Infliximab and adalimumab are highly effective in Crohn's Disease (CD). This is supported by clinical trials and open-label studies using either infliximab or adalimumab, thus not allowing a proper comparison between these anti-TNFs in CD.
To evaluate the efficacy and safety of infliximab and adalimumab in active CD.
In a longitudinal study, CD patients with indication for anti-TNFs were treated with infliximab or adalimumab.
Ninety-three patients were treated with infliximab (n = 44) or adalimumab (n = 49). In the infliximab group, the induction was completed by 77.3% of patients, due to no response (n = 2), delayed hypersensitivity reactions (DHR) or infusion reactions (n = 8). Maintenance with infliximab was completed by 60% of patients, due to clinical worsening or loss of efficacy (n = 5), DHR or infusion reactions (n = 5). In the adalimumab group, all patients completed the induction, while maintenance was completed by 67% of patients, due to clinical worsening or loss of efficacy (n = 8), DHR (n = 1), other causes (n = 7). In both groups, the CDAI significantly reduced at baseline vs. each visit (P < 0.04). The Kaplan–Meier survival analysis performed to evaluate the risk of steroid-free remission in patients treated with infliximab vs. adalimumab detected no differences (log-rank test P = 0.4). Cox proportional-hazards regression identified two predictors of steroid-free remission using anti-TNFs: no smokers [HR = 2.94 (1.52–5.70), P = 0.001] and non stricturing non penetrating behaviour [HR = 3.116 (1.06–9.13), P = 0.03826].
Infliximab and adalimumab showed a similar efficacy. No smoking and non-stricturing non-penetrating behaviour were predictors of steroid-free remission.
Crohn's disease (CD) is a chronic IBD of unknown aetiology. A dysregulation of the mucosal immune response appears to be involved in the pathogenesis of tissue damage. Experimental evidences concordantly support a key role of tumour necrosis factor alpha (TNF-α) in the induction and perpetuation of the inflammatory process in CD tissue.[2-4] On the basis of these observations, a chimerical monoclonal antibody specifically targeting TNF-α has been developed, showing a dramatic efficacy in CD. Since then, infliximab has been widely used in CD, including patients refractory to conventional treatments.[6-8] Mucosal healing has also been reported in subgroups of responsive patients. Limits of infliximab are mainly represented by possible incomplete response, loss of response and/or side effects. These unfavourable outcomes may be related to the murine component of the drug and to the possible development of antibodies against infliximab, which may be associated with infusion reactions (6.9–19%). Due to the proven efficacy of this anti-TNF, other molecules providing the same immunomodulatory effect have been developed. Adalimumab, a fully humanised anti-TNF monoclonal antibody showed a proven efficacy for inducting and maintaining remission in CD.[12-14] As for infliximab, mucosal healing has been described in preliminary reports after 1 year treatment with adalimumab. Adalimumab therefore appears to represent an additional effective anti-TNF treatment in CD. Switching anti-TNF treatments is also indicated in patients showing loss of response or side effects when using one of the two anti-TNFs. Adalimumab may be effective and well tolerated also in CD patients refractory or intolerant to infliximab.[16, 17] The humanised adalimumab also induces a lower frequency of side effects related to the murine component of infliximab, although no comparative studies using these anti-TNFs are available. All these observations indeed derive from comparisons between clinical trials including different study populations and study designs, thus not allowing a proper comparison between these highly effective drugs in CD. To our knowledge, no observations reported in the same study the safety and efficacy of infliximab and adalimumab in patients with CD. Primary end point of this study was to investigate, in a consecutive series of CD patients, the efficacy of infliximab and adalimumab, as evaluated by the assessment of clinical activity, steroids-free remission and quality of life. Secondary end point included the assessment of the safety profile of infliximab and adalimumab.
Materials and methods
From January 2007 to December 2010 all CD patients under regular follow-up in our tertiary IBD centre, with indication for anti-TNF treatment (infliximab or adalimumab), were consecutively enrolled. Inclusion criteria were: (i) Diagnosis of CD made ≥12 months before enrolment (to exclude differential diagnoses), as assessed by conventional criteria; (ii) Compliance and feasibility to complete both the anti-TNF treatment and the scheduled visits up to 76 weeks; (iii) Age ≥18 years and (iv) Indications for anti-TNFs, according to current guidelines.[10, 19]
In particular, indications included moderate to severe CD defined only by a CDAI ≥ 220, active perianal disease, steroid-dependent or refractory disease (defined as patients with active CD not responsive to conventional treatments, including steroids, thiopurins and/or methotrexate). Exclusion criteria included conditions contraindicating the use of anti-TNFs, according to current guidelines.
Eligible patients were screened for latent infections (i.e. tuberculosis by using quantiferon and chest-X-ray, HBV, HCV). Patients were also tested for EBV and CMV infections.[10, 19] The presence of perianal or abdominal abscesses was excluded according to conventional clinical and radiological criteria, when needed.[10, 19] Demographics, prior surgery, smoking habits, previous exposure to anti-TNFs, CD location, phenotype and duration, assessed according to the Montreal classification were recorded. At baseline and at each visit, clinical assessment was made according to the CDAI. The clinical outcome of patients with indication for anti-TNFs was classified in four categories: (i) response (ΔCDAI ≥70 points); (ii) remission (CDAI < 150); (iii) capability to wean off steroids and (iv) loss of response (patients with an initial response to anti-TNFs showing no improvement or worsening ≥2 weeks after the last anti-TNF treatment).
