Biological interventions for maintenance of mucosal healing in Crohn’s disease
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
The objectives of this systematic review are to determine the efficacy and safety of biologics for maintaining mucosal healing in CD.
Background
Description of the condition
Crohn’s disease (CD) is a relapsing‐remitting inflammatory disorder of the gastrointestinal tract. Ongoing transmural inflammation frequently leads to structural complications, such as strictures and fistulae (Cosnes 2002), and 50% of patients require surgery within 10 years of diagnosis (Peyrin‐Biroulet 2010). Medical therapy targeting clinical remission has not altered the natural history of CD (Cosnes 2005). Furthermore, clinical symptoms, as reflected by clinical disease activity indices, such as the Crohn’s Disease Activity Index (CDAI), correlate poorly with endoscopic findings (De Cruz 2013). Alternatively, there is evidence that mucosal healing (MH) is associated with improved long‐term outcomes in CD, including a trend toward reduced surgery (Shah 2016). MH has emerged as an attractive alternative therapeutic endpoint in CD. Most recent clinical trials evaluating medical therapies in CD have used endoscopic indices as an outcome measure. While no validated definition for MH exists (Sandborn 2002), MH is generally perceived to imply resolution of mucosal ulcerations observed at baseline endoscopy. However, the extent to which partial MH is associated with improved clinical outcomes is unknown. The two most widely used indices for grading endoscopic activity in CD are the Crohn’s Disease Endoscopic Index of Severity (CDEIS) (Mary 1989), and the Simple Endoscopic Score for Crohn’s Disease (SES‐CD) (Daperno 2004). The creators of the CDEIS, which ranges from 0 to 44, empirically defined complete endoscopic remission as a score of less than three and endoscopic remission as a score of less than six (Mary 1989). CDEIS scores of less than 5, 5 to 10 and greater than 15 represent mild, moderate and severe inflammation, respectively (Geboes 2005). The SES‐CD ranges from 0 to 60 and has been empirically divided into the following categories by investigators: 0 to 2 (remission), 3 to 6 (mild inflammation), 7 to 16 (moderate inflammation) and greater than 16 (severe inflammation) (Daperno 2004). There is no definite consensus regarding the definition for endoscopic response with either of these tools and different definitions have been put forward (Khanna 2014; Khanna 2016; Levesque 2015).
Description of the intervention
The introduction of biologic agents over the past 15 years revolutionized the treatment of inflammatory bowel disease (IBD). The advent of these highly efficacious therapies transformed the target of MH into a more attainable goal and spurred, to a large extent, the increased attention paid to MH. Relatively little is known about the ability of biologics to achieve MH in CD compared to the ability to attain clinical endpoints, which have been the focus of most large trials until only recently.
How the intervention might work
The chronic inflammation in CD is thought to stem from dysregulation of the immune system. Biologic therapies used for the treatment of CD inhibit specific molecular immune mediators of disease, including pro‐inflammatory cytokines and adhesion molecules that enable leukocyte migration. Infliximab, a chimeric immunoglobulin G1 (IgG1) monoclonal antibody directed against the cytokine tumour necrosis factor alpha (TNFα), was the first biologic used in CD. Additional anti‐TNFα agents include the humanized monoclonal antibodies adalimumab, certolizumab and golimumab. Ustekinumab is a fully humanized IgG1 monoclonal antibody directed against the common p40 subunits of the cytokines interleukin‐12 and interleukin‐23 (IL‐12/23p40). Tocilizumab inhibits IL‐6. Natalizumab, vedolizumab and etrolizumab disrupt the adhesive interactions between endothelial cells and circulating leukocytes, thereby hindering leukocyte trafficking.
Why it is important to do this review
Given the expanding armamentarium of biologics, sound evidence on which to base treatment decisions is paramount. While MH as an outcome measure in IBD is not a novel concept (Williams 1990), it has garnered much more attention as of late, coincident with the introduction of biologic therapy. MH is considered by many to be the most clinically relevant therapeutic endpoint currently available because of its association with improved long‐term outcomes. As biologic agents are fairly new and modulate the immune system, concerns regarding safety have been raised. A systematic review of biologics for maintaining MH in CD, including risk of bias assessment, has not yet been performed. This systematic review will summarize the existing evidence derived from randomized controlled trials (RCTs) regarding the efficacy and safety of biologic therapies for maintaining MH in CD.
Objectives
The objectives of this systematic review are to determine the efficacy and safety of biologics for maintaining mucosal healing in CD.
Methods
Criteria for considering studies for this review
Types of studies
RCTs will be considered for inclusion. This will include abstracts. All other study designs will be excluded, including studies with a randomized withdrawal design. Only studies with a clear, reproducible definition for MH will be eligible for inclusion.
