Pfizer's Oxbryta is advised as a treatment option for haemolytic anaemia caused by sickle cell disease in individuals aged 12 and above, whether used with or without hydroxycarbamide. This recommendation applies under the conditions that individuals cannot tolerate or are ineligible for hydroxycarbamide, or if hydroxycarbamide alone proves insufficiently effective.
Clinical effectiveness
The HOPE trial provided data over a period of 72 weeks, with an additional 48 weeks of data from its open-label extension. However, this trial did not demonstrate the long-term impact of voxelotor on sickle cell disease complications. Furthermore, no significant differences were observed between voxelotor and placebo for certain short-term outcomes, including the number and frequency of vaso-occlusive crises, health-related quality of life, and the need for acute transfusions. The company clarified that the HOPE trial was not designed to assess these outcomes.
Clinical experts noted the challenge in determining voxelotor's potential to reduce long-term complications, as there is currently no clinical evidence to support this. Nonetheless, they suggested that because voxelotor increases haemoglobin levels, it might reduce the risk of long-term complications associated with haemolytic anaemia. They also highlighted the silent damage caused by chronic haemolytic anaemia in sickle cell disease, which can lead to end-organ damage over time. There is growing evidence that chronic haemolytic anaemia can impact cardiac function and bone density due to the increased workload on the heart. The committee acknowledged the difficulties in providing long-term evidence for voxelotor's efficacy in reducing complications but noted that according to NICE's health technology evaluation manual, surrogate outcomes should have strong evidence showing that the technology's effect on these endpoints predicts the final outcome. This evidence is ideally obtained from randomised controlled trials, or if not feasible, from epidemiological or observational studies. The company emphasized the plausible biological link between lower haemoglobin levels and worse outcomes, supported by epidemiological studies, which aligns with a level 2 surrogate relationship per the NICE manual. The committee agreed that while it was plausible that voxelotor could reduce long-term complications in sickle cell disease, the lack of evidence resulted in significant uncertainty about the extent and nature of its effects.
Economic Model and Company’s Approach
The company submitted a discrete event simulation model to evaluate the cost-effectiveness of voxelotor compared to standard care for treating haemolytic anaemia in sickle cell disease. This model considers events on a time-to-event basis. The committee found this approach methodologically valid for estimating the cost-effectiveness of medicines, appreciating its flexibility in incorporating disease history and competing risks.
Regular Transfusion Therapy with Standard Care
In response to initial guidance consultations, the company presented results from consultations with nine UK haematologists conducted between March and April 2023, estimating the rate of regular transfusions under standard care. This rate was higher than what was used in the company's base case. Additionally, an estimated rate from UK clinicians in 2020 was lower than the base case rate (exact rates are confidential). The company argued that the modified Delphi panel's rate used in the base case was reasonable, falling between various clinical estimates. The Evidence Assessment Group (EAG) acknowledged the difficulties in estimating regular transfusion rates under standard care and found the company's preference for the modified Delphi panel’s rates reasonable based on the presented data. Despite the uncertainty, the committee concluded that the most plausible transfusion rate, considering the evidence, was that from the company's modified Delphi panel.
Quality of Life and the HOPE Trial
In the HOPE trial, no significant difference was observed in EQ-5D scores between the voxelotor and standard-care groups at 72 weeks. The company expressed concerns about the effective use of EQ-5D in the trial, citing missing data and unexpectedly high baseline EQ-5D values for sickle cell disease patients. They also noted that the trial did not capture the impact of long-term complications on quality of life. Instead of using direct trial data, the company analyzed EQ-5D data from a patient journey survey to assess the relationship between haemoglobin levels and quality of life. They estimated a utility benefit per 1 g/dl increase in haemoglobin, applying this benefit in the model for both arms (the exact benefit is confidential).
Patient and clinical experts commented that EQ-5D might not fully capture the true quality of life for sickle cell disease patients, given the chronic nature of the condition. The committee noted the clinical expert's expectation of improved haemoglobin levels within 1-2 weeks of voxelotor treatment. They acknowledged that EQ-5D scores from earlier in the HOPE trial did not show significant differences between groups and that the European Medicines Agency reported no beneficial effect on endpoints reflecting disease burden and patient well-being.
The committee recognized the potential impact of a 1 g/dl increase in haemoglobin on quality of life but acknowledged the uncertainty due to the trial's results. They suggested exploring alternative approaches to reduce this uncertainty, such as reviewing EQ-5D scores from other sources or considering other health-related quality-of-life measures like SF-36. In response, the company tried to exclude higher-than-expected EQ-5D scores from the HOPE trial but found the dataset too small. They highlighted improvements in the Clinical Global Impression of Change from HOPE, with a higher percentage of people in the voxelotor group reporting significant improvement compared to the placebo group. The company conducted a literature review to explore the impact of a 1 g/dl increase in haemoglobin on quality of life, identifying studies in other diseases like chronic kidney disease and anaemias related to cancer. They presented a range of utility benefits and provided scenario analyses using different utility benefits associated with a 1 g/dl increase in haemoglobin.
Despite the uncertainty, the committee agreed that an increase in haemoglobin was likely associated with a quality-of-life improvement, acknowledging the minimal impact on cost-effectiveness results from varying utility values. They concluded that the utility benefit used in the company's base case was appropriate for decision-making, though it remained highly uncertain.
Recommendation
NICE's health technology evaluations consider cost-effectiveness but do not base decisions solely on it. The committee, recognizing the need for effective sickle cell disease treatments and aiming to reduce health inequalities, accepted a higher cost-effectiveness estimate despite uncertainties. They stressed cautious application of this flexibility to avoid displacing funding from more cost-effective treatments elsewhere in the NHS. Observing that the HOPE trial did not fully represent the NHS population and had high uncertainties, particularly regarding transfusion rates, the committee accepted the uncertainties and additional benefits. With adjusted assumptions, the ICER was deemed acceptable for NHS resources, leading to voxelotor's recommendation for routine use.
Commentaires