CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2012;10:1002–1007
Patients Enrolled in Randomized Controlled Trials Do Not Represent theInflammatory Bowel Disease Patient Population
CHRISTINA HA,* THOMAS A. ULLMAN,‡ COREY A. SIEGEL,§ and ASHER KORNBLUTH‡
*Division of Gastroenterology, The Johns Hopkins School of Medicine, Baltimore, Maryland; ‡Dr. Henry D. Janowitz Division of Gastroenterology, Mount Sinai Schoolof Medicine, New York, New York; and §Division of Gastroenterology, Dartmouth-Hitchcock Medical Center, Hanover, New Hampshire
This article has an accompanying continuing medical education activity on page e78. Learning Objectives—At the endof this activity, the successful learner will be able to identify the potential limitations of clinical trial data when
translated to a real-world clinical practice.
itort
See editorial on page 1008.
BACKGROUND & AIMS: Multiple randomized controlledrials (RCTs) have been conducted to determine therapeutic effi-acy of the biological agents for the inflammatory bowel diseasesIBD). However, the external validity of findings from RCTs mighte compromised by their stringent selection criteria. We investi-ated the proportion of patients encountered during routine clin-cal practice who would qualify for enrollment into a pivotal RCTf biological agents for IBD. METHODS: We performed aetrospective cohort study of adult patients with moderate–severeBD who presented to a tertiary referral center. Inclusion andxclusion criteria were extracted from published RCTs of biologicspproved by the Food and Drug Administration and applied tohe study population. RESULTS: Only 31.1% of 206 patients
with IBD (34% with Crohn’s disease [CD], 26% with ulcerativecolitis) would have been eligible to participate in any of the se-lected RCTs. Patients would have been excluded because they hadstricturing or penetrating CD, took high doses of steroids, hadcomorbidities or prior exposure to biologics, or received topicaltherapies. Of the trial-ineligible patients with ulcerative colitis,23.3% had colectomies, and 31.7% received infliximab, with a 63.2%response rate. Approximately half (49.4%) of the 82 trial-ineligiblepatients with CD received biological therapies, with lower responserates (60%) than trial-eligible patients (89%; P � .03). CONCLU-
IONS: Most patients with moderate–severe IBD evaluatedn an outpatient practice would not qualify for enrollment in
pivotal RCT of biological reagents; this finding raises im-ortant questions about their therapeutic efficacy beyond thelinical trial populations. Additional evaluation of the trans-arency of RCT design and selection criteria is needed toetermine whether trial results can be generalized to theopulation.
Watch this article’s video abstract and others at http://tiny.cc/bz9jv.
Scan the quick response (QR) code to the left withyour mobile device to watch this article’s video ab-stract and others. Don’t have a QR code reader? Getone at mobiletag.com/en/download.php.
Crohn’s disease (CD) and ulcerative colitis (UC) are chronicinflammatory diseases of the gastrointestinal tract char-
acterized by symptomatic relapses often requiring escalation ofmedical therapy. The introduction of the biologic anti–tumornecrosis factor alpha (anti-TNF) agents has transformed themanagement of the inflammatory bowel diseases (IBD) duringthe past decade. The evidence supporting their use for patientswith moderate–severe IBD is compelling, with demonstratedefficacy at inducing and maintaining remission for CD and UCcompared with placebo.1,2 In addition to symptom improve-ment, these agents also might promote mucosal healing andother favorable longer-term clinical outcomes such as decreasedhospitalizations, surgeries, and improved quality of life.3–10
There is also evidence supporting the use of anti-alpha4 integ-rin therapy for the treatment of patients with moderate–severeCD.11,12 Since 2002, there have been at least 8 major random-zed controlled trials (RCTs) for the Food and Drug Adminis-ration (FDA) approved biologics. The results from these piv-tal RCTs form the basis for regulatory agency approval, expertecommendations, and evidence-based practice guidelines forhe management of patients with clinically active IBD.13,14
RCTs are currently the best approach available to studytreatment effect. By means of prespecified patient enrollmentcriteria, these trials aim to minimize the potential bias ofconfounding variables to preserve the internal validity of thestudy results. It is assumed that the findings obtained fromthese studies carry a high level of external validity with appli-cability to general clinical practice. However, by limiting the
