• Crohn's disease;
  • adalimumab;
  • remission;
  • extraintestinal manifestations


  1. Top of page
  2. Abstract
  6. Acknowledgements
  8. Supporting Information


Data regarding the effectiveness of anti-tumor necrosis factor (TNF) agents for resolution of extraintestinal manifestations (EIMs) are scarce. The CARE study evaluated clinical effectiveness, EIM resolution, and safety of adalimumab in a large pan-European cohort of patients with moderate to severe Crohn's disease (CD).


In all, 945 patients with a Harvey–Bradshaw Index (HBI) ≥7 enrolled in this multicenter, open-label phase IIIb trial. Patients received subcutaneous adalimumab, 160/80 mg at weeks 0/2, then 40 mg every other week. Dose adjustments were allowed for CD-related concomitant medications (from week 8) and adalimumab (from week 12). Clinical endpoints were analyzed through week 20 for all patients, and after stratification by prior infliximab exposure and by reason for discontinuing infliximab (primary nonresponse [PNR] or other).


The remission rate (HBI <5) at week 20 was 52% (95% confidence interval, 49%–55%) overall, and was higher for infliximab-naïve versus infliximab-exposed patients (62% versus 42%, P < 0.001). Remission rates were similar for PNR (37%) and other reasons (43%; P = 0.278). Of 497 patients with baseline EIMs, 51% were free of EIM signs and symptoms at week 20. Serious infectious adverse events were reported in 5% of patients. Opportunistic infections and malignancies were rare (≤1%). There was one case of demyelinating disease, but no occurrences of lupus, tuberculosis, or death.


In this large cohort of patients, adalimumab treatment resulted in rates of clinical remission and EIM resolution exceeding 50%, and substantial rates of effectiveness in patients who had PNR to infliximab. Adalimumab was well tolerated, with safety consistent with prior reports. (Inflamm Bowel Dis 2011)

Crohn's disease (CD) is a chronic inflammatory disease of the digestive tract, most frequently involving the ileum and colon. Hallmark symptoms of CD are chronic diarrhea and abdominal pain, and, in moderate to severe patients, weight loss and general malaise. Extraintestinal manifestations (EIMs) of CD are seen in 25%–40% of patients,1–3 and can be associated with significant morbidity, contributing to impaired quality of life in patients with CD.4 Large studies on the treatment of EIMs are lacking. Other complications of CD include fistulae, abscesses, and strictures, which may require surgical management.5

Treatment of CD aims to reduce inflammation, and includes aminosalicylates, corticosteroids, and immunosuppressants (e.g., thiopurines or methotrexate), and tumor necrosis factor (TNF) antagonists.5 Anti-TNF agents approved for CD in Europe include the chimeric monoclonal antibody infliximab and the human monoclonal antibody adalimumab.

Adalimumab has demonstrated efficacy in inducing6, 7 and maintaining8 response and remission in patients with moderate to severe CD in anti-TNF-naïve patients and in patients with prior loss of response or intolerance to infliximab. To date there have been no large-scale prospective clinical trials measuring the clinical effectiveness of an anti-TNF agent in a population including patients who had never exhibited an initial response to another anti-TNF agent.

Here we report the results of CARE (Crohn's Treatment with Adalimumab: Patient Response to a Safety and Efficacy Study), a phase IIIb, open-label clinical trial conducted in Europe. CARE evaluated clinical effectiveness, EIM resolution, fistula healing, and safety in a large clinical cohort of patients with moderate to severe CD, including those who had never responded to infliximab.


  1. Top of page
  2. Abstract
  6. Acknowledgements
  8. Supporting Information

CARE was a 20-week, multicenter (187 sites in 17 European countries) open-label trial (, NCT00409617), conducted from December, 2006 to July, 2008.


CARE included adults 18 to 75 years of age with a radiologic or endoscopic diagnosis of CD for more than 4 months. Patients had to have a Harvey–Bradshaw Index9 (HBI) of ≥7 points at screening. Female patients of childbearing potential were required to use an adequate method of contraception. Exclusion criteria are described in the Supporting Material.

