The Department of Health has asked the National Institute for Health and Care Excellence (NICE) to produce guidance on using rifaximin-a in the NHS in England and Wales. The Appraisal Committee has considered the evidence submitted by the manufacturer and the views of non-manufacturer consultees and commentators, and clinical specialists and patient experts.

This document has been prepared for consultation with the consultees. It summarises the evidence and views that have been considered, and sets out the draft recommendations made by the Committee. NICE invites comments from the consultees and commentators for this appraisal (see section 9) and the public. This document should be read along with the evidence base (the evaluation report).

The Appraisal Committee is interested in receiving comments on the following:

• Has all of the relevant evidence been taken into account?
• Are the summaries of clinical and cost effectiveness reasonable interpretations of the evidence?
• Are the provisional recommendations sound and a suitable basis for guidance to the NHS?
• Are there any aspects of the recommendations that need particular consideration to ensure we avoid unlawful discrimination against any group of people on the grounds of race, gender, disability, religion or belief, sexual orientation, age, gender reassignment, pregnancy and maternity?

Note that this document is not NICE's final guidance on this technology. The recommendations in section 1 may change after consultation.

After consultation:

• The Appraisal Committee will meet again to consider the evidence, this appraisal consultation document and comments from the consultees.
• At that meeting, the Committee will also consider comments made by people who are not consultees.
• After considering these comments, the Committee will prepare the final appraisal determination (FAD).
• Subject to any appeal by consultees, the FAD may be used as the basis for NICE’s guidance on using rifaximin-a in the NHS in England and Wales.

For further details, see the Guides to the technology appraisal process.

The key dates for this appraisal are:

Closing date for comments: 18 December 2013

Second Appraisal Committee meeting: TBC

Details of membership of the Appraisal Committee are given in section 8, and a list of the sources of evidence used in the preparation of this document is given in section 9.

1  Appraisal Committee’s preliminary recommendations

1.1  The Committee is minded not to recommend rifaximin-a within its marketing authorisation, that is, for reducing the recurrence of episodes of overt hepatic encephalopathy in people aged 18 years or older.

1.2  The Committee recommends that NICE requests the following further analyses from the manufacturer, which should be available for the next Appraisal Committee meeting:

• estimates of utilities from SF-36 quality-of-life data collected from the RFHE3001 trial and mapped directly to the EQ-5D, using all observations in each relevant health state and adjusted for baseline utility
• results of scenario analyses reflecting a clinically plausible diminishing mortality benefit with rifaximin-a over time and excluding mortality benefit in the remission state
• useof a lifetime time horizon to capture all costs and benefits associated with rifaximin-a and lactulose
• results of a full probabilistic sensitivity analysis for the revised analyses, including the incremental cost-effectiveness ratios.

2  The technology

2.1  Rifaximin-a (Targaxan, Norgine) is a semi-synthetic derivative of the antibiotic rifamycin, which inhibits ribonucleic acid (RNA). Rifaximin-a decreases intestinal production and absorption of ammonia, which is thought to be responsible for the neurocognitive symptoms of hepatic encephalopathy, thereby delaying the recurrence of acute episodes. Rifaximin-a has a UK marketing authorisation ‘for the reduction in recurrence of episodes of overt hepatic encephalopathy in patients aged 18 years or older’. The summary of product characteristics highlights that 91% of the patients in the pivotal study were using concomitant lactulose.

2.2  The summary of product characteristics lists the following common adverse reactions for rifaximin-a: depression, dizziness, headache, dyspnoea, abdominal pain upper, abdominal distension, diarrhoea, nausea, vomiting, ascites, rashes, pruritus, muscle spasms, arthralgia, anaemia, oedema peripheral and pyrexia. For full details of adverse reactions and contraindications, see the summary of product characteristics.

2.3  Rifaximin-a is available as 550 mg film-coated tablets at a net price of £259.23 per 56-tablet pack (excluding VAT; ‘Medicines Index of Medicinal Specialties [MIMS]). It is administered orally at a recommended dose of 550 mg twice daily. The manufacturer estimated an average cost of £1689.65 for 6 months of treatment. Costs may vary in different settings because of negotiated procurement discounts.

3  The manufacturer’s submission

The Appraisal Committee (section 8) considered evidence submitted by the manufacturer of rifaximin-a and a review of this submission by the Evidence Review Group (ERG; section 9).

Clinical-effectiveness evidence

3.1 The manufacturer conducted a systematic literature review and identified 3 relevant studies including rifaximin-a. Of these, 2 studies were excluded because the doses used were different from those in the UK marketing authorisation for rifaximin-a. The submission therefore included 1 study (Bass et al. 2010) which reported results from the pivotal trial – RFHE3001. This was a 6 month, international, multicentre, randomised, double-blind, placebo-controlled trial comparing rifaximin-a with placebo for maintaining remission in patients with recurrent, overt, episodic hepatic encephalopathy resulting from chronic liver disease. Patients receiving lactulose at baseline were allowed to continue its use during the study period and dose changes were allowed as needed. The manufacturer also identified a second trial – RFHE3002. This was an international, multicentre, single-arm, open-label study, which assessed the long-term safety and tolerability of rifaximin-a in patients with a history of hepatic encephalopathy. It was an extension to RFHE3001 for patients who completed RFHE3001 and for new patients. The manufacturer stated that no peer reviewed publications of RFHE3002 were available at the time of the evidence submission to NICE.

RFHE3001 trial

3.2  Patients were enrolled in RFHE3001 if they had a Conn score of 0 or 1 and were in remission after 2 or more documented recurrent episodes of overt hepatic encephalopathy resulting from chronic liver disease, or portal hypertension equivalent to a Conn score of 2 or more in the 6 months before screening. The Conn score measures the severity of impaired mental status on a scale of 0–4, with higher scores indicating more severe impairment. Patient characteristics at baseline were generally similar between the 2 treatment groups. Patients were only enrolled if they had a Model End Stage Liver Disease score (MELD; used to predict survival to prioritise liver transplant) of less than 25, therefore patients with more severe liver disease (MELD score of 25 or more) were excluded from the study. After screening, 299 patients were randomised to receive either rifaximin-a 550 mg (rifaximin-a group; n=140) or matching placebo (placebo group; n=159) twice daily, plus ongoing treatment with lactulose. In the rifaximin-a and placebo groups, 91.4% and 91.2% of patients took concomitant lactulose respectively. The mean duration of treatment was 130.3 days in the rifaximin-a group and 105.7 days in the placebo group. The rate of adherence, defined as the use of at least 80% of dispensed tablets, was high in both treatment groups (84.3% in the rifaximin-a group and 84.9% in the placebo group). Patients could discontinue treatment if they had an adverse event with an unacceptable risk to them, developed any condition meeting the exclusion criteria, had a breakthrough episode of overt hepatic encephalopathy, became pregnant or asked to be withdrawn.

3.3  The primary outcome in RFHE30001 was time to first breakthrough episode of overt hepatic encephalopathy, defined as an increase in the Conn score from 0 or 1 to 2 or more or an increase in Conn and asterixis score of 1 grade each for patients who entered the study with a Conn score of 0. The asterixis score measures worsening neurological impairment in terms of flapping tremor, which is determined by the patient extending their arms with wrists flexed backwards and fingers open for 30 seconds or more. This is also measured on a scale of 0 to 4, with higher scores indicating more flapping motions. Key secondary outcomes included time to first hepatic encephalopathy-related hospital admission, time to any increase from baseline in Conn score, time to any increase from baseline in asterixis score, mean change from baseline in fatigue domain score on the Chronic Liver Disease Questionnaire (CLDQ) at end of treatment and mean change from baseline in venous ammonia concentration at end of treatment. The CLDQ was used by patients to measure their level of fatigue on a 7-point scale, with ‘1’ representing a high degree of fatigue and ‘7’ representing minimal fatigue. Results were based on the intention-to-treat (ITT) population, that is, all randomised patients who received at least 1 dose of the study drug or placebo.In addition to the CLDQ, health-related quality of life was also assessed using SF-36 and the Epworth Sleepiness Scale.

3.4  There was a statistically significant reduction in the risk of a breakthrough episode with rifaximin-a compared with placebo during the 6-month study period, with a hazard ratio (HR) of 0.42; 95% confidence interval [CI] 0.28 to 0.64, p<0.001. Patients who stopped early for reasons other than a breakthrough episode of overt hepatic encephalopathy were contacted 6 months after randomisation to check if they had a breakthrough episode. The manufacturer stated that therefore breakthrough episodes of overt hepatic encephalopathy were captured completely for up to 6 months after randomisation. Patients who did not have a breakthrough episode during the study period were followed up and assessed after the study was discontinued. Their results were similar to the results for the 6-month study period, with a statistically significant reduction in the risk of a breakthrough episode with rifaximin-a compared with placebo, with a hazard ratio of 0.46; 95% CI 0.31 to 0.69, p<0.0001.