Abbreviations used in this paper: ACCENT 1, A Crohn’s DiseaseClinical Study Evaluating Infliximab in a New Long-Term TreatmentRegimen; ACT 1 and 2, Active Ulcerative Colitis Trials; anti-TNF, anti–tumor necrosis factor alpha; CD, Crohn’s disease; CHARM, Crohn’sTrial of the Fully Human Antibody Adalimumab for Remission Mainte-nance; CLASSIC 1, Clinical Assessment of Adalimumab Safety andEfficacy Studied as Induction Therapy in Crohn’s Disease; ENACT,Efficacy of Natalizumab as Active Crohn’s Therapy; ENCORE, Efficacyof Natalizumab in Crohn’s Disease Response and Remission; FDA,Food and Drug Administration; HBI, Harvey–Bradshaw Index; IBD,inflammatory bowel diseases; PRECISE1, Pegylated Antibody Frag-ment Evaluation in Crohn’s Disease; RALES, Randomized AldactoneEvaluation; RCT, randomized controlled trial; SONIC, Study of Biologicand Immunomodulator Naive Patients in Crohn’s Disease; UC, ulcer-ative colitis.
September 2012 EXTERNAL VALIDITY OF IBD CLINICAL TRIALS 1003
available pool of potential study patients with strict inclusionand exclusion criteria in an effort to maximize internal validity,trial results might have limited external validity. The generaliz-ability of these studies might be further compromised by theheterogeneity of the IBD patient population. The success orfailure of clinical trials informs the clinician about the effect ofthat agent on the included patients but is unable to commenton the probable benefits to an individual patient. In addition,many patients with clinically active disease present with symp-toms and circumstances that would preclude entry into a clin-ical trial, yet there is little information available regarding theapplicability of the individual trial data for these patients.
With these potential limitations of RCTs in mind, we hy-pothesized that only a small percentage of patients seen in aconsultative IBD practice would have been eligible for studyparticipation. Therefore, the aim of this study was to estimatethe percentage of IBD patients with moderate–severe diseasewho would have been eligible for enrollment into a pivotalbiologic RCT and to assess the enrollment criteria most likelyto impact trial eligibility.
MethodsStudy Population and Data CollectionWe performed a retrospective chart review of consecu-
tive adult IBD patients with moderate–severe disease activitypresenting to the Mount Sinai Medical Center for an escalationor adjustment of their medical therapy from January 2008 toJune 2009. Moderate–severe disease activity was defined as aHarvey–Bradshaw Index (HBI) score between 8 and 16 for CDand a Mayo UC disease activity index score between 6 and 12 forUC.7,15 Patients with indeterminate colitis, suspected but notstablished IBD, mild symptoms, symptoms thought to benrelated to IBD, disease remission, or with incomplete records
or review were excluded. Because there is currently only oneDA-approved biological therapy for UC, UC patients currentlyn or previously exposed to biologics were also excluded from
Figure 1. Clinical trial selectionand study enrollment algorithm.
he study. The Program for the Protection of Human Subjects
at the Mount Sinai School of Medicine in New York approvedthis study.
Data collection included routine demographics, IBD pheno-type, extent, disease activity assessments, comorbidities, me-dication, and surgical histories. Additional data collectedincluded subsequent medical treatments and surgical interven-tions as well as therapeutic response determined by physician’sglobal assessment by 4 –12 weeks after the initial visit.
Trial SelectionFor the CD patients, we selected 7 published RCTs of
biological therapy for patients with moderate–severe diseaseactivity: ACCENT 1 (A Crohn’s Disease Clinical Study Evaluat-ing Infliximab in a New Long-Term Treatment Regimen),16
CLASSIC 1 (Clinical Assessment of Adalimumab Safety andEfficacy Studied as Induction Therapy in Crohn’s Disease),17
CHARM (Crohn’s Trial of the Fully Human Antibody Ada-limumab for Remission Maintenance),18 PRECISE1 (Pegylated
ntibody Fragment Evaluation in Crohn’s Disease),19 ENCORE(Efficacy of Natalizumab in Crohn’s Disease Response andRemission),11 ENACT (Efficacy of Natalizumab as ActiveCrohn’s Therapy),12 and SONIC (Study of Biologic and Immu-
omodulator Naïve Patients in Crohn’s Disease).20 For the UCpatients, we selected the ACT 1 and 2 trials (Active UlcerativeColitis Trials) for infliximab in moderate–severe UC7 (Figure 1).
