Abstract
Objective:
To determine the short-term costs per sustained remission and sustained response of three tumor necrosis factor inhibitors (infliximab, adalimumab, and golimumab) in comparison to conventional therapy for the treatment of moderately-to-severely active ulcerative colitis.
Methods:
A probabilistic Markov model was developed. This included an 8-week induction period, and 22 subsequent 2-week cycles (up to 1 year). The model included three disease states: remission, response, and relapse. Costs were from a Canadian public payer perspective. Estimates for the additional cost per 1 year of sustained remission and sustained response were obtained.
Results:
Golimumab 100 mg provided the lowest cost per additional remission ($935) and cost per additional response ($701) compared with conventional therapy. Golimumab 50 mg yielded slightly higher costs than golimumab 100 mg. Infliximab was associated with the largest additional number of estimated remissions and responses, but also higher cost at $1975 per remission and $1311 per response. Adalimumab was associated with the largest cost per remission ($7430) and cost per response ($2361). The cost per additional remission and cost per additional response associated with infliximab vs golimumab 100 mg was $14,659 and $4753, respectively.
Conclusions:
The results suggest that the additional cost of 1 full year of remission and response are lowest with golimumab 100 mg, followed by golimumab 50 mg. Although infliximab has the highest efficacy, it did not exhibit the lowest cost per additional remission or response. Adalimumab produced the highest cost per additional remission and response.
Key words::
Background
Response and remission are among the most clinically relevant outcomes for patients treated for ulcerative colitis (UC)Citation1. Randomized controlled trials (RCTs) have shown superior response and remission rates among patients treated with tumor necrosis factor α inhibitors (anti-TNFs) compared to conventional therapy (i.e., corticosteroids, aminosalicylates, and/or immunosuppressants) for moderately-to-severely active UCCitation2–4. Since no RCTs have assessed anti-TNFs head-to-head, indirect treatment comparisons (ITCs) have been conductedCitation5–8. The first of these analyses presented in the peer-reviewed literature indicated that the odds of achieving remission and response after treatment induction were approximately twice as high with infliximab than with adalimumabCitation7. Two ITCs assessed the treatments adalimumab, infliximab, and golimumabCitation6,Citation8. Both of these analyses demonstrated similar trends; however, only one accounted for the differences in trial designCitation6. The most recent ITC including the treatment vedolizumab (which is not of interest in the current study due to its pending regulatory approval) found results similar to the previous three ITCs; however, this ITC did not account for differences in trial designCitation5.
In addition to different efficacy profiles among the anti-TNFs, the costs of these treatments also vary. Golimumab and adalimumab have similar treatment costs, whereas infliximab is more expensive than either of the two treatmentsCitation9. There have been several cost-utility analyses previously conducted to determine the impact of anti-TNFs on overall quality-of-lifeCitation10–14. However, there has only been one cost-effectiveness analysis designed to specifically assess remission and response. This analysis only assessed marginal and not incremental cost-effectivenessCitation15. There has yet to be a cost-effectiveness analysis assessing comparative remission and response in all anti-TNFs currently approved for treatment for UC.
Therefore, this study determined the short-term (i.e., 1 year) costs per sustained remission and sustained response associated with infliximab, adalimumab, and golimumab compared with conventional therapy (i.e., corticosteroids, aminosalicylates, and/or immunosuppressants) for the treatment of moderately-to-severely active UC. This cost-effectiveness analysis makes use of treatment costs in Canada and the ITC that includes all approved treatments while also accounting for differences in trial designCitation6.
Methods
Model overview
A probabilistic Markov state-transition model was developed to determine the cost per additional 1-year remission and additional 1-year response with each of the three anti-TNFs compared to conventional therapy. Five treatment groups were included in the model: conventional therapy (i.e., corticosteroids, aminosalicylates, and/or immunosuppressants), which was set as the anchor, infliximab 5 mg/kg, adalimumab 40 mg, golimumab 50 mg, and golimumab 100 mg. Clinical response was defined as a decrease from baseline in the Mayo score ≥30% and ≥3 points, accompanied by either a rectal bleeding sub-score of 0 or 1 or a decrease from baseline in the rectal bleeding sub-score ≥1. Clinical remission was defined as a Mayo score ≤2 points, with no individual sub-score >1Citation1,Citation3.
