Traditional corticosteroids for induction of remission in Crohn's disease
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
The primary objective is to systematically review the efficacy of traditional (systemically‐absorbed, non‐topical) corticosteroids for inducing remission in CD. The secondary objective is to evaluate the adverse effects associated with use of such corticosteroids to induce remission in CD.
Background
Crohn's disease (CD) is characterized by chronic transmural inflammation of the gastrointestinal tract (Bousvaros 2007). A wide range of clinical symptoms occur with some patients remaining chronically clinically active and others experiencing a series of relapses and remissions. CD is currently thought to be caused by a cascade of immunologic reaction triggered by environmental factors in a genetically‐predisposed host. Historically, prior to the advent of biologic therapies, corticosteroids have been the most effective class of medication for treatment of acute flares of CD in adults (Baumgart 2007) and children (Hyams 2005). Corticosteroids downregulate production of inflammatory cytokines such as interleukin (IL)‐1, IL‐6, and tumor necrosis factor (TNF)‐alpha by inhibiting transcription of specific genes involved in their production (Hyams 2000). Corticosteroids also inhibit protein synthesis by affecting the stability of messenger RNA (Barnes 1993). The interaction between corticosteroid receptors and NF‐ΚB results in downregulation of NF‐ΚB and therefore a blunting of inflammatory response (Yang 2002). Corticosteroids have been used for the treatment of inflammatory bowel disease (IBD) since the 1950s (Truelove 1954). Unfortunately, systemically‐administered corticosteroids are associated with adverse effects such as moon facies, acne, infection (increased risk of abdominal and pelvic abscess in CD patients), ecchymoses, hypertension, diabetes mellitus, osteoporosis, cataracts, glaucoma and growth failure in children (Baumgart 2007).
More recently, non‐traditional topically‐active formulations of corticosteroids (such as budesonide or topical enema therapy) have been developed in order to reduce systemic absorption and side effects while maintaining efficacy. Studies examining non‐traditional corticosteroids have been subject to previous meta‐analysis (Otley 2005) and will therefore not be included in this review.
Aminosalicylates and Crohn's disease
Another mainstay medication used for the past forty years in the treatment of CD are the 5‐aminosalicylate (5‐ASA) drugs, including sulfasalazine (Akobeng 2005). 5‐ASA drugs act through a variety of mechanisms including inhibiton of the NF‐ΚB pathway, inhibition of apoptosis induced by oxidative stress, modulation of prostaglandin metabolism, and inhibition of colonic production of leukotrienes (Desreumaux 2006). Although considered to be modestly effective for induction of remission in CD, 5‐ASA drugs are used by many practitioners. One systematic review examined both placebo‐controlled and corticosteroid‐controlled clinical trials and concluded that 5‐ASA was less effective for induction of remission than corticosteroids in patients with CD (Feagan 1998). A number of forms of 5‐ASA have been developed for the treatment of IBD. Sulfasalazine, the earliest form of 5‐ASA, maintains its bond between sulphapyridine and 5‐ASA until interacting with bacteria in the colon, where the bond is cleaved, releasing the 5‐ASA medication (Azad Khan 1977). One gram of sulfasalazine is reported to contain 400 mg of 5‐ASA (Sandborn 2003). Other forms of 5‐ASA such as mesalazine/mesalamine (Asacol®, Salofalk®, Rowasa®) and olsalazine are formulated to interact with the pH environment of the colon and therefore may have maximal activity distal to the ileum (Desreumaux 2006). By contrast, Pentasa® (mesalamine) was formulated for release throughout the small intestine and colon (Hardy 1993). The choice of 5‐ASA formulation is left to the individual practitioner, who must base the choice on CD location, previous experience, patient response and adverse effect profile.