Types of participants
Individuals eligible for inclusion will be CD patients with endoscopically documented MH (i.e., resolution of ulcers seen on baseline endoscopy) 4 to 26 weeks after initiating a biologic treatment, with or without fibrostenotic or fistulizing complications. Crohn’s disease will be defined as per conventional clinical, radiological and endoscopic criteria. Patients with isolated upper gastrointestinal CD or isolated perianal disease without mucosal involvement prior to initiating a biologic will not be eligible for inclusion. No restrictions will be applied for age, sex, disease duration or previous medication exposure. Patients with surgically‐induced remission will not be included. The authors acknowledge that endoscopy is best suited to assess patients with colonic disease and sub‐optimal for evaluating small bowel disease as only the terminal ileum is visualized during ileocolonoscopy.
Types of interventions
Studies that assess the efficacy of a biologic, alone or in combination with a second agent, for maintaining MH in humans with CD, compared to placebo or active medical treatment, will be included, For the purpose of this review, biologic therapy will refer strictly to monoclonal antibodies. Studies of small molecules, such as tofacitinib, will be excluded. Interventions involving a monoclonal antibody, alone or in combination with another agent, for a minimum duration of 6 months, at any dose or frequency and by any route of administration, will be considered for inclusion. The monoclonal antibodies currently approved for CD and thus to be included in this review are the anti‐TNFα agents, infliximab, adalimumab, certolizumab and golimumab, the integrin antagonists, natalizumab and vedolizumab, and the anti‐IL12/23p40 agent, ustekinumab.
Types of outcome measures
Primary outcomes
The primary outcome measure will be maintenance of MH, defined as continued absence of mucosal ulcerations during ileocolonoscopy performed at least six months after biologic initiation, after endoscopically‐confirmed induction of MH (resolution of baseline ulcers within 4 to 26 weeks of biologic initiation).
Secondary outcomes
Secondary outcome measures will include:
endoscopic response, defined as a relative or absolute decrease in either the CDEIS or SES‐CD compared to baseline, assessed at least 6 months after biologic initiation, among patients who experienced at least endoscopic response to induction;
adverse events;
serious adverse events; or
disease‐specific quality of life (QOL), using the IBD Questionnaire or Short IBD Questionnaire.
Given the lack of an agreed upon definition for endoscopic improvement in CD, the definition of endoscopic response used in this review will not entail a specific magnitude of change in CDEIS or SES‐CD but, rather, the definition applied by the authors of the included studies will be used.
Search methods for identification of studies
Electronic searches
We will search the following databases from inception to date:
1. MEDLINE;
2. EMBASE;
3. Cochrane Central Register of Controlled Trials (CENTRAL); and
4. The Cochrane IBD Group Specialized Trials Register.
The search strategies are reported in Appendix 1.
Searching other resources
References of relevant publications will be manually searched to identify additional studies that may have been missed by the computer‐assisted search strategy. The proceedings from major gastroenterology meetings from 2002 to the present will be manually searched as well. We will contact leaders in the field (B. Feagan and W. Sandborn) and representatives of pharmaceutical companies involved in drug development for CD to identify additional studies. We will search the trial database clinicaltrials.gov to identify ongoing studies.
Data collection and analysis
Selection of studies
The titles and abstracts of publications identified by the search strategy will be assessed independently by two authors (AR and HHS) and studies meeting the inclusion criteria will be retained. Where the eligibility of a study is unclear, the full text will be retrieved and reviewed. Disagreements will be resolved by discussion and consensus with the input of a third author (CHS) as required.
Data extraction and management
Data will be extracted independently by two authors (AR and PC) using standardized extraction forms. We will resolve disagreements will by consensus and discussion with the input of a third author (CHS) as necessary. Study authors will be contacted for additional information as needed. Authors extracting data will not be blinded to author, institution or journal. Data will be entered into Review Manager 5.3. The following data will be abstracted:
General study information, including title, first author and year of publication;
Participant population details, including the number of patients screened, number randomized to each arm, number completing each arm and number lost to follow‐up, as well as geographic location (country) and year of patient enrolment;
Participant characteristics, including age, sex, disease location, severity, disease duration, previous therapy and previous resection;
Intervention details, including the biologic agent, dose, route of administration, frequency and duration;
Control group details, including the placebo or active medical agent, dose, route of administration, frequency and duration;
Details pertaining to concomitant medications received in each study arm;
Outcome details, including the precise definition of MH used, timing of assessment, number/proportion of patients maintaining MH in each study arm, number and proportion of patients with endoscopic response in each study arm, number and proportion experiencing adverse and serious adverse events in each study arm, and quality of life outcomes in both study arms; and
Study methodology, including details pertaining to randomization, allocation concealment, blinding and inclusion and exclusion criteria.