Study Procedures

At the baseline visit (week 0), all patients received a loading dose of adalimumab (160 mg) subcutaneously, followed by 80 mg at week 2. Starting at week 4, all subjects began receiving adalimumab 40 mg every other week. At or after week 12, patients who experienced a disease flare (defined by an HBI increase of ≥3 points compared with week 4 and a total HBI score ≥7) or who exhibited nonresponse to study medication (defined as a decrease in the HBI <3 compared with baseline) were permitted to increase the adalimumab dose to 40 mg weekly. Patients with continued lack of improvement could be withdrawn from the study. Patients residing in a country where the marketing application for adalimumab in CD had not been approved at week 20 were permitted to continue on study therapy until marketing approval, with the exception of Norway and Denmark, where patients were mandated to complete the study at week 20.

Concomitant Medications

Patients continued baseline stable doses of immunosuppressants (azathioprine, 6-mercaptopurine, or methotrexate), aminosalicylates, and CD-related antibiotics. Use of corticosteroids (≤40 mg of prednisone or equivalent or ≤9 mg of budesonide) was permitted, but concomitant use of both budesonide and a systemic corticosteroid was not permitted. Starting at week 8, modifications in CD-related concomitant treatments were allowed.

Effectiveness Assessments

Patients were assessed at weeks 0, 2, 4, 8, 12, and 20, and every 12 weeks thereafter. At each visit, physical examination and laboratory analyses were performed and HBI was recorded. The presence of EIMs was evaluated by the investigator at baseline and at each visit through week 20. EIMs assessed were: oral aphthous ulcers, erythema nodosum, pyoderma gangrenosum, iritis, uveitis, arthritis, and arthralgia. Sacroiliitis, ankylosing spondylitis, CD-related hepatic disease, thrombosis, and nephrolithiasis were also assessed, although a simple clinical evaluation is insufficient for these and the results concerning these manifestations should be interpreted with caution. A count of the number of draining (upon gentle compression) cutaneous fistulae was collected during each visit. Fistula healing was defined as the absence of drainage from all fistulae.

Analytical Plan

Qualitative baseline characteristics were compared between subgroups using chi-square or Fisher's exact test while quantitative variables were compared with the unpaired t-test. The percentages of patients with clinical remission (HBI <5) and clinical response (HBI reduction ≥3) were determined at weeks 4 and 20 and compared using nonresponder imputation (NRI), with 95% confidence intervals (CI). Rates of fistula healing were calculated for the patients with fistulae at baseline. Patients were stratified by previous infliximab exposure (yes/no) and by reason for discontinuation of infliximab (primary nonresponse [PNR], or other: loss of response in initial responders, intolerance or adverse events [AEs], or missing reason, as determined by the investigator). Response and remission were compared using the chi-square test, and fistula healing was compared using Fisher's exact test. Clinical response and remission were assessed by duration of CD, and remission was assessed by baseline corticosteroid or immunosuppressant use. The median HBI score at each visit was determined using as-observed analysis. Percentages of clinical response and remission in patients who increased to weekly dosing during the study were determined at week 20 using NRI.

The presence of EIMs at baseline and at week 20 was calculated in the entire study population as well as in the subset of patients with an EIM at baseline using the last observation carried forward (LOCF) for patients with missing values. Presence of specific EIMs was compared using the sign test. Rates of EIM resolution in the subset with an EIM present at baseline were compared in the following subgroups: HBI remission status at week 20 and prior infliximab exposure, using the chi-square test. Steroid-free remission rates at weeks 12 and 20 were calculated for the patients on steroids at baseline (not calculated at week 4 because concomitant medications could not be changed until week 8).

Crude odds ratios (OR) with 95% CI and P values were calculated for the following potential predictors: sex, age, current nicotine use, baseline C-reactive protein (CRP) (<0.5 mg/dL versus ≥0.5 mg/dL), duration of CD, history of CD-related surgery, extent of CD involvement (≥2 segments versus 1), prior infliximab use, steroid or immunosuppressant use at baseline, EIMs or draining fistulae at baseline, and baseline HBI. Logistic regression modeling with backward elimination (based on a selection level of 0.05) was performed to identify variables associated with remission (using as-observed values) at weeks 4 and 20. OR with 95% CI and P value are presented.

