Ciclosporin for treating dry eye disease that has not improved despite treatment with artificial tears

Clinical effectiveness

4.3 The Committee discussed the clinical effectiveness evidence for ciclosporin. It noted that the company and the ERG considered SANSIKA to be more relevant than SICCANOVE because SICCANOVE included people with moderate to severe dry eye disease and SANSIKA only included people with severe dry eye disease. The Committee noted that ciclosporin has a marketing authorisation in the UK for treating severe keratitis in adult patients with dry eye disease, which has not improved despite treatment with tear substitutes. Therefore, it concluded that SANSIKA was more relevant than SICCANOVE for its decision‑making.

4.4 The Committee discussed the use of the vehicle as a comparator in the trials. The Committee heard from the company that ciclosporin contains the active ingredient ciclosporin and the excipient (cetalkonium chloride), whereas the vehicle only contains the excipient. The Committee heard from the clinical experts that the excipient is used to help the ciclosporin eye drop stay on the eye surface for longer. Both the company and the ERG stated that the vehicle may have some beneficial effects on its own, which could affect the relative clinical effectiveness of ciclosporin plus artificial tears compared with the vehicle plus artificial tears. The Committee heard from the clinical experts that the vehicle alone is not commercially available as a treatment although the formulation used in ciclosporin (Ikervis) is similar to an artificial tear (Cationorm), but which is also not available in the UK. The Committee considered that it was possible that the vehicle could have had an effect on the relative results of the clinical trials and acknowledged that the vehicle is part of the ciclosporin formulation. The Committee considered that its use as a comparator in the trials limited the interpretation of the results and that the appropriate comparison should have been ciclosporin plus corticosteroids (if needed) and artificial tears compared with corticosteroids (if needed) and artificial tears, and that it would have liked to have seen an indirect comparison for this. The Committee acknowledged that in response to the Committee's request the company presented the results from an updated systematic review and that the company and the ERG concluded it was not possible to do a robust indirect treatment comparison. The Committee concluded that it had not been presented with evidence on the relative clinical effectiveness of ciclosporin compared with established clinical practice that is, corticosteroids (if needed) and artificial tears.

4.5 The Committee considered the primary end point in SANSIKA, namely Corneal Fluorescein Staining score – Oxford Surface Disease Index (CFS‑OSDI) response, which was a composite outcome of individual measures for signs (CFS) and symptoms (OSDI). It heard from the clinical experts that there is no established and standardised measure of response in severe dry eye disease and that several measures of signs and symptoms are used in clinical practice in the NHS, including both CFS and OSDI. The Committee noted that in SANSIKA, ciclosporin plus artificial tears did not show a statistically significant difference compared with the vehicle plus artificial tears in CFS‑OSDI response rate, and that the only statistically significant difference between ciclosporin plus artificial tears and the vehicle plus artificial tears was shown in changes in CFS over time and in human leukocyte antigen‑DR (HLA‑DR). The Committee noted that ciclosporin plus artificial tears did not show any differences compared with the vehicle plus artificial tears in any measure for symptoms. It heard from the company that this could be because of the well‑known poor correlation between signs and symptoms and because of the possible beneficial effect of the vehicle on its own. The Committee noted that, based on the evidence presented, ciclosporin had not shown superior clinical effectiveness to the vehicle.

4.6 The Committee considered comments from the clinical experts that severe dry eye disease is an inflammatory disease associated with long‑term disease progression. The clinical experts also stated that some people with severe dry eye disease might be close to having complete corneal blindness and that any treatment which offers a benefit in terms of reducing inflammation should be considered clinically relevant. The clinical experts explained that improvements in signs of dry eye disease will generally translate into benefits in symptoms in the long term. The Committee noted that ciclosporin showed a statistically significant difference in reducing HLA‑DR, a measure of inflammation, and in change in CFS, a measure of corneal damage, and concluded that these outcomes were clinically relevant.

