Downloaded from gut.bmj.com on June 30, 2014 - Published by group.bmj.com
Gut Online First, published on June 26, 2014 as 10.1136/gutjnl-2014-307126 Leading article
Systematic review: monotherapy with antitumour necrosis factor α agents versus combination therapy with an immunosuppressive for IBD Parambir S Dulai,1 Corey A Siegel,1 Jean-Frederic Colombel,2 William J Sandborn,3 Laurent Peyrin-Biroulet4 INTRODUCTION The treatment of IBD has evolved significantly over the past 15 years. Historically, steroids and immunosuppressive drugs, such as methotrexate (MTX), azathioprine (AZA) and 6-mercaptopurine (6-MP) had been the mainstay of therapy. When infliximab (IFX) was approved for the treatment of Crohn’s disease, it began the era of trying to understand how best to optimise our available treatment options. Initially, antitumour necrosis factor (anti-TNF) agents were used after immunosuppressives failed. Then, evidence showed that using AZA together with IFX early in the disease course was more effective than the typical sequential treatment algorithm.1 With the addition of newer anti-TNF drugs and development of different treatment regimens,2–5 there are still significant questions related to the most effective therapeutic strategy. Perhaps the most controversial of these questions is when to use anti-TNF monotherapy versus combination therapy with 6-MP, AZA or MTX. When considering patients with rheumatoid arthritis, where anti-TNF therapy also has a demonstrated efficacy, it is clear from randomised controlled trials (RCT) that the use of combination therapy improves clinical outcomes.6 Although it is logical that a similar benefit should exist in IBD, the evidence base upon which to make decisions is less robust, leading to patient and provider concerns regarding the balance between efficacy and toxicity. Providers have two sources of data to rely upon when addressing this concern with IBD patients, RCTs 1 Inflammatory Bowel Disease Center, Dartmouth Hitchcock Medical Center, Lebanon, New Hampshire, USA; 2Icahn School of Medicine at Mount Sinai, New York, New York, USA; 3University of California San Diego, La Jolla, California, USA; 4Inserm U954, Department of Hepato-Gastroenterology, Université de Lorraine, Nancy, France Correspondence to Dr Corey A Siegel, Inflammatory Bowel Disease Center, Dartmouth Hitchcock Medical Center, One Medical Center Drive, Lebanon, NH 03756, USA; [email protected]
and observational data. RCTs are more rigorous, generally have better patient follow-up, and more often include prospective measurement of study end-points using well-accepted and often validated outcome measures. For this reason, RCTs result in a relatively robust assessment of key outcome measures such as clinical response, remission and mucosal healing. Although the data from RCTs may not directly translate to the ‘real world’,7 whatever generalisability is lost due to the application of strict inclusion and exclusion criteria is offset by the minimisation of bias achieved through randomisation. Subgroup analyses of RCTs, on the other hand, carry a significant potential for bias, as patients were not randomised to the treatment assignment being evaluated, these studies were not designed to address the question being asked, and these questions were often not predefined. Therefore, the primary and secondary endpoints from RCTs represent the best available evidence in regards to treatment efficacy. The strict selection of patients, and exclusion of high-risk populations (elderly, comorbidities),8–10 makes the assessment of treatment safety from these data difficult. Providers must, therefore, turn to ‘real life’ observational data to better understand the risks of therapy and identify populations at relatively greater risk. These data, however, are limited by unmeasured confounders, variability in outcome measurements, and the absence of a control group. Ultimately, providers must weigh the incremental benefits of combination therapy in RCTs against the concerns for serious adverse events noted in observational data.11–13 The purpose of this review is to discuss the efficacy (clinical remission, mucosal healing and pharmacokinetics) and the risks (focusing on serious infection and lymphoma) of anti-TNF monotherapy versus combination therapy with an immunosuppressive in patients with IBD. We will examine the differences in outcomes between RCTs evaluating monotherapy versus combination therapy,
subgroup analyses of RCTs of anti-TNF therapy versus placebo stratified for concomitant immunosuppressive therapy, and observational data when considering Crohn’s disease and UC, and among the various anti-TNF agents. We will also explore differences between using AZA or 6-MP versus MTX as the second agent, and highlight key points to consider when approaching this question in high-risk populations. The goal of this review is to help providers better understand the currently available literature, thereby enhancing the shared decision-making process with patients as they decide together on the most appropriate treatment regimen.
