Aflibercept for treating visual impairment caused by macular oedema secondary to central retinal vein occlusion

The Committee considered the populations in COPERNICUS and GALILEO. The Committee acknowledged that the trials did not exclude people with ischaemia or severe ischaemia. The Committee heard from the manufacturer that aflibercept was effective across the full trial populations of COPERNICUS and GALILEO. However, the Committee heard from the clinical specialists and the ERG

that the proportions of patients with ischaemia or severe ischaemia in the trials were uncertain because different definitions of ischaemia and severe ischaemia exist. The Committee accepted that aflibercept could be considered effective for all of the population included in the trials.

The Committee considered the evidence presented by the manufacturer on the clinical effectiveness of aflibercept. It acknowledged that the main sources of evidence came from the COPERNICUS and GALILEO randomised controlled trials, which compared aflibercept with sham injection in people with CRVO. The Committee noted that in both COPERNICUS and GALILEO, aflibercept was associated with statistically significantly more eyes gaining 15 or more letters at 24 weeks compared with sham injection. The Committee was aware that people in the sham groups could receive aflibercept after 24 weeks in COPERNICUS and after 52 weeks in GALILEO. The Committee agreed that aflibercept resulted in greater visual gains when it was given to patients soon after diagnosis. The Committee concluded that aflibercept is a clinically effective treatment for visual impairment caused by macular oedema secondary to CRVO compared with sham injection.

The Committee considered the evidence for the clinical effectiveness of aflibercept compared with ranibizumab and dexamethasone. It noted that evidence from direct comparisons was not available, and that a network meta-analysis was presented by the manufacturer. The Committee heard from the ERG that the methods used in the network meta-analysis were appropriate and that the analysis was well conducted. The Committee agreed that in the absence of a direct comparison, the results could be used to inform decisions about the clinical effectiveness of aflibercept compared with ranibizumab and dexamethasone. The Committee agreed that given the nature of the evidence, there was some uncertainty about the clinical effectiveness of aflibercept compared with ranibizumab and dexamethasone, but concluded that there was no evidence that aflibercept was not as clinically effective as ranibizumab or dexamethasone.

The Committee considered the evidence for adverse effects associated with aflibercept. It noted that the overall frequency of adverse events in the COPERNICUS and GALILEO trials was low. The Committee heard from the clinical specialists that the safety profile of aflibercept is similar to that of ranibizumab, which is already licensed for use in this condition (see NICE's technology appraisal guidance on ranibizumab for treating visual impairment caused by macular oedema secondary to retinal vein occlusion) . The Committee concluded that treatment with aflibercept had a similar adverse event profile to ranibizumab.

The Committee considered the manufacturer's economic model and the critique and exploratory analyses performed by the ERG. The Committee noted that the manufacturer had presented a comparison of aflibercept with ranibizumab for its base-case analysis and a comparison of aflibercept with dexamethasone as a scenario analysis. The Committee acknowledged that the ERG had concerns about some of the assumptions made by the manufacturer in the base-case analysis. In particular, the ERG queried why:

• any net gain in visual acuity at 24 weeks was assumed to persist through the lifetime of the model

• the relative risk of losing 15 or more letters at 24 weeks was not included in the model

• the duration of treatment was assumed to be 1 year

• EQ-5D data from GALILEO was used as a source of utility values

• the costs of adverse events were not included in the model

• a stopping rule was not included in the model

• the estimated administration costs of aflibercept and ranibizumab were high

• the estimated costs of blindness were low.
  
  The Committee considered each of these issues in turn, as detailed below.

The Committee discussed the manufacturer's assumption that the benefits of treatment at 24 weeks would continue indefinitely throughout the lifetime of the model. The Committee heard from the ERG that the assumption may be optimistic and that it would exaggerate the quality-adjusted life year (QALY) gain of aflibercept over ranibizumab and dexamethasone. The Committee acknowledged that a sensitivity analysis around the duration of benefit of treatment was not undertaken by the manufacturer or the ERG. The Committee heard from the ERG that if the benefits of treatment at 24 weeks had not been assumed to continue indefinitely it was unlikely to change the overall results of the manufacturer's base-case analysis because aflibercept would still dominate (be less expensive and more effective than) ranibizumab. The Committee accepted that it was not appropriate to assume that the duration of treatment benefit at 24 weeks would continue indefinitely but concluded that it was likely to have little impact on the cost-effectiveness estimates.