Safety was evaluated through 70 days after the last adalimumab injection for all patients who received at least one dose of adalimumab. Adverse events were summarized for all patients for the duration of their participation in the trial and reported as overall incidence. Additionally, AEs of interest with anti-TNF agents (serious infections, malignancies, opportunistic infections, lupus, tuberculosis, and demyelinating disease) were recorded.

Ethical Considerations

The Institutional Review Board or independent Ethics Committee for each site approved the protocol and all patients provided written informed consent.


  1. Top of page
  2. Abstract
  6. Acknowledgements
  8. Supporting Information


Of 945 patients enrolled, 785 (83%) completed 20 weeks. Adverse events and lack of efficacy were the most common reasons for study discontinuation (Supporting Fig. 1). A total of 131 patients (14%) changed to weekly dosing prior to week 20; 112 of these patients completed 20 weeks. Patients with previous exposure to infliximab were more likely to be female (P = 0.005) and to have anal/perianal disease (P < 0.001), fistulae (P < 0.001), higher CRP (P = 0.075), higher baseline HBI (P < 0.001), longer disease duration (P < 0.001), and a history of CD-related surgery (P < 0.001). They were less likely to be on steroids (P = 0.035) or aminosalicylates (P = 0.011) at baseline (Table 1).

Table 1. Baseline Characteristics
 All PatientsNo Prior IFXPrior IFX
  • *

    P < 0.001;

  • P < 0.05, comparing IFX exposure groups.

  • a

    Multiple locations possible.

  • b

    Anal/perianal vs. nonanal/perianal.

  • c

    Includes balsalazide, mesalamine, olsalazine, sulfasalazine.

  • d

    Includes budesonide (n=117; recorded median daily dose 9 mg, range 3-12 mg) and other systemic steroids (beclomethasone, cortisone, deflazacort, hydrocortisone, methylprednisolone, prednisone/prednisolone; n=283 recorded median daily dose in prednisolone equivalents 15 mg; range 0.5-80 mg). Two patients were recorded as taking both budesonide and another steroid at baseline.

  • e

    Includes ciprofloxacin, cotrimazole, metronidazole, miconazole, and ofloxacin.

  • f

    Loss of response, intolerance/AE, other/missing.

Female, n (%)568 (60.1)266 (55.6)302 (64.7)
White, n (%)926 (98.0)468 (97.9)458 (98.1)
Body wt (kg), mean (SD)67.9 (16.3)68.8 (16.8)67.1 (15.7)
Involved intestinal area,an (%)   
 Colonic704 (74.5)343 (71.8)361 (77.3)
 Ileal661 (69.9)341 (71.3)320 (68.5)
 Anal/perianal228 (24.1)87 (18.2)141 (30.2)*b
 Rectum190 (20.1)90 (18.8)100 (21.4)
 Jejunum47 (5.0)28 (5.9)19 (4.1)
 Gastroduodenal37 (3.9)21 (4.4)16 (3.4)
 Other34 (3.6)19 (4.0)15 (3.2)
Draining fistulae at baseline, n (%)171 (18.1)63 (13.2)108 (23.1)*
HBI, median (range)10.0 (3-47)9.0 (3-27)10.0 (7-47)*
Current smoker, n (%)399 (42.3)202 (42.3)197 (42.3)
Duration of CD (y), median (range)7.8 (0.4-45.4)6.3 (0.4-45.4)9.1 (0.5-40.5)*
High sensitivity CRP (mg/dL), median (range)0.9 (0.0-98.5)0.8 (0.0-33.8)1.1 (0.0-98.5)
CRP ≥ 0.5 mg/dL, n (%)575 (62.6)275 (59.5)300 (65.8)
Previous CD-related surgery518 (54.8)222 (46.4)296 (63.4)*
Concomitant medication use at baseline, n (%)   
 Aminosalicylatesc170 (18.0)101 (21.1)69 (14.8)
 Immunosuppressants517 (54.7)262 (54.8)255 (54.6)
 Any corticosteroidd401 (42.4)219 (45.8)182 (39.0)
 Corticosteroids and/or immunosuppressants706 (74.7)363 (75.9)343 (73.4)
 CD-related antibioticse47 (5.0)23 (4.8)24 (5.1)
Reason for infliximab discontinuation, n (%)   
 PNR  89 (19.1)
 Otherf  378 (80.9)

