IOLOGIS IN MODERATE TO SEVERE ROHN’S … American journal of gastroenterology 104, ... Ustekinumab...
1
Introduction Objective • Crohn’s disease (CD) is a chronic, progressive, non-infectious inflammatory disease affecting different parts of the gastrointestinal tract and is characterised by periods of symptoms remission and relapse [1]. • The use of biologics (infliximab, adalimumab, vedolizumab) in CD has been reserved for patients who have failed, have contraindications, or are intolerant to conventional therapy. Phase III trials indicate these therapies are significantly more effective than standard of care in inducing and maintaining response and remission [2]. • Ustekinumab is fully human IgG1κ monoclonal antibody directed against IL12/23 for which phase III results in Crohn's disease recently came available [3,4,5]. • Trials in CD are similarly designed placebo-controlled induction and maintenance studies in which patients responding to induction with the specific biologic are randomised to active treatment or withdrawn to placebo (randomised withdrawal) [6]. • A lack of comparability between placebo arms in different biologic trials has previously been identified in the literature [7]. The lack of comparability between maintenance “placebo” arms suggests that traditional evidence synthesis methods may not be appropriate to capture the true relative treatment effects of ustekinumab. Conclusions A CKNOWLEDGEMENT & REFERENCES Pacou M 1 , Mesana, L 2 , Gauthier A 2 , Naessens D 3 , Sloan S 4 , Danese S 5 , Bonovas S 5 , Abrams K 6 1 Amaris, Paris, France, 2 Amaris, London, United Kingdom, 3 Janssen Pharmaceutica N.V., Beerse, Belgium, 4 Janssen Global Services, LLC, 5 Humanitas Clinical and Research Center, Rozzano, Milan, Italy, 6 Department of Health Sciences, University of Leicester, Leicester, UK INDIRECT TREATMENT COMPARISON OF USTEKINUMAB VERSUS OTHER BIOLOGICS IN MODERATE TO SEVERE CROHN’S DISEASE: A 1-YEAR TREATMENT SEQUENCE ANALYSIS • Indirect evidence is needed to better inform decision-makers on the clinical efficacy of ustekinumab in CD. In order to overcome important methodological challenges linked to the feasibility of a standard indirect treatment comparison (ITC) due to conceptual heterogeneity and low placebo arm transitivity potential, an ITC of one-year efficacy data was conducted, taking into account the entire treatment pathway of response to induction and maintenance of therapy. • As a standard ITC was unfeasible, this novel approach addresses issues of similarity and transitivity and seems most appropriate to compare one year efficacy of biologics in CD. • The entire treatment sequence of induction followed by maintenance suggests a higher likelihood of response or remission at year 1 with ustekinumab compared to vedolizumab or adalimumab. FC: Conventional therapy failure, FA: Anti-TNF failure, in orange: imputed placebo rates Odds ratios (for ustekinumab to perform better) and their associated credible intervals are presented on forest plots. The corresponding Bayesian probabilities are also given. Results are displayed in figures 3 and 4, respectively. Methods • A systematic literature review identified randomized controlled trials for induction and maintenance of clinical efficacy evaluating ustekinumab, infliximab, adalimumab, or vedolizumab. • A treatment sequence analysis was conducted: the relative probability of achieving response (Crohn’s Disease Activity Index (CDAI) reduction of 70 points) at the end of induction was multiplied by the conditional probability of maintaining response or remission after one year (figure 1). Discussion [1] Lichtenstein, G. R., Hanauer, S. B. & Sandborn, W. J. Management of Crohn's disease in adults. The American journal of gastroenterology 104, 465-483 (2009). [2] Hazlewood, G. S. et al. Comparative effectiveness of immunosuppressants and biologics for inducing and maintaining remission in Crohn's disease: a network meta-analysis. Gastroenterology 148, 344-354 e345; quiz e314-345, doi:10.1053/j.gastro.2014.10.011 (2015). [3] Sandborn, W. et al. O-001 A Multicenter, Double-Blind, Placebo-Controlled Phase3 Study of Ustekinumab, a Human IL-12/23P40 mAB, in Moderate-Service Crohn's Disease Refractory to Anti-TFNalpha: UNITI-1. Inflamm Bowel Dis 22 Suppl 1, S1, doi:10.1097/MIB.0000000000000720 (2016). [4]Janssen. A Phase 3, Randomized, Double-blind, Placebo-controlled, Parallel-group, Multicenter Study to Evaluate the