4.7 The Committee discussed the company's post hoc analyses for SANSIKA. In particular, it considered the results from the post hoc analysis for the primary end point CFS‑OSDI response, for which the company adopted a more stringent definition of response (improvement in CFS score of 3 or more). The Committee noted that ciclosporin showed statistically significant differences compared with the vehicle alone in this post hoc analysis. However, it was aware that the ERG considered that the clinical relevance of this revised definition of response was unclear and that it excluded the level of benefit which most favoured the vehicle group. The clinical experts stated that in clinical practice there is no clear definition for response and non‑response, but that the greater the benefit in CFS the more likely this would have a beneficial effect in slowing disease progression and thus, in improving quality of life. The Committee had reservations about all the post hoc analyses presented by the company and considered that these analyses were not sufficiently robust. It concluded that the original CFS‑OSDI response data were more appropriate to assess the relative clinical effectiveness of ciclosporin compared with the vehicle.

4.8 The Committee discussed the results of the company's meta‑analysis for the subgroup of people with Sjögren's syndrome and severe dry eye disease. It heard from the clinical experts that people with Sjögren's syndrome and severe dry eye disease have a lifelong disease which is difficult to treat and needs careful management. The clinical experts stated that because dry eye disease was associated with other autoimmune diseases including Sjögren's syndrome, this subgroup was clinically relevant and it would benefit most from treatment with ciclosporin. The Committee noted that the results of the meta‑analysis showed that the CFS‑OSDI response rate at month 6 was statistically significantly higher with ciclosporin plus artificial tears compared with the vehicle plus artificial tears and that, although the numbers of patients included in the analysis was small, this subgroup was clinically relevant. The Committee concluded that ciclosporin plus artificial tears showed greater benefits compared with the vehicle plus artificial tears in the subgroup of people with Sjögren's syndrome and severe dry eye disease.

Cost effectiveness

4.9 The Committee considered the cost‑effectiveness evidence presented by the company for ciclosporin plus artificial tears compared with artificial tears alone. It noted that the company used the results from the vehicle group in SANSIKA as a proxy to model the results of artificial tears alone and that the company stated that the response or reduction in the use of artificial tears in the vehicle group was viewed as a regression to the mean. The Committee noted the ERG's concerns highlighting that the SANSIKA results could not be used directly to inform an economic evaluation because the comparator in the model was the vehicle. The Committee concluded that the company's original model was only of limited relevance because it failed to show the cost effectiveness of ciclosporin compared with established clinical practice in the NHS, that is corticosteroids (if needed) plus artificial tears.

4.10 The Committee noted that the company had provided the amendments it requested in the appraisal consultation document (see section 3.37), by presenting an updated economic model that compared ciclosporin plus corticosteroids (if needed) and artificial tears with vehicle plus corticosteroids (if needed) and artificial tears. The Committee discussed that this updated model included corticosteroids as a cost parameter only, and so any potential advantage of lower stopping rates in the ciclosporin group because of corticosteroids (see section 4.1) was not explored. Moreover, results from the vehicle group in SANSIKA were still used as a proxy for the comparator group in the model. The Committee concluded that both models provided by the company were only of limited relevance for its decision‑making.

4.11 The Committee nevertheless explored the results from both the company's original and updated models. It noted that there were 3 main drivers of the results and discussed them in turn. It noted that the company used the post‑hoc analysis for CFS‑OSDI response in its base‑case analysis in its original and updated models, in which ciclosporin showed a statistically significantly higher response than the vehicle and which excluded the level of benefit that most favoured the vehicle group. The Committee was aware that when using the original CFS‑OSDI response data, the incremental cost‑effectiveness ratio (ICER) for ciclosporin plus artificial tears compared with vehicle plus artificial tears increased substantially, from £14,500 to £45,600 per quality‑adjusted life year (QALY) gained (see section 3.41). However, it also noted that this did not occur if a 3‑month (as opposed to 6‑month) stopping rule for ciclosporin was applied. This resulted in an ICER of £33,400 per QALY gained when using the CSF‑OSDI post‑hoc response definition and £24,700 per QALY gained when using the original CFS‑OSDI response definition (see sections 3.42 and 4.12). The Committee noted that these differences were carried through to the results from the updated model. The Committee restated its concerns about the company's post‑hoc analyses of SANSIKA and concluded it was more appropriate to use the original CFS‑OSDI response data in the model (see section 4.7).