3.5  Age, MELD score, duration of current verified remission and number of prior hepatic encephalopathy episodes were identified as significant prognostic factors. To control for these factors because of chance imbalances between treatment groups, multivariate analysis was performed using the Cox proportional hazards model. This resulted in a statistically significant reduction in the risk of a breakthrough episode of overt hepatic encephalopathy with rifaximin-a compared with placebo (HR 0.40; 95% CI 0.26 to 0.62, p<0.0001).

3.6  Rifaximin-a was associated with statistically significant reductions compared with placebo for the secondary outcomes of time to first hepatic encephalopathy-related hospital admission (HR 0.50; 95% CI 0.29 to 0.87, p=0.01) and time to any increase from baseline in Conn score (HR 0.46; 95% CI 0.31 to 0.69, p<0.0001). However, the reduction in time to any increase from baseline in asterixis score with rifaximin-a did not reach statistical significance when compared with placebo (HR 0.65; 95% CI 0.41 to 1.01, p=0.0523). The differences in the changes from baseline in CLDQ fatigue scores were minimal (3.28 compared with 3.34 at baseline and 3.57 compared with 3.51 at the end of treatment for the rifaximin-a and placebo groups respectively). The manufacturer stated that this was because patients were not able to complete the CLDQ assessment during a breakthrough episode because of altered mental and neuromotor status. Therefore, the CLDQ results for these patients would be similar to baseline levels because mental status was closer to baseline levels at the end of treatment assessments. The manufacturer also stated that there were no consistent differences between the rifaximin-a and placebo groups in change from baseline using the SF-36 and Epworth Sleepiness Scale.The rifaximin-a group had greater reductions in venous ammonia levels compared with the placebo group, although the difference was not statistically significant (p=0.0818). The manufacturer stated that the mortality data from the trial were not mature enough to address the impact of rifaximin-a on survival.

3.7  Pre-planned subgroup analyses indicated that the effect of rifaximin-a in reducing the risk of breakthrough episodes of overt hepatic encephalopathy compared with placebo during the 6-month study period was not statistically significant in the subgroup with MELD scores of 19 to 24 (26 of the 299 patients, p=0.21) and the subgroup who did not receive lactulose at baseline (26 of the 299 patients, p=0.33).

3.8  The safety population (n=299) was described by the manufacturer as patients who received at least 1 dose of study drug and provided at least 1 post-baseline safety assessment. In the rifaximin-a group, 80% of patients had adverse events during the study compared with 79.9% in the placebo group. Most adverse events were mild or moderate. The rifaximin-a group had higher incidences of anaemia (rifaximin-a compared with placebo; 7.9%: 3.8%), peripheral oedema (15%: 8.2%), pyrexia (6.4%: 3.1%), arthralgia (6.4%: 2.5%), and dizziness (12.9%: 8.2%) than the placebo group. Treatment-related adverse events occurred in 19.3% of patients in the rifaximin-a group compared with 21.4% in the placebo group. Approximately 11.4% of patients in the rifaximin-a group and 21.4% in the placebo group had breakthrough episodes of overt hepatic encephalopathy that were considered serious adverse events (for example, needing hospitalisation). A total of 10 (7.1%) patients in the rifaximin-a group and 11 (6.9%) patients in the placebo group died during the study, mainly because of conditions associated with disease progression, including hepatic cirrhosis, decompensated liver cirrhosis or hepatic failure. Adverse events leading to study discontinuation occurred in 21.4% of patients in the rifaximin-a group compared with 28.3% of patients in the placebo group. The manufacturer stated that most of the study discontinuations from adverse events were caused by hepatic encephalopathy events. Adverse events of special interest based on known potential side effects of systemic antibiotics and prior experience with rifaximin-a occurred similarly between the treatment groups, with diarrhoea being the most common (10.7% with rifaximin-a and 13.2 % with placebo).

RFHE3002 trial

3.9  A total of 322 patients were enrolled in the RFHE3002 single-arm, open label study. Of these, 152 continued from RFHE3001 (70 patients from the rifaximin-a group and 82 from the placebo group) and 170 patients were newly enrolled. All patients had a Conn score of 0 to 2 at enrolment and the newly enrolled patients had 1 or more verifiable hepatic encephalopathy episodes in the 12 months before screening. All patients received rifaximin-a 550 mg twice daily and were followed up for at least 24 months, during which time treatment was still ongoing on an outpatient basis until regulatory approval of rifaximin-a or until the sponsor closed the study, whichever came first. All concomitant drugs, including those from RFHE3001, were maintained at stable doses whenever possible. The criteria for study discontinuation were the same as those for RFHE3001, although patients who had an episode of recurrent hepatic encephalopathy during the study were not automatically withdrawn and were allowed to continue on medication.

3.10  Most patients had baseline Conn scores of either 0 (66%) or 1 (30%) and asterixis scores of 0 (71%) or 1 (24%). The time since the most recent verified hepatic encephalopathy episode was shorter in the new rifaximin-a group compared with the continuing rifaximin-a group. The 2 groups of patients were also different in terms of the number of hepatic encephalopathy episodes experienced before screening for RFHE3002 because of the differences in the number of hepatic encephalopathy episodes needed for inclusion in the 2 studies. The manufacturer indicated that the rest of the baseline characteristics were commercial in confidence.

3.11  The main efficacy outcomes assessed were change from baseline in Conn scores and asterixis scores over time. However, the manufacturer indicated that these results were commercial in confidence. The manufacturer stated that the profiles of the time to first breakthrough episode of overt hepatic encephalopathy demonstrated long-term maintenance of remission in new rifaximin‑a and continuing rifaximin-a patients. In addition, 60 rifaximin-a patients from RFHE3001 who had not had an episode of hepatic encephalopathy were followed during RFHE3002. The incidence of breakthrough episodes for these patients was lower than in the RFHE3001 placebo group, after adjusting for the different exposure time between rifaximin-a and placebo. The manufacturer also stated that the all-cause hospitalisation rate was similar to that seen for rifaximin-a during the shorter double-blind trial. However, hazard ratios were not presented.

3.12  A total of 300 patients (93.2%) reported an adverse event in RFHE3002 and approximately 56% of patients receiving rifaximin-a had severe adverse events. Overall, 67 deaths occurred during the study or within 30 days after the last dose of rifaximin-a and 8 patients died more than 30 days after their last dose. The manufacturer stated that none of these deaths were related to rifaximin-a. The manufacturer indicated that the data for treatment-related adverse events, severe adverse events and adverse events resulting in study discontinuation were commercial in confidence.

Cost-effectiveness evidence

3.13  The manufacturer conducted a systematic review of the literature but stated that none of the 3 studies identified were relevant because the patient population and outcomes did not match those of the current decision problem. The manufacturer stated that a comparison with neomycin was excluded because it is not routinely used in clinical practice, no clinical data are available for the use of neomycin in this indication, it is associated with risks of ototoxicity, nephrotoxicity and hepatic impairment and it is also not licensed for the indication being appraised.  

3.14  The manufacturer carried out a de novo analysis of the cost effectiveness of rifaximin-a plus concomitant lactulose compared with placebo plus concomitant lactulose, given that approximately 91% of patients in each arm of RFHE3001 received concomitant lactulose. In the original submission, a Markov cohort model consisting of 3 states (remission, overt and dead) was developed to reflect the clinical pathway of hepatic encephalopathy. Given that time to subsequent episodes was not available from RFHE3001, the manufacturer assumed that the risk of having a subsequent breakthrough episode was independent of the risk of preceding episodes and the time spent in the remission state. It was also assumed that the risk reduction for the first breakthrough episode could also be applied to subsequent episodes, based on clinical expert opinion, and this was assumed to be constant over time for subsequent episodes. However, in response to Committee’s considerations in the first appraisal consultation document that this model structure oversimplified the nature and course of the disease, the manufacturer submitted revised analyses, splitting the remission and overt states, effectively resulting in 5 health states: remission, breakthrough overt episode, subsequent remissions, subsequent overt episodes and death. The model had a time horizon of 5 years consisting of monthly cycles, did not include a half-cycle correction, and both costs and benefits were discounted at 3.5%. The analysis was performed from the perspective of the NHS and personal social services.