Safety and Efficacy of Ustekinumab Induction Therapy in Subjects with Moderately to Severely Active Crohn’s Disease (UNITI 2). Research & Development (2015). [5] Janssen. A Phase 3, Randomized, Double-blind, Placebo-controlled, Parallel-group, Multicenter study to Evaluate the Safety and Efficacy of Ustekinumab Maintenance Therapy in Subjects with Moderately to Severely Active Crohn's Disease (44-Week Clinical Study Report). Research & Development (2015). Results (continued) Figure 1. Treatment sequence pathway for generation of inputs • The treatment sequence analysis is based on previous publications addressing the unfeasibility of standard ITCs due to a lack of true common comparator during maintenance [8,9]. Randomization of each trial is preserved up to the end of induction but not after entry into maintenance. • Placebo-to-placebo rates used were only available for ustekinumab in IM- UNITI and were used across maintenance phases for different comparators. These rates were adjusted for remission rates at the end of induction in order to account for variations in response criteria across the included trials. • Comparisons with infliximab are limited due to the fact that the trial by Targan et al. was fairly old and the induction phase in ACCENT I fairly small, with a high number of missing data in its placebo arm and inverse dose- specific response rates observed in the active treatment arms. • Comparisons to adalimumab patients who failed anti-TNF therapy are limited due to differences in the inclusion of prior treatment failures across trials. • Sensitivity analyses conducted to account for the additional uncertainty associated with the imputed placebo-to-placebo rates for the vedolizumab, adalimumab and infliximab trials revealed no changes in conclusions. The two other sensitivity analyses conducted (using data for the truly naïve subgroup of IM-UNITI patients, conducting all analyses within a frequentist framework) did not impact the conclusions of the base case analysis either. Results Figure 2. Global network of evidence • Placebo-to-placebo maintenance response and remission rates had to be imputed using data from the ustekinumab maintenance trial IM-UNITI across all biologics’ maintenance phases as they were only available for IM-UNITI (Placebo-to-placebo: patients received placebo in both induction & maintenance phases). • Separate analyses were conducted for CDAI-100 clinical response reduction of 100 points and clinical remission (CDAI under 150) in patients having either failed conventional or anti-TNF therapy. • Analyses were performed in WinBUGS, using a Bayesian framework. • Outputs were Bayesian posterior distribution probabilities for ustekinumab to perform better than each comparator over the entire treatment pathway. • The number needed to treat (NNT) was calculated, applying the ORs obtained through the treatment sequence analysis on the IM-UNITI placebo rates to generate absolute treatment response and remission rates for each comparator. Failed conventional Failed anti-TNF UST Q8W UST Q12W ADA eow VDZ Q8W IFX 5 UST Q8W UST Q12W ADA eow VDZ Q8W IFX 5 CDAI-100 4.7 4.9 10.6 9.9 NA 10.1 11.2 14.9 95.8 NA CDAI<150 4.5 5.4 7.4 7.2 3.2 14.7 16.6 21.6 45.5 NA • Significant levels of heterogeneity were detected by a chi square test, suggesting that placebo maintenance arms are not appropriate common comparators across biologic trials and challenging the feasibility of a traditional network meta-analysis (NMA) in the maintenance phase. • In total, 12 trials were included in the treatment sequence analysis. In patients having failed conventional therapy, six trials reported induction results, and four reported maintenance results. In patients having failed anti-TNF therapy, six trials reported induction results and three reported maintenance data. Figure legend: In red: conventional therapy failure trials In blue: anti-TNF failure trials (I):Induction (M):Maintenance ADA: Adalimumab, UST: Ustekinumab, VDZ: Vedolizumab, eow: every other week, qXw: every X week Table 1. Inputs for the treatment sequence data Treatment Induction response % Maintenance Sequence % CDAI-100 % CDAI<150 % CDAI-100 % CDAI<150 FC Placebo-Placebo 39% 60% 47% 23% 18% UST 6-UST 90 q12w [4,5] 65% 67% 57% 43% 37% UST 6-UST 90 q8w [4,5] 65% 68% 63% 44% 41% Placebo-Placebo 38% 58% 37% 22% 14% VDZ 300- VDZ 300 q8w [12] 53% 60% 50% 32% 26% Placebo-Placebo 35% 58% 40% 20% 14% ADA 160/80- ADA 40 eow [13-17] 