The model structure is presented in . The model time horizon was 1 year, which consisted of an 8-week induction period, and a 44-week maintenance period. The 8-week induction period was the first cycle in the Markov model. The 44-week maintenance period was made up of 22 2-week cycles. The model included three states: remission, response, and relapse (i.e., neither remission nor response). However, since costs of treatments (i.e., anti-TNFs plus concomitant therapies) are determined by the proportion of patients either in remission or response (i.e., not in relapse), remission and response were modeled as one state. After the induction period, patients could enter into any one of these states. If patients in any of the anti-TNF groups were not in remission or response, they would be discontinued from the anti-TNF (and only incur the cost of conventional therapy for the remainder of the year). Patients on conventional therapy were continued on the same treatment for the full year, regardless of their response status.
Figure 1. Disease state transitions over 1 year.
The model was developed from the perspective of the Canadian publicly-funded healthcare system. Costs included in this perspective were the interventions, concomitant therapies, and inpatient, outpatient, and laboratory fees covered under public medical service plans.
Health states and transition probabilities
Transition probabilities were derived from an indirect comparison based on the key five RCTs of infliximab, adalimumab, and golimumabCitation6. For each anti-TNF treatment group, the estimated proportion of patients in response after treatment induction (8 weeks) and the probabilities of patients sustaining their response at 1-year were used as the foundations for deriving the probability of sustaining response at the end of every 2-week cycle in the maintenance period. One should note that, in the PURSUIT trial, the induction period was 6 weeks; however, the model estimated all costs under the assumption of an 8-week induction period. The estimated probability of response at 8 weeks, Pr(8 weeks), the estimated probability of sustaining response until the end of the 1-year maintenance therapy, Pr(1 year), and the isolated probability of sustaining a response at any 2-week cycle, Pr(current cycle | previous cycle), can be determined as follows:
The 8-week and sustained 1-year probability of response and remission for conventional therapy were obtained as the average across the five key trials included in the indirect comparisonCitation2–4,Citation16. The calculated transition probabilities are presented in .
Table 1. One year remission probabilities and response probabilities associated with conventional therapy, infliximab, adalimumab, and golimumab.
Costs
All cost data obtained are presented in . Costs of the anti-TNFs were obtained from the Régie de l’assurance maladie du Québec Manuel des médecins specialists (RAMQ). Infliximab, adalimumab, and golimumab unit costs were $940, $729, and $1447, respectively. Golimumab 50 mg and 100 mg have the same cost per unit in the current context.
Table 2. Obtained costs of anti-TNF therapies, concomitant therapies, laboratory tests, and physician and specialist visits.
Concomitant therapies used as an adjunct were also priced according to RAMQ. Clinical data on frequency of physician and specialist visits and laboratory testing required among those on treatment were included in the model and were also obtained from the five key clinical trialsCitation2–4,Citation16. The proportional use of concomitant medications were multiplied to the cost of each medication and added up over cycles towards the end of the 1-year time horizon. Costs for anti-TNF related support (in the form of physician/specialist visits and laboratory testing) were obtained by a combination of RAMQ, Ontario Case Costing Initiative (OCCI), as well as clinical experts. Infusion costs are excluded as they are borne by the drug manufacturers.
Among the most common adverse events associated with anti-TNF treatment for UC, serious infections (which include hospitalizations) are generally considered the most expensive to treat, as reported by the OCCI. The model incorporated the 1-year risk of serious infections, which is ∼3% for all anti-TNFs, as reported in the clinical trialsCitation2–4,Citation16. The cost of treating a serious infection was assumed to be $8200 per event in accordance with the OCCI. Other adverse events, such as nausea, rash, fever, fatigue, and headache, were ignored due to their negligible impact on costs (minimal or no treatment required). All calculated cumulative costs for patients undergoing 1 year of anti-TNF therapy are presented in Appendix A. All model costs are presented in 2013 Canadian dollars and are incremental (i.e., always relative to another treatment). Conversion rates to $US, £, and € are 0.89, 0.55, and 0.71, respectively.