The Committee discussed how the manufacturer's model did not incorporate the relative risk of losing 15 or more letters. The Committee heard from the ERG that incorporating the relative risk of losing 15 or more letters into the model would ensure that the transition probabilities were correctly calculated. The ERG highlighted that including the relative risk of losing 15 or more letters in its exploratory analysis, which included the confidential discount applied to the list price for aflibercept, did not affect the dominance of aflibercept over ranibizumab except when a 50% reduction was applied to the list price of ranibizumab (see section 3.43). The Committee accepted that the relative risk of losing 15 or more letters should have been included in the manufacturer's model but concluded that including it was unlikely to change the dominance of aflibercept over ranibizumab.

The Committee discussed the manufacturer's assumption that the duration of aflibercept treatment was 1 year. The Committee heard from the clinical specialists that most patients still receive treatment with anti-VEGF therapy in year 2 and around a third of patients receive treatment in year 3. The clinical specialists also noted that data on the difference in the length of treatment with ranibizumab or aflibercept for macular oedema secondary to CRVO are not yet available. The Committee considered the manufacturer's scenario analyses in which the treatment duration of aflibercept was extended to 2 years or 4 years (see section 3.25) and the ERG's exploratory sensitivity analyses in which treatment duration was extended to 2 or 5 years and which included the confidential discount applied to the list price for aflibercept and a range of discounts from 0 to 50% applied to the list price of ranibizumab (see section 3.43). The Committee noted that when treatment duration was extended to 5 years, aflibercept was dominant over ranibizumab regardless of the discount applied to the ranibizumab list price. The Committee concluded that the duration of aflibercept treatment was likely to be longer than 1 year, and that increasing the duration of treatment up to 5 years did not change the dominance of aflibercept over ranibizumab.

The Committee considered the source of health-related quality-of-life data used in the manufacturer's model. The Committee acknowledged that EQ-5D data from the European subset of the GALILEO population were used as a source of utility values in the manufacturer's base-case analysis, which meets NICE's reference case. The Committee also considered the manufacturer's scenario analyses, which used Czoski-Murray utility values, and included the confidential discount applied to the list price for aflibercept, noting that this did not affect the dominance of aflibercept over ranibizumab. The Committee noted that the ERG also carried out an exploratory sensitivity analysis using utility values from Czoski-Murray and Brown, which included the confidential discount applied to the list price for aflibercept and a range of discounts from 0 to 50% applied to the list price of ranibizumab. The Committee noted that the use of these utility values did not affect the dominance of aflibercept over ranibizumab unless there was a 50% reduction in the list price of ranibizumab. The Committee concluded that the source of the utility values did not substantially affect the cost-effectiveness estimates of aflibercept compared with ranibizumab.

The Committee considered how the cost of adverse events had not been included in the manufacturer' base-case analysis. It noted that the reason given in the manufacturer's submission was that anti-VEGFs have similar safety profiles to each other. The Committee also acknowledged that the manufacturer presented a scenario analysis that included adverse events and that adverse events were also included in the ERG's exploratory analyses (see sections 3.25 and 3.43). The Committee noted that aflibercept continued to dominate ranibizumab in the manufacturer's scenario analysis and in the ERG's exploratory analysis if adverse events were included and a discount of 0 to 45% in the list price of ranibizumab was applied. The Committee was aware that the incremental cost-effectiveness ratio (ICER) for aflibercept was less than £3,500 per QALY gained when a 50% discount was applied to the list price of ranibizumab. The Committee concluded that even if adverse events had been included in the manufacturer's base-case analysis, it was likely that aflibercept would continue to dominate ranibizumab.

The Committee discussed the fact that the manufacturer's model did not include a stopping rule. The Committee noted that the summary of product characteristics for aflibercept states that continued treatment is not recommended if there is no improvement in visual and anatomic outcomes over the course of the first 3 injections. It was also aware that the summary of product characteristics for ranibizumab for macular oedema secondary to retinal vein occlusion also states that continued treatment is not recommended if there is no improvement in visual acuity after 3 injections. The Committee heard from the clinical specialists that they would not continue treatment if there was no improvement in visual acuity after 3 injections. The Committee concluded that a stopping rule should have been included in the manufacturer's model.