Importance of this review
Although a number of systematic reviews examine the efficacy of corticosteroids for induction of remission in CD (Yang 2002; Bebb 2004; Lichtenstein 2006), only one meta‐analysis has been published (Salomon 1992). No studies have compared corticosteroids to 5‐ASA using meta‐analytic techniques. Despite the high incidence of adverse effects, oral and intravenous corticosteroids continue to be commonly used to induce clinical remission in active CD. Additionally, corticosteroids have been considered standard treatment in both clinical practice (Lichtenstein 2006) and randomized‐controlled clinical trials involving newer CD treatments such as biologics (Kamm 2006). It is important to carefully examine the efficacy and safety of corticosteroids, in order to provide an evidence‐based approach to assessment of the risk/benefit profile of this class of medication.
Objectives
The primary objective is to systematically review the efficacy of traditional (systemically‐absorbed, non‐topical) corticosteroids for inducing remission in CD. The secondary objective is to evaluate the adverse effects associated with use of such corticosteroids to induce remission in CD.
Methods
Criteria for considering studies for this review
Types of studies
Only randomized controlled trials of oral or intravenous corticosteroids, administered for the induction of remission of CD in adults or children, published in any language, will be included. Studies published in abstract form only will be included if enough data are provided to assess outcome. The control arm of studies assessed must include either placebo or an active comparator including 5‐aminosalicylic acid (5‐ASA) or sulfasalazine.
Types of participants
Participants will include patients of any age with CD defined by conventional clinical, radiological and/or endoscopic criteria, which is categorized as being acutely active (i.e. active clinical symptoms, Crohn's disease activity index (CDAI) >150 OR Pediatric Crohn's Disease Activity Index (PCDAI) >15), OR a validated severity scale indicating active disease (e.g. Harvey‐Bradshaw Index, Van Hess Index)).
Types of interventions
Trials will be included if the primary intervention is traditional, systemically‐active corticosteroid medication in any form (oral or parenteral). Topically‐released corticosteroids will not be considered for this review (e.g. enteric‐coated budesonide, skin creams/ointments, inhaled fluticasone, enema therapy, etc.). Co‐interventions will be allowed only if they are balanced between the study groups, including but not restricted to immunomodulators, biologic therapy, and dietary therapy.
Types of outcome measures
The primary outcome measure will be the number of patients achieving remission as defined by an absence of clinical symptoms (determined by the investigator), a CDAI <150 or a PCDAI <15 at follow‐up points at weeks 4‐6 (early), weeks 10‐12 (middle), and weeks 15 or later (late) following initiation of therapy. Secondary outcomes will include clinical response, change in mean CDAI, presence of adverse events, and withdrawal rate of participants among the intervention and control groups (for toxicity and adverse events).
Search methods for identification of studies
See: Cochrane Inflammatory Bowel Disease and Function Bowel Disorders (IBD/FBD) Group methods used in reviews.
Electronic databases
An on‐line database literature search will be performed for human studies, without language restrictions, using the following databases: MEDLINE (NLM, National Library of Medicine, Bethesda; 1950 to May 2007), and EMBASE (Elsevier, NY; 1980 to May 2007) on OVID, as well as the Cochrane Central Register of Controlled Trials (The Cochrane Collaboration, UK; 2007, issue 2) and the Cochrane IBD/FBD Review Group Specialized Register (The Cochrane Collaboration, UK; May 11 2007). Ongoing trials will be assessed using ClinicalTrials.gov (NLM, National Library of Medicine, Bethesda, May 2007). Proceedings from major gastrointestinal meetings (American Gastroenterology Association, British Society of Gastroenterology and North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition) will also be manually searched from 2002 to 2007 in order to identify recent unpublished studies.