Inclusion and exclusion criteria were extracted from thepublished manuscripts of these CD and UC trials and appliedto our study population to determine overall percentage of trialeligibility. Exclusion criteria from the studies were further sub-categorized on the basis of disease complications, comorbidconditions, and concomitant medications to identify variablesthat most affect trial eligibility.
Outcomes and AnalysisThe primary outcome of interest was to determine the
RCT eligibility of our study population. Secondary outcomes ofinterest included identification of the selection criteria that
more commonly impact trial eligibility among outpatient refer-
acdsp
ybT
c5t
H
1004 HA ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 10, No. 9
ral practices and biological response rate among trial-ineligiblestudy participants.
We report dichotomous variables as frequency counts andpercentages with comparisons performed by using Fisher exacttest. Continuously reported variables are presented as meanswith standard deviations, and comparisons were tested by usingStudent unpaired t test. All tests performed were two-sided witha significance level of P � .05 used in all analyses.
ResultsPatient CharacteristicsTwo hundred six patients with moderate–severe IBD
met enrollment criteria during the 18-month study period. Ofthe 125 CD patients, 34% of patients (n � 43) would havequalified for enrollment into at least 1 of the 7 included RCTs.Trial eligibility for the RCTs for infliximab, adalimumab, cer-tolizumab pegol, and natalizumab ranged from 8% for theSONIC trial to 27% for the CHARM and PRECISE1 trials. Only25% of the UC patients (n � 21) would have qualified forenrollment into the ACT trial.
There were no significant differences in age, gender, diseaseduration, or location between CD patients who would vs wouldnot have qualified for clinical trial participation. The onlynotable difference between the groups related to disease behav-ior was that 40% of the trial-ineligible CD patients (n � 51) had
symptomatic stricture or abscess (nonperianal), an exclusionriterion for almost all of the CD trials (Table 1). Diseaseuration, extent, and Mayo score were similar between thetudy patients who would and would not have been eligible forarticipation in the ACT trial for infliximab in UC (Table 2).
Factors Impacting Trial EligibilityThe most common reasons for trial ineligibility for CD
were the presence of symptomatic strictures or documentedabscesses on imaging (n � 51, 62.2%), recent exposure to anti-TNF therapy or nonresponse to prior anti-TNF therapy (n �42, 51.2%), high-dose steroid use (n � 15, 18.3%), and comorbidconditions such as poorly controlled cardiovascular or pulmo-
Table 1. CD Patient Demographics and DiseaseCharacteristics
GI, gastrointestinal; SD, standard deviation.aP � .0001. P � NS for all other comparisons.
nary disease and malignancies (n � 21, 25.6%). Subgroup anal-
ses of trial eligibility among patients on the basis of prioriological exposure and disease behavior are summarized inable 3.
The most common reason for trial ineligibility for UC wasurrent rectal therapy usage (n � 34, 56.7%) because topical-aminosalicylates or corticosteroids were not allowed duringhe 2 weeks before the initial visit for ACT trial enrollment.7
However, 16 of these patients (47.1%) had other reasons for trialexclusion in addition to concurrent topical therapy. Other ex-clusion criteria included patients who were steroid and immu-nomodulator naive (n � 27, 45.0%), had new diagnoses of UC(n � 10, 16.7%), or needed probable colectomy because of age,comorbidity, or concomitant dysplasia found during colonos-copy (n � 9, 15.0%).