3.15  In the original submission, time to first breakthrough episode of hepatic encephalopathy was extracted and combined from the Kaplan-Meier survival curves in RFHE3001 and RFHE3002. Parametric survival distributions were fitted to the combined data set for rifaximin-a patients in order to extrapolate the pivotal study results beyond the final 168 day observation point in RFHE3001 using a log normal distribution. However, in its revised analysis the manufacturer took into account the Committee’s comments that the use of a combined data set was not appropriate. It used data from RFHE3001 only to model time to first breakthrough episode of hepatic encephalopathy for the rifaximin-a and placebo arms and survival times were censored at day 170. The survival curves were then extrapolated beyond 170 days for both groups using a log normal distribution and a proportional hazards assumption. Because there were nodata on subsequent episodes from RFHE3001, the manufacturer used data from the new patients in RFHE3002 only to model subsequent hepatic encephalopathy episodes in order to avoid potential bias from patients who had been in RFHE3001 and did not experience a breakthrough overt episode in RFHE3001. Given that RFHE3002 was a single-arm study (rifaximin-a only), the manufacturer applied the treatment effect for rifaximin-a on first breakthrough episode from RFHE3001 for all subsequent episodes to estimate a placebo arm. All subsequent episodes were modelled together as a single episode and the survival curves were also extrapolated to 5 years using the log normal distribution and applying an assumption of proportional hazards. The manufacturer considered the log normal distribution to provide the best model fit in both cases based on the log likelihood values of the distributions tested and visual assessment.

3.16  The manufacturer also presented 3 separate analyses to justify using RFHE3002 data in the economic model. The analyses included:

• comparing time to first observed breakthrough hepatic encephalopathy episode between the RFHE3002 subgroups, that is, rollover (previous RFHE3001 rifaximin-a patients), crossover (previous RFHE3001 placebo patients) and new patients, from RFHE3002 baseline
• comparing time to first observed breakthrough hepatic encephalopathy episode between RFHE3001 treatment groups and RFHE3002 subgroups, from each respective study baseline
• comparing time to first breakthrough episode between new rifaximin-a RFHE3002 patients and the combined RFHE3001 and RFHE3002 data for rollover and crossover patients.

The results of these analyses showed that time to first breakthrough hepatic encephalopathy episode was comparable between RFHE3001 and RFHE3002 patients given the similarity in their baseline characteristics, except in the number of prior hepatic encephalopathy episodes before trial entry (see sections 3.2 and 3.10).

3.17  In the original submission, hepatic encephalopathy-specific mortality was estimated from other external data sources (Bustamante et al. 1999 and Shawcross et al. 2011) rather than RFHE3001. The manufacturer stated that this was because the trial did not provide evidence on patients at the more severe end of the disease spectrum even though it reflected the range of patients who would present with hepatic encephalopathy in clinical practice, and that data from the trial were not sufficiently mature for an analysis of mortality. However, the manufacturer presented revised mortality estimates after the first appraisal consultation document. In the updated economic analysis, the manufacturer modelled mortality using data for all RFHE3002 patients, given the small number of deaths observed, unlike the subsequent hepatic encephalopathy episodes when only data for the new RFHE3002 patients were used in the model. Time to death for the first remission state was estimated from the first dose of rifaximin-a in RFHE3002, whereas for subsequent remission states it was estimated from day 31 after each overt episode (assumed to be when patients returned to the remission state after an overt episode). The survival curves were then extrapolated based on an assumption of proportional hazards using the log normal distribution for the first remission state and the Weibull distribution for subsequent remissions taking into consideration the log likelihood and visual assessment of the best model fit. The probability of death for the overt states was estimated as the number of deaths within 30 days of the onset of the first observed overt episode in RFHE3002 and the number of deaths within 30 days of the onset of any subsequent overt episode.

3.18  As part of the revised analysis, the manufacturer also carried out a literature review of mortality in people with hepatic encephalopathy and an analysis of the mortality data for people with hepatic encephalopathy using the Clinical Practice Research Datalink (CPRD). The literature review and CPRD data showed that mortality was highest in the first 6 months after diagnosis of hepatic encephalopathy, which could be associated with the first overt hepatic encephalopathy episode. The study by Neff et al. (2012), which was an abstract identified in the literature review, showed that people treated with lactulose alone had a higher 6-month mortality; 40% compared with 35% for people treated with rifaximin-a plus lactulose and 24% for people treated with rifaximin-a alone. The manufacturer also noted from the CPRD data that people who had survived 6 months after the first overt episode had similar mortality to that observed in RFHE3001 and RFHE3002, in which patients had been in remission after at least 1 overt episode in the 12 months before randomisation.

3.19  The overt state also included hospital admissions caused by hepatic encephalopathy episodes. In RFHE3001, 19 of the 31 patients (61.29%) in the rifaximin-a group and 36 of the 73 patients (49.32%) in the placebo group were hospitalised because of a breakthrough episode. In the original submission, these 6-month probabilities of hospital admission (61.29% and 49.32%) were converted to monthly probabilities of 14.63% and 10.71% respectively (assuming a constant hazard over time) and applied to the patients predicted to reach the overt health state. However, in the revised analysis, the manufacturer applied the aggregated 6-month probability of 52.88% to both arms of the model. The manufacturer used the aggregated 6-month probability rather than individual probabilities of 61.29% for rifaximin-a and 49.32% for placebo on the basis that the rates of hepatic encephalopathy hospital admissions in the trial were not statistically significantly different between the 2 arms.

3.20  The utility values used in the original model were derived from 200 randomly selected members of the general public, using the time-trade-off and standard gamble approach. Noting the Committee’s concerns around this approach in the first appraisal consultation document and the Committee’s preference for quality-of-life data collected from RFHE3001, the manufacturer presented alternative results in the revised analysis. The manufacturer used a validated algorithm developed by Gray et al. (2006) to map the SF-36 utility scores collected in RFHE3001 to EQ-5D utility values. The manufacturer also included CLDQ scores which were derived from the quality of life-study by Sanyal et.al. (2011) in the analysis. The study reported CLDQ data only for North American and Canadian patients from RFHE3001 (219 of 299 patients) because there was no validated Russian translation of CLDQ to be administered to Russian patients. It then carried out regression analysis to derive the relationship between overall CLDQ scores from RFHE3001 and EQ-5D utility, which showed that 59% of the variation in the predicted EQ-5D utility was explained by the overall CLDQ score. The manufacturer also carried out a repeated measures analysis to quantify the relationship. The resulting parameter estimate for CLDQ was applied to the baseline CLDQ score, the increment in the CLDQ score for the rifaximin-a arm while in remission and the CLDQ scores for the overt state. The resulting utility value estimated for the baseline remission health state was used for the placebo arm in the updated analysis, and the resulting utility increment of 0.106 was applied to the utility associated with remission in the placebo arm to estimate the utility associated with remission in the rifaximin-a arm. The resulting utility values for the overt state were assumed to be the same for the rifaximin-a and placebo groups. The utility values used in the model were designated academic in confidence by the manufacturer.

3.21  The average monthly cost of rifaximin-a (£281.80) used in the model was based on the recommended dosing schedule of 550 mg twice daily at a unit price of £259.23 per 56-tablet pack. Average monthly lactulose costs (£6.64 and £7.31) for the rifaximin-a and placebo groups respectively were obtained from the ‘British national formulary’ (BNF) 64 based on the mean dose of concomitant lactulose in RFHE3001 (3.14 cups for the rifaximin-a group and 3.51 cups for the placebo group). No administration costs were included in the model because rifaximin-a and lactulose are taken orally; therefore the total costs of treatment per month were £288.34 for the rifaximin-a group compared with £7.31 for the placebo group. The manufacturer estimated a total monthly cost of £36.89 for an outpatient visit in the remission state and £359.73 compared with £292.96 for the overt state (outpatient visit plus hospital admission for 5 days) for the rifaximin-a and placebo groups respectively. No costs were included in the model for adverse events because there were no statistically significant differences between the adverse events reported with rifaximin-a and placebo in RFHE3001 and there was limited evidence available for disutilities associated with the adverse events. This approach was validated by the manufacturer’s clinical experts.

3.22  The updated base-case results comparing rifaximin-a plus concomitant lactulose with placebo plus concomitant lactulose showed an incremental cost of £10,962 and an incremental quality-adjusted life year (QALY) gain of 0.527. This resulted in an incremental cost-effectiveness ratio (ICER) of £20,799 per QALY gained compared with the ICER of £23,186 per QALY gained estimated in the original analysis. The manufacturer stated that the mortality rates predicted by the revised model at 6 months (rifaximin-a: 6.68%; placebo: 10.28%) reflected the trial data and CPRD data better, and these increased to 47.66% and 55.53% respectively after 5 years. In a scenario analysis the manufacturer explored the use of alternative distributions for extrapolating breakthrough and subsequent hepatic encephalopathy episodes. It was noted that varying the distributions for the breakthrough episode only produced a small range of ICERs (from £18,909 per QALY gained using the Weibull distribution to £19,687 per QALY gained using the exponential distribution). Varying the distributions for subsequent episodes only produced a wider range of ICERs (from £13,779 per QALY gained using the exponential distribution to £21,380 per QALY gained using the log-logistic distribution). When a 10-year time horizon was used, the base-case ICER increased to £23,659 per QALY gained.

