Bortezomib for induction therapy in multiple myeloma before high-dose chemotherapy and autologous stem cell transplantation

Clinical effectiveness

4.4 The Committee noted that the manufacturer presented evidence on the clinical effectiveness of bortezomib from the PETHEMA, GIMEMA, HOVON and IFM trials. The Committee was aware that the HOVON trial included the bortezomib, doxorubicin and dexamethasone regimen, which was not included in the bortezomib marketing authorisation. It therefore excluded it from its clinical and cost-effectiveness discussions. It noted that the trials included different regimens of bortezomib and had different study designs. The Committee noted that in the GIMEMA trial patients only received 3 cycles of bortezomib, thalidomide and dexamethasone whereas in the PETHEMA trial patients received 6 cycles, which was more in line with the marketing authorisation. Moreover, the Committee noted that the PETHEMA and GIMEMA trials differed in the number of stem cell transplants given and the type of consolidation (intensification therapy to sustain remission before lower-dose maintenance therapy) and maintenance treatments used after transplant, and that consolidation and maintenance treatment was not standard clinical practice in the NHS. The Committee was aware that none of the trials included a comparison with cyclophosphamide, thalidomide and dexamethasone, which is standard clinical practice in England.

4.5 The Committee discussed the results from the PETHEMA and IFM trials. It noted that in the PETHEMA trial, bortezomib, thalidomide and dexamethasone was associated with a statistically significant gain in overall response rate after induction compared with thalidomide and dexamethasone (84.6% compared with 61.4%, p<0.001) and that this was maintained after stem cell transplant (77.7% compared with 56.7%, p<0.001). The Committee also noted that, in the IFM trial, bortezomib and dexamethasone was associated with a similar gain in overall response rate compared with vincristine, doxorubicin and dexamethasone (77.1% compared with 60.7%, p<0.001) but a statistically significant difference was not shown after stem cell transplantation (79.6% compared with 74.4%, p=0.179). It noted that progression-free survival was longer in the bortezomib, thalidomide and dexamethasone arm of the PETHEMA trial than in the thalidomide and dexamethasone arm, and that the difference was statistically significant (hazard ratio [HR] 0.65, 95% confidence interval [CI] 0.45 to 0.92, p=0.015). It further noted that progression-free survival was longer in the bortezomib and dexamethasone arm of the IFM trial than in the vincristine, doxorubicin and dexamethasone arm, but the difference was not statistically significant (HR 0.88, 95% CI 0.70 to 1.11, p value not reported). The Committee agreed that induction treatment with bortezomib and dexamethasone was associated with statistically significant improvements in post-induction overall response rates compared with vincristine, doxorubicin and dexamethasone, and that induction treatment with bortezomib, thalidomide and dexamethasone resulted in statistically significant improvements in overall response rates (post-induction and post-stem cell transplant) and progression-free survival compared with thalidomide and dexamethasone. However, it concluded that no direct evidence was available to compare the efficacy of bortezomib, thalidomide and dexamethasone or bortezomib and dexamethasone against cyclophosphamide, thalidomide and dexamethasone, the comparator regimen considered to be current standard care in the UK and therefore the most relevant comparator for the Committee's decision-making.

4.6 The Committee considered the unadjusted overall survival reported in the manufacturer's submission for the PETHEMA and IFM trials. It noted that median overall survival was not reached in either the PETHEMA trial (bortezomib, thalidomide and dexamethasone compared with thalidomide and dexamethasone, hazard ratio 0.80, 95% CI 0.48 to 1.34, p=0.393) or the IFM trial (bortezomib and dexamethasone compared with vincristine, doxorubicin and dexamethasone, HR 0.8, 95% CI 0.54 to 1.19, p value not reported) and there was no statistically significant difference in overall survival between the treatment arms in each study. The Committee heard from the manufacturer and from the clinical specialists that the duration of the trials was too short to allow differences in overall survival to be seen between treatment groups. The clinical specialists also stated that given the differences in trial design relating to the numbers of stem cell transplants and types of maintenance treatment received it was not possible to draw firm conclusions. The Committee concluded that although there was uncertainty in the magnitude of overall survival associated with bortezomib, it was plausible that bortezomib's impact on induction response could be associated with improved overall survival.