Therapeutic Response to Biologics of theTrial-Ineligible PatientsApproximately 50% of the CD study patients who
would not have qualified for a clinical trial enrollment (n � 40)were started on anti-TNF therapies as monotherapy or in com-bination with immunomodulators, had a dose escalation oftheir current biologics, or switched to a different anti-TNFagent. Ultimately, almost 50% of these trial-ineligible patientsunderwent surgery either as a primary therapy or because ofinadequate response to biologics or immunomodulators. Themajority of studied UC patients who would not have qualifiedfor the ACT trial were started on immunomodulators, hadaltered dosing of immunomodulators, were started on inflix-imab outside the confines of a clinical trial, or were treated with
Table 2. UC Patient Demographics and DiseaseCharacteristicsa
September 2012 EXTERNAL VALIDITY OF IBD CLINICAL TRIALS 1005
cyclosporine. Approximately one-fourth of these trial-ineligiblepatients went for colectomy at 4 –12 weeks (Table 4).
Among the study patients with CD who were started onbiologics, the majority of patients who would have qualified forenrollment (89%) had a subjective clinical response to therapycompared with only 60% of the trial-ineligible CD patients 4 –12weeks after the initial visit (P � .03). Of the 40% who were
iological nonresponders in the trial-ineligible group, 10 ofhese 16 patients required surgery. There were no significantifferences in biological response rates or colectomy rates be-ween ACT trial-eligible and trial-ineligible UC patients (Figure
2). Six trial-ineligible UC patients received cyclosporine, with 2patients responding to therapy and the remaining 4 patientsrequiring colectomy.
DiscussionThe biologics are among the most prescribed medica-
tions for symptomatic IBD patients. The evidence supportingtheir increasing utilization is largely based from RCT data.
able 4. Therapeutic Changes Among Trial-Ineligible CD andUC Patients
Outcome n (%)
DStarted or optimized dosing of immunomodulators 12 (14.8)Anti-TNF therapy 40 (49.4)Initiated anti-TNF monotherapy 22 (24.7)Dose escalation or change in anti-TNF 13 (16.0)Initiated combination anti-TNF � immunomodulator 5 (6.2)Surgery 40 (49.4)
UCChanged dosing of 5-aminosalicylates 6 (10.0)Started or optimized dosing of immunomodulators 16 (26.7)Started anti-TNF therapy 19 (31.7)Started cyclosporine 6 (10.0)Colectomy 14 (23.3)
Figure 2. Response rates to bi-logics among trial-eligible andrial-ineligible CD and UC pa-
ients.
However, our study suggests that IBD patients enrolled in themajor RCTs of the FDA-approved biological agents might notbe completely representative of patients encountered duringroutine clinical practice where anti-TNF therapy is often con-sidered as the next step of management. Compared with thepatients enrolled in the major IBD biological trials, our studypopulation appeared to be fairly similar in terms of age, gender,disease distribution, and activity. However, if applying the RCTinclusion and exclusion criteria to our study subjects, only 34%of CD patients and only 26% of UC patients would have beeneligible for study enrollment. Our findings mirror similar stud-ies of the external validity of clinical trials in other chronicdiseases such as rheumatoid arthritis, chronic obstructive pul-monary diseases, and congestive heart failure.21–23
There are other factors to consider when interpreting thefindings of these randomized controlled IBD trials. First is theheterogeneity of disease behavior and presentation. Many mod-erate–severe CD patients have more advanced disease; therefore,their CD activity might not reflect solely inflammatory diseasebut the stricturing and penetrating sequelae of longstandinginflammation. Approximately 40% of our CD study populationhad symptomatic strictures or radiologically confirmed ab-scesses, which was a factor for 60% of the patients who wouldnot have qualified for enrollment into any of the biologicaltrials. Although the trial data do not represent these patients,they are still defined as having moderate–severe disease. Be-cause these patients would have been excluded from enrollmentinto the majority of these pivotal RCTs, we cannot assumesimilar response rates among these patients with more complexmanifestations of CD. However, many of these patients weregiven biologics, which might not have been the optimal first-line treatment strategy because the potential for adverse reac-tions or nonresponse might be higher. Compared with patientswho were deemed trial-eligible, response rates were significantlylower among those patients who would not have qualified forthe RCTs. Although our data suggest that these trial-ineligiblepatients had a 60% response rate to biological therapy, response
lscfiemor
aafialwcaapwtcfb
i
lgpamti
1006 HA ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 10, No. 9
was not determined by objective markers but was based onglobal assessment after only a brief period of observation. Thelonger-term durability of biological response and safety amongtrial-ineligible patients remain to be determined.