The Committee discussed the administration costs of aflibercept and ranibizumab included in the manufacturer's model. The Committee noted that the manufacturer used a weighted average cost, based on the assumption that 52.38% of patients would be treated in an outpatient setting and the remaining in a day-case setting. The Committee heard from the clinical specialists that most patients would be treated as outpatients; however, it also heard that not all units have the facilities to perform the treatment as an outpatient procedure. The Committee noted that the ERG presented lower costs of administration of aflibercept and ranibizumab in its report. The Committee concluded that the ERG's assumptions about the costs of administration were likely to be more realistic than those used by the manufacturer and therefore it was uncertain of the impact on the cost-effectiveness analyses.

The Committee discussed the costs of blindness included in the manufacturer's model. The Committee was aware that the source of the estimated costs of depression associated with blindness used by the ERG was more recent than the source used by the manufacturer. The Committee also noted that the costs of blindness presented by the ERG in their report were higher than those in the base-case analysis. The Committee recognised that if the costs presented by the ERG had been used, aflibercept would continue to dominate ranibizumab. The Committee concluded that the ERG's assumption about the costs of blindness were likely to be more in line with clinical practice than those used by the manufacturer.

The Committee considered the ICERs for aflibercept compared with ranibizumab estimated by the manufacturer and the ERG. It noted that these analyses incorporated the discount agreed in the patient access scheme for aflibercept and a range of discounts applied to the list price of ranibizumab. The Committee was aware of the actual discount agreed in the patient access scheme for ranibizumab (this was submitted as commercial in confidence and therefore cannot be presented). It agreed that the analyses undertaken by the manufacturer and the ERG captured the discount agreed in the patient access scheme for ranibizumab. The Committee noted that in the manufacturer's base-case analysis aflibercept dominated ranibizumab when the discounted price of aflibercept was used. The Committee considered the concerns raised by the ERG about the manufacturer's model and acknowledged the ERG's amendments to the manufacturer's model (see section 3.40). The Committee was aware that the ERG's exploratory analysis resulted in slightly more cost savings for aflibercept, and that aflibercept continued to dominate ranibizumab. The Committee also discussed the ERG's exploratory analysis around the list price of ranibizumab, incorporating the confidential discount on the list price for aflibercept. The Committee noted that a reduction of 0 to 45% in the list price of ranibizumab did not affect the dominance of aflibercept over ranibizumab. It also considered that when a 50% reduction was applied to the list price of ranibizumab, with the exception of the scenario of including the relative risk of losing 15 or more letters, the ICERs for aflibercept compared with ranibizumab ranged from £750 to £9,300 per QALY gained. Taking into account the exact magnitude of the discounts agreed in the patient access schemes for aflibercept and ranibizumab, the Committee concluded that aflibercept was a cost-effective use of NHS resources compared with ranibizumab for treating people with visual impairment caused by macular oedema secondary to CRVO.

The Committee discussed the manufacturer's scenario analysis comparing aflibercept with dexamethasone, and the ERG's exploratory analyses. The Committee noted that the ERG's exploratory analysis, which included the confidential discount applied to the list price for aflibercept, resulted in an ICER of £12,300 per QALY gained for aflibercept compared with dexamethasone. The Committee considered that using Czoski-Murray utility values resulted in an ICER of £12,900 per QALY gained and using Brown utility values resulted in an ICER of £16,800 per QALY gained when applied to the 'worse-seeing eye'. The Committee also acknowledged that even using the Brown utilities for the 'better-seeing eye', that is to say, the 'worst case scenario', the ICER was below the top end of the range that would normally be considered a cost-effective use of NHS resources (£20,000 to 30,000 per QALY gained). The Committee concluded that aflibercept was a cost-effective use of NHS resources compared with dexamethasone for treating people with visual impairment caused by macular oedema secondary to CRVO.

The Committee discussed how innovative aflibercept is in its potential to make a significant and substantial impact on health-related benefits. It agreed that aflibercept as well as other anti-VEGF treatments were a substantial improvement over previous treatments, and considered that this improvement applied to the class of drugs. In addition there were no substantial benefits of aflibercept over its comparators that were not already captured in the QALY estimation in the modelling.