Vortioxetine for treating major depressive episodes

The committee reviewed the clinical trial evidence submitted by the company, and agreed that the REVIVE trial comparing vortioxetine with agomelatine was of good quality. However, it noted that a key issue highlighted by the evidence review group (ERG) was the generalisability of the results from REVIVE to people whose major depressive episode had responded inadequately to a course of SSRI antidepressants (that is, the second‑line population on which the company focused its original evidence submission). The committee heard from the clinical expert that agomelatine was a reasonable comparator for vortioxetine in a trial setting because it is not sedative. The committee understood that agomelatine is not widely used in clinical practice in the NHS, but is used as an alternative treatment for some people for whom mirtazapine is not appropriate because it is associated with weight gain. The committee agreed that, because of agomelatine's limited use, the comparison of vortioxetine with agomelatine was of limited relevance to clinical practice in England. The committee considered whether the previous treatments received by the REVIVE population were generalisable to clinical practice in England. The committee was aware that over 20% of patients in REVIVE received initial treatment with an SNRI rather than an SSRI as recommend by NICE's guideline on depression in adults, and agreed that this did not reflect clinical practice in England. The committee noted that the proportion of people recruited to the REVIVE trial from the UK was small (about 7%), and agreed that the variation in managing major depressive disorder across countries may limit the applicability of the trial results to patients in England. The committee concluded that the results from the REVIVE trial were not generalisable to most patients in routine clinical practice in England.

The committee considered the results from the REVIVE trial. The committee heard from the company that it used a 'full analysis set' rather than an intention‑to‑treat analysis to assess the outcomes, in accordance with the committee for Medicinal Products for Human Use guidelines for non‑inferiority trials. The committee commented that it preferred to see outcomes analysed using an intention‑to‑treat analysis but it was aware that few patients were excluded from the 'full analysis set' analysis in the REVIVE trial. The committee noted that the primary outcome in REVIVE was the change in severity of depressive symptoms measured by the mean change from baseline in the Montgomery–Åsberg Depression Rating Scale (MADRS) at 8 weeks, and that this was 2.16 points lower with vortioxetine than with agomelatine. The committee also noted that vortioxetine showed a statistically significant improvement in both response and remission rates (secondary outcomes) compared with agomelatine. The committee discussed what size of changes in depressive symptom severity scores clinicians and patients consider clinically important. The committee heard from the clinical expert that the mean change from baseline in total MADRS score was not a useful outcome measure for judging whether a clinically important difference was observed because the MADRS included 10 items for measuring depressive symptoms. The clinical expert explained that a reduction in 1 item of the MADRS by 2 or more points would generally be considered clinically meaningful in clinical practice. The committee noted that this was disputed in some comments it received on the appraisal consultation document. However, the committee agreed that achieving remission and avoiding relapse were much more useful outcomes than the mean change in a person's depressive symptom severity score for measuring success of treatment in clinical practice.

The committee discussed the company's indirect treatment comparison presented for the second‑line population. It was concerned that the evidence network only consisted of 4 trials and only included 1 trial for each treatment comparison. The committee was also aware that 1 of these trials (Kasper et al. 2013) included people who may not have been changing to another treatment for a major depressive episode but starting first‑line treatment for a recurrent major depressive episode. The company acknowledged that the population in Kasper may not be comparable with the other populations in the evidence network, or consistent with the population specified in its decision problem. The committee considered that the patient populations in the trials differed in baseline severity of depression. It was aware that the company's indirect treatment comparison reported remission rates and the proportion of people stopping treatment because of adverse events, both of which depend on trial duration, which differed between the trials in the network. The committee concluded that, because of the evidence base, the company's indirect treatment comparison was not sufficiently robust for estimating the clinical effectiveness of vortioxetine compared with other antidepressants for second‑line treatment.