4.7 The Committee considered the manufacturer's indirect comparison. It acknowledged the manufacturer's rationale that a network could not be formed to conduct an indirect comparison between the available trials and assumptions to overcome the network limitations would generate considerable uncertainties and unreliable results. At the appraisal consultation stage, the Committee requested further analysis comparing single arms of the available trials, adjusting for the differences between the trial designs and baseline characteristics of the patients included in each study. The Committee concluded that although there would be limitations to this approach, further analysis from the manufacturer would provide useful comparative data to draw conclusions on the relative effectiveness of bortezomib, thalidomide and dexamethasone, and bortezomib and dexamethasone, compared with the most relevant comparator cyclophosphamide, thalidomide and dexamethasone, or an alternative comparator in situations in which the cyclophosphamide, thalidomide and dexamethasone regimen is considered inappropriate.

4.8 The Committee considered the adverse events associated with using a bortezomib-containing regimen. It heard from the clinical specialists that intravenously administered bortezomib had been associated with peripheral neuropathy, but that rapid dose reductions could effectively manage it. In addition, the clinical specialists highlighted that although the evidence presented was for intravenously administered bortezomib, the introduction of a subcutaneous formulation has substantially reduced the side effects related to peripheral neuropathy and also reduced the need for thromboprophylaxis. The Committee concluded that the adverse event profile of bortezomib was manageable (for full details of adverse reactions and contraindications, see the summary of product characteristics).

Cost effectiveness

4.9 The Committee considered the structure, assumptions and results of the manufacturer's economic model, which was based on data from the PETHEMA and IFM trials for the bortezomib, thalidomide and dexamethasone and the bortezomib and dexamethasone regimens respectively. The Committee was aware that the model only provided comparisons for bortezomib, thalidomide and dexamethasone against thalidomide and dexamethasone and for bortezomib and dexamethasone against vincristine, doxorubicin and dexamethasone, which was not in line with current clinical practice in the UK, and that the vincristine, doxorubicin and dexamethasone regimen was outside of the scope of this appraisal. The Committee acknowledged that there was no direct evidence available to compare the bortezomib, thalidomide and dexamethasone and the bortezomib and dexamethasone regimens with cyclophosphamide, thalidomide and dexamethasone, and it asked the manufacturer to further explore this by conducting a further indirect comparison using single arms from the relevant clinical trials.

4.10 The Committee considered the manufacturer's approach to using post-induction rates from the PETHEMA and IFM trials in the economic model for the bortezomib, thalidomide and dexamethasone and the bortezomib and dexamethasone regimens respectively. It noted the ERG's comments that data from the MRC Myeloma VII trial used by the manufacturer to estimate long-term survival related to maximal response to treatment rather than post-induction response rate and therefore using post-stem cell transplant response rates from the PETHEMA and IFM trials might better predict progression-free survival and overall survival. The clinical specialists stated that post-stem cell transplant response rates would be more appropriate as long as they were based on an intention-to-treat analysis that included all people who were randomised in the trials regardless of whether they received treatment or not; however, if they were based only on patients who received a transplant, post-induction response rates would be more meaningful. The ERG confirmed that post-stem cell transplant rates in both the PETHEMA and GIMEMA trials were based on an intention-to-treat analysis. The Committee concluded that using post-stem cell transplant response rates rather than post-induction response rates was more appropriate.

4.11 The Committee considered the way in which long-term survival had been extrapolated in the manufacturer's model. The Committee was aware that that the model extrapolated level of response after induction therapy to long-term survival and time to progression based on data from the MRC Myeloma VII trial. It noted the ERG's comments that the MRC Myeloma VII trial was not very recent because it had recruited patients between 1993 and 2003, and that its rates for overall survival and progression-free survival were likely to be lower than would be seen in current clinical practice. The Committee was also aware that although long-term survival end points had not been reached in the PETHEMA and IFM trials, the data available in these trials suggested that the manufacturer's model, using data from MRC Myeloma VIl, underestimated overall survival. It noted that other data were available (for example, from a study by Alvares and the NMSG 5/94 study) that fit better with the data observed in the PETHEMA and IFM trials. The Committee concluded that the ERG's exploratory analysis using data from the Alvares and NMSG 5/94 studies was appropriate (see section 3.34) for estimating long-term survival and should be considered together with the analysis based on the MRC Myeloma VII study.