Search terms
The following keywords will be used, including both text word [tw] and Medical Subject Heading (MeSH) terms (using the subheadings described) where appropriate and revised according to the database used:
1. inflammatory bowel diseases/ [MeSH] OR crohn disease/ [MeSH] OR (crohn$ OR (inflammatory ADJ2 bowel ADJ2 disease$)).mp. [mp=title, original title, abstract, name of substance word, subject heading word] [tw]
2. glucocorticoids/ad, tu [MeSH] OR dexamethasone/ad, tu [MeSH] OR dexamethasone isonicotinate/ad, tu [MeSH] OR fluprednisolone/ad, tu [MeSH] OR methylprednisolone/ad, tu [MeSH] OR methylprednisolone hemisuccinate/ad, tu [MeSH] OR prednisolone/ad, tu [MeSH] OR prednisone/ad, tu [MeSH] OR hydrocortisone/ad, tu [MeSH] OR cortisone/ad, tu [MeSH] OR methylprednisole (nm) [MeSH] OR METHYLPREDNISOLONE HEMISUCCINATE/ [MeSH] OR PREDNISOLONE, THERAPEUTIC/ [MeSH] OR PREDNISONE, THERAPEUTIC/ [MeSH] OR cortisone, therapeutic/ [MeSH]
3. 1 AND 2
4. exp Placebos/ [MeSH] OR (placebo$ OR sham$ OR dummy$).mp. [mp=title, original title, abstract, name of substance word, subject heading word] [tw]
5. exp MESALAMINE/ [MeSH] OR ("5‐aminosalicylic$" OR "5‐aminosalicylate$" OR "5‐ASA" OR 5aminosalicylic$ OR 5aminosalicylate$ OR 5ASA OR pentasa OR mesalamine OR asacol).mp. [tw]
6. Sulfasalazine/ [MeSH] OR (sulfasalazine$ OR salazopyrin$ OR salazosulfapyridine$ OR asulfidine$ OR azulfadine$ OR azulfidine$).mp. [tw]
7. OR/4‐6
8. 3 AND 7
9. limit 3 to randomized controlled trial
10. 9 NOT 8
11. from 10 keep 37,45‐46,52
12. (random$ OR RCT OR RCTs OR ((singl$ OR doubl$ OR tripl$ OR trebl$) ADJ25 (mask$ OR blind$))).ti,ab. [tw]
13. 3 AND 12
14. 13 NOT (8 OR 9)
15. 8 OR 11
Other sources
Additional citations will be identified by manually searching the reference lists of articles retried from the computerized databases and relevant review articles. Unpublished studies will be sought by contacting experts in the field. The authors of published abstracts will be contacted to complete missing data, and sensitivity analysis is planned in case they do not respond.
Data collection and analysis
Selection of studies
Abstracts of all articles meeting the above search strategy will be screened for eligibility. Full text studies will be retrieved if they are potentially eligible for inclusion or if they are relevant review articles, for manual reference search. The retrieved full text articles will then be independently reviewed by EIB and CHS for eligibility. Any disagreements will be solved by consensus with content experts (AHS and AMG).
Data extraction and management
Two authors (EIB and CHS) will independently complete a data extraction form for each eligible study. The following data will be retrieved:
1. General information: title, journal, year, published/unpublished.
2. Study information: design (e.g. who was blinded), years of enrollment, crossover or not, methods used to ensure adequacy of randomization, concealment of allocation and blinding, power calculation (a priori and post hoc).
3. Intervention: form and dose of corticosteroid, conversion to prednisone‐equivalents, type of comparison group, co‐intervention.
4. Eligibility: inclusion/exclusion criteria, total number screened and randomized.
5. Baseline characteristics (in each group): age, sex, race, disease severity (and how evaluated), concurrent medications used, disease location, prior surgery, time since last surgery, CDAI/PCDAI, length of symptoms prior to randomization.
6. Follow‐up: length of follow‐up, assessment of compliance, withdrawals and loss to follow‐up.
7. Outcome: cumulative remission rate in each group at weeks 2, 4, 6, 8, 10, 12, 15, 16, 17, 18, and 24 following initiation of treatment or placebo, adverse events, and drug interaction rate and details.