Many of our study patients were on high-dose steroids forCD or rectal therapies for UC, which would have impacted theirability to participate in trials. Although these medication regi-mens could have been modified in the subsequent weeks toallow for study enrollment, particularly the topical-based ther-apies, most patients had more than one reason for being ex-cluded from these RCTs. However, these factors also suggestthese patients had higher disease activity and might have ben-efited more from aggressive therapy than patients on lowersteroid doses or without the rectal symptoms that prompttopical therapy usage. These fixed exclusion criteria might cre-ate a selection bias by eliminating patients with more complexor aggressive disease who might not respond as well to thestudied therapy. By excluding these patients from trial partici-pation, the potential effect of the intervention might be in-creased.
Clinical decision-making is largely based on observationalexperience that relies on more subjective, patient-directed de-terminants of response or from RCT data, which are intention-ally more stringent and regimented. The goal of an RCT is tooptimally study a single variable and outcome to yield a quan-titative assessment of therapeutic effect. By having specificinclusion and exclusion criteria, the study group has beenpredetermined to identify the patients who are most likely tohave a clinically and numerically significant treatment out-come.24 However, by having such strict study criteria, the chal-enge rests in the ability to translate these findings to a larger-cale group of people who might also be impacted byomorbidity, polypharmacy, and disease burden. Applying thesendings to a more at-risk population who would have beenxcluded from original trial participation might have compro-ised treatment efficacy, although by including a broader range
f patients seen in practice, the results would have more accu-ately measured effectiveness.
Many clinical trials report similar adverse event profilesmong patients receiving study drug vs placebo. Therefore, thessumption is that the treatment has a comparable safety pro-le to placebo. However, this could be because these patientsre preselected on the basis of the study-specific criteria to haveower adverse event risks by excluding patients who are older,ho have multiple comorbidities, or who are concurrently re-
eiving more aggressive immunosuppression. The potential fordverse events might be greater when these treatments arepplied to these “higher risk” patients who might be morerone to medication-related complications. As an example,hen the study results from the RALES (Randomized Aldac-
one Evaluation) trial were applied to a more at-risk group ofongestive heart failure patients who would have been excludedrom the original study, rates of hyperkalemia-associated mor-idity and mortality increased significantly.25 This issue is of
particular importance for IBD because these are immune-me-diated therapies and might potentially negatively impact asicker, more debilitated patient compared with the relativelyhealthier study patient.
The issue of external validity because of stringent studyselection criteria is a primary criticism of RCT design.26 Limit-
ng selection criteria would require larger study populations,
onger durations of follow-up observation, and substantiallyreater cost.24 Shifting from an explanatory RCT-based ap-roach to study treatment effect to a more pragmatic-basedpproach, which is aimed to measure effectiveness with mini-al study exclusions, might allow for more generalizability to
he office-based practice setting and corroborate RCT findingsn a more real-world setting.27
There are several inherent limitations to our study findings.First, this is a retrospective study of patients who were seen ina tertiary care referral center where patients with newer diag-noses of IBD and relatively naive to therapies are not as fre-quently encountered compared with community-based prac-tices. Therefore, the study population might have been biasedtoward more trial ineligibility because of disease complexity ormedical refractoriness. Another limitation is that this studyonly looked at trial eligibility as a single cross-sectional obser-vation in time. If patients were studied longitudinally with theinclusion and exclusion criteria applied at the time patientswere first being considered for immunomodulators or biologicsor at the time of diagnosis, a larger percentage of patients mighthave been eligible for trial participation. In a study of clinicaltrial eligibility for disease-modifying antirheumatic drugsamong rheumatoid arthritis patients, only 38% of patients inthe cross-sectional study met criteria for trial enrollment. How-ever, when followed longitudinally, 68% of patients would havebeen trial-eligible at least once during the follow-up period.28
This is also suggested by our subgroup analyses of the CDpatients because the percentage of trial-eligible patients in-creases when including only anti-TNF naive patients or patientswith inflammatory disease activity only.