Pending country marketing authorization for adalimumab, 476 patients remained in the study at week 32 and 297 at week 44. The median duration of adalimumab treatment during the trial was 185 days (range 15–544 days).

Clinical Effectiveness

The percentage of patients in remission for the all-adalimumab group at week 4 was 43% (95% CI: 40%–46%) and increased to 52% (95% CI: 49%–55%) at week 20 (Fig. 1A). The percentage of patients in remission at weeks 4 and 20 was significantly lower in the infliximab-exposed subgroup than in the infliximab-naïve subgroup. The percentages of patients in remission at weeks 4 and 20 were similar in the infliximab-exposed subgroup stratified by primary reason for infliximab discontinuation (Fig. 1B). In all groups, rates of remission were slightly higher at week 20 than at week 4. A substantial decline in observed HBI was seen as early as week 2 and was sustained through week 20 (Fig. 1C).

thumbnail image

Figure 1. (A) Percentage of patients in remission, overall population and by prior infliximab use. (B) Percentage of patients previously treated by infliximab in remission, by primary reason for infliximab discontinuation. (C) HBI (median and interquartile range) at each visit. At Week 20, 13 patients who discontinued the study provided HBI values. (D) Percentage of patients in remission by disease duration.

Download figure to PowerPoint

Shorter disease duration (<2 years and 2 to ≥5 years) was associated with higher rates of clinical remission at week 4 compared with disease duration >5 years (50%, 52%, and 38%, respectively, P < 0.001; Fig. 1D), although by week 20 remission rates were not markedly different between groups (58%, 56%, and 50%, respectively; P = 0.136). High (≈75%) rates of clinical response were achieved in all subgroups by week 4 and were sustained through week 20 (data not shown).

Rates of clinical response were generally sustained from week 4 to week 20 in the overall group and in the subgroups (Fig. 2A). The percentage of patients achieving clinical response by previous infliximab use was similar at week 4 and slightly lower in the exposed cohort at week 20. There were no meaningful differences at weeks 4 and 20 in rates of clinical response stratified by reason for infliximab discontinuation (Fig. 2B).

thumbnail image

Figure 2. (A) Percentage of patients with clinical response, overall population and by prior infliximab use. (B) Percentage of patients previously treated with infliximab with clinical response, by primary reason for infliximab discontinuation.

Download figure to PowerPoint

EIM Resolution

Incidences of EIMs at baseline and week 20 in the entire CARE population (N = 945) are shown in Table 2. Overall, 53% (497/945) of patients had at least one EIM present at baseline, compared with 30% at week 20. Of these 497 patients, at week 20, 79% had resolution of at least one EIM and 51% were free of EIM signs and symptoms. The rates of resolution of at least one EIM were higher in patients in HBI remission at week 20 (88% for remitters versus 71% for nonremitters, P < 0.001). The EIM resolution rates were similar regardless of prior infliximab use (76% for infliximab-exposed, 82% for infliximab-naïve, P = 0.100), and were similar in patients with PNR to infliximab (73%) and those who discontinued for other reasons (77%, P = 0.625).

Table 2. Incidence of EIMs at Baseline and Week 20 in the Overall CARE Population
 Baseline (N=945)Week 20 (N=942a)
EIMn (%)n (%)
  • a

    Percentages calculated on non-missing values; 3 patients had no EIM assessments after baseline so these patients' data were not carried forward; no patients had thrombosis at baseline or week 20.