4.12 The Committee considered the costs used in the manufacturer's economic model. It was aware that the clinical trials were all conducted using intravenously administered bortezomib, and it heard from the manufacturer that this was also assumed in the economic model and was therefore associated with a day-patient cost. The Committee was, however, aware that bortezomib is available as a subcutaneous formulation which is widely used and would therefore only be associated with an outpatient cost. The Committee also noted the comments from the clinical specialists that the subcutaneous formulation could reduce the risk of peripheral neuropathy and also reduce the need for thromboprophylaxis. The Committee considered that these issues combined might reduce the total cost of bortezomib in the model. The manufacturer updated the model to reflect this during consultation to reduce costs from £203 to £197 for first attendance and from £284 to £211 for subsequent visits.

4.13 The Committee noted that the manufacturer's comparison of bortezomib and dexamethasone with vincristine, doxorubicin and dexamethasone was outside the NICE scope because the comparator did not contain thalidomide, and in addition, the clinical specialists commented that the vincristine, doxorubicin and dexamethasone regimen is no longer used as an induction therapy for multiple myeloma in the UK. The Committee decided, therefore, that it was not appropriate to consider the results from this comparison. However, the Committee also noted that the bortezomib and dexamethasone regimen had a lower acquisition cost than the bortezomib, thalidomide and dexamethasone regimen as it did not include thalidomide, and heard from clinical specialists that it would provide a valuable treatment option, especially for patients who cannot tolerate thalidomide. The Committee was aware that the ERG's exploratory analyses also included comparisons of bortezomib and dexamethasone against thalidomide and dexamethasone, and comparisons of bortezomib and dexamethasone against cyclophosphamide, thalidomide and dexamethasone. Based on the analyses incorporating the Alvares survival data and post-stem cell transplant response rates that were preferred by the Committee for the bortezomib, thalidomide and dexamethasone analysis, the Committee noted that the ERG's exploratory ICER for bortezomib and dexamethasone compared with thalidomide and dexamethasone was £26,700 per QALY gained. It also noted that in the comparison of bortezomib and dexamethasone with cyclophosphamide, thalidomide and the dexamethasone, the bortezomib and dexamethasone regimen was dominated by (that is, it was more costly and less effective than) cyclophosphamide, thalidomide and dexamethasone. The Committee acknowledged the lack of an appropriate comparison in the manufacturer's submission and the caveats surrounding the ERG's exploratory analysis, and asked the manufacturer to further explore the cost effectiveness of the bortezomib and dexamethasone (see section 4.15).

4.14 The Committee noted the manufacturer's original comparison of bortezomib, thalidomide and dexamethasone with thalidomide and dexamethasone, which resulted in a deterministic incremental cost-effectiveness ratio (ICER) of £20,500 per QALY gained, and that the manufacturer had not presented probabilistic ICERs. The Committee then discussed the results of the ERG's exploratory analyses. The Committee noted that using post-stem cell transplant response rates instead of post-induction response rates (see section 4.10) resulted in an ICER of £26,300 per QALY gained for bortezomib, thalidomide and dexamethasone compared with thalidomide and dexamethasone. The Committee also noted that using data from the Alvares study instead of the MRC Myeloma VII study to model long-term survival resulted in an ICER of £30,400 per QALY gained for bortezomib, thalidomide and dexamethasone compared with thalidomide and dexamethasone. The clinical specialists raised concerns that the Alvares study was retrospective in design and also quite old and therefore a more recent study would be more appropriate. The ERG confirmed that it had also conducted an exploratory analysis based on data from the NMSG 5/94 study and that this resulted in an ICER higher than £30,400 per QALY gained, that but this did not include the discontinuation rule and was therefore not presented (see section 3.34). The Committee noted that given the disparity between the overall survival results from the trials and those in the model, using an alternative data source such as the Alvares study, which was a better fit to the trial data, was appropriate. The Committee noted that incorporating post-stem cell transplant rates and using the Alvares study to inform overall survival together resulted in an ICER of £39,000 per QALY gained for bortezomib, thalidomide and dexamethasone compared with thalidomide and dexamethasone. The Committee concluded that based on the analyses that were available before the appraisal consultation document was released, the ERG's exploratory analyses were appropriate and that £39,000 per QALY gained was an appropriate starting point for discussion on the most plausible ICER for bortezomib, thalidomide and dexamethasone compared with thalidomide and dexamethasone.