Evidence Review Group comments

3.23  The ERG stated that it was likely that the manufacturer’s systematic review, updated after clarification, contained all the relevant studies. The ERG identified a 6-month trial of rifaximin-a compared with neomycin reported by Miglio et al. (1997) which was not presented in the manufacturer’s submission. In its clarification response, the manufacturer stated that this trial had been identified but excluded because it was not considered appropriate for this appraisal in terms of the population, treatment regimens and outcomes included. The ERG acknowledged the manufacturer’s justification for excluding the study. However, it remained concerned about excluding neomycin from the analysis because clinical expert opinion indicated that it works in the same way as rifaximin-a and is sometimes used in clinical practice, especially for people not having a liver transplant, even though it is not as well tolerated as rifaximin-a.

3.24  The ERG stated that the evidence submitted by the manufacturer generally reflected the decision problem adequately even though the population and comparator differed from those specified in the final scope. Whereas the scope referred to adults who have had episodes of hepatic encephalopathy, the manufacturer’s submission only considered adults with chronic liver disease, excluding hepatic encephalopathy caused by acute liver disease. Patients with more severe liver disease (MELD score of 25 or more) were also excluded from the analysis. However, the ERG noted the manufacturer’s assertion that the results would apply to this population as well. The ERG was concerned about the validity of this assumption, given that the treatment effect of rifaximin-a compared with placebo was not statistically significant in the subgroup with MELD scores of 19 to 24 (the more severe MELD score category in the trial). In addition, it noted that the study by Hassest et al. (2001), which was provided as evidence for the effectiveness of rifaximin-a in patients with MELD scores of 20 or more, was a poor quality descriptive study. The final scope referred to a comparison of rifaximin-a with lactulose, neomycin or neomycin plus lactulose, but the ERG noted that the analysis presented by the manufacturer was based on rifaximin-a plus concomitant lactulose compared with placebo plus lactulose, in line with the pivotal clinical trial and UK clinical practice. The ERG stated that their clinical expert agreed that UK current practice involved using concomitant lactulose.

3.25  The ERG indicated that the quality of RFHE3001 was high. It noted that the treatment groups were similar in terms of patient characteristics, although there were more men in the placebo group (67%) than in the rifaximin-a group (54%). The ERG stated that the outcomes assessed were appropriate and in line with those specified in the scope. Overall, the ERG considered the clinical effectiveness data and the statistical approaches in the manufacturer’s submission to be of good validity. However, it recognised that RFHE3002 was an unpublished study and although assessments of Conn scores and asterixis scores were conducted to monitor hepatic encephalopathy status, only summary statistics were presented for these analyses.

3.26  The ERG was satisfied with the manufacturer’s modelling approach and agreed that the health states in the model appropriately captured disease progression over time. Clinical opinion obtained by the ERG confirmed that the manufacturer’s assumption of concomitant lactulose use in both arms of the model was appropriate, but emphasised that the effectiveness results were based on 91.3% of patients using rifaximin-a with concomitant lactulose and only 8.7% taking rifaximin-a alone. The ERG stated that a 5-year time horizon was appropriate in the manufacturer’s original model to capture the relevant costs and benefits, but stated that in the revised analysis a lifetime time horizon would have been appropriate given that approximately 52% of patients in the rifaximin-a arm and 45% of patients in the placebo arm were alive after 5 years. In an exploratory analysis, the ERG increased the time horizon to lifetime (40 years, when all patients in the model had died) and this resulted in an ICER of £22,069 per QALY gained.

3.27  The ERG reviewed the manufacturer’s updated analysis and stated that it was unclear why the manufacturer used different time points to censor survival times when modelling breakthrough overt hepatic encephalopathy episodes; data censoring took place at day 168 in the manufacturer’s original analysis and at day 170 in the updated analysis. The ERG highlighted that censoring patients at day 170 resulted in different regression coefficients when extrapolating RFHE3001 data, leading to different transition probabilities. It noted, however, that censoring at day 168 rather than day 170 in the updated analysis would only increase the revised base-case ICER to £21,329 per QALY gained. The ERG stated that the manufacturer should have carried out some validity tests to assess the assumption of a proportional treatment effect between rifaximin-a and placebo when fitting the parametric distributions to the trial data. It noted that assuming a proportional treatment effect implied that people will continue to receive the drug until death, and this was confirmed by the ERG’s clinical experts who indicated that people were generally kept on rifaximin-a until death or liver transplant.

3.28  The ERG stated that clinical opinion indicated that the manufacturer’s revised approach of applying the treatment effect from RFHE3001 to model subsequent overt episodes in the placebo arm was reasonable. The manufacturer’s use of new patients only from RFHE3002 to avoid potential enrichment bias when modelling subsequent overt episodes was also considered reasonable. The ERG stated, based on clinical expert opinion, that the manufacturer’s original assumption of a constant probability of subsequent episodes over time did not reflect reality. The ERG stated that the manufacturer’s updated approach of modelling subsequent episodes as dependent on previous episodes was a significant improvement from the original analysis. However, it was concerned that combining all subsequent episodes into 1 health state does not take into account the number of episodes and does not fully address the issues around the complexity of disease progression. The ERG noted from the manufacturer’s exploratory analysis (see section 3.22) that the choice of distribution for extrapolating hepatic encephalopathy episodes had less impact on the resulting ICERs compared with the original. However, it stated that the manufacturer should have taken further steps to assess the goodness of fit of the distributions used and that the gamma distribution should have been included for completeness.

3.29  The ERG stated that the manufacturer did not justify using different approaches to model mortality, that is, using survival analysis to model mortality in the remission states and using simple proportions to estimate the probability of death in the overt states. It also stated that the manufacturer did not also explain how it calculated the probability of death in the overt states. It noted that the Kaplan-Meier curves were not provided for visual comparison of the different distributions used to model mortality in the remission state and stated that the manufacturer should have taken further steps to assess the goodness of fit of the distributions used to extrapolate mortality. The ERG also highlighted the inconsistency in the use of RFHE3002 data in the model, in which data for all patients in RFHE3002 were used to model mortality, whereas only data for new patients in RFHE3002 were included in the modelling of subsequent episodes.

3.30  The ERG compared mortality from RFHE3001, RFHE3002, the CPRD data and the revised model. It noted that the model over-estimated mortality for the placebo arm at 6 months (10%) compared with RFHE3001 (7%), whereas mortality estimated for the rifaximin-a arm at 6 months was similar to that in RFHE3001 (7%). It also noted that 2-year mortality for the rifaximin-a arm was similar to that observed in RFHE3002 and that 5-year mortality estimated from the model showed an incremental survival benefit of 8% for rifaximin-a compared with placebo. The ERG stated that clinical opinion indicated that, although the survival benefit predicted was reasonable, the overall mortality seemed low for patients in both arms of the model, and also that a higher number of liver transplants would be expected among people with hepatic encephalopathy (only 1 liver transplant was reported in RFHE3001). The ERG explored the impact of assuming no differential mortality between the rifaximin-a and placebo arms by excluding the survival benefit from the remission states only, the overt states only and both states together. This resulted in ICERs of £22,700, £26,120 and £30,200 per QALY gained, respectively, which shows that mortality is still a significant driver of the results.

3.31  The ERG was satisfied with the manufacturer’s use of the 6-month probabilities to model hepatic encephalopathy-related hospital admissions rather than the monthly probabilities used inthe original analysis. However, it stated that clinical opinion did not support the use of an aggregated probability for the rifaximin-a and placebo arms, even though the manufacturer showed that the differences between the individual probabilities were not statistically significant. The ERG stated that the more conservative approach of using the individual probabilities should have been taken and, when this was explored in a scenario analysis, the ICER increased slightly to £21,389 per QALY gained. The ERG noted that, although more patients had an episode of hepatic encephalopathy in the placebo arm than in the rifaximin-a arm, the rate of hospital admission for those who had an episode was higher in the rifaximin-a arm. Therefore, the overall effect of the costs of hospital admission on the ICER was neutralised to some extent, indicating that this was not a key driver of the cost-effectiveness results.

3.32  The ERG stated that it was unclear why the manufacturer used the EQ-5D utility values estimated from the condition-specific CLDQ in the model given that it was possible to incorporate values directly from mapping the SF-36 utilities onto the EQ-5D. It was concerned that the estimated quality-of-life increment with rifaximin-a compared with placebo in the remission state was in contrast to the manufacturer’s original analysis in which the impact on quality of life was linked to movement between the overt and remission health states only. The ERG stated that there was uncertainty in the validity of the utility increment estimated for the rifaximin-a arm given that the manufacturer appeared to have measured the value in centimetres directly from the study by Sanyal et.al rather than using the actual published value of difference in least square means. It also highlighted that applying the increment to the remission state only further increased the uncertainty because the incremental value reported in the study represents the CLDQ data collected for the whole duration of treatment in the trial. The ERG stated that the utility values used in the model were a key driver of the results given that excluding the utility increment from the rifaximin-a arm of the remission state increased the ICER to £59,421 per QALY gained.