4.12 The Committee considered the ERG's concerns about how stopping treatment had been modelled in the company's original base‑case analysis. The Committee noted that in the original model, the company had used the probabilities for continuing treatment beyond the end of the trial from different time periods for each treatment group (6–12 months for ciclosporin and 0–6 months for the vehicle). It heard from the ERG that it considered it to be more appropriate to use Kaplan–Meier analyses for time to stopping treatment, which accurately take into account the moment when the patient stopped treatment. The Committee was aware that when using the ERG's approach, there was a higher rate of people stopping treatment in the ciclosporin group during the first month and this rate subsequently remained stable. In contrast, the rates of people stopping treatment were lower in the vehicle group with no evidence of an initial higher rate of people stopping treatment. The ERG suggested that this could be related to a higher rate of people stopping treatment with ciclosporin because of adverse effects. The Committee heard from the clinical experts that in clinical practice treatment is not stopped because of adverse effects. However the ERG stated that in the trials, the majority of people who stopped treatment did so because of treatment‑related adverse effects. The clinical experts explained that because ciclosporin is given with intermittent corticosteroids in clinical practice, the rates of stopping treatment were expected to be lower than in the trial where corticosteroids were not used, because corticosteroids allow treatment with ciclosporin to be given for longer. The Committee was aware that this was a parameter that had a large effect on the ICER for ciclosporin plus artificial tears compared with vehicle plus artificial tears. The Committee also noted that the company, in its updated model, provided the results of the cost‑effectiveness analysis assuming a different stopping rule at 3 months (based on CFS‑OSDI assessment at 3 months instead of 6 months, as per its original analysis) and that the ICER for ciclosporin plus corticosteroids (if needed) and artificial tears compared with vehicle plus corticosteroids (if needed) and artificial tears was lower when this assumption was applied (see section 4.11). The Committee concluded that it was unclear when treatment with ciclosporin would be stopped in clinical practice because corticosteroids' potential effect on stopping rates had not been included in the company's updated model.

4.13 The Committee discussed the utility values used in the original model. It noted that the company used pooled EuroQoL 5D questionnaire (EQ‑5D) data from SANSIKA for both response and non‑response. However, in its exploratory analyses the ERG applied different utility values for response by treatment and people in the vehicle group showed a larger utility benefit based on response compared with people in the ciclosporin group. The ERG suggested that the differences in utility values between treatments could be because of the additional adverse effects in people having ciclosporin. The Committee heard from the clinical experts that in clinical practice, adverse events were mild and transient and would not have an effect on quality of life. The ERG noted that adverse effects such as instillation site irritation were likely to occur at each instillation and these adverse effects had not been captured anywhere in the model. The Committee recognised that the utility value for response was the parameter that had the biggest effect on the ICER and that assuming different utility values for response by treatment group led to ciclosporin plus artificial tears being dominated by (that is, it was both more costly and less effective than) vehicle plus artificial tears. The Committee noted that the company still used pooled utility values in its updated model and that the company stated that any differences in utility values between treatment groups were circumstantial. The Committee also noted comments from a clinical expert stating that adverse effects with ciclosporin are transient and have limited impact on quality of life. The Committee recognised that the cost‑effectiveness results varied substantially when applying treatment‑specific utility values, but was also aware that the analyses did not capture corticosteroids' potential to mitigate adverse effects (see section 4.2). The Committee concluded that this added additional uncertainty to the results presented by the company.

4.14 The Committee noted that in its original cost‑effectiveness analyses, the company did not present a subgroup analysis for people with Sjögren's syndrome and severe dry eye disease. The Committee was aware that the clinical experts highlighted that people with Sjögren's syndrome and severe dry eye disease would be most likely to benefit from treatment with ciclosporin (see section 4.8). The Committee noted that the company provided a subgroup analysis for people with Sjögren's syndrome in its updated model. However, this subgroup analysis incorporated the same assumptions as the analysis for all patients (see section 4.10) and the Committee concluded that these results also lacked relevance for its decision-making.