The committee discussed whether evidence from the first‑line treatment population was relevant for informing the relative effectiveness of vortioxetine compared with other antidepressants for people having second and subsequent lines of treatment. The committee heard from the company that, although there is a paucity of evidence for vortioxetine used second line, the company chose not to include data from its trials of vortioxetine as a first‑line treatment, because it claimed that the relative effectiveness changes across lines of treatment. The committee was aware that the ERG considered that the company did not provide sufficient evidence that the relative effectiveness differs between non‑SSRIs within each line of treatment, but the ERG accepted that the absolute effectiveness may change between each line of treatment. This was confirmed by the clinical expert who stated that, in clinical practice, the absolute effectiveness of each antidepressant is likely to decline with each subsequent line of treatment because there are people whose major depressive episode is difficult‑to‑treat (that is, treatment‑resistant) and therefore unlikely to remit or respond. However, the clinical expert noted that the relative effectiveness of the antidepressants compared with one another may also change at each subsequent line of treatment. The clinical expert explained that depression which does not respond to 1 or 2 SSRIs may be mediated by different receptors, so the relative effectiveness of treatments with a different mechanism may differ across subsequent lines of treatment. The ERG acknowledged that the relative effectiveness may reduce in clinical practice at second or later lines of treatment compared with first‑line treatment, particularly for SSRIs compared with antidepressants of a different class. However, it emphasised that the company had not provided sufficient evidence, either in its original submission or in its response to the appraisal consultation document, to support a declining relative effect of treatment between non‑SSRIs within each line of treatment. The committee was also aware that NICE's guideline on depression in adults concluded that 'the evidence for the relative advantage of switching either within or between classes is weak' and 'that evidence from primary efficacy studies of existing treatments should also be considered' when making decisions about second and subsequent lines of treatment. On balance, the committee concluded that evidence from trials in the first‑line population was relevant to informing the relative effectiveness of vortioxetine compared with other antidepressants for second and subsequent lines of treatment.

The committee discussed alternative sources (Pae et al. 2015 and Llorca et al. 2014) to estimate the relative effectiveness of vortioxetine compared with other antidepressants. The committee was aware that these meta‑analyses included populations being treated first line. It noted that the absolute remission rates for vortioxetine were lower than for some of the other antidepressants included in Pae and Llorca (see table 10). It noted that this was not consistent with the company's indirect treatment comparison, which estimated vortioxetine to be the most effective treatment option (see table 10). The committee appreciated that the 2 studies took different methodological approaches (see section 3.29). It heard from the ERG that each analysis had a number of biases (for example, Pae included trials with active reference arms), and that the ERG considered Llorca to be the most credible. The committee was aware that Llorca included more treatment options and trial evidence than Pae, and also used indirect evidence to inform the estimates of relative effectiveness (rather than only direct evidence as carried out by Pae). The committee was aware from the response to the appraisal consultation document that the company was concerned that the results from Llorca for remission and response were potentially biased because several of the included trials did not present data for these outcomes. The committee heard from the ERG that there was no evidence to suggest that the Llorca analysis was affected by reporting bias (see section 3.48). The committee concluded that the estimates of relative effectiveness in each analysis were subject to uncertainty but, of the available sources, Llorca had the fewest weaknesses for informing the relative effectiveness of vortioxetine compared with other antidepressants.

The committee discussed the relative effectiveness evidence available for vortioxetine compared with other antidepressants. The committee noted that the published meta‑analyses were consistent with the interpretation that vortioxetine was neither better nor worse than other treatments. Specifically, the committee acknowledged that none of the analyses it had seen (that is, the company's indirect treatment comparison, Pae et al. 2015, Llorca et al. 2014), showed statistically significant differences between vortioxetine and the other antidepressants for achieving remission (other than compared with agomelatine, a comparator not widely used in the NHS). The committee was aware that Pae (not sponsored by the company) concluded that vortioxetine was 'more effective than placebo but the difference was of doubtful clinical significance', and that Llorca (sponsored by company) concluded that vortioxetine had 'comparable or favourable' efficacy and tolerability compared with other antidepressants. Furthermore, the committee noted that the evidence for vortioxetine in people having second‑line treatment included trials only of short duration, so the treatment effect of vortioxetine after 12 weeks was uncertain. The committee concluded that no convincing evidence existed to show that vortioxetine was more or less effective than other antidepressants.