METHODS To construct this review article, we performed a systematic review of the literature to identify relevant studies comparing anti-TNF monotherapy to combination therapy with an immunosuppressive for IBD (see online supplementary material methods). Results are presented below as summaries stratified by anti-TNF agent: IFX, adalimumab (ADA), certolizumab (CTZ) and golimumab; and study type: RCT evaluating monotherapy versus combination therapy, subgroup analyses of RCTs of anti-TNF therapy versus placebo stratified for concomitant immunosuppressive therapy, and observational data.
CLINICAL EFFICACY: RESPONSE, REMISSION, MUCOSAL HEALING AND PHARMACOKINETICS RCTs evaluating anti-TNF monotherapy versus combination therapy Comparative effectiveness RCT data demonstrates that the combination of IFX and AZA is more efficacious than IFX monotherapy in regards to clinical response and mucosal healing in Crohn’s disease (SONIC14), but the use of MTX as the concomitant immunosuppressive has not been demonstrated to be superior to IFX monotherapy (COMMIT15) (table 1). When interpreting this disparity between immunosuppressive agents, providers must recognise several key differences between the SONIC and COMMIT trials. First, the investigators of COMMIT did not use a minimum Crohn’s disease activity index (CDAI) score for inclusion which resulted in the recruitment of a patient population with a much milder baseline disease activity and, potentially, a more benign natural disease course. Second, the primary outcome of interest in COMMIT was time to treatment failure which was defined as failure to achieve corticosteroid-free remission at week 14,
Dulai PS, et al. Gut Month 2014 Vol(or 0 Notheir 0 1 Copyright Article author employer) 2014. Produced by BMJ Publishing Group Ltd (& BSG) under licence.
Downloaded from gut.bmj.com on June 30, 2014 - Published by group.bmj.com
Leading article Table 1 Comparative effectiveness randomised controlled trials comparing anti-TNF monotherapy to combination therapy with an immunosuppressive Crohn’s disease SONIC
Patient characteristics Participants, n Treatment regimen Male, % Mean age, years Disease characteristics Mean duration, years Extensive*, % Prior surgery, % Prior IS therapy, % Baseline steroids, % Clinical disease activity score†, mean Clinically active disease at baseline‡, % Endoscopic lesions at baseline, % CRP, mean Outcomes Primary CFREM¶, % Week 50 CFREM, % Mucosal healing, % Antidrug antibody, % IFX concentration, mean mg/mL Serious infection, %
Ulcerative colitis
14
COMMIT
15
UC-SUCCESS16
IFX mono
IFX combo
IFX mono
IFX combo
IFX mono
IFX combo
169 IFX+PBO 50 35
169 IFX+AZA 52 34
63 IFX+PBO 59 39
63 IFX+MTX 54 40
78 IFX+PBO 54 39
80 IFX+AZA 60 38
2.2 39 28 0 31 285 100 57 1.1
2.2 44 26 0 28 290 100 66 1.0
9.6 56 46 25 100 208 73 n/a 6.0
10.9 60 57 24 100 208 68 n/a 3.0§
6.3 19 n/a 10 40 6.0 100 99 n/a
5.2 34 n/a 10 48 6.3 100 100 n/a
44 35 30 14.6 1.6 4.9
57§ 46§ 44‡ 0.9 3.5§ 3.9
78 57 n/a 20.4 3.8 n/a
76 56 n/a 4.0 6.4 n/a
22 n/a 55 19 n/a 0.01
40§ n/a 63** 3 n/a 0
‡ p=0.06. *Extensive disease definition for SONIC and COMMIT is ileocolitis at baseline and for UC-SUCCESS is Mayo endoscopic subscore of three at baseline. †Clinical disease activity score for SONIC and COMMIT is the Crohn’s Disease Activity Index (CDAI) and for UC-SUCCESS is the partial Mayo score. ‡Clinical remission at baseline was defined as a CDAI score of
Gut Online First, published on June 26, 2014 as 10.1136/gutjnl-2014-307126 Leading article
Systematic review: monotherapy with antitumour necrosis factor α agents versus combination therapy with an immunosuppressive for IBD Parambir S Dulai,1 Corey A Siegel,1 Jean-Frederic Colombel,2 William J Sandborn,3 Laurent Peyrin-Biroulet4 INTRODUCTION The treatment of IBD has evolved significantly over the past 15 years. Historically, steroids and immunosuppressive drugs, such as methotrexate (MTX), azathioprine (AZA) and 6-mercaptopurine (6-MP) had been the mainstay of therapy. When infliximab (IFX) was approved for the treatment of Crohn’s disease, it began the era of trying to understand how best to optimise our available treatment options. Initially, antitumour necrosis factor (anti-TNF) agents