3.33  The ERG was satisfied with the calculation of costs in the model. However, it noted that the cost of outpatient attendance for non-hospitalised hepatic encephalopathy events used in the updated analysis (£176.27) was different from the cost of £110.68 used in the original analysis and that using the cost from the original analysis increased the ICER to £23,543 per QALY gained. The ERG noted that the manufacturer did not include costs for adverse events because they were comparable between the rifaximin-a and placebo groups in the pivotal trial, but stated that including relative risks (or risk difference) and 95% confidence intervals for each adverse event would have strengthened this justification. However, the ERG also stated that including adverse events in the costs and QALY calculations would not have a significant impact on the ICER.

3.34  Full details of all the evidence are in the manufacturer’s submission and the ERG report.

4  Consideration of the evidence

The Appraisal Committee reviewed the data available on the clinical and cost effectiveness of rifaximin-a, having considered evidence on the nature of hepatic encephalopathy and the value placed on the benefits of rifaximin-a by people with the condition, those who represent them, and clinical specialists. It also took into account the effective use of NHS resources.

4.1  The Committee considered the clinical management of maintaining remission from episodes of hepatic encephalopathy. The Committee heard from the clinical specialists that after an episode of overt hepatic encephalopathy and treatment to address the precipitating factor, it was important to prevent or reduce the recurrence of overt episodes, which may be fatal. It heard that rifaximin-a was the only licensed treatment available to maintain remission from hepatic encephalopathy episodes. The patient expert stated that hepatic encephalopathy has a profound impact on patients’ daily activities and quality of life with symptoms including personality changes, reduced levels of consciousness and altered neuromuscular activity, with a resulting impact on caregivers. The Committee noted comments from consultation that treatment with rifaximin-a would improve quality of life, prevent readmissions to hospital and reduce morbidity and carer burden. The Committee heard that current standard practice included treatment with lactulose, or other laxatives, primarily because hepatic encephalopathy episodes are often triggered by constipation. However, the Committee heard that lactulose was not well tolerated; when used inlarge doses it can cause diarrhoea, although doses were titrated to improve their tolerability. The patient expert emphasised that this treatment is considered undignified and there is a need for alternative treatment. The patient expert also emphasised that adding rifaximin-a to current treatment with lactulose may result in patients needing reduced doses of lactulose and this could improve adherence to treatment. The clinical specialists also stated that neomycin was not used routinely in clinical practice because of the significant toxicity associated with its long-term use. In particular, its association with deafness made it an unacceptable alternative. The Committee recognised the need for alternative treatment options in maintaining remission from episodes of hepatic encephalopathy and also concluded that neomycin was not an appropriate comparator for this appraisal.

Clinical effectiveness

4.2  The Committee discussed the pivotal RFHE3001 trial. The Committee was satisfied that the trial was well conducted and that relevant outcomes were assessed in line with the scope of the appraisal, including health-related quality of life using the Chronic Liver Disease Questionnaire (CLDQ) and SF-36 questionnaire. However, the Committee noted that mortality was not reported because the data were not considered mature enough to assess mortality. The Committee noted that patients with more severe liver disease (Model End Stage Liver Disease [MELD] score of 25 or more) were excluded from the trial, but that the manufacturer had suggested that the trial results may apply to this group. It noted the Evidence Review Group’s (ERG) concerns that this was unlikely because the effectivenessof rifaximin-a in the subgroup of patients with MELD scores of 19 to 24 was not statistically significantly better than placebo, although the Committee recognised that this subgroup comprised only 26 of the 299 patients in RFHE3001. However, the Committee heard from the clinical specialists that, in practice, patients with a MELD score greater than 24 have a poor prognosis and are often hospitalised awaiting liver transplant. The clinical specialists stated that rifaximin-a would not generally be prescribed for these patients because it was unlikely to provide any meaningful benefit. The Committee was aware that 91.4% of patients in the rifaximin-a arm and 91.2% of patients in the placebo arm received concomitant lactulose and queried whether this reflected normal practice. The clinical specialists confirmed that this was in line with UK clinical practice and there was limited evidence on the efficacy of rifaximin‑a monotherapy (8.6% of patients in RFHE3001).The Committee concluded that RFHE3001 was appropriately conducted and relevant to UK clinical practice.

4.3  The Committee examined the clinical-effectiveness data from RFHE3001,which compared rifaximin-a with placebo. It was aware that the efficacy of rifaximin-a was based on approximately 91% of the patients in each arm receiving concomitant lactulose. The Committee noted that there was a statistically significant increase in time to first breakthrough episode of overt hepatic encephalopathy compared with placebo (HR 0.42; 95% CI 0.28 to 0.64, p<0.001) for the intention-to-treat (ITT) population. It also noted that rifaximin-a was associated with statistically significant reductions in hepatic encephalopathy-related hospital admissions and time to any increase from baseline in Conn scores. However, the improvements in asterixis score and venous ammonia levels were not statistically significant, and the difference in changes in CLDQ fatigue scores, SF-36 scores and Epworth Sleepiness scale scores between the rifaximin-a arm and placebo arm was minimal. The Committee also heard from the clinical specialists that apart from reducing blood ammonia levels, rifaximin-a also influenced endotoxin production and reduced systemic inflammation, which plays an important role in the development of hepatic encephalopathy. The Committee questioned whether this was in line with the proposed mode of action of rifaximin-a, but the manufacturer explained that the pathogenesis remains speculative and is an area of ongoing research. The clinical specialists stated that both the ammonia-lowering and endotoxin-lowering properties of rifaximin-a were relevant in producing a treatment effect. The Committee also discussed RFHE3002, which was performed as an open label, follow-up study to assess the long-term safety and tolerability of rifaximin-a. It noted that not all patients from RFHE3001 continued in RFHE3002, which may be a potential source of selection bias. RFHE3002 provided only exploratory effectiveness data. The Committee also recognised that RFHE3002 was unpublished, although it considered it to be informative to some extent. The Committee considered the pivotal RFHE3001 trial to be the main source of evidence for determining the clinical efficacy of rifaximin‑a, and concluded that rifaximin-a was effective in maintaining remission from episodes of hepatic encephalopathy in the trial population.

4.4  The Committee considered the adverse event profile associated with rifaximin-a in RFHE3001. It noted that hepatic encephalopathy adverse events leading to study discontinuation occurred less frequently in the rifaximin-a group than in the placebo group. However, anaemia, peripheral oedema, pyrexia, arthralgia and dizziness occurred more frequently in the rifaximin-a group than in the placebo group (see section 3.8). The Committee noted that approximately 56% of patients had severe adverse events in RFHE3002. The Committee also noted from the Medicines and Healthcare Products Regulatory Agency (MHRA) public assessment report that cases of Clostridium difficile associated diarrhoea have been reported with the use of rifaximin-a. It understood that this was a potential ongoing safety concern with the use of antibacterial agents including rifaximin-a. It heard from the manufacturer that the MHRA public assessment report stated that there were no new safety concerns with rifaximin-a and the benefit-risk profile was considered to be positive because therapeutic value has been demonstrated. It noted the patient expert’s statement that the potential side effects of rifaximin-a were considered to be more tolerable than the physical and psychological side effects associated with current treatment (such as diarrhoea associated with lactulose treatment), which have a detrimental effect on quality of life for patients and carers. The Committee concluded that the current evidence indicates that rifaximin-a has an acceptable adverse event profile.

 Cost effectiveness

4.5  The Committee considered the manufacturer’s economic model and the ERG’s critique of the model.It accepted the exclusion of neomycin from the analysis because it was not routinely used in clinical practice. The Committee noted that after consultation, the manufacturer amended the way it estimated the risk of subsequent episodes and that the revised model assumed that the risk of subsequent hepatic encephalopathy episodes depended on time since the first episode. The Committee was aware that all subsequent episodes were combined in 1 health state, that is, subsequent remission state and subsequent overt state; thereby ignoring the number of episodes. However, it noted that including more health states for subsequent episodes would increase the complexity of the model and would be difficult to populate with the current evidence base. The Committee therefore concluded that, although the number of episodes was not considered, the revised model was an improvement on that used in the manufacturer’s original analysis, which oversimplified the nature and course of the disease.

4.6  The Committee discussed the appropriateness of the 5-year time horizon used in the model. It noted that 52% of people in the rifaximin-a group and 45% in the placebo group were predicted as being still alive in the model after 5 years. The Committee was aware that the NICE reference case (Guide to the methods of technology appraisal 2013) indicates a preference for a lifetime time horizon when alternative technologies lead to differences in survival or benefits that persist for the remainder of a person’s life. It noted that the ERG had conducted an exploratory analysis with a lifetime time horizon where all patients died after 35 years. However, it heard from the clinical specialist that people with hepatic encephalopathy would not survive for up to 35 years because of their underlying liver cirrhosis. The clinical specialist stated that life expectancy usually ranged between 5 and 15 years depending on disease severity, stating that it was even lower for people with severe disease who had not had a liver transplant. The Committee noted that a time horizon of 35 years did not imply that all patients survived for 35 years. It considered that the prolonged survival of 35 years in the model could be a result of extrapolating outcomes using the log normal distribution, which has a long tail and could lead to implausible survival results. The Committee concluded that the manufacturer’s choice of a 5-year time horizon was not in line with the NICE reference case and that a lifetime time horizon would be more appropriate to capture all relevant costs and benefits. It recommended that further analyses using a lifetime time horizon be requested.