The committee discussed the adverse effects associated with vortioxetine and the other antidepressants. It noted that the company's indirect treatment comparison, Pae et al. (2015), and Llorca et al. (2014) measured the odds of stopping treatment because of adverse events. The committee was aware that some patients may stop treatment for reasons other than adverse events, and that some patients tolerate adverse events and do not stop treatment. The committee understood that vortioxetine's recommended starting dose may be increased and that the long‑term safety data suggested that the overall incidence of adverse reactions was higher in people taking 15–20 mg of vortioxetine daily compared with 5 mg of vortioxetine daily. The committee was aware that the TAK318 trial, which the company did not include in its indirect comparison or modelling, showed that vortioxetine improved sexual function in people with sexual dysfunction more than escitalopram. The committee agreed that the long‑term adverse effect profile of vortioxetine compared with commonly used antidepressants in England was uncertain. However, it accepted that the available evidence suggested vortioxetine leads to a lower probability of stopping treatment and fewer adverse effects than most other antidepressants in the short term. The committee concluded that, based on the available (albeit sparse) evidence, vortioxetine may have a better overall safety profile than other antidepressants.

The committee discussed the company's revised economic model received in response to the appraisal consultation document. The committee highlighted that it could not make a recommendation for vortioxetine to treat all people included in the marketing authorisation. At the first appraisal committee meeting, the committee was not convinced that vortioxetine offered a cost‑effective use of NHS resources as a second‑line treatment option. The committee was also concerned about the structure of the company's original economic model. The company's revised cost‑effectiveness results were for vortioxetine as a third‑line treatment and the committee was satisfied that the company's revised economic model had addressed several structural uncertainties. The committee concluded that it was now able to assess the cost effectiveness of vortioxetine compared with other antidepressants for people whose condition has responded inadequately to 2 antidepressants within the current major depressive episode.

The committee discussed the costs and resource‑use included in the company's economic model. The committee noted that the dose of third‑line treatment was increased after the acute phase in the company's economic model, and was aware that this may reflect clinical practice in people who tolerate, and whose depression responds to, treatment. Moreover, the committee understood from the ERG's exploratory analysis that assuming that the dose of third‑line treatment did not increase after the acute phase had little impact on the incremental cost‑effectiveness results. The committee noted that continuing treatment in the company's revised model was based on whether a person's depression remits, responds (but does not remit) or does not respond. The committee understood from the clinical expert that this reflected how clinicians decide when to continue treatment in clinical practice. The clinical expert explained that people with a major depressive episode whose condition responds after 8 to 10 weeks of treatment, but does not remit, would generally be treated for a further 4 weeks. The committee considered that, because people whose condition responds to treatment but does not remit have a lower health‑related quality of life, it was appropriate for the company to assume that their condition costs more to treat than people whose condition does remit. The committee concluded that the company appropriately modelled continuing treatment. The committee also accepted that the company appropriately modelled the time at which people change treatments. The committee acknowledged that people who switch treatment because of adverse reactions were likely to switch earlier than people who switch treatment because of a lack of response, in line with the company's approach. The committee was also aware that the company had provided scenarios with either 6 or 22 months maintenance therapy (that is, continued treatment for up to 2 years in people at high risk of relapse), in line with the recommendations in NICE's guideline on depression in adults and clinical practice. The committee was also aware that the company had provided separate scenarios for people being treated either in primary care or in secondary care. The committee understood from the clinical expert that vortioxetine is likely to be used predominantly in secondary care. It noted a comment received on the appraisal consultation document, which highlighted that people treated for major depressive disorder in primary care were likely to be completely different from those treated in secondary care. The committee commented that the company's approach to only changing the unit cost of a healthcare professional visit from a GP visit (primary care) to a psychiatrist visit (secondary care) was unlikely to reflect the change in healthcare resource use between primary and secondary care. The committee therefore noted that the company's cost‑effectiveness results for the secondary care analysis should be interpreted with caution. However, the committee concluded that overall the cost and resource use included in the company's model generally reflected the pathway of care for people for whom vortioxetine would be considered appropriate.