were used after immunosuppressives failed. Then, evidence showed that using AZA together with IFX early in the disease course was more effective than the typical sequential treatment algorithm.1 With the addition of newer anti-TNF drugs and development of different treatment regimens,2–5 there are still significant questions related to the most effective therapeutic strategy. Perhaps the most controversial of these questions is when to use anti-TNF monotherapy versus combination therapy with 6-MP, AZA or MTX. When considering patients with rheumatoid arthritis, where anti-TNF therapy also has a demonstrated efficacy, it is clear from randomised controlled trials (RCT) that the use of combination therapy improves clinical outcomes.6 Although it is logical that a similar benefit should exist in IBD, the evidence base upon which to make decisions is less robust, leading to patient and provider concerns regarding the balance between efficacy and toxicity. Providers have two sources of data to rely upon when addressing this concern with IBD patients, RCTs 1 Inflammatory Bowel Disease Center, Dartmouth Hitchcock Medical Center, Lebanon, New Hampshire, USA; 2Icahn School of Medicine at Mount Sinai, New York, New York, USA; 3University of California San Diego, La Jolla, California, USA; 4Inserm U954, Department of Hepato-Gastroenterology, Université de Lorraine, Nancy, France Correspondence to Dr Corey A Siegel, Inflammatory Bowel Disease Center, Dartmouth Hitchcock Medical Center, One Medical Center Drive, Lebanon, NH 03756, USA; [email protected]
and observational data. RCTs are more rigorous, generally have better patient follow-up, and more often include prospective measurement of study end-points using well-accepted and often validated outcome measures. For this reason, RCTs result in a relatively robust assessment of key outcome measures such as clinical response, remission and mucosal healing. Although the data from RCTs may not directly translate to the ‘real world’,7 whatever generalisability is lost due to the application of strict inclusion and exclusion criteria is offset by the minimisation of bias achieved through randomisation. Subgroup analyses of RCTs, on the other hand, carry a significant potential for bias, as patients were not randomised to the treatment assignment being evaluated, these studies were not designed to address the question being asked, and these questions were often not predefined. Therefore, the primary and secondary endpoints from RCTs represent the best available evidence in regards to treatment efficacy. The strict selection of patients, and exclusion of high-risk populations (elderly, comorbidities),8–10 makes the assessment of treatment safety from these data difficult. Providers must, therefore, turn to ‘real life’ observational data to better understand the risks of therapy and identify populations at relatively greater risk. These data, however, are limited by unmeasured confounders, variability in outcome measurements, and the absence of a control group. Ultimately, providers must weigh the incremental benefits of combination therapy in RCTs against the concerns for serious adverse events noted in observational data.11–13 The purpose of this review is to discuss the efficacy (clinical remission, mucosal healing and pharmacokinetics) and the risks (focusing on serious infection and lymphoma) of anti-TNF monotherapy versus combination therapy with an immunosuppressive in patients with IBD. We will examine the differences in outcomes between RCTs evaluating monotherapy versus combination therapy,
subgroup analyses of RCTs of anti-TNF therapy versus placebo stratified for concomitant immunosuppressive therapy, and observational data when considering Crohn’s disease and UC, and among the various anti-TNF agents. We will also explore differences between using AZA or 6-MP versus MTX as the second agent, and highlight key points to consider when approaching this question in high-risk populations. The goal of this review is to help providers better understand the currently available literature, thereby enhancing the shared decision-making process with patients as they decide together on the most appropriate treatment regimen.