  • b

    As determined by investigator.

  • *

    P < 0.001 vs. baseline.

  • P = 0.016 vs. baseline.

  • Incidence too small for comparison.

Arthralgia445 (47.1)252 (26.8)*
Arthritis82 (8.7)20 (2.1)*
Oral aphthous ulcers49 (5.2)20 (2.1)*
Sacroiliitis34 (3.6)18 (1.9)
Erythema nodosum23 (2.4)4 (0.4)*
Ankylosing spondylitis16 (1.7)15 (1.6)
Nephrolithiasis8 (0.8)8 (0.8)
Iritis7 (0.7)2 (0.2)
Pyoderma gangrenosum4 (0.4)2 (0.2)
Uveitis3 (0.3)3 (0.3)
CD-related hepatic diseaseb1 (0.1)1 (0.1)

Fistula Healing

Of the 171 patients with fistulae present at baseline, 35% had multiple fistulae (20% had two, 6% had three, and 9% had ≥ four). Complete fistula closure was observed in 19% of patients at week 4 and 26% at week 20. Fistula closure rates were numerically higher in the infliximab-naïve group at weeks 4 and 20 (22% and 33%, respectively) than in the infliximab-experienced group (17% and 22%, respectively; P = 0.369 at week 4, and P = 0.275 at week 20). Similar rates of closure were seen in the PNR (15% at week 4, 19% at week 20) and the other reason for infliximab discontinuation groups (17% at week 4, 23% at week 20; P = 0.840 at week 4, P = 0.973 at week 20).

Changes in Corticosteroids or Immunosuppressants

Of the 401 patients taking corticosteroids at baseline, 37% discontinued them through week 20; 44 (11%) and 56 (14%) achieved steroid-free remission at weeks 12 and 20, respectively. Immunosuppressants were discontinued for 7% of the 517 patients taking them at baseline. From week 8 to week 20, of the patients not using them at baseline, 7% of 544 patients initiated corticosteroids and <1% of 428 patients started immunosuppressants.

Effect of Concomitant Corticosteroids and Immunosuppressants on Remission

Patients on steroids at baseline had similar rates of clinical remission as those not taking steroids at week 4 (45% versus 41%, P = 0.314) and week 20 (52% in both groups; P = 0.976). At week 4, patients on immunosuppressants at baseline had a slightly higher rate of remission than those not taking immunosuppressants (46% versus 39%, respectively; P = 0.025). By week 20 the rates of clinical remission were 55% and 49%, respectively (P = 0.052).

Effect of Dose Escalation

Of the patients who received weekly dosing before week 20, 46 of 131 (35%) were in clinical remission at week 20 (using NRI). The corresponding value for clinical response was 58%. The median time to dose escalation was 92 days (range: 70–137 days).

Predictors of Remission at Weeks 4 and 20

Crude ORs for predictors of remission are shown in Supporting Table 1. Results of logistic regression with backward elimination are shown in Table 3.

Table 3. Predictors of Remission at Weeks 4 and 20, Observed Values (Logistic Regression with Backward Elimination)
 Odds Ratio95% CIP-value
Week 4   
 Baseline CRP < 0.5 mg/dL1.461.09–1.970.013
 Duration of CD (y)0.960.94–0.98<0.001
 History of CD-related surgery0.590.44–0.79<0.001
 Prior infliximab use0.680.51–0.910.009
 Higher baseline HBI0.820.78–0.87<0.001
Week 20   
 Age (y)0.970.96–0.99<0.001
 History of CD-related surgery0.540.39–0.74<0.001
 Prior infliximab use0.490.36–0.67<0.001
 Higher baseline HBI0.870.83–0.91<0.001