62% 46% 41% 29% 26% Placebo-Placebo 17% N/A 36% N/A 6% IFX 5 – IFX 5 [10,11] 81% N/A 28% N/A 23% FA Placebo-Placebo 30% 39% 35% 12% 11% UST 6-UST 90 q12w [3,5] 45% 47% 39% 21% 18% UST 6-UST 90 q8w [3,5] 45% 48% 41% 22% 18% Placebo-Placebo 31% 38% 34% 12% 10% VDZ 300- VDZ 300 q8w [12] 44% 29% 28% 13% 12% Placebo-Placebo 34% 36% 30% 12% 10% ADA 160/80- ADA 40 eow [14-16] 53% 36% 31% 19% 16% Figure 4. Achievement of remission (CDAI<150) after one year [6] D'Haens, G. et al. Challenges to the design, execution, and analysis of randomized controlled trials for inflammatory bowel disease. Gastroenterology 143, 1461-1469, doi:10.1053/j.gastro.2012.09.031 (2012). [7] Su C, L. G., Krok K, Brensinger CM, and Lewis JD. A Meta-Analysis of the Placebo Rates of Remission and Response in Clinical Trials of Active Crohn’s Disease. GASTROENTEROLOGY 126, 1257-1269 (2004). [8] Thorlund, K., Druyts, E., Toor, K. & Mills, E. J. Comparative efficacy of golimumab, infliximab, and adalimumab for moderately to severely active ulcerative colitis: a network meta-analysis accounting for differences in trial designs. Expert Rev Gastroenterol Hepatol 9, 693-700, doi:10.1586/17474124.2015.1024657 (2015). [9] Abrams, K. R., Cooper, N. J., Williams, K. S., Gillies, C. L., Lambert, P. C., Turner, D., ... & McGrother, C. W. Evaluating complex interventions using multi-phase nested trial designs-Management of urinary incontinence across the primary/secondary care interface. Controlled Clinical Trials 24, 57S-58S (2003). [10] Targan et al. (1997) NEJM 15:1029-35 [11] Hanauer et al. (2002) ACCENT I. Lancet 359(9317):1541-9 [12] Sandborn et al (2013) GEMINI II & III. NEJM 369(8):711-21 [13] Hanauer et al (2006) CLASSIC I. gastroenterology 130(2):323-33 [14] Watanabe et al (2012) J. Crohns Colitis 6(2):160-73 [15] Colombel et al (2007) Gastroenterology 132(1):52-65 [16] Sandborn et al (2009) Ann Intern Med 146:829-838 This project was funded by Janssen. Table 2. Results from the NNT analysis Figure 3. Maintenance of response (CDAI-100) after one year
Transcript of IOLOGIS IN MODERATE TO SEVERE ROHN’S … American journal of gastroenterology 104, ... Ustekinumab...
Introduction
Objective
• Crohn’s disease (CD) is a chronic, progressive, non-infectious inflammatorydisease affecting different parts of the gastrointestinal tract and ischaracterised by periods of symptoms remission and relapse [1].
• The use of biologics (infliximab, adalimumab, vedolizumab) in CD has beenreserved for patients who have failed, have contraindications, or areintolerant to conventional therapy. Phase III trials indicate these therapiesare significantly more effective than standard of care in inducing andmaintaining response and remission [2].
• Ustekinumab is fully human IgG1κ monoclonal antibody directed againstIL12/23 for which phase III results in Crohn's disease recently came available[3,4,5].
• Trials in CD are similarly designed placebo-controlled induction andmaintenance studies in which patients responding to induction with thespecific biologic are randomised to active treatment or withdrawn toplacebo (randomised withdrawal) [6].
• A lack of comparability between placebo arms in different biologic trials haspreviously been identified in the literature [7]. The lack of comparabilitybetween maintenance “placebo” arms suggests that traditional evidencesynthesis methods may not be appropriate to capture the true relativetreatment effects of ustekinumab.
Conclusions
ACKNOWLEDGEMENT & REFERENCES
Pacou M1, Mesana, L2, Gauthier A2, Naessens D3, Sloan S4, Danese S5, Bonovas S5, Abrams K6
1Amaris, Paris, France, 2Amaris, London, United Kingdom,3Janssen Pharmaceutica N.V., Beerse, Belgium, 4Janssen Global Services, LLC, 5Humanitas Clinical and Research Center, Rozzano, Milan, Italy, 6Department of Health Sciences, University of Leicester, Leicester, UK
INDIRECT TREATMENT COMPARISON OF USTEKINUMAB VERSUS OTHER BIOLOGICS IN MODERATE TO SEVERE CROHN’S DISEASE: A 1-YEAR TREATMENT SEQUENCE ANALYSIS
• Indirect evidence is needed to better inform decision-makers on the clinicalefficacy of ustekinumab in CD. In order to overcome importantmethodological challenges linked to the feasibility of a standard indirecttreatment comparison (ITC) due to conceptual heterogeneity and lowplacebo arm transitivity potential, an ITC of one-year efficacy data wasconducted, taking into account the entire treatment pathway of responseto induction and maintenance of therapy.