4.7  The Committee discussed the assumptions used to model clinical outcomes. It noted that, after consultation, the manufacturer used data from RFHE3001 only to model breakthrough overt hepatic encephalopathy episodes for both arms of the model, rather than the combined data set from RFHE3001 and RFHE3002 used to model the rifaximin-a arm in the original analysis. The Committee also noted the ERG’s comment that data censoring when modelling time to first breakthrough overt episode was inconsistent between the revised and original analysis, that is, at day 170 and day 168 respectively. However, it considered the revised analysis at day 170 to be more appropriate given that the clinical study report for RFHE3001 reports data censoring at day 170. It also noted that the manufacturer included only new patients from RFHE3002 when modelling subsequent overt episodes to avoid potential enrichment bias from rollover and crossover patients who did not experience an overt episode in RFHE3001. The Committee was satisfied with this approach given the evidence presented by the manufacturer showing that time to first breakthrough overt episode was not statistically significantly different between the RFHE3002 subgroups (rollover, crossover and new patients). The Committee noted that there was less variation in the incremental cost-effectiveness ratios (ICERs) from the revised analysis (see section 3.22) than those of the original analysis when different distributions were used to model time to first breakthrough and subsequent overt episodes. The Committee concluded that the manufacturer’s revised modelling of hepatic encephalopathy episodes was appropriate.

4.8  The Committee considered the manufacturer’s approach to utility estimation. It noted that, in response to the first appraisal consultation document, the utilities used were based on quality-of-life data collected from RFHE3001 rather than derived from the general public in a study commissioned by the manufacturer in the original submission. However, the Committee expressed several concerns around the manufacturer’s revised approach. First, the Committee was unsure why the manufacturer had not presented results based on mapping SF-36 data collected from RFHE3001 directly to EQ-5D. The Committee considered that incorporating further modelling steps after the mapping exercise to quantify the relationship with CLDQ scores unnecessarily introduced further uncertainties in the analysis. Second, the Committee was concerned that the manufacturer used utilities estimated from the quality-of-life study by Sanyal et al. (2011), which reported CLDQ data from a post-hoc analysis of RFHE3001. It noted that the study included quality-of-life data for North American and Canadian patients only (219 of 299 patients) from RFHE3001, and that Russian patients were excluded from the analysis because there was no validated Russian translation of CLDQ, resulting in loss of data. It heard from the manufacturer that quality-of-life data collected from RFHE3001 using the CLDQ and SF-36 were collected at discrete time points and did not reflect the decrement in quality of life of hepatic encephalopathy episodes because patients were unable to rate their health during an overt episode; therefore the data collected only reflected the remission period.The manufacturer stated that it used the CLDQ values from the post-hoc analysis because quality of life was assessed as a time-weighted average, which it considered to be a more reasonable approach for capturing important outcomes. The Committee was, however, aware that the area under the curve methodology could have been applied directly with the SF-36 mapped scores. The Committee also noted that there was no significant difference in the CLDQ fatigue domain scores between rifaximin-a and placebo in RFHE3001, whereas the post-hoc analysis reported statistically significant improvement with rifaximin-a across all items in the fatigue domain. In particular, the Committee noted that the post-hoc analysis indicated that there were significant differences in the drowsiness and sleepiness score of the CLDQ fatigue domain, whereas the specific Epworth Sleepiness Scale used in the trial indicated no consistent differences between the placebo and rifaximin-a groups in change from baseline. The Committee concluded that the methodology used by the manufacturer to estimate utilities was not appropriate. 

4.9  The Committee considered the utilities applied by the manufacturer in the economic model. The Committee noted that the utilities estimated showed an increment of 0.106 with rifaximin-a compared with placebo for the remission state, whereas the manufacturer’s original submission indicated that there were no significant differences in the SF-36 scores between the rifaximin-a and placebo groups. The Committee was concerned that a substantial benefit in utility of 0.106 was estimated for rifaximin-a when people are in remission, that is, Conn scores 0 and 1. However, it heard from the clinical specialist that remission (in particular, Conn score 1) is also known as minimal hepatic encephalopathy and patients still experience mild neurological abnormalities, reversal of the sleep-wake cycle and reduced quality of life. The clinical specialist also stated that Conn scores were based on a physical examination, lower Conn scores do not indicate normal quality of life and that rifaximin-a improves quality of life substantially and enables people to return to their normal activities with less dependence on carers. The Committee also expressed concern regarding the face validity of the utilities used in the model. It noted that the difference between the remission and overt state utility values was less than the benefit assumed for rifaximin-a compared with placebo for the remission state. The Committee considered that although a utility benefit with rifaximin-a was very plausible, a utility benefit of 0.106 in the remission state was likely to be an overestimation. In addition, the Committee considered that the utility difference between the remission and the overt state was likely to be underestimated. The Committee was aware that both these elements would push the ICERs in different directions. The Committee noted the ERG’s exploratory analysis excluding the utility increment for rifaximin-a from the remission state, which resulted in an ICER of approximately £59,400 per quality-adjusted life year (QALY) gained. Given the concerns around methodology and the inconsistencies identified in the results available, the Committee considered that the utility estimates driving the model were not robust. The Committee concluded that it would be more appropriate to use utility values estimated by mapping SF-36 data from the entire RFHE3001 population directly to EQ-5D and it recommended that further analyses incorporating this data be requested.

4.10  The Committee considered the costs used in the model. It was aware that, based on the literature, length of hospital stay was assumed to be equal in both treatment arms. It also noted that length of stay was assumed to be 5 days in the model, whereas data from Hospital Episode Statistics for encephalopathy admissions suggests mean stays of around 17 days. The clinical specialist stated that an assumption of 5 days was reasonable.The Committee considered it would have been useful for this information to have been collected in the trial so that differences in length of stay could have been appropriately captured in the model. The Committee noted that the manufacturer used the aggregated probability for hepatic encephalopathy-related hospital admissions in the model on the basis that the difference between the rates of hospital admissions for the rifaximin-a and placebo groups was not statistically significantly different. The Committee considered this approach to be selective and that the individual rates would have been more appropriate, although it noted that the impact on the ICER was minimal when the ERG used the individual rates in a scenario analysis (see section 3.31). It also noted that adverse events were excluded from the cost calculation on the basis that there were no statistically significant differences between the treatment arms in RFHE3001 and there was limited evidence for disutilities associated with the adverse events. It heard from the clinical specialists that adverse events were minimal and primarily related to lactulose use. The ERG stated that including adverse events was not likely to have a large impact on the cost-effectiveness results. The Committee was uncertain of the impact on the ICER and concluded that an attempt should have been made to capture some of the differences in the costs of adverse events between the rifaximin-a and placebo arms.

4.11  The Committee considered the mortality estimates from the manufacturer’s updated analysis. It noted that mortality predicted by the model was similar to the CPRD, RFHE3001 and RFHE3002 data, in contrast to the original analysis in which approximately 50% of the model population died over a 6-month period compared with 7% (21 deaths) in RFHE3001. However, the Committee was concerned that the model predicted a difference of 3.6% in the risk of mortality with rifaximin-a compared with placebo during the first 6 months of the model and a difference of up to 8% over the 5-year time horizon. It noted that no differences in mortality between rifaximin-a and placebo were observed in RFHE3001. The Committee noted the data showing a differential mortality benefit for rifaximin-a in the additional evidence presented by the manufacturer, but was aware that this was from a retrospective observational study (Neff et.al (2012) published in abstract form only. It also noted that 6-month data from this study indicated significantly higher mortality with rifaximin-a plus lactulose and lactulose only (35% and 40% respectively) than occurred in RFHE3001. Further, the Committee noted that mortality with rifaximin-a plus lactulose in the study was 11% higher than with rifaximin-a alone and considered that this was unexpected. The Committee was not persuaded that these data provided robust evidence for a differential mortality benefit for rifaximin-a for preventing recurrence of overt hepatic encephalopathy. The Committee accepted that no additional mortality data were available beyond those presented by the manufacturer in its revised analysis, and concluded that these data reflected the trial and UK clinical practice better than the original estimates.

