Nintedanib for treating idiopathic pulmonary fibrosis

4.1 The Committee understood that idiopathic pulmonary fibrosis is a distressing illness that limits physical activity because of breathlessness, and can lead to hypoxia, pulmonary hypertension, heart failure and death. The Committee heard from clinical experts that 80% of people with idiopathic pulmonary fibrosis die from the condition or from respiratory failure. The median overall survival in the clinical trials, which excluded people with the most severe disease, was 3 to 4 years. It heard that there is no cure for idiopathic pulmonary fibrosis, although the Committee acknowledged that pirfenidone and lung transplant (when drug treatment is not appropriate) are options available to manage the condition. It heard however that these are not appropriate for many people because of the tolerability of pirfenidone, the severity of the person's condition, or other comorbidities. Patient experts explained that a key aim of treatment is to slow the progression of disease. They stated that people with idiopathic pulmonary fibrosis would benefit from alternative treatment options. The Committee concluded that idiopathic pulmonary fibrosis is associated with substantial morbidity and mortality, and that there are few treatment options.

4.2 The Committee discussed the effect of acute exacerbations in idiopathic pulmonary fibrosis, which it understood to be more serious than exacerbations in other respiratory diseases. It heard from clinical experts that half of people with idiopathic pulmonary fibrosis will die within 30 days of an acute exacerbation. People who survive an exacerbation have permanent and substantially reduced lung function (up to a 20% decline in percent predicted forced vital capacity [FVC]). Clinical experts explained that acute exacerbations can be difficult to define in clinical trials, because they can be confused with respiratory infections, but are clearly recognised in clinical practice by experienced clinicians. The Committee heard from patient experts that preventing or delaying acute exacerbations is an important way to maintain quality of life. The Committee concluded that exacerbations are an important clinical event, but can be difficult to define, particularly in clinical trials.

4.3 The Committee considered how clinicians assess lung function in people with idiopathic pulmonary fibrosis. It understood that clinicians use a number of measures of lung function, and heard that they routinely use percent predicted FVC to guide clinical decision‑making. Clinical experts noted some disadvantages with using percent predicted FVC. For example, the equations used to calculate percent predicted FVC (adjusting for age, sex and height) extrapolate data from a middle‑aged white male population and may under‑ or overestimate the expected lung volume for current clinical practice in England. Clinical experts explained that it is difficult to know how a person's lung function would have progressed without treatment. However, the Committee acknowledged that this is not unique to idiopathic pulmonary fibrosis. The Committee heard from clinical experts that other measures of lung function (such as the 6‑minute walk test distance and diffusion capacity of the lung for carbon monoxide) are less reliable than FVC. The company stated that emphysema commonly co‑exists with idiopathic pulmonary fibrosis, and that in people with both conditions, percent predicted FVC can be less sensitive (that is, it could be high despite significant pulmonary disease) and can change from one day to the next. The Committee recognised the limitations of FVC but understood that in clinical practice the wider patient characteristics would be taken into account in interpreting percent predicted FVC. Clinical experts noted that they follow the stopping rule in NICE's technology appraisal guidance on pirfenidone for treating idiopathic pulmonary fibrosis, but explained that before withdrawing treatment they retest FVC to confirm that the 10% drop is not temporary, which might happen with an infection. The Committee concluded that, although it has some limitations, percent predicted FVC is the most reliable and widely used measure of lung function in clinical practice.

4.4 The Committee discussed the treatment options for people with idiopathic pulmonary fibrosis.

4.5 The Committee discussed the clinical trial evidence for nintedanib and heard that the trials reflected current clinical practice. It understood that there were inconsistencies in some of the results across the 2 phase III trials. However, it noted that a pre‑planned pooled analysis showed statistically significant differences between nintedanib and placebo for loss of lung function and a non‑significant benefit in favour of nintedanib for reducing acute exacerbations. The Committee heard that, based on pre‑planned and post‑hoc subgroup analyses, nintedanib was effective regardless of the baseline FVC. It noted that the mean baseline percent predicted FVC was approximately 80% across all 3 nintedanib trials, indicating that the trials provided evidence for treating idiopathic pulmonary fibrosis in people with a percent predicted FVC above 80%. The Committee concluded that the trials provided an appropriate basis for its decision‑making, and showed that nintedanib is more effective than placebo in all subgroups.