4.15 The Committee summarised its considerations about the company's original and updated models. Firstly, it considered that the models lacked relevance because the comparator used was vehicle rather than corticosteroids (if needed) and artificial tears, which is considered established clinical practice (see sections 4.9 and 4.10). The Committee considered that the vehicle's independent characteristics may have affected the results of the model, because vehicle alone showed benefits in terms of CFS‑OSDI response rate (see sections 4.4 and 4.5). The Committee acknowledged that the vehicle may be associated with some benefit but highlighted that it was not commercially available on its own in the UK. Secondly, the Committee noted that 3 parameters had a substantial effect on the cost‑effectiveness results (namely, using the original or post‑hoc CFS‑OSDI response definition, a 3- or 6‑month stopping rule, and pooled or different utility values for treatment groups) and that changing them led to very variable results (see sections 4.11–4.13). The Committee considered it was difficult to draw any conclusions from a model subject to these assumptions and with a non‑relevant comparator.

4.16 The Committee agreed that it was relevant to consider ciclosporin (Ikervis) in comparison with other ciclosporin formulations available. It therefore discussed the ERG's and the company's cost‑minimisation analyses comparing ciclosporin (Ikervis) with the other ciclosporin formulations. It noted that based on the ERG's analysis, ciclosporin (Ikervis) was less costly than Restasis and that there was not a big cost difference compared with the other ciclosporin formulations. It also noted that based on the company's analysis, ciclosporin was the least expensive and that the differences between the company's and the ERG's estimates were based on different assumptions in the number of vials and tubes of ciclosporin needed per month. The Committee understood that other ciclosporin formulations that do not have a marketing authorisation in the UK require additional monitoring, whereas this is not needed with ciclosporin (Ikervis; see section 4.2). The Committee noted comments from a clinical expert and the company that if ciclosporin (Ikervis) were not recommended for use in the NHS, other ciclosporin formulations that do not have marketing authorisation in the UK (and are associated with higher costs) would continue to be used. The Committee restated its previous conclusion that it was reasonable to assume that the different ciclosporin formulations would show similar efficacy (see section 4.1) and considered that, based on the cost‑minimisation analyses presented by the company and the ERG, the cost of ciclosporin (Ikervis) was reasonable compared with the other ciclosporin formulations. Therefore, the Committee concluded that, on balance, ciclosporin (Ikervis) was a cost‑effective use of NHS resources for people with severe keratitis in adult patients with dry eye disease, which has not improved despite treatment with artificial tears.

4.17 The Committee discussed the innovative nature of ciclosporin. The Committee noted that ciclosporin was not a novel technology, but it heard from the clinical experts that because people with severe dry eye disease were close to complete corneal blindness and there were no other effective treatments available and licensed in the UK, there was a high unmet medical need. The Committee also noted that this was even more important for people with severe dry eye disease and Sjögren's syndrome because of the need for long‑term management of the condition with an effective treatment that would help to delay disease progression. The company highlighted that ciclosporin was particularly beneficial because it is administered as 1 eye drop per day compared with other treatments that need to be provided several times per day. The company stated that the benefits in terms of administration had not been appropriately captured in the QALY calculation. The Committee also noted that the use of several artificial tears per day can have a detrimental impact on quality of life and that the fact that ciclosporin helps to reduce the number of artificial tears needed was particularly important to patients. The Committee concluded that even though ciclosporin is used for treating severe dry eye disease which has not improved despite treatment with artificial tears in people who have a high unmet need, the new formulation of ciclosporin could not be considered an innovative technology.

4.18 The Committee was aware of the NICE's position statement about the Pharmaceutical Price Regulation Scheme (PPRS) 2014, and in particular the PPRS Payment Mechanism. It acknowledged '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 about the relevance of the PPRS to this appraisal of ciclosporin. It therefore concluded that the PPRS Payment Mechanism was irrelevant for the consideration of the cost effectiveness of ciclosporin.

Summary of Appraisal Committee's key conclusions