Overall, the majority of patients seen in an outpatient IBDreferral practice would not have qualified for enrollment into amajor biologic RCT. Although the initial anti-TNF responserates among these trial-ineligible patients were somewhat sim-ilar to reported clinical trial response rates, the definitions ofresponse were largely subjective and not directly comparable topublished trial data. Additional longitudinal studies, as eitherobservational cohort studies or pragmatic trials of therapiesincluding both community and referral center patients, areneeded to support study findings and assess long-term thera-peutic response of trial-ineligible patients to truly determinethe external validity of these pivotal clinical trial findings. Prag-matic trials of the major IBD therapeutics would not only serveto validate RCT findings but also provide additional insightregarding medication safety across a broader patient popula-tion. Also, further exploration into study design is necessary toimprove the transparency of RCTs and justification of exclusioncriteria for these studies.
Supplementary MaterialNote: To access the supplementary material accompa-
nying this article, visit the online version of Clinical Gastroenter-ology and Hepatology at www.cghjournal.org, and at http://dx.doi.org/10.1016/j.cgh.2012.02.004.
References
1. Talley NJ, Abreu MT, Achkar JP, et al. An evidence-based system-atic review on medical therapies for inflammatory bowel disease.Am J Gastroenterol 2011;106(Suppl 1):S2–S25.
2. Ford AC, Sandborn WJ, Khan KJ, et al. Efficacy of biological
September 2012 EXTERNAL VALIDITY OF IBD CLINICAL TRIALS 1007
therapies in inflammatory bowel disease: systematic review andmeta-analysis. Am J Gastroenterol 2011;106:644–659.
3. Schnitzler F, Fidder H, Ferrante M, et al. Long-term outcome oftreatment with infliximab in 614 patients with Crohn’s disease:results from a single-centre cohort. Gut 2009;58:492–500.
4. Loftus EV, Feagan BG, Colombel JF, et al. Effects of adalimumabmaintenance therapy on health-related quality of life of patientswith Crohn’s disease: patient-reported outcomes of the CHARMtrial. Am J Gastroenterol 2008;103:3132–3141.
5. Feagan BG, Panaccione R, Sandborn WJ, et al. Effects of ada-limumab therapy on incidence of hospitalization and surgery inCrohn’s disease: results from the CHARM study. Gastroenterol-ogy 2008;135:1493–1499.
6. Lichtiger S, Binion DG, Wolf DC, et al. The CHOICE trial: ada-limumab demonstrates safety, fistula healing, improved quality oflife and increased work productivity in patients with Crohn’sdisease who failed prior infliximab therapy. Aliment PharmacolTher 2010;32:1228–1239.
7. Rutgeerts P, Sandborn WJ, Feagan BG, et al. Infliximab for induc-tion and maintenance therapy for ulcerative colitis. N Engl J Med2005;353:2462–2476.
8. Rutgeerts P, Diamond RH, Bala M, et al. Scheduled maintenancetreatment with infliximab is superior to episodic treatment for thehealing of mucosal ulceration associated with Crohn’s disease.Gastrointest Endosc 2006;63:433–442.
9. D’Haens G, Baert F, van Assche G, et al. Early combined immu-nosuppression or conventional management in patients withnewly diagnosed Crohn’s disease: an open randomised trial.Lancet 2008;371:660–667.
0. D’Haens G, Van Deventer S, Van Hogezand R, et al. Endoscopicand histological healing with infliximab anti-tumor necrosis factorantibodies in Crohn’s disease: a European multicenter trial. Gas-troenterology 1999;116:1029–1034.
1. Targan SR, Feagan BG, Fedorak RN, et al. Natalizumab for thetreatment of active Crohn’s disease: results of the ENCORE trial.Gastroenterology 2007;132:1672–1683.