4.6 The Committee considered whether the company network meta‑analysis was robust. It heard from the ERG that the company had included all relevant trials, but had explored heterogeneity in the results by excluding trials in sensitivity analyses. The Committee understood that the results of the network meta‑analysis informed the relative effectiveness of nintedanib and pirfenidone in the company model. It heard that the company used the results of different sensitivity analyses (that is, using data from different sets of trials) for different outcomes in the model (see section 3.4). The Committee agreed that this introduced a potential bias in favour of nintedanib because the results of the analyses chosen by the company were more favourable to nintedanib than the results from analyses including all trials. It concluded that the same trials should be included for all outcomes. The Committee agreed with excluding one of the Japanese studies (SP2) because it considered it to be an outlier. The company clarified that not all trials reported evidence for all outcomes considered in the model, and presented an updated analysis excluding the SP2 study from the relevant outcomes (overall survival and acute exacerbations). The Committee concluded that the updated network meta‑analysis provided a more appropriate basis for its decision‑making. The Committee considered the effectiveness of nintedanib and discussed the results of the company network meta‑analysis and concluded that the clinical effectiveness of nintedanib is similar to pirfenidone.

4.7 The Committee considered whether the company model, in which health states were based on percent predicted FVC and occurrence of acute exacerbations, accurately represents the progression of idiopathic pulmonary fibrosis. It heard from patient experts that idiopathic pulmonary fibrosis is a progressive disease which does not improve, and noted that the model reflects this. The Committee heard from the company that exacerbations increased the rate of disease progression (that is, loss of lung function) in the model, which the Committee considered appropriate. However, taking into account the clinical experts' comments about the substantial impact of exacerbations on quality of life (see section 4.2), the Committee was concerned that the results of the model were not sensitive to changes in the rate of exacerbations. The ERG stated that it considered the methods applied in the company's economic analyses to be generally appropriate. The Committee concluded that a model based on percent predicted FVC and exacerbations was appropriate for decision‑making.

4.8 The Committee discussed the population included in the economic model. The Committee appreciated that its remit was to compare nintedanib with current NHS treatment which, for people with percent predicted FVC of 50–80%, is pirfenidone or best supportive care. For people with percent predicted FVC of more than 80%, current NHS treatment is best supportive care. It noted that the company analyses included a population with a percent predicted FVC of more than 50%, and that the ERG modelled a restricted population with a percent predicted FVC of 50–79.9%. The Committee recognised that the ERG had assumed that people with a percent predicted FVC of 80% or more do not present in clinical practice. However, the Committee heard from the clinical experts that people with a percent predicted FVC of 80% or more represent a third of people with idiopathic pulmonary fibrosis in specialist clinical practice (see section 4.4), and that the relevant comparator for this population is best supportive care. The Committee agreed that to compare nintedanib with best supportive care, it would have preferred to see a model representing only people with a percent predicted FVC of 80% or more. After consultation, the ERG provided an analysis in people with a percent predicted FVC of 80% or more. The Committee noted the ERG's concerns that subgroup analyses were subject to limitations because parameters such as hazard ratios, mortality rates, and rates of stopping treatment may differ between the subgroups and the whole population. In the absence of subgroup‑specific parameters, the Committee concluded that the company model was appropriate for its decision‑making.

4.9 The Committee considered the extrapolations of overall survival based on the company model. It noted that the company and the ERG had used the log logistic curve in their base‑case analyses. The Committee discussed whether this curve was appropriate given that it has a long tail, meaning that the model might overestimate life expectancy and therefore underestimate the incremental cost‑effectiveness ratio (ICER) for nintedanib. The Committee noted that using other methods to extrapolate survival substantially increased the ICER for nintedanib compared with best supportive care. The Committee heard from the clinical experts that the median survival estimated with the log logistic curve generally reflects the natural history of treated disease. The Committee agreed that the survival modelling was uncertain, but noted that it had little effect on the ICER when comparing nintedanib with pirfenidone because the company had assumed equal survival with these drugs. Any differences in quality‑adjusted life year (QALY) gain between nintedanib and pirfenidone were therefore derived from differences in quality of life. The Committee concluded that the log logistic curve was sufficient for decision‑making, but recognised that the ICER for nintedanib compared with best supportive care, when using the log logistic curve, may be an underestimate.

4.10 The Committee discussed the utility values in the company's model. It approved of the company using trial‑based EQ‑5D data to estimate health‑state utility values. The Committee expressed some concern that the company did not include a disutility for diarrhoea in the model, because this is a common adverse event with nintedanib that it considered would worsen quality of life. It heard from the company that including a diarrhoea‑related disutility would not affect the model results because the event was mild‑to‑moderate and led to less than 5% of people stopping treatment in the nintedanib clinical trials. However, the Committee did not agree that diarrhoea would have no clinical impact.