• As a standard ITC was unfeasible, this novel approach addresses issues ofsimilarity and transitivity and seems most appropriate to compare one yearefficacy of biologics in CD.
• The entire treatment sequence of induction followed by maintenancesuggests a higher likelihood of response or remission at year 1 withustekinumab compared to vedolizumab or adalimumab.
FC: Conventional therapy failure, FA: Anti-TNF failure, in orange: imputed placebo rates
Odds ratios (for ustekinumab to perform better) and their associated credibleintervals are presented on forest plots. The corresponding Bayesianprobabilities are also given. Results are displayed in figures 3 and 4,respectively.
Methods
• A systematic literature review identified randomized controlled trials forinduction and maintenance of clinical efficacy evaluating ustekinumab,infliximab, adalimumab, or vedolizumab.
• A treatment sequence analysis was conducted: the relative probability ofachieving response (Crohn’s Disease Activity Index (CDAI) reduction of 70points) at the end of induction was multiplied by the conditional probabilityof maintaining response or remission after one year (figure 1).
Discussion
[1] Lichtenstein, G. R., Hanauer, S. B. & Sandborn, W. J. Management of Crohn's disease in adults. The American journal of gastroenterology 104, 465-483 (2009).[2] Hazlewood, G. S. et al. Comparative effectiveness of immunosuppressants and biologics for inducing and maintaining remission in Crohn's disease: a network meta-analysis. Gastroenterology 148, 344-354 e345; quiz e314-345, doi:10.1053/j.gastro.2014.10.011 (2015).[3] Sandborn, W. et al. O-001 A Multicenter, Double-Blind, Placebo-Controlled Phase3 Study of Ustekinumab, a Human IL-12/23P40 mAB, in Moderate-Service Crohn's Disease Refractory to Anti-TFNalpha: UNITI-1. Inflamm Bowel Dis 22 Suppl 1, S1, doi:10.1097/MIB.0000000000000720 (2016).[4]Janssen. A Phase 3, Randomized, Double-blind, Placebo-controlled, Parallel-group, Multicenter Study to Evaluate the Safety and Efficacy of Ustekinumab Induction Therapy in Subjects with Moderately to Severely Active Crohn’s Disease (UNITI 2). Research & Development (2015).[5] Janssen. A Phase 3, Randomized, Double-blind, Placebo-controlled, Parallel-group, Multicenter study to Evaluate the Safety and Efficacy of Ustekinumab Maintenance Therapy in Subjects with Moderately to Severely Active Crohn's Disease (44-Week Clinical Study Report). Research & Development (2015).
Results (continued)
Figure 1. Treatment sequence pathway for generation of inputs
• The treatment sequence analysis is based on previous publications addressingthe unfeasibility of standard ITCs due to a lack of true common comparatorduring maintenance [8,9]. Randomization of each trial is preserved up to theend of induction but not after entry into maintenance.
• Placebo-to-placebo rates used were only available for ustekinumab in IM-UNITI and were used across maintenance phases for different comparators.These rates were adjusted for remission rates at the end of induction in orderto account for variations in response criteria across the included trials.
• Comparisons with infliximab are limited due to the fact that the trial byTargan et al. was fairly old and the induction phase in ACCENT I fairly small,with a high number of missing data in its placebo arm and inverse dose-specific response rates observed in the active treatment arms.
• Comparisons to adalimumab patients who failed anti-TNF therapy are limiteddue to differences in the inclusion of prior treatment failures across trials.
• Sensitivity analyses conducted to account for the additional uncertaintyassociated with the imputed placebo-to-placebo rates for the vedolizumab,adalimumab and infliximab trials revealed no changes in conclusions. The twoother sensitivity analyses conducted (using data for the truly naïve subgroupof IM-UNITI patients, conducting all analyses within a frequentist framework)did not impact the conclusions of the base case analysis either.
Results
Figure 2. Global network of evidence
• Placebo-to-placebo maintenance response and remission rates had to beimputed using data from the ustekinumab maintenance trial IM-UNITI acrossall biologics’ maintenance phases as they were only available for IM-UNITI(Placebo-to-placebo: patients received placebo in both induction &maintenance phases).
• Separate analyses were conducted for CDAI-100 clinical response reductionof 100 points and clinical remission (CDAI under 150) in patients havingeither failed conventional or anti-TNF therapy.
• Analyses were performed in WinBUGS, using a Bayesian framework.
• Outputs were Bayesian posterior distribution probabilities for ustekinumabto perform better than each comparator over the entire treatment pathway.