4.12  The Committee considered whether it was clinically plausible to expect a differential mortality benefit with rifaximin-a plus lactulose compared with lactulose only. It noted that the manufacturer’s updated analysis demonstrated that the time to subsequent overt hepatic encephalopathy episodes decreased as the number of episodes increased. The Committee expressed concern that the model predicted an increasing differential mortality benefit for rifaximin-a when clinical experience indicates decreasing clinical efficacy over time in preventing overt hepatic encephalopathy episodes. The clinical specialist stated that it was plausible that rifaximin-a could have an impact on mortality and the modelled differential seemed reasonable but acknowledged that, with disease progression, the effect in preventing overt hepatic encephalopathy episodes would decrease and that there would be additional causes of death, independent of hepatic encephalopathy, including renal failure, sepsis and bleeding from oesophageal varices. The Committee also noted that the ERG’s exploratory analysis indicated a modelled mortality benefit for rifaximin-a in the remission state that was independent of the effect in preventing episodes of overt hepatic encephalopathy. The Committee noted from the ERG’s exploratory analyses that excluding a differential mortality benefit with rifaximin-a from the remission state increased the ICER to £22,700 per QALY gained and to £30,200 per QALY gainedwhen  mortality benefit was excluded from both health states. The Committee concluded that although an initial mortality benefit with rifaximin-a resulting from a reduction in overt hepatic encephalopathy episodes was plausible, the magnitude of the effect is uncertain and the benefit will likely diminish over time with disease progression. It recommended that further analyses be requested excluding mortality benefit in the remission state and incorporating a clinically plausible diminishing mortality benefit over time.

4.13  The Committee considered whether rifaximin-a reflected a cost-effective use of NHS resources. The Committee noted that the manufacturer’s revised base-case ICER was £20,800 per QALY gained for rifaximin-a plus lactulose compared with placebo plus lactulose. The Committee considered that addressing the issues around the time horizon (see section 4.6) and the use of aggregated probability for hepatic encephalopathy-related hospital admissions would increase the ICER further, but agreed that the impact would not be substantial. The Committee was aware that the key drivers of the cost-effectiveness results were the utility and mortality estimates. It noted that excluding the utility increment with rifaximin-a in the remission state increased the ICER to £59,400 per QALY gained. Whereas the Committee was persuaded that rifaximin-a would be associated with some quality-of-life benefit, it did not consider the magnitude estimated by the manufacturer to be plausible. Therefore, it considered that the base-case ICER of £20,800 per QALY gained was an underestimate, but it was uncertain as to what the most plausible ICER would be given that the most appropriate data for estimating the utilities were not used. It also noted that excluding the mortality benefit with rifaximin-a increased the ICER to £30,200 per QALY gained. The Committee was concerned about the incremental benefit with rifaximin-a in the remission state and the increasing mortality benefit with disease progression. The Committee was also concerned that the manufacturer did not report a probabilistic ICER that would take account of all the uncertainties in the model parameters simultaneously. The Committee concluded that these were important limitations of the updated economic analysis, even though the analysis was an improvement on the original analysis. Based on the evidence presented, the Committee was minded not to recommend rifaximin-a as a cost-effective use of NHS resources for reducing the recurrence of episodes of overt hepatic encephalopathy. It requested  further analyses including:

• estimates of utilities from SF-36 quality-of-life data collected from the RFHE3001 trial and mapped directly to the EQ-5D, using all observations in each relevant health state and adjusted for baseline utility
• results of scenario analyses reflecting a clinically plausible diminishing mortality benefit with rifaximin-a over time and excluding mortality benefit in the remission state
• use of a lifetime time horizon to capture all costs and benefits associated with rifaximin-a and lactulose
• results of a full probabilistic sensitivity analysis for the revised analysis, including the ICERs.

4.14  The Committee noted the manufacturer’s comment regarding the innovative nature of rifaximin-a. The manufacturer stated that rifaximin-a was expected to offer a step change in the management of hepatic encephalopathy by significantly reducing breakthrough episodes and hospital admissions, while maintaining health-related quality of life. The Committee heard from the clinical specialists that the use of rifaximin-a could potentially reduce the number of people on the liver transplant list, which would reduce the burden of expensive procedures on the healthcare system. The patient expert also emphasised that availability of rifaximin-a would reduce the doses of lactulose given to people, thereby improving their wellbeing. The Committee noted that rifaximin-a for treating hepatic encephalopathy was a new indication for the antibiotic and it is well tolerated, but considered that there were no additional gains in health-related quality of life over those already included in the QALY calculations. However, the Committee also acknowledged that hepatic encephalopathy has a far-reaching effect on family and carers, including loss of income in order to become full-time carers. It heard from the clinical specialist that symptoms of hepatic encephalopathy are similar to dementia and can also cause features of Parkinson’s disease. Patients develop depression, wander at night, are unable to work and require constant supervision either from family or professional carers. The Committee agreed that these costs could not be built into the model and factoring them in would reduce the ICER.

Summary of Appraisal Committee’s key conclusions

TAXXX	Appraisal title: Rifaximin-a for maintaining remission from episodes of hepatic encephalopathy		Section
Key conclusion
The Committee is minded not to recommend rifaximin-a within its marketing authorisation, that is, for reducing the recurrence of episodes of overt hepatic encephalopathy in people aged 18 years or older. The Committee recommends that NICE requests the following further analyses from the manufacturer, which should be available for the next Appraisal Committee meeting: estimates of utilities from SF-36 quality-of-life data collected from the RFHE3001 trial and mapped directly to the EQ-5D, using all observations in each relevant health state and adjusted for baseline utility results of scenario analyses reflecting a clinically plausible diminishing mortality benefit with rifaximin-a over time and excluding mortality benefit in the remission state use of a lifetime time horizon to capture all costs and benefits associated with rifaximin-a and lactulose results of a full probabilistic sensitivity analysis for the revised analysis, including the incremental cost-effectiveness ratios.			1.1               1.2
Current practice
Clinical need of patients, including the availability of alternative treatments		The clinical specialists stated that after an episode of overt hepatic encephalopathy and treatment to address the precipitating factor, it was important to prevent or reduce the recurrence of overt episodes, which may be fatal. The Committee heard that lactulose, which is standard treatment, was not well tolerated when used in large doses because it can cause diarrhoea and that neomycin was not used routinely in clinical practice because of the significant toxicity associated with its long-term use. The Committee recognised the need for alternative treatment options in maintaining remission from episodes of hepatic encephalopathy and also concluded that neomycin was not an appropriate comparator for this appraisal.	4.1
The technology
Proposed benefits of the technology How innovative is the technology in its potential to make a significant and substantial impact on health-related benefits?		The Committee concluded that rifaximin-a was effective in maintaining remission from episodes of hepatic encephalopathy in the trial population. The Committee considered that there were no additional gains in health-related quality of life over those already included in the QALY calculations.	4.3                       4.14
What is the position of the treatment in the pathway of care for the condition?		Rifaximin-a has a UK marketing authorisation ‘for the reduction in recurrence of episodes of overt hepatic encephalopathy in patients aged 18 years or older’. The Committee heard that current standard practice included treatment with lactulose, or other laxatives, and that neomycin was not an appropriate comparator.	4.1
Adverse reactions		The Committee concluded that the current evidence indicates that rifaximin-a has an acceptable adverse event profile.	4.4
Evidence for clinical effectiveness
Availability, nature and quality of evidence		The Committee was satisfied that the RFHE3001 trial was well conducted and that relevant outcomes were assessed in line with the scope of the appraisal, including health-related quality of life using the CLDQ and SF-36 questionnaire. The Committee noted that the open label, follow-up study, RFHE3002, provided only exploratory effectiveness data. The Committee also recognised that RFHE3002 was unpublished, although it considered it to be informative to some extent.	4.2                               4.3
Relevance to general clinical practice in the NHS		The Committee heard that patients with a MELD score greater than 24 (who were excluded from the trial) have a poor prognosis; and in practice would not generally be prescribed rifaximin-a because it was unlikely to provide any meaningful benefit. The Committee also heard that the concomitant use of lactulose by 91.3% of patients in the trial was in line with UK clinical practice and concluded that RFHE3001 was relevant to UK clinical practice.	4.2
Uncertainties generated by the evidence		The Committee noted that not all patients from RFHE3001 continued in RFHE3002, which may be a potential source of selection bias. RFHE3002 provided only exploratory effectiveness data.	4.3
Are there any clinically relevant subgroups for which there is evidence of differential effectiveness?		None	–
Estimate of the size of the clinical effectiveness including strength of supporting evidence		The Committee noted that there was a statistically significant increase in time to first breakthrough episode of overt hepatic encephalopathy compared with placebo (HR 0.42; 95% CI 0.28 to 0.64, p<0.001). It also noted that rifaximin-a was associated with statistically significant reductions in hepatic encephalopathy-related hospital admissions and time to any increase from baseline in Conn scores. However, the improvements in asterixis score and venous ammonia levels were not statistically significant, and the difference in changes in CLDQ fatigue scores, SF-36 scores and Epworth Sleepiness scale scores between the rifaximin-a arm and placebo arm was minimal. The Committee concluded that rifaximin-a was effective in maintaining remission from episodes of hepatic encephalopathy in the trial population.	4.3
Evidence for cost effectiveness
Availability and nature of evidence		The Committee concluded that, although the number of hepatic encephalopathy episodes was not considered, the revised model was an improvement on that used in the manufacturer’s original analysis, which oversimplified the nature and course of the disease.	4.5
Uncertainties around and plausibility of assumptions and inputs in the economic model		The Committee concluded that the manufacturer’s choice of a 5-year time horizon was not in line with the NICE reference case and that a lifetime time horizon would be more appropriate to capture all relevant costs and benefits. Given the concerns around methodology and the inconsistencies identified in the results available, the Committee considered that the utility estimates driving the model were not robust. The Committee concluded that it would be more appropriate to use utility values estimated by mapping SF-36 data from the entire RFHE3001 population directly to EQ-5D. The Committee accepted that no additional mortality data were available beyond those presented by the manufacturer in its revised analysis, and concluded that these data reflected the trial and UK clinical practice better than the original estimates. The Committee also concluded that although an initial mortality benefit with rifaximin-a resulting from a reduction in overt hepatic encephalopathy episodes was plausible, the magnitude of the effect is uncertain and the benefit will likely diminish over time with disease progression.	4.6                          4.9                                  4.11, 4.12
Incorporation of health-related quality-of-life benefits and utility values Have any potential significant and substantial health-related benefits been identified that were not included in the economic model, and how have they been considered?		The Committee considered that there were no additional gains in health-related quality of life over those already included in the QALY calculations. However, the Committee also acknowledged that hepatic encephalopathy has a far-reaching effect on family and carers and agreed that these costs could not be built into the model and factoring them in would reduce the ICER.	4.14,
Are there specific groups of people for whom the technology is particularly cost effective?		None identified.	–
What are the key drivers of cost effectiveness?		The Committee was aware that the key drivers of the cost-effectiveness results were the utility and mortality estimates.	4.13
Most likely cost-effectiveness estimate (given as an ICER)		The Committee considered that the base-case ICER of £20,800 per QALY gained was an underestimate, but it was uncertain as to what the most plausible ICER would be given that the most appropriate data for estimating the utilities were not used.	4.13
Additional factors taken into account
Patient access schemes (PPRS)		None	–
End-of-life considerations		N/A	–
Equalities considerations and social value judgements		No potential equality issues were identified.	-