The committee was aware that the company did not assume that treating depression lowered the risk of suicide, so any modelled gains in quality‑adjusted life years (QALYs) reflected only a difference in health‑related quality of life. The committee agreed it was appropriate for the company to use the utility values from REVIVE for all phases of the economic model, rather than using 2 separate sources of evidence for the utility values as applied in its original economic model. The committee concluded that it preferred the EQ‑5D data from REVIVE because it represented the best evidence available and was more internally consistent.

The committee discussed the company's approach to modelling adverse events. The committee was aware that the company based adverse event rates for vortioxetine and its comparators on absolute rates reported from individual trials. The committee noted that it accepted that safety data would be transferable across lines of treatment. However, it was uncertain whether the company's approach to modelling adverse events was appropriate, given differences in the baseline severity of depression in the trials' populations for vortioxetine and its comparators. On balance, the committee recognised from the evidence currently available that vortioxetine was likely to lead to fewer adverse events than other antidepressants. The committee noted that the company assigned no decrease in health‑related quality of life because of several adverse events (for example, dry mouth, dizziness), and agreed that this was unlikely to reflect reality. However, the committee was aware that, for these adverse events, the incidence rates were generally lower for vortioxetine than the other antidepressants, so the company's approach underestimated the benefits of vortioxetine. The committee also noted that the company had not considered adverse events in people receiving fourth and subsequent lines of treatment. The committee agreed that, because a substantial proportion of people receive therapy after third‑line use in the company's model, this led to further uncertainty around the cost‑effectiveness results. The committee concluded that it would have preferred the company to justify its approach for modelling adverse reactions, but appreciated that data for antidepressants in comparable populations were likely to be sparse.

The committee discussed the company's approach to modelling remission and relapse rates for people having fourth and subsequent lines of treatment. The committee heard from the clinical expert that there was limited evidence available on the prognosis for people having subsequent lines of treatment, and that the STAR*D trial provided the best available data. The committee accepted this, but understood that STAR*D included treatments that did not reflect those commonly used in England, and that the population was different from the population in REVIVE. The committee also appreciated that the effectiveness of subsequent lines of treatment was independent of the initial treatment strategy, but the proportion of people that subsequently switched to fourth‑line treatment differed depending on the effectiveness of the third‑line treatment. The company noted that the company used the absolute rates of remission from the STAR*D trial for subsequent lines of treatment. However, the committee considered it more appropriate to apply a proportionate reduction in the rates of remission for fourth and subsequent lines of treatment, as seen in the STAR*D trial, to the remission data used for third‑line treatment. The committee noted that the company assumed that the rate of relapse did not differ between initial treatments but did differ between subsequent lines of treatment. The committee considered it more appropriate to assumed that the rate of relapse was independent of treatment line. The committee concluded that fuller exploration of alternative scenarios for modelling remission and relapse rates for people having subsequent lines of therapy would have been helpful.

The committee discussed the cost‑effectiveness results presented by the company and the ERG's exploratory analyses. The committee noted that the company's base‑case results were not sensitive to changes to most parameters. The committee was aware that, when relative effectiveness was estimated by Llorca et al. (2014) or Pae et al. (2015), or when vortioxetine was assumed to be as effective as other antidepressants, the incremental cost-effectiveness ratios (ICERs) for vortioxetine were extremely unstable because of the very small differences in incremental QALYs (that is, highly sensitive to the parameters used for the rates of remission and relapse). The committee understood from the ERG that the small differences between the incremental QALYs were driven partly by the short time horizon of the model and partly by the data suggesting that vortioxetine was not more or less effective than other antidepressants (but reflecting the small observed differences in absolute effects). The ERG explained that, given that there were no substantial differences in depressive severity symptoms scores between the antidepressants reported in the trials included in the company's indirect treatment comparison (or Pae or Llorca), it was not surprising that the incremental QALYs were equally small. The committee concluded that it needed to take into account the instability of the ICERs in its decision‑making.