• The number needed to treat (NNT) was calculated, applying the ORsobtained through the treatment sequence analysis on the IM-UNITI placeborates to generate absolute treatment response and remission rates for eachcomparator.
Failed conventional Failed anti-TNF
UST Q8W
UST Q12W
ADA eow
VDZ Q8W
IFX 5UST
Q8WUST
Q12WADA eow
VDZ Q8W
IFX 5
CDAI-100 4.7 4.9 10.6 9.9 NA 10.1 11.2 14.9 95.8 NA
CDAI<150 4.5 5.4 7.4 7.2 3.2 14.7 16.6 21.6 45.5 NA
• Significant levels of heterogeneity were detected by a chi square test,suggesting that placebo maintenance arms are not appropriate commoncomparators across biologic trials and challenging the feasibility of atraditional network meta-analysis (NMA) in the maintenance phase.
• In total, 12 trials were included in the treatment sequence analysis. Inpatients having failed conventional therapy, six trials reported inductionresults, and four reported maintenance results. In patients having failedanti-TNF therapy, six trials reported induction results and three reportedmaintenance data.
Figure legend:In red: conventional therapy failure trialsIn blue: anti-TNF failure trials(I):Induction(M):Maintenance
ADA: Adalimumab, UST: Ustekinumab, VDZ: Vedolizumab, eow: every other week, qXw:every X week
Table 1. Inputs for the treatment sequence data
TreatmentInduction
response %
Maintenance Sequence
% CDAI-100 % CDAI<150 % CDAI-100 % CDAI<150
FC
Placebo-Placebo 39% 60% 47% 23% 18%UST 6-UST 90 q12w [4,5] 65% 67% 57% 43% 37%UST 6-UST 90 q8w [4,5] 65% 68% 63% 44% 41%Placebo-Placebo 38% 58% 37% 22% 14%VDZ 300- VDZ 300 q8w [12] 53% 60% 50% 32% 26%Placebo-Placebo 35% 58% 40% 20% 14%ADA 160/80- ADA 40 eow [13-17] 62% 46% 41% 29% 26%Placebo-Placebo 17% N/A 36% N/A 6%IFX 5 – IFX 5 [10,11] 81% N/A 28% N/A 23%
FA
Placebo-Placebo 30% 39% 35% 12% 11%UST 6-UST 90 q12w [3,5] 45% 47% 39% 21% 18%UST 6-UST 90 q8w [3,5] 45% 48% 41% 22% 18%Placebo-Placebo 31% 38% 34% 12% 10%VDZ 300- VDZ 300 q8w [12] 44% 29% 28% 13% 12%Placebo-Placebo 34% 36% 30% 12% 10%ADA 160/80- ADA 40 eow [14-16] 53% 36% 31% 19% 16%
Figure 4. Achievement of remission (CDAI<150) after one year
[6] D'Haens, G. et al. Challenges to the design, execution, and analysis of randomized controlled trials for inflammatory bowel disease. Gastroenterology 143, 1461-1469, doi:10.1053/j.gastro.2012.09.031 (2012).[7] Su C, L. G., Krok K, Brensinger CM, and Lewis JD. A Meta-Analysis of the Placebo Rates of Remission and Response in Clinical Trials of Active Crohn’s Disease. GASTROENTEROLOGY 126, 1257-1269 (2004).[8] Thorlund, K., Druyts, E., Toor, K. & Mills, E. J. Comparative efficacy of golimumab, infliximab, and adalimumab for moderately to severely active ulcerative colitis: a network meta-analysis accounting for differences in trial designs. Expert Rev GastroenterolHepatol 9, 693-700, doi:10.1586/17474124.2015.1024657 (2015).[9] Abrams, K. R., Cooper, N. J., Williams, K. S., Gillies, C. L., Lambert, P. C., Turner, D., ... & McGrother, C. W. Evaluating complex interventions using multi-phase nested trial designs-Management of urinary incontinence across the primary/secondary care interface. Controlled Clinical Trials 24, 57S-58S (2003).[10] Targan et al. (1997) NEJM 15:1029-35[11] Hanauer et al. (2002) ACCENT I. Lancet 359(9317):1541-9[12] Sandborn et al (2013) GEMINI II & III. NEJM 369(8):711-21[13] Hanauer et al (2006) CLASSIC I. gastroenterology 130(2):323-33[14] Watanabe et al (2012) J. Crohns Colitis 6(2):160-73[15] Colombel et al (2007) Gastroenterology 132(1):52-65[16] Sandborn et al (2009) Ann Intern Med 146:829-838 This project was funded by Janssen.
Table 2. Results from the NNT analysis
Figure 3. Maintenance of response (CDAI-100) after one year