The 945 patients who received at least 1 injection of adalimumab in CARE represented 564.1 patient-years of adalimumab exposure. Adverse events were reported in 80% of patients, and 11% discontinued the study as a result of an AE (Table 4). Serious events were reported by 19% of patients. The overall AE profile was similar in subgroups stratified by previous exposure to infliximab. There was one case of demyelination in a 36-year-old male with a history of previous azathioprine and infliximab use, which presented as blurred vision on study day 42. The event led to treatment discontinuation on day 89 and was ongoing as of day 166. There were no cases of tuberculosis, congestive heart failure, or lupus or lupus-like reactions, and no deaths. A listing of serious adverse events (SAEs) occurring in ≥3 patients is reported in the Supporting Material. The most common SAE category was “gastrointestinal disorders,” occurring in 93 patients (10%), and the most commonly reported individual SAE was “Crohn's disease,” indicating a worsening of the patient's underlying disease, occurring in 54 patients (6%).

Table 4. Adverse Events of Interest: All Patients with at Least 1 Injection of Adalimumab
 All PatientsWithout Prior IFXWith Prior IFX
 N = 945N = 478N = 467
  • All data presented as n (%).

  • a

    Serious infections: abscess (2), anal abscess (2), Clostridial infection (2), diverticulitis (2; 1 of these patients also had abdominal abscess), gastroenteritis (2; 1 of these patients also had abscess intestinal), gastroenteritis viral (2; 1 of these patients also had influenza), rectal abscess (2), abdominal wall abscess (1), abscess intestinal (1; patient also had gastroenteritis), anogenital warts (1), bacterial infection (1), bronchiectasis (1), bronchitis (1), bronchopneumonia (1), Campylobacter intestinal infection (1), Clostridrium difficile colitis (1), enteritis infectious (1; patient also had vaginal infection), enterocolitis bacterial (1; patient also had Klebsiella sepsis), Epstein-Barr virus infection (1), folliculitis (l), infection (1), influenza (1; patient also had gastroenteritis viral), Klebsiella sepsis (1; patient also had enterocolitis bacterial), esophageal candidiasis (1), pelvic abscess (1), perineal abscess (1), perirectal abscess (1), pharyngitis (1), pneumonia (1), pseudomembranous colitis (1), pyelonephritis (1), sepsis (1), urinary tract infection (1), vaginal infection (1; patient also had enteritis infectious), viral infection (1).

  • b

    Malignancies: follicular thyroid cancer (1), oral neoplasm (1), small intestinal carcinoma (1).

  • c

    Nonmelanoma skin cancers were basal cell carcinoma (3 events in 2 patients).

  • d

    Opportunistic infections: oral candidiasis (5), candidiasis (3), serious esophageal candidiasis (1), and pharyngeal candidiasis (1).

Any AE754 (79.8)379 (79.3)375 (80.3)
 At least possibly related to study drug447 (47.3)221 (46.2)226 (48.4)
 Leading to discontinuation of study drug100 (10.6)44 (9.2)56 (12.0)
 Serious AE181 (19.2)85 (17.8)96 (20.6)
 Infection432 (45.7)213 (44.6)219 (46.9)
  Serious infectiona50 (5.3)23 (4.8)27 (5.8)
 Malignancy (excluding lymphoma and NMSC)b3 (0.3)1 (0.2)2 (0.4)
 NMSCc2 (0.2)2 (0.4)0
 Injection-site pain38 (4.0)17 (3.6)21 (4.5)
 Opportunistic infectiond10 (1.1)4 (0.8)6 (1.3)
 Demyelinating disease1 (0.1)01 (0.2)


  1. Top of page
  2. Abstract
  6. Acknowledgements
  8. Supporting Information

Anti-TNF agents have demonstrated efficacy in induction6, 7, 10, 11 and maintenance8, 12–14 of remission in controlled, blinded clinical trials, some of which included patients who lost response to or could not tolerate previous anti-TNF agents. Published data on the effect of any anti-TNF agent on EIM resolution or clinical effectiveness in patients who did not achieve an initial response to another anti-TNF agent are limited. The CARE study is the largest clinical trial ever conducted in CD. It was designed to more closely reflect clinical practice than registration trials in order to gather data in a large cohort of patients in many countries.