The committee noted that the company's revised ICERs were estimated from probabilistic analyses based on pairwise comparisons. The committee concluded that it preferred probabilistic ICERs estimated for all comparators simultaneously, but acknowledged that the company had addressed other structural uncertainties in the economic model.

The committee discussed whether it could recommend vortioxetine as a third‑line treatment option for treating major depressive episodes. The committee acknowledged that the company's revised economic model generally reflected the pathway of care for people for whom vortioxetine would be offered, and had used EQ‑5D utility data as preferred by NICE in its guide to the methods of technology appraisal (2013). Furthermore, the committee emphasised that there was no convincing evidence to show that vortioxetine was more or less effective than other antidepressants (see section 4.10). The committee noted that the results of the SOLUTION trial provided by the company in its response to the appraisal consultation document supported the general conclusion of Llorca et al. (2014) that vortioxetine's effectiveness was comparable with that of other antidepressants (in this trial, venlafaxine). The committee stated that there were likely to be small differences between the antidepressants, but it was satisfied based on all the evidence that a scenario assuming equal efficacy could be considered for the purposes of assessing the cost effectiveness of vortioxetine compared with other third‑line antidepressants. The committee noted that, although the emerging evidence on adverse effects for vortioxetine compared with other antidepressants was relatively immature, the available evidence suggested that people tolerate vortioxetine better than other options. The committee based its decision‑making on the company's cost‑effectiveness analyses that excluded SSRIs (see sections 4.3 and 4.4). The committee highlighted that, across all of the company's scenarios using its revised economic model, and when assuming equal efficacy between treatments, the ICERs for vortioxetine compared with other antidepressants were £9000 per QALY gained or below. Therefore, the committee agreed that treatment with vortioxetine was a cost‑effective use of NHS resources compared with other antidepressants. The committee concluded that vortioxetine could be recommended as an option for treating major depressive episodes in adults whose condition has responded inadequately to 2 antidepressants within the current episode.

The committee was aware that the company's evidence included people with first and recurrent major depressive episodes. It recognised that it was appropriate to change treatments between the first episode and any subsequent episodes, and to not use any drug, or a specific sequence of drugs, for a recurrent episode that had not worked during a previous episode. The committee also noted that NICE's guideline on depression in adults states that 'treatment choice should be influenced by: previous treatment history, including the consequences of a relapse, residual symptoms, response to previous treatment, any discontinuation symptoms, and the person's preference'. The committee therefore agreed that a flexible approach was needed in clinical practice when treating recurrent episodes of depression.

The committee discussed whether vortioxetine could be considered innovative, and whether the company's economic analysis had captured all changes in health‑related quality of life. The committee noted that the company considered vortioxetine innovative because: it reduces cognitive dysfunction independent of its effect on MADRS; it minimises impact on social relationships; it reduces symptoms associated with stopping treatment; and it provides benefits related to health‑related quality of life underestimated by the EQ‑5D instrument. The committee acknowledged that vortioxetine may be a valuable treatment option for people with a major depressive disorder experiencing cognitive dysfunction. However, it noted that the EQ‑5D data from REVIVE reported for the vortioxetine and agomelatine groups did not suggest that the average utility for remission and non‑remission was notably different between treatments. The committee also acknowledged that, in general, the benefits of mental health conditions relative to other conditions may be underestimated by the EQ‑5D instrument. However, the committee considered that any shortcomings in the EQ‑5D would impact each treatment option included in the company's economic analysis similarly, particularly because there was no convincing evidence to suggest that vortioxetine was more or less effective than its comparators. The committee concluded that all benefits were sufficiently captured in the company's economic modelling.