Assessment of methodological quality of included studies
To assess methodological quality a five point ordinal scale (Jadad 1996) will be independently completed by EIB and CHS for all eligible articles. Articles with poor quality (score of 0‐2) will be considered for subgroup analysis of low‐quality studies. Since reliability of the Jadad score is not high (Clark 1999; Juni 2001), the final score will be used only as a general guideline and decision on eligibility will be accomplished by the mutual agreement of the authors based on the adequacy of concealment, blinding of intervention and outcome, and completeness of follow‐up (Higgins 2005; Juni 2001). Studies will be classified as "low risk of bias" (i.e. all above criteria met), "moderate risk of bias" (i.e. one or more criteria partly met) and "high risk of bias" (i.e. one or more criteria not met). The latter group will be considered for subgroup analysis of low‐quality studies. When insufficient information is provided to determine the methodological quality, study authors will be contacted to provide further details on the above criteria. Any disagreement between authors (EIB and CHS) will be solved by consensus with content experts (AHS and AMG).
STATISICAL STATISTICAL ANALYSIS
Data will be analyzed using Review Manager (RevMan 4.2.10, Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2006). Categorical data (as remission or not at weeks 2, 4, 6, 8, 10, 12, 15, 16, 17, 18, and 24 following initiation of treatment or placebo) for each study will be transformed into 2x2 tables. The latest follow‐up time point for each study will be used for each of the three primary outcomes: Early remission (weeks 4‐6), middle remission (weeks 10‐12) and late remission (weeks 15 or later). For example if a study follows patients for 5 weeks following initiation of therapy, its outcomes will be pooled with all studies following patients for 4‐6 weeks. The last reported follow‐up time point for each study will be combined in order to account for differing patient follow‐up lengths among studies.
Measures of treatment effect
The proportions of patients in remission will be calculated and reported as relative risk and 95% confidence interval (95% CI). The number needed to treat (NNT) and absolute risk reduction (ARR) will also be calculated where appropriate. Analysis of studies using placebo or 5‐ASA/sulfasalzine in the control arms will be conducted separately. Secondary outcomes such as presence of adverse effects and drug interactions will also be reported as above.
Meta‐analysis
A random or fixed effects model will be used to incorporate studies depending on clinical and statistical heterogeneity. Weighting of the studies will be performed using the Mantel‐Haenszel method (for fixed effects models) or the DerSimonian‐Laird method (for random effects models).
Subgroup analysis and assessing for heterogeneity
Studies will be independently reviewed for any clinical and methodological heterogeneity and possible reasons for heterogeneity will be explored. The decision of whether to pool studies will be aided by calculating the I2 measure, interpreted as low heterogeneity (25%), moderate heterogeneity (50%) and high heterogeneity (75%) (Higgins 2003). The Cochran's Chi‐Square test for homogeneity (Q test) will be performed with P <0.10 being considered to be statistically significant due to its low sensitivity. The following a priori subgroup analyses are planned, governed by the number of identified studies: dose of corticosteroid used (in prednisone equivalents), method of administration (intravenous versus oral), pediatric versus adult patients, different formulations of corticosteroids (e.g. prednisone, prednisolone, methylprednisolone, hydrocortisone, etc.) or 5‐ASA (sulfasalazine, mesalamine, mesalazine, etc.), and disease location (Silverberg 2005). Studies with balanced co‐interventions will also be analyzed separately from studies without co‐interventions. Subgroups were chosen based on the possibility that differing doses, formulations and disease location may impact on success of treatment. This is particularly true of certain 5‐ASAs which have been suggested to be more successful at treating colonic disease given their predicted controlled‐release characteristics.
Sensitivity analysis
In order to assess the robustness of the eligibility criteria, a sensitivity analysis is planned to exclude poor quality studies, studies published in an abstract form, studies not reporting methods to assess compliance and small studies (<50 patients). Publication bias will be assessed using a funnel plot, depending on the number of eligible studies.