Cumulative calculations
The treatment algorithms for infliximab 5 mg/kg, adalimumab 40 mg, golimumab 50 mg, and golimumab 100 mg are presented in Appendix B. Since infliximab dosing is weight-based, an additional calculation was required. The average cost of infliximab was calculated assuming 25%, 45%, and 30% of patients falling within the weight categories <70 kg, 71–90 kg, and >91 kg (i.e., requiring use of 2, 3, or 4 vials). These proportions were approximated from the reported baseline weight data from two key RCTs; that is, by using the reported mean weight and reported standard deviation and assuming a normal distributionCitation2,Citation3. The cumulative cost of anti-TNF therapy was calculated by summing the costs across all cycles. For the first 8-week cycle (i.e., induction period), all patients receiving anti-TNF therapy received treatment. For the subsequent 2-week cycles (i.e., maintenance period), the proportion of patients still in response or remission for a given anti-TNF treatment was multiplied to the 2-week cost of that treatment. No discounting was applied due to the short time horizon.
The proportion of patients taking each of the common concomitant medications were derived from the five key RCTs that examined infliximab, adalimumab, and golimumabCitation2–4,Citation16. The proportions of patients using concomitant therapies were very similar across these trials, and, so, we assumed that 35% of patients were concomitantly receiving an immunomodulator (azathioprine or 6-mercaptopurine), 60% of patients were concomitantly receiving aminosalicylates, and 60% of patients were concomitantly receiving oral prednisone.
The 1-year 3% risk of a serious infection was dispersed across all 2-week cycles with a 0.11% probability of having a serious infection at each 2-week cycle, with costs accumulating for all treatments.
Analysis
Probabilistic analysis was carried out using a Monte Carlo simulation in Microsoft Excel 2011. Two input fields were varied over 10,000 iterations of the model: the proportion of patients undergoing background therapy (which consisted of immunomodulators, aminosalicylates, and oral steroids), and the proportion of patients in remission or response at the end of induction. As this analysis is in the context of a Canadian public payer perspective, treatment prices remained fixed. Secondly, credibility intervals from the indirect comparison were used to estimate variance in the induction proportionsCitation6. Induction proportions and background therapy proportions were sampled from a binomial beta distribution. Uncertainty intervals were calculated by taking the 2.5th percentile and 97.5th percentile values from the model simulations.
The outcome of interest was total (cumulative) cost of one full year of remission as well as the total cost of one full year of response. One full year of remission (or response) was defined as sustaining the remission (or response) that was achieved both at the induction period and the end of the 1-year maintenance period (i.e., until the end of week 52). We refer to the cost of one full year of remission as the cost per remission and the cost of one full year of response as the cost per response (or cost per responder).
As anti-TNF therapy is added to conventional therapy, the cost per remission and cost per response for the anti-TNFs should first be compared against the cost per remission and cost per response of background conventional therapy alone. In accordance with conventional health economic analytic practices, the differences in cost per remission and cost per response are calculated as the ratio between the difference in cost (between two interventions) and the difference in proportion of remissions or responsesCitation17. In particular, letting CA and CB denote the total 1-year cost of treatment course including treatment A and B, respectively, and letting RA and RB denote the proportion of patients with one full year of either sustained remission or sustained response on treatment A and B, respectively, the cost per additional remission or cost per additional response (CPR) is calculated as follows:
These calculations resulted in the additional cost for one full year of remission and response comparisons between all anti-TNFs and conventional therapy as well as between the most ‘cost-effective’ anti-TNF and the other anti-TNFs.