The overall proportions of patients achieving clinical response and remission in CARE were higher than those seen in the CHARM pivotal study,7 with a majority of patients achieving clinical response and over 40% achieving clinical remission by week 4. These values were sustained, with over 50% of patients in clinical remission at week 20. As in CHARM,7 patients with previous infliximab exposure in CARE had somewhat lower, but still clinically meaningful, rates of response and remission compared with infliximab-naïve patients. Additionally, CARE patients who had PNR to infliximab had similar rates of response and remission as patients who had discontinued infliximab for other reasons. The mechanism for PNR to infliximab and subsequent response to adalimumab is unclear; differences in clinical efficacy between anti-TNF agents may be due to differences in pharmacokinetic properties or tissue penetration.15 Classification of study subjects as “primary nonresponders” to infliximab was determined by the investigator and not verified objectively in CARE. It is thus possible that some of these patients did not have true PNR but may have received an inadequate regimen or trial of infliximab. However, information regarding induction dosing and duration of trial of infliximab for patients classified as having PNR was not collected.

As was seen in a subanalysis of CHARM,16 the CARE study found a numerically higher rate of clinical remission in patients with shorter duration of CD (<2 years versus >5 years), which suggests that there is value in early intervention with biologic therapy for selected patients, especially those who have a poor long-term prognosis. Response and remission also occurred later in patients with longer disease duration, suggesting that clinicians may need to wait longer before declaring treatment failure in these patients.

The published efficacy of anti-TNF agents on EIM resolution has to date been mostly limited to case series and small, open-label trials.4, 17–20 A small (N = 30) randomized controlled trial comparing infliximab to placebo for the treatment of pyoderma gangrenosum found a higher rate of clinical improvement at week 2 in the infliximab recipients (46% versus 6%, P = 0.025),21 although not all of the patients in this series had inflammatory bowel disease. A recent publication of an observational analysis of adalimumab use for CD in a referral center22 found a meaningful rate of complete resolution of EIMs at week 4 (31.5%) and week 12 (41%). The EIM response was associated with response of luminal CD at both times (OR at week 4 12.5, 95% CI 3.7–42.3; P < 0.0001). In CARE, more than half of patients with an EIM at baseline were completely free of EIM signs and symptoms at week 20. Since the presence of EIMs is a component of the HBI, a higher rate of EIM resolution in patients in HBI remission would be expected; however, a substantial number of patients in CARE who did not achieve clinical remission still had meaningful rates of EIM resolution. Some EIMs (erythema nodosum, aphthous ulcers, peripheral arthritis) are reported to be associated with underlying intestinal disease activity, whereas others (pyoderma gangrenosum, ankylosing spondylitis, and uveitis) may have an independent course.23 This pattern was seen in CARE, where the incidence of arthralgias, arthritis, oral aphthous ulcers, and erythema nodosum were all reduced with adalimumab treatment. The clinical manifestations of sacroiliitis were also reduced, whereas those related to ankylosing spondylitis were not; the remaining EIMs occurred infrequently but generally did not have meaningful reductions. As EIMs were evaluated only by physical exam and patient interview, the true incidence of those who are not diagnosed or evaluated on clinical grounds only (i.e., sacroiliitis, ankylosing spondylitis, thrombosis, hepatic disease, nephrolithiasis) may not be accurately reflected, and results for these should be interpreted with caution.

The rates of complete fistula closure and steroid-free remission reported in CARE were somewhat lower than those reported in the pivotal phase III CHARM trial.8 The lower rates in CARE may be explained by differences in treatment duration and in steroid-tapering schedule.

The rates of clinical remission were slightly higher in patients taking concomitant immunosuppressants at baseline at both week 4 and week 20; at week 20 the proportions of patients in clinical remission by baseline immunosuppressant use were not significantly different. The lack of a clear and consistent difference in remission in patients with or without concomitant immunosuppressants is similar to the data seen in CHARM,7 as well as in a recent observational analysis of adalimumab use in a referral center.22 The latter analysis found that concomitant immunosuppressant therapy did not affect treatment outcome, influence adalimumab trough serum concentrations, or decrease the development of antibodies against adalimumab, although the median time to dose escalation in this cohort of patients (all of whom had previously failed to respond to infliximab) was longer in the combination therapy group (17.0 weeks versus 12.0 weeks; P = 0.008). Neither CARE nor CHARM was designed to establish the effects of combined immunosuppression. The optimal treatment strategy for patients on immunosuppressants who are started on an anti-TNF agent added remains to be seen.