Assessment of risk of bias in included studies
The Cochrane risk of bias tool will be used to assess the methodological quality of the included studies (Higgins 2011a). We will assess the following study characteristics:
Random sequence generation (i.e., the allocation sequence was adequately generated);
Allocation concealment (i.e., allocation concealment was adequately concealed);
Blinding of participants, personnel and outcome assessors (i.e. knowledge of intervention allocation was appropriately prevented);
Incomplete outcome data (i.e. attrition was appropriately addressed);
Selective outcome reporting (i.e. all outcomes outlined in the methods were included in the analysis); and
Other potential sources of bias (i.e. the study appeared to be free of other problems that might place it at high risk of bias).
Based on the above, studies will be deemed to have 'low' (if low risk for all six factors), 'high' (if high risk for one or more factors) or 'unclear' (if unclear risk for one or more factors) risk of bias. The risk of bias assessment will be performed independently by two authors (AR and PC). We will resolve disagreements by discussion and consensus with the input of a third author (CHS) as necessary. Study authors will be contacted for additional information as needed. Quality assessors will not be blinded to author, institution or journal.
We will use the GRADE approach to assess the overall quality of evidence for the primary outcome and selected secondary outcomes (Schünemann 2011). Outcomes from pooled RCTs begin as high quality evidence, but may be downgraded as a result of: high risk of bias (methodological quality), indirectness of evidence, inconsistency (unexplained heterogeneity), imprecision (sparse data), and reporting (publication) bias. All elements will be considered in determining the quality of the evidence, which will be graded as high (i.e. further research is very unlikely to influence the degree of confidence in the effect estimate); moderate (i.e. further research is likely to have an important influence on the degree of confidence in the effect estimate and may change the estimate); low (i.e. further research is very likely to have an important influence on the degree of confidence in the effect estimate and is likely to change the estimate); or very low (i.e. very uncertain about the estimate).
Measures of treatment effect
We will analyze data using Review Manager (RevMan version 5.3, Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014). Data will be abstracted from the original publications and transformed into two by two tables. Treatment effects will be expressed separately for each comparison. Dichotomous outcomes, including the primary outcome, will be expressed as a risk ratio (RR) with corresponding 95% confidence interval (CI). An intention‐to‐treat analysis will be used. Continuous outcomes will be expressed as the mean difference (MD) with corresponding 95% CI if the same instrument and scale was used across studies for a given outcome, or as the standardized mean difference (SMD) with corresponding 95% CI, if this was not the case.
Unit of analysis issues
If we encounter any cluster randomized trials, we will be perform approximate analyses as appropriate, by calculating effective sample sizes (Higgins 2011b). In the case of cross‐over studies, only data pertaining to the first treatment period will be considered. For trials with multiple intervention or control groups, study groups will be combined, as appropriate, to create single pair‐wise comparisons (Higgins 2011b).
Dealing with missing data
In the case of missing data despite attempts to contact study authors, we will use a conservative intention‐to‐treat analysis (i.e. the worst outcome will be assumed) for dichotomous outcomes. We will not make any assumptions about missing continuous missing data and only patients with available data will be included in analyses.
Assessment of heterogeneity
We plan to assess clinical and statistical heterogeneity. We will assess statistical heterogeneity using the Chi2 test and I2 statistic. A P value of less than 0.10 will be considered indicative of statistically significant heterogeneity for the Chi2 test. We will interpret I2 values as follows: 0% to 40% ‐ might not be important; 30% to 60% ‐ may represent moderate heterogeneity; 50% to 90% ‐ may represent substantial heterogeneity; and 75% to 100% ‐ considerable heterogeneity. If the I2 statistic shows a moderate to high degree of heterogeneity and the Chi2 test is statistically significant, we will assess heterogeneity by visually inspecting the forest plots for obvious outliers. If applicable, a sensitivity analysis will then be performed excluding the outlier to see if this explains the heterogeneity. If possible, subgroup analyses will be performed to explore sources of clinical and methodological heterogeneity.
Assessment of reporting biases
If there are sufficient studies (i.e., ≥10), publication bias will be assessed by means of a funnel plot, as per Egger’s method (Egger 1997).
Data synthesis
We will pool data where it is reasonable to assume studies were estimating the same underlying treatment effect. As such, it was decided a priori to not combine outcomes for different medications. Where appropriate, pooling of results will be performed, using the Mantel‐Haenszel statistical method and a random‐effects model, with weights determined using the DerSimonian and Laird method. For dichotomous outcomes, we will calculate the pooled RR with corresponding 95% CI. We will calculate the pooled MD or SMD with corresponding 95% CI as appropriate for continuous outcomes.
Subgroup analysis and investigation of heterogeneity
The following subgroup analyses will be performed, if possible, to investigate sources of clinical heterogeneity:
patients previously exposed to biologic therapy and biologic‐naive patients.
Sensitivity analysis
In addition to the sensitivity analysis described above to investigate heterogeneity, the following sensitivity analyses will be performed, if possible:
low risk of bias studies; and
studies in which MH was assessed using a validated endoscopic scoring tool.