Sensitivity analyses
Induction in clinical practice is often 12 weeks rather than the 8 weeks used in the source RCTs. For this reason, a sensitivity analysis assuming a 12-week induction period was performed. That is, the model was run under the premise that all patients were receiving an anti-TNF for the first 12 weeks rather than only for the first 8 weeks. The relative decrease in response and remission commenced for the 7th cycle (week 13–14) and, thus, the derivation of the conditional probability of maintaining remission or response at any 2-week cycle was calculated accordingly. In a sensitivity analysis, we assumed that all patients were kept on a biologic for the first 12 weeks. For this analysis, we used the same 2-week probabilities of maintaining response and remission as those used in the primary analysis; however, the relative decrease in 2-week probabilities of maintaining respond did not commence until the transition from week 12 to 14.
Results
presents the results of the primary probabilistic analyses of cost per additional remission and cost per additional response for each anti-TNF vs conventional therapy and for each anti-TNF compared to one another. graphically displays results for the Markov Chain Monte Carlo simulations of the cost per additional remission and cost per additional response vs conventional therapy. Appendix C presents the results of the sensitivity analyses when the induction period was 12 weeks instead of 8 weeks. Cost-effectiveness acceptability curves are presented in .
Figure 2. Cost per additional 1 year of remission (a) and cost per additional 1 year of response (b) simulation results depicted in relation to the anchor (conventional therapy); 8 week induction.
Figure 3. Cost-effectiveness acceptability curves for remission (a) and response (b); 8-week induction.
Table 3. Results of the primary probabilistic analysis; 8-week induction.
When compared with conventional therapy, golimumab 100 mg produced the lowest cost per additional remission ($935, 95% uncertainty interval [UI] = $634–$1605) and cost per additional response ($701, 95% UI = $573–$890). Golimumab 50 mg produced the second lowest cost per additional remission and cost per additional response, followed by infliximab. Adalimumab had the highest cost per additional remission ($7430) and cost per additional response ($2361).
Since golimumab 100 mg produced the lowest cost per additional remission and cost per additional response when compared to conventional therapy, all other anti-TNFs were then compared against golimumab 100 mg. Golimumab 50 mg costs less and produces less additional remissions and additional responses when compared to golimumab 100 mg. Infliximab produced a relatively high cost per additional remission ($14,659) and cost per additional response ($4753). Adalimumab had the highest cost per additional remission and cost per additional response when compared to golimumab 100 mg. Results for the sensitivity analyses are presented in Appendices C and D. The results of the sensitivity analysis mirrored those of the primary analysis.
Discussion
Our results indicate that the cost of one full year of remission or response is lowest with golimumab 100 mg compared to conventional therapy. Infliximab 5 mg/kg is both more costly and effective than both golimumab 50 mg and golimumab 100 mg, with a substantial increment in the cost per remission and response. Adalimumab 40 mg is less effective than the other anti-TNFs, and the cost per additional remission and cost per additional response were found to be substantially higher for adalimumab 40 mg compared with golimumab 50 mg, golimumab 100 mg, and infliximab 5 mg/kg.
Our analysis has several strengths and limitations. Our efficacy input parameters are based on the best available evidence from a recently conducted indirect treatment comparison meta-analysis that included all key RCTs of infliximab, adalimumab, and golimumabCitation2–4,Citation6,Citation7,Citation16. Our model inputs for concomitant medication use, clinical monitoring, and risk of serious infections further reflect patient populations in key RCTsCitation2–4,Citation16. However, it is possible that these risks are not constant over a 1-year period. For example, it could be argued that the longer a patient has been on an anti-TNF (and other immunomodulators), the more likely that patient is to suffer from an infection due to the suppression of the immune system. Cost of surgery was consciously omitted from this analysis, as the time horizon was limited to 1 year. Several studies have shown that the proportion of patients on biologics that undergo colectomy at the 1-year time point ranges between 0–3%Citation18–22. As such, the overall cost of surgery is negligible in comparison to treatment and concomitant costs. In the context of cost per remission and cost per responder calculations, the average patient cost of infections is small compared to the average cost of anti-TNF treatment as well as conventional therapy, and, as such, the assumption of constant risk of adverse events and discontinuations are unlikely to bias the results. In all RCTs informing the analyses, treatment discontinuations due to adverse events and other reasons occurred. As the efficacy inputs were based on intention-to-treat analyses we did not factor in treatment discontinuations into the Markov model as this would be inconsistent with the ITT assumption. However, it is possible that sensitivities exist around this assumption and that per-protocol based efficacy inputs incorporating discontinuations would have yielded different comparative cost per remission and cost per responder profiles. Lastly, our models and analyses only considered a 1-year time horizon. Generalization to a longer time horizon is likely valid. However generalizations to longer time horizons such as those used in cost-utility analyses (e.g., 10 years or longer) should not be considered valid.