2. Sandborn WJ, Colombel JF, Enns R, et al. Natalizumab inductionand maintenance therapy for Crohn’s disease. N Engl J Med2005;353:1912–1925.
3. Lichtenstein GR, Hanauer SB, Sandborn WJ, et al. Managementof Crohn’s disease in adults. Am J Gastroenterol 2009;104:465–483.
4. Kornbluth A, Sachar DB, Practice Parameters Committee of theAmerican College of Gastroenterology. Ulcerative colitis practiceguidelines in adults: American College of Gastroenterology, Prac-tice Parameters Committee. Am J Gastroenterol 2010;105:501–523.
5. Harvey RF, Bradshaw JM. A simple index of Crohn’s diseaseactivity. Lancet 1980;1:514.
6. Hanauer SB, Feagan BG, Lichtenstein GR, et al. Maintenanceinfliximab for Crohn’s disease: the ACCENT I randomised trial.Lancet 2002;349:1541–1549.
7. Hanauer SB, Sandborn WJ, Rutgeerts P, et al. Human anti-tumornecrosis factor monoclonal antibody (adalimumab) in Crohn’sdisease: the CLASSIC-I trial. Gastroenterology 2006;130:323–
333. T
8. Colombel JF, Sandborn WJ, Rutgeerts P, et al. Adalimumab formaintenance of clinical response and remission in patients withCrohn’s disease: the CHARM trial. Gastroenterology 2007;132:52–65.
9. Sandborn WJ, Feagan BG, Stoinov S, et al. Certolizumab pegol forthe treatment of Crohn’s disease. N Engl J Med 2007;357:228–238.
0. Colombel JF, Sandborn WJ, Reinisch W, et al. Infliximab, azathio-prine, or combination therapy for Crohn’s disease. N Engl J Med2010;362:1383–1395.
1. Sokka T, Pincus T. Eligibility of patients in routine care for majorclinical trials of anti-tumor necrosis factor alpha agents in rheu-matoid arthritis. Arthritis Rheum 2003;48:313–318.
2. Herland K, Akselsen JP, Skjønsberg OH, et al. How representa-tive are clinical study patients with asthma or COPD for a larger“real life” population of patients with obstructive lung disease?Respir Med 2005;99:11–19.
3. Masoudi FA, Havranek EP, Wolfe P, et al. Most hospitalized olderpersons do not meet the enrollment criteria for clinical trials inheart failure. Am Heart J 2003;146:250–257.
4. Van Spall HG, Toren A, Kiss A, et al. Eligibility criteria of random-ized controlled trials published in high-impact general medicaljournals: a systematic sampling review. JAMA 2007;297:1233–1240.
5. Juurlink DN, Mamdani MM, Lee DS, et al. Rates of hyperkalemiaafter publication of the Randomized Aldactone Evaluation Study.N Engl J Med 2004;351:543–551.
6. Rothwell PM. External validity of randomised controlled trials: “towhom do the results of this trial apply?” Lancet 2005;365:82–93.
7. Treweek S, Zwarenstein M. Making trials matter: pragmatic andexplanatory trials and the problem of applicability. Trials 2009;10:37.
8. Kingsley GH, Khoshaba B, Smith CM, et al. Are clinical trials inrheumatoid arthritis generalizable to routine practice? A re-eval-uation of trial entry criteria. Rheumatology (Oxford) 2005;44:629–632.
Reprint requestsAddress requests for reprints to: Christina Ha, MD, Division of Gastro-
enterology, The Johns Hopkins School of Medicine, 1830 East MonumentStreet, Suite 434, Baltimore, Maryland 21287. e-mail: [email protected].
onflicts of interestThese authors disclose the following: Asher Kornbluth has served on
he Advisory board for Igen/Biodec, Abbott, Elan, and Centocor; pro-ided research support for Abbott and Centocor; and has been on thepeaker’s Bureau for Abbott and Centocor. Corey A. Siegel has beenn the Consultant/Advisory board for Abbott, Elan, UCB; CME lectureror Abbott and Janssen; and received research funding from Abbott.homas Ullman has been a Consultant for Abbott, UCB, and Centocor.
