undertaken (data not presented here).

Strengths

The NO16966 trial was of reasonable methodological quality (with some limitations) and measured a range of outcomes that were as appropriate and clinically relevant as possible. The ERG believed that the modelling structure employed was appropriate.

Weaknesses

Despite no evidence to suggest that the statistical validity of the factorial approach was methodologically inappropriate, the validity of simply pooling data from essentially two different study designs (i.e. a two-arm design and a 2 × 2 factorial design) without accounting for between-study variability is inappropriate. Unweighted (for uncertainty) pooling of results from different studies is not advisable as there are almost certainly differences between trials that, if not accounted for, are likely to lead to biased estimates of effect. The appropriateness of combining data from the two parts of the study was also questioned by the European Medicines Agency. 10 The resulting pooled data (manufacturer’s primary pooled analysis of superiority and non-inferiority) should therefore be treated with caution. Additionally it is unclear whether patients with prior adjuvant chemotherapy should be excluded from the analysis.

The restriction to the trial data from the 2 × 2 part of the NO16966 study, the pooling of the XELOX and FOLFOX arms, and the restriction to the data of patients without prior adjuvant chemotherapy all had a large impact on the resulting ICERs.

The MS did not make use of the range of utility values identified from the literature review and did not explain why these values were not used. The sources of the utility values used in the MS were poorly referenced, resulting in the ERG being unable to verify them. The distributions used for the utility values in the probabilistic sensitivity analyses (PSA) reflected the uncertainty relating to the specific values used but underestimated the uncertainty relating to the selection of utility values. The ERG noted that using wider distributions for utility values would significantly increase the CIs around the mean ICERs from the PSA, and reducing the utility values by 20% markedly increased the ICERs.

Chemotherapy can be administered intermittently or continuously, but the difference in cost and effectiveness between intermittent and continuous treatment is unclear. Current care in England is often intermittent treatment with chemotherapy, but the trial and the model both represent continuous treatment chemotherapy. It is unclear how this difference may impact the ICERs but, as an example, if intermittent treatment was cheaper than continuous treatment whilst having a similar efficacy, then the ICER for continuous treatment with bevacizumab versus intermittent treatment would be greater than the ICER for continuous treatment with bevacizumab versus continuous treatment.

In clinical practice, treatment with non-oxaliplatin chemotherapy components may continue beyond oxaliplatin cessation although in the N016966 trial this was rarely seen. Because of the structure of the PAS (in which oxaliplatin is received free of charge), the incremental cost of continuing bevacizumab after oxaliplatin cessation is almost three times the incremental cost of adding bevacizumab to oxaliplatin. Hence the impact of continuing bevacizumab treatment on the ICERs could be considerable.

Under the PAS, bevacizumab has a fixed price per cycle, but for calculations without the PAS it is important that drug wastage should be included for both oxaliplatin and bevacizumab. The MS ‘without PAS’ ICERs did not include drug wastage within the base case although bevacizumab wastage was included within one analysis as stated in Table 1. The inclusion of drug wastage resulted in higher ICERs.

Disclaimers

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