There has been one other study regarding cost per remission of anti-TNFs for the treatment of UCCitation15. However, this study only estimated the cost per remission associated with each anti-TNF rather than the cost per additional remission compared with conventional therapy. By neglecting the comparison with conventional therapy, the high individual cost per remission for anti-TNFs can create an impression that anti-TNFs are unrealistically expensive. However, as observed in our analysis, the low proportion of patients in remission (and response) with conventional therapy offsets this difference. In fact, the additional cost per remission and response appears cost-efficient, especially with regards to golimumab. The previous cost per remission analysis used the same five major trial publications for inputs on efficacy parametersCitation2–4,Citation16. However, the absolute trial proportions were used as efficacy inputs. In contrast, our efficacy inputs were based on estimates from an indirect treatment comparison meta-analysis assuming a control group proportion equal to the average across the included trials. Since the control group proportions differ to some extent between trials, using crude trial data as efficacy input results in confounded efficacy estimates. Our indirect comparison provides results more relevant to clinical practice as it takes control group differences into account by creating a singular reference group, which all treatments would be compared to. This makes our comparison of treatments to the anchor more reliable, while also providing more accurate comparisons between treatments.
In summary, the average cost of achieving and maintaining 1 year of remission or response in patients with moderate-to-severe UC are lowest with golimumab 100 mg, followed by golimumab 50 mg. Infliximab produced the highest number of remissions and responses at the highest cost. Adalimumab produced the highest cost per additional remission and response when compared to conventional therapy.
Transparency
Declaration of funding
This study was funded by Janssen Inc. Canada based on a submitted protocol.
Declaration of financial/other relationships
ED has received an award for doctoral training from the Canadian Institutes of Health Research (CIHR) Drug Safety & Effectiveness Network. KT has received salary support from the CIHR Drug Safety & Effectiveness Network to develop methods and educational materials on network meta-analysis. KT and JJ have previously consulted to Boehringer Ingleheim, Merck, Pfizer, Novartis, Janssen, Roche, Novo Nordisk, UCB, and Gilead.
References
- D'Haens G, Sandborn WJ, Feagan BG, et al. A review of activity indices and efficacy end points for clinical trials of medical therapy in adults with ulcerative colitis. Gastroenterology 2007;132:763-86
- Reinisch W, Sandborn WJ, Hommes DW, et al. Adalimumab for induction of clinical remission in moderately to severely active ulcerative colitis: results of a randomised controlled trial. Gut 2011;60:780-7
- Rutgeerts P, Sandborn WJ, Feagan BG, et al. Infliximab for induction and maintenance therapy for ulcerative colitis. N Engl J Med 2005;353:2462-76
- Sandborn WJ, Feagan BG, Marano C, et al. Subcutaneous golimumab maintains clinical response in patients with moderate-to-severe ulcerative colitis. Gastroenterology 2014;146:96-109.e101
- Danese S, Fiorino G, Peyrin-Biroulet L, et al. Biological agents for moderately to severely active ulcerative colitis: a systematic review and network meta-analysis. Ann Intern Med 2014;160:704-11
- Thorlund K, Druyts E, Eapen S, et al. Comparative efficacy and safety of golimumab, infliximab and adalimumab for the treatment of moderate to severe ulcerative colitis: a Bayesian indirect treatment comparison meta-analysis. Montreal, Canada: International Society for Pharmacoeconomics and Outcomes Research, 2014