In CARE, a history of CD-related surgery, prior infliximab usage, and higher baseline HBI were associated with lower rates of HBI remission at weeks 4 and 20. Longer disease duration was associated with a lower rate of clinical remission only at week 4. Baseline CRP <0.5 mg/dL was associated with a higher rate of remission only at week 4. Higher age was associated with a lower rate of clinical remission only at week 20. The lower rate of remission seen with higher age and/or history of CD-related surgery may reflect a progression of CD to a more fibrotic phenotype over time24 or reflect a disease with a greater propensity to develop transmural lesions requiring surgery, which may be less responsive to treatment with an anti-inflammatory agent such as an anti-TNF. It is also possible that the pattern of immune response in CD changes over time, as evidenced by the demonstration of lower T-helper (Th) type 1 and exacerbated Th17 pathway activation in later CD25, 26; this may influence the ability of antiinflammatory agents to affect disease activity. The association of a lower baseline CRP with week 4 remission is unexpected, as the rates of remission after 160/80 mg dosing in the phase III CLASSIC I induction trial tended to be higher in the subgroup with higher baseline CRP.5 This may reflect a greater early response in patients with lower inflammatory disease activity in the present open-label uncontrolled trial than in a randomized, blinded trial, and could potentially represent a higher early “placebo” response in these patients.

The safety profile seen in CARE is similar to that seen in CHARM,8 with no new safety signals detected. The 19% rate of SAEs may be due to the nature of the moderate to severe CD patient population, as reflected by the high rate of AEs in the “gastrointestinal disorder” category, and the fact that the most commonly reported SAE was CD. The most common serious infections were abscesses, as was also seen in a recent comprehensive safety review of the entire adalimumab CD clinical development program.27 There was one case of demyelinating disease in CARE. The calculated event rate (in events/100 patient-years [PY] of exposure) for this type of AE in CARE is 0.2 E/100 PY, which is consistent with the rate (0.2 E/100 PY) reported in the aforementioned pooled clinical trial safety analysis of adalimumab.27

Limitations of this study include the open-label and uncontrolled study design, relatively short duration of follow-up, ability to adjust concomitant medications after week 8 (which may have influenced the effectiveness evaluations at later visits), and lack of validation of reasons for infliximab discontinuation. In addition, the analytical plan specified stratification by previous infliximab use, not previous anti-TNF use, reflecting the lack of commercial availability of certolizumab pegol in Europe. Three patients in the infliximab-naïve cohort had received an anti-TNF agent (either etanercept or certolizumab pegol) before enrolling in CARE. However, this small number is unlikely to have an effect on the results.

In summary, the CARE trial provides data supporting clinical, EIM, and fistula effectiveness and safety in patients with moderate to severe CD receiving adalimumab in a large-scale cohort treated over a broad range of countries.


  1. Top of page
  2. Abstract
  6. Acknowledgements
  8. Supporting Information

The authors thank Pamala Mitchell, Pharm.D., of Abbott, for study and data management; Helmut Latscha, of Abbott GmbH & Co KG, for data management support; Monika Engemann, of Abbott GmbH & Co KG, for programming support; and Laurinda Cooker, PhD, of Abbott, for medical writing services in the development and revision of the article.


  1. Top of page
  2. Abstract
  6. Acknowledgements
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  6. Acknowledgements
  8. Supporting Information

Additional supporting information may be found in the online version of this article.

IBD_21663_sm_SuppFig1.tif51KSupporting Information Figure 1. Patient disposition
IBD_21663_sm_SuppTable1.doc68KSupporting Information Table 1.
IBD_21663_sm_SuppMaterials.doc30KSupporting Information Materials.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.