4.15 The Committee was aware of the ERGs further exploratory analyses on bortezomib, thalidomide and dexamethasone compared with cyclophosphamide, thalidomide and dexamethasone. The Committee understood that there were limitations with this approach because the data were drawn from a range of heterogeneous studies containing different comparators and different study designs and therefore could not be directly compared. Moreover, differences in trial design and baseline characteristics had not been taken into account. The Committee noted that this approach resulted in an ICER of £228,200 per QALY gained for bortezomib, thalidomide and dexamethasone compared with cyclophosphamide, thalidomide and dexamethasone when using the manufacturer's base-case assumptions. Applying post-stem cell transplant response rates and survival data from the Alvares study resulted in an ICER of £82,000 per QALY gained for bortezomib, thalidomide and dexamethasone compared with cyclophosphamide, thalidomide and dexamethasone. The Committee agreed that although there was considerable uncertainty associated with such an approach, the ICER based on the analyses that were available before the appraisal consultation document was released for bortezomib, thalidomide and dexamethasone compared with cyclophosphamide, thalidomide and dexamethasone was likely to be higher than the ICER of £39,000 per QALY gained compared with thalidomide and dexamethasone. The Committee considered that further analyses were needed to explore the cost effectiveness of the bortezomib and dexamethasone regimen and asked the manufacturer to present:

4.16 The Committee considered the manufacturer's response to the appraisal consultation document that included changes to the economic models for all the bortezomib regimens (see section 3.38). The Committee considered the manufacturer's revised modelling assumption that incorporated an 11.8 month survival benefit for people who received a stem cell transplant. The Committee noted that the model included survival curves for complete responders in addition to the assumption of survival benefit from receiving a stem cell transplant and discussed whether this amounted to double counting. It heard from the clinical specialists that incorporating the additional survival benefit was not double counting because the purpose of a stem cell transplant is to increase the depth of response, which could provide additive effect resulting in a survival benefit of up to a year. The Committee heard from the manufacturer that the additional survival benefit assumption was not a key driver of cost effectiveness because removing this benefit had minimal effect on the ICERs. The Committee concluded that the manufacturer's approach to modelling stem cell transplant benefit was acceptable for its decision making.

4.17 The Committee was aware that the manufacturer used the stem cell transplant rates from the PETHEMA trial to inform the comparison of bortezomib, thalidomide and dexamethasone against thalidomide and dexamethasone (80.8% and 61.4% respectively) but that other bortezomib trials (GIMEMA, HOVON and IFM) indicated rates were above 80% regardless of the treatment arm. The Committee was also aware that transplant rates for the comparison of bortezomib and dexamethasone (without thalidomide) against cyclophosphamide, thalidomide and dexamethasone, were incorporated from the IFM trial for the bortezomib and dexamethasone arm (89.1%) and from the MRC Myeloma IX trial for the cyclophosphamide, thalidomide and dexamethasone arm (66.7%). The Committee noted the ERG's concerns that this was inconsistent with transplant rates taken from the control arm of the IFM trial (81.8%, which was directly comparable and would be expected to be no better than a cyclophosphamide, thalidomide and dexamethasone control arm). The Committee noted that the ERG had explored the impact of incorporating a stem cell transplant rate of 81.8%, which resulted in the manufacturer's revised base-case ICER increasing from £20,588 to £36,700 per QALY gained. The Committee discussed whether the differences in the transplant rates were too wide for both comparisons. It heard from the clinical specialists that in clinical practice cyclophosphamide, thalidomide and dexamethasone was associated with a stem cell transplant rate of approximately 50% and this was corroborated by 2 large population-based studies. The clinical specialists stated that bortezomib regimens were likely to be associated with stem transplant rates of 60–65%. The clinical specialists also suggested that stem cell transplant rates should better reflect complete response rates than the control arms of the other trials suggest. The Committee agreed that although a bortezomib regimen might be expected to improve the rate of stem cell transplant, it would be to a lesser extent than was modelled by the manufacturer, and more closely linked to response in clinical practice. However, the Committee also agreed that the differences in transplant rates between treatment arms in the models were plausible. The Committee concluded that although the impact of stem cell transplant rates included in the model on cost-effectiveness results was uncertain, it was unlikely to undermine the manufacturer's base-case cost-effectiveness results.