- Thorlund K, Druyts E, Mills EJ, et al. Adalimumab versus infliximab for the treatment of moderate to severe ulcerative colitis in adult patients naive to anti-TNF therapy: An indirect treatment comparison meta-analysis. J Crohn’s Colitis 2014;80:571-81
- Stidham RW, Lee TC, Higgins PD, et al. Systematic review with network meta-analysis: the efficacy of anti-tumour necrosis factor-alpha agents for the treatment of ulcerative colitis. Alimen Pharmacol Therapeut 2014;39:660-71
- Régie de l'assurance maladie du Québec (RAMQ). Liste de médicaments (no 44). Quebec: Gouvernement due Quebec, 2013
- Brereton N, Bodger K, Kamm MA, et al. A cost-effectiveness analysis of MMX mesalazine compared with mesalazine in the treatment of mild-to-moderate ulcerative colitis from a UK perspective. J Med Econ 2010;13:148-61
- Chaudhary MA, Fan T. Cost-effectiveness of infliximab for the treatment of acute exacerbations of ulcerative colitis in the Netherlands. Biol Ther 2013;3:45-60
- Park KT, Tsai R, Perez F, et al. Cost-effectiveness of early colectomy with ileal pouch-anal anastamosis versus standard medical therapy in severe ulcerative colitis. Ann surg 2012;256:117-24
- Punekar YS, Hawkins N. Cost-effectiveness of infliximab for the treatment of acute exacerbations of ulcerative colitis. Eur J Health Econ: HEPAC: Health Econ Prevent Care 2010;11:67-76
- Ung V, Thanh NX, Wong K, et al. Real-life treatment paradigms show infliximab is cost-effective for management of ulcerative colitis. Clin Gastroenterol Hepatol Off Clin Pract J Am Gastroenterol Assoc 2014;12:1871-8.e1878
- Lofland JH, Mallow P, Rizzo J. Cost-per-remission analysis of infliximab compared to adalimumab among adults with moderate-to-severe ulcerative colitis. J Med Econ 2013;16:461-7
- Sandborn WJ, van Assche G, Reinisch W, et al. Adalimumab induces and maintains clinical remission in patients with moderate-to-severe ulcerative colitis. Gastroenterology 2012;142:257-65 e251–3
- Briggs A, Claxton K, Sculpher M. Decision modelling for health economic evaluation. New York, NY: Oxford University Press, 2006
- Aratari A, Papi C, Clemente V, et al. Colectomy rate in acute severe ulcerative colitis in the infliximab era. Digest Liver Dis Off J Italian Soc Gastroenterol Italian Assoc Study Liver 2008;40:821-26
- Gies N, Kroeker KI, Wong K, et al. Treatment of ulcerative colitis with adalimumab or infliximab: long-term follow-up of a single-centre cohort. Alimentary Pharmacol Therapeut 2010;32:522-8
- Kaplan GG, Seow CH, Ghosh S, et al. Decreasing colectomy rates for ulcerative colitis: a population-based time trend study. Am J Gastroenterol 2012;107:1879-87
- Kim IK, Park KJ, Kang GH, et al. Risk factors for complications after total colectomy in ulcerative colitis. Turk J Gastroenterol Off J Turk Soc Gastroenterol 2012;23:515-22
- Targownik LE, Singh H, Nugent Z, et al. The epidemiology of colectomy in ulcerative colitis: results from a population-based cohort. Am j gastroenterol 2012;107:1228-35
- Régie de l’assurance maladie du Québec (RAMQ). Manuel des médecins spécialistes (no 150). 2012
Appendix A: Cumulative costs used as input parameters to calculate cost per additional remission and cost per additional response over 1 year
Appendix B: Schematics of the approved treatment algorithms for infliximab, adalimumab, and golimumab
Appendix C: Results of the probabilistic analysis (12-week induction)
Appendix D: Cost per additional 1 year of remission (a) and cost per additional 1-year of response (b) simulation results depicted in relation to the anchor (conventional therapy); 12-week induction