METHODS To construct this review article, we performed a systematic review of the literature to identify relevant studies comparing anti-TNF monotherapy to combination therapy with an immunosuppressive for IBD (see online supplementary material methods). Results are presented below as summaries stratified by anti-TNF agent: IFX, adalimumab (ADA), certolizumab (CTZ) and golimumab; and study type: RCT evaluating monotherapy versus combination therapy, subgroup analyses of RCTs of anti-TNF therapy versus placebo stratified for concomitant immunosuppressive therapy, and observational data.
CLINICAL EFFICACY: RESPONSE, REMISSION, MUCOSAL HEALING AND PHARMACOKINETICS RCTs evaluating anti-TNF monotherapy versus combination therapy Comparative effectiveness RCT data demonstrates that the combination of IFX and AZA is more efficacious than IFX monotherapy in regards to clinical response and mucosal healing in Crohn’s disease (SONIC14), but the use of MTX as the concomitant immunosuppressive has not been demonstrated to be superior to IFX monotherapy (COMMIT15) (table 1). When interpreting this disparity between immunosuppressive agents, providers must recognise several key differences between the SONIC and COMMIT trials. First, the investigators of COMMIT did not use a minimum Crohn’s disease activity index (CDAI) score for inclusion which resulted in the recruitment of a patient population with a much milder baseline disease activity and, potentially, a more benign natural disease course. Second, the primary outcome of interest in COMMIT was time to treatment failure which was defined as failure to achieve corticosteroid-free remission at week 14,
Dulai PS, et al. Gut Month 2014 Vol(or 0 Notheir 0 1 Copyright Article author employer) 2014. Produced by BMJ Publishing Group Ltd (& BSG) under licence.
Downloaded from gut.bmj.com on June 30, 2014 - Published by group.bmj.com
Leading article Table 1 Comparative effectiveness randomised controlled trials comparing anti-TNF monotherapy to combination therapy with an immunosuppressive Crohn’s disease SONIC
Patient characteristics Participants, n Treatment regimen Male, % Mean age, years Disease characteristics Mean duration, years Extensive*, % Prior surgery, % Prior IS therapy, % Baseline steroids, % Clinical disease activity score†, mean Clinically active disease at baseline‡, % Endoscopic lesions at baseline, % CRP, mean Outcomes Primary CFREM¶, % Week 50 CFREM, % Mucosal healing, % Antidrug antibody, % IFX concentration, mean mg/mL Serious infection, %
Ulcerative colitis
14
COMMIT
15
UC-SUCCESS16
IFX mono
IFX combo
IFX mono
IFX combo
IFX mono
IFX combo
169 IFX+PBO 50 35
169 IFX+AZA 52 34
63 IFX+PBO 59 39
63 IFX+MTX 54 40
78 IFX+PBO 54 39
80 IFX+AZA 60 38
2.2 39 28 0 31 285 100 57 1.1
2.2 44 26 0 28 290 100 66 1.0
9.6 56 46 25 100 208 73 n/a 6.0
10.9 60 57 24 100 208 68 n/a 3.0§
6.3 19 n/a 10 40 6.0 100 99 n/a
5.2 34 n/a 10 48 6.3 100 100 n/a
44 35 30 14.6 1.6 4.9
57§ 46§ 44‡ 0.9 3.5§ 3.9
78 57 n/a 20.4 3.8 n/a
76 56 n/a 4.0 6.4 n/a
22 n/a 55 19 n/a 0.01
40§ n/a 63** 3 n/a 0
‡ p=0.06. *Extensive disease definition for SONIC and COMMIT is ileocolitis at baseline and for UC-SUCCESS is Mayo endoscopic subscore of three at baseline. †Clinical disease activity score for SONIC and COMMIT is the Crohn’s Disease Activity Index (CDAI) and for UC-SUCCESS is the partial Mayo score. ‡Clinical remission at baseline was defined as a CDAI score of