 

5  Implementation

NICE has developed tools [link to www.nice.org.uk/guidance/TAXXX] to help organisations put this guidance into practice (listed below). [NICE to amend list as needed at time of publication]

• Slides highlighting key messages for local discussion.Costing template and report to estimate the national and local savings and costs associated with implementation.
• Implementation advice on how to put the guidance into practice and national initiatives that support this locally.
• A costing statement explaining the resource impact of this guidance.
• Audit support for monitoring local practice.

6  Related NICE guidance

There is no related guidance for this technology.

7  Proposed date for review of guidance

7.1  NICE proposes that the guidance on this technology is considered for review by the Guidance Executive in July 2017. NICE welcomes comment on this proposed date. The Guidance Executive will decide whether the technology should be reviewed based on information gathered by NICE, and in consultation with consultees and commentators.

Gary McVeigh
Chair, Appraisal Committee
November 2013

8  Appraisal Committee members and NICE project team

8.1  Appraisal Committee members

The Appraisal Committees are standing advisory committees of NICE. Members are appointed for a 3-year term. A list of the Committee members who took part in the discussions for this appraisal appears below. There are 4 Appraisal Committees, each with a chair and vice chair. Each Appraisal Committee meets once a month, except in December when there are no meetings. Each Committee considers its own list of technologies, and ongoing topics are not moved between Committees.

Committee members are asked to declare any interests in the technology to be appraised. If it is considered there is a conflict of interest, the member is excluded from participating further in that appraisal.

The minutes of each Appraisal Committee meeting, which include the names of the members who attended and their declarations of interests, are posted on the NICE website.

Professor Gary McVeigh (Chair)
Professor of Cardiovascular Medicine, Queens University Belfast and Consultant Physician, Belfast City Hospital

Dr Lindsay Smith (Vice Chair)
General Practitioner, West Coker Surgery, Somerset

Dr Aomesh Bhatt
Regulatory and Medical Affairs Director Europe and North America, Reckitt Benckiser

Dr Andrew Black
General Practitioner, Mortimer Medical Practice, Herefordshire

Dr Matthew Bradley
Therapy Area Leader, Global Health Outcomes, GlaxoSmithKline

Dr Ian Davidson
Lecturer in Rehabilitation, University of Manchester

John Dervan
Lay Member

Professor Simon Dixon
Professor of Health Economics, University of Sheffield

Dr Martin Duerden
Assistant Medical Director, Betsi Cadwaladr University Health Board, North Wales

Christopher Earl
Surgical Care Practitioner, Wessex Neurological Centre at Southampton University Hospital

Gillian Ells
Prescribing Advisor – Commissioning, NHS Hastings and Rother and NHS East Sussex Downs and Weald

Professor Paula Ghaneh
Professor and Honorary Consultant Surgeon, University of Liverpool

Dr Susan Griffin
Research Fellow, Centre for Health Economics, University of York

Professor Carol Haigh
Professor in Nursing, Manchester Metropolitan University

Professor John Henderson
Professor of Paediatric Respiratory Medicine, University of Bristol and Bristol Royal Hospital for Children

Dr Paul Hepple
General Practitioner, Muirhouse Medical Group

Professor John Hutton
Professor of Health Economics, University of York

Professor Peter Jones
Emeritus Professor of Statistics, Keele University

Professor Steven Julious
Professor of Medical Statistics, University of Sheffield

Dr Tim Kinnaird
Lead Interventional Cardiologist, University Hospital of Wales, Cardiff

Warren Linley
Senior Research Fellow, Centre for Health Economics and Medicines Evaluation, Bangor University

Professor Jonathan Michaels
Professor of Clinical Decision Science, University of Sheffield

Professor Femi Oyebode
Professor of Psychiatry and Consultant Psychiatrist, The National Centre for Mental Health

Dr John Radford
Director of Public Health, Rotherham Primary Care Trust and MBC

Dr Brian Shine
Consultant Chemical Pathologist, John Radcliffe Hospital, Oxford

Dr Murray Smith
Associate Professor in Social Research in Medicines and Health, University of Nottingham

Paddy Storrie
Lay member

Charles Waddicor
Chief Executive, NHS Berkshire West

8.2  NICE project team

Each technology appraisal is assigned to a team consisting of 1 or more health technology analysts (who act as technical leads for the appraisal), a technical adviser and a project manager.

Nwamaka Umeweni
Technical Lead

Raisa Sidhu
Technical Adviser

Kate Moore
Project Manager

9  Sources of evidence considered by the Committee

A. The Evidence Review Group (ERG) report for this appraisal was prepared by Peninsula Technology Assessment Group (PenTAG):

• Bacelar M, et al. The clinical and cost-effectiveness of rifaximin for maintaining remission from episodes of hepatic encephalopathy: A critique of the submission from Norgine, (April, 2013)

B. The following organisations accepted the invitation to participate in this appraisal as consultees and commentators. They were invited to comment on the draft scope, the ERG report and the appraisal consultation document (ACD). Organisations listed in I were also invited to make written submissions. Organisations listed in II and III had the opportunity to give their expert views. Organisations listed in I, II and III also have the opportunity to appeal against the final appraisal determination.

I. Manufacturer/sponsor:

• Norgine

II. Professional/specialist and patient/carer groups:

• British Liver Trust
• British Infection Association
• British Society of Gastroenterology
• Royal College of Nursing
• Royal College of Pathologists
• Royal College of Physicians

III. Other consultees:

• Department of Health
• South Essex PCT Cluster
• Welsh Government

IV. Commentator organisations (did not provide written evidence and without the right of appeal):

• Commissioning Support Appraisals Service
• Department of Health, Social Services and Public Safety for Northern Ireland
• Healthcare Improvement Scotland
• Foundation for Liver Research
• National Institute for Health Research Health Technology Assessment programme
• Peninsula Technology Assessment Group (PenTAG)

C. The following individuals were selected from clinical specialist and patient expert nominations from the consultees and commentators. They gave their expert personal view on rifaximin-α by attending the initial Committee discussion and providing written evidence to the Committee. They are invited to comment on the ACD.

• Dr Sulleman Moreea, consultant Gastroenterologist/Hepatologist, nominated by the British Society of Gastroenterology – clinical specialist
• Dr Debbie Shawcross, Senior Lecturer and Honorary Consultant in Hepatology, nominated by Norgine – clinical specialist
• Andrew Langford, Chief Executive of the British Liver Trust, nominated by British Liver – patient expert

D. Representatives from the following manufacturer/sponsor attended Committee meetings. They contributed only when asked by the Committee chair to clarify specific issues and comment on factual accuracy.

• Norgine