4.18 The Committee discussed the overall survival modelling informing the manufacturer's revised ICERs for the comparison of bortezomib, thalidomide and dexamethasone against thalidomide and dexamethasone. It noted that the manufacturer preferred the MRC Myeloma VII data as the source for long-term survival, which resulted in an ICER of £17,800 per QALY gained. The Committee heard from the ERG that the MRC Myeloma VII data did not fit the observed PETHEMA data. The Committee was aware that incorporating data from the ERG's preferred Alvares and NMSG 5/94 studies resulted in ICERs of £22,700 and £39,600 per QALY gained respectively. The manufacturer stated that caution should be taken in interpreting the fit with PETHEMA trial data beyond 30 months because of the level of censoring. The Committee questioned why data from the PETHEMA trial had not been used directly. The manufacturer argued that the data were immature. The Committee also heard from the clinical specialists that because patients in the PETHEMA trial could receive bortezomib as maintenance treatment (which is not standard practice in the UK) there would be substantial convergence between the treatment arms when more long-term survival data become available. The Committee recognised that even though there was little face validity in the modelling of survival, it appreciated that the Kaplan−Meier curve from the PETHEMA trial was confounded by post-induction treatment. The Committee noted that incorporating only the NMSG 5/94 data resulted in ICERs that would not normally be considered to be cost effective (above £30,000 per QALY gained). The Committee considered concerns raised by the manufacturer and consultees that the NMSG study did not provide fully relevant data because it did not report median overall survival for partial and non-responders. The Committee concluded that using data from the NMSG 5/94 study would represent a pessimistic scenario and that the ICERs based on survival data from the MRC Myeloma VII (£17,800 per QALY gained) and Alvares (£22,700 per QALY gained) studies were appropriate for its decision making.

4.19 The Committee considered the manufacturer's ICERs for bortezomib and dexamethasone compared with cyclophosphamide, thalidomide and dexamethasone, as requested in the appraisal consultation document. It noted that using the MRC Myeloma VII, Alvares and NMSG 5/94 data sources to inform overall survival in the model resulted in ICERS of £20,600, £24,300 and £33,400 per QALY gained respectively. It noted that only using the NMSG 5/94 study data resulted in ICERs above £30,000 per QALY gained. Having discussed the concerns around the NMSG 5/94 data (see section 4.18), the Committee considered that results based on these data were likely to represent a pessimistic scenario. The Committee also considered the analyses presented by manufacturer when cyclophosphamide, thalidomide and dexamethasone may not be suitable, noting that vincristine, doxorubicin and dexamethasone was included as the most appropriate comparator in this situation. The Committee noted that the ICERs ranged from £18,900 per QALY gained using MRC Myeloma VII survival data to £25,600 using survival data from the Alvares study. The Committee concluded that most plausible ICERs based on survival data from the MRC Myeloma VII and Alvares studies were appropriate for its decision making.

4.20 The Committee remained concerned that the modelling was subject to uncertainties, and that the manufacturer had also not provided sufficient external and internal validity. However, the Committee acknowledged that bortezomib regimens had a clear advantage with respect to induction response and that a link between improved response and survival was plausible. In particular, the Committee considered that there were people with clinically aggressive disease, with organ function at risk, or at risk of irreversible renal damage, who would benefit from a fast response associated with treatment with bortezomib, but heard that this group could not be categorically defined. Taking into account its consideration that the uncertainty around stem cell transplant rates was not likely to have a substantial impact on the ICER (section 4.16) and taking into consideration ICERs based on survival data from the MRC Myeloma VII and Alvares studies, the Committee concluded that, on balance, the ICERs for bortezomib, thalidomide and dexamethasone compared with thalidomide and dexamethasone, and for bortezomib and dexamethasone compared with cyclophosphamide, thalidomide and dexamethasone and compared with vincristine, doxorubicin and dexamethasone, were likely to be below £30,000 per QALY gained. Therefore both bortezomib regimens could be considered a cost-effective use of NHS resources.

Summary of Appraisal Committee's key conclusions