4.11 The Committee considered the changes to adverse‑event‑related disutilities suggested by the ERG. It heard from clinical experts that the ERG's preferred estimate of rash‑related disutility (based on data from the RECAP study) was inaccurate because the RECAP study underestimated the incidence of rash by 50%. Clinical experts agreed with the ERG's choice of a lower incidence of photosensitivity with pirfenidone, and stated that it was reasonable to assume people have adverse events in the model for approximately 1 month rather than for 1 year (as in the company base case). The Committee concluded that the estimate of rash‑related disutility in the company's original base case was more appropriate than the ERG's estimate, but that the ERG provided more accurate estimates of other disutilities which the company incorporated in its revised base case (see sections 3.21 and 3.22).

4.12 The Committee compared the company model with the model submitted for NICE's technology appraisal guidance on pirfenidone for treating idiopathic pulmonary fibrosis, for external validation. It noted that the nintedanib model produced different results from the pirfenidone model, when comparing pirfenidone with best supportive care. The ERG could not fully compare the 2 models because of confidentiality, but explained some key differences. For example, treatment effect in the pirfenidone model was based on FVC and the 6‑minute walk test distance. However, the Committee understood from clinicians that the 6‑minute walk test distance was an unreliable measure. The pirfenidone model did not include acute exacerbations, which the Committee understood to be an important clinical event in idiopathic pulmonary fibrosis. The pirfenidone model used a mapping algorithm to calculate utility values, whereas the nintedanib model included trial‑based EQ‑5D data. The Committee concluded that it could not compare the models fully, but that the current model for nintedanib was robust, and appropriate for decision‑making.

4.13 The Committee considered whether it should take into account the consequences of the Pharmaceutical Price Regulation Scheme (PPRS) 2014, and in particular the PPRS payment mechanism, when appraising nintedanib. The Appraisal Committee noted NICE's position statement in this regard, and accepted the conclusion 'that the 2014 PPRS payment mechanism should not, as a matter of course, be regarded as a relevant consideration in its assessment of the cost effectiveness of branded medicines'. The Committee heard nothing to suggest that there is any basis for taking a different view with regard to the relevance of the PPRS to this appraisal of nintedanib. It therefore concluded that the PPRS payment mechanism was not applicable for considering the cost effectiveness of nintedanib.

4.14 The Committee discussed the most plausible ICER. It noted that the most plausible scenario would include its preferred assumptions for:

4.15 The Committee considered the population with a percent predicted FVC of more than 80%, for whom the comparator is best supportive care. The Committee would have preferred to see a model only of people with a percent predicted FVC of more than 80%, but in the absence of appropriate subgroup‑specific parameters (see section 4.8) the Committee considered that the company model of people with a percent predicted FVC of more than 50% was appropriate for decision‑making. It noted that the ICER for nintedanib compared with best supportive care was substantially greater than £30,000 per QALY gained. The Committee acknowledged that the ICER based on the ERG's exploratory analysis including only people with a percent predicted FVC of 80% or more was also substantially greater than £30,000 per QALY gained. NICE cannot report the ICERs because the patient access schemes are confidential. The Committee noted comments that nintedanib is clinically effective in all subgroups and that it is not clinically beneficial to delay treatment until the condition worsens. However, the Committee was aware of the high ICERs estimated for nintedanib compared with best supportive care. The Committee concluded that the ICERs for nintedanib compared with best supportive care in people with a percent predicted FVC of more than 80% were not within the range considered to be a cost‑effective use of NHS resources.

4.16 The Committee heard from patient experts and Committee members that nintedanib was innovative in its potential to make a significant and substantial impact on health‑related benefits. The patient expert emphasised the better tolerability profile associated with nintedanib which significantly improved their quality of life. For example, the patient expert valued their opportunity to pursue outdoor activities while receiving nintedanib, which they had been unable to do when receiving pirfenidone because of the associated photosensitivity. The Committee acknowledged that nintedanib is associated with adverse events that are not commonly associated with pirfenidone, such as serious gastrointestinal events, and that diarrhoea is more common with nintedanib than pirfenidone, but heard from clinical and patient experts that people may tolerate nintedanib better than pirfenidone. The Committee also noted that people do not need to take nintedanib as often as pirfenidone but heard differing views about the value of this, and considered this a small advantage. On balance, the Committee concluded that there are benefits associated with nintedanib that the economic model does not fully capture, such as the impact on patients' lives of nintedanib's better tolerability profile and reduced dosing frequency, compared with pirfenidone. The Committee concluded that there may be some additional gains in health‑related quality of life over those already included in the QALY calculations, supporting its recommendation that nintedanib should be offered to people as an alternative to pirfenidone.