Perianal disease was also assessed at each visit in relation to the presence of absence of local discharge after gentle compression, according to previous studies. The Inflammatory Bowel Disease Questionnaire (IBDQ) was completed by a subgroup of compliant patients before and after the induction and maintenance treatment with infliximab or adalimumab.
Anti-TNF treatment was chosen on individual basis according to a possible history of adverse events to either infliximab or adalimumab and to patients' compliance to receive either an i.v. infusion requiring a Day Hospital recovery or a self-administered s.c. injection.
Scheduled infliximab treatment was performed according to standard protocols, including i.v. infusion (5 mg/kg) at week 0 (week 0), week 2 (week 2) and week 6 (week 6) for induction and subsequently every 8 weeks up to 54 weeks (week 54) for maintenance treatment. Before each infusion, methyl-prednisolone (40 mg i.v.) was used as premedication.
Scheduled adalimumab treatment was also performed according to standard protocols, including s.c. injection every 2 weeks, at a dose of 160 mg (week 0), 80 mg (week 2) and 40 mg (week 4) for induction and subsequently 40 mg every other week up to 54 week (week 54) for maintenance.
Patients treated with either infliximab or adalimumab underwent clinical assessment at baseline (week 0), at the end of the induction (week 6 for infliximab and week 4 for adalimumab) and subsequently at the first maintenance treatment (week 12 for infliximab and week 8 for adalimumab), at 6 months (week 24) and at 12 months (week 54). The study protocol also included clinical assessment 6 months (week 76) after the end of the maintenance treatment. Concomitant medications at enrolment were continued during the study follow-up, while steroids could be tapered or discontinued according to clinical assessment and response to anti-TNFs.
The safety of infliximab and adalimumab was assessed by reporting any adverse event (AE) occurring either during anti-TNF administration or at any time during the study period. With this purpose, patients were instructed to contact the referral centre if any illness or AE.
Continuous variables were expressed as median and range, while categorical variables were expressed as absolute value and percentage both in the text and tables. Continuous variables were expressed as median with interquartile range in all figures.
Differences in terms of characteristics between patients treated with infliximab or adalimumab were assessed by the t-test and Wilcoxon-test (continuous variables), and by the χ2 test and Fisher test (categorical variables). The statistical analysis was first carried out to assess the efficacy of infliximab and adalimumab separately. With this purpose, the analysis was performed according to the per-protocol principle, describing the outcome of patients who adhered to the study protocol. In the per-protocol analysis, the CDAI and IBDQ values were analysed using the Kruskal–Wallis One-Way anova by Ranks and the pair wise comparison with a Wilcoxon-test with Bonferroni correction in each treatment's group.
Moreover, we analysed data with an intention-to-treat analysis, detecting the cumulative probabilities of steroid-free remission (CDAI < 150) survival, by using the Kaplan–Meier method. Analysis included all patients with no steroid-free remission at week 0.
Time to event was analysed from the date of the first anti-TNF treatment to the date of the last known follow-up. As patients were not randomised, a univariate analysis was performed. To assess the predictive factors of steroid-free remission, among clinical baseline characteristics (Table 1), univariate analysis with log-rank test was used. Baseline characteristics were binary categorical variables (i.e. male gender) or continuous variables (i.e. CD duration before anti-TNFs). To analyse independent variables predictive of steroid-free remission, all significant variables in the univariate analysis were inserted into Cox proportional-hazards regression using a forward selection method. We used the threshold based on the significance detected by the univariate analysis, P = 0.3, to define the initial set of candidate variables for the multivariate Cox regression model. The model selection was implemented without the use of any automatic procedure, carefully checking, at any inclusion or exclusion step, the stability of the model, taking into account aspects like multicollinearity. All variables not included in the final model would have a significance level P > 0.3, if included. Results were reported as hazard ratio (HR) with 95% CI and P <0.05 was considered statistically significant.
Table 1. Demographic and clinical characteristics of the 93 patients enrolled
INFLIXIMAB (n = 44)
ADALIMUMAB (n = 49)
NS, not significant.
Gender Male n (%)
P = NS
Age median (range), years
P = NS
CD duration median (range), years
P = 0.02
Previous surgery n (%)
P = NS
Smoking Habits n (%)
P = NS
Age at diagnosis (A)
A1 ≤ 16 years
P = NS
A2 17–40 years
A3 > 40 years
L1 terminal ileum
P = NS
Upper GI modifier (L4)
3 (L1); 1 (L2); 1 (L3)
2 (L2); 1 (L3)
B1 nonstrict. nonpenetrant.
P = NS
Perianal disease modifier (p)
13 (B1); 1 (B2); 2 (B3)
6 (B1); 2 (B2); 4 (B3)
Previous exposure to anti-TNF-α, n (%)
P = NS
Indication for anti-TNF-α, n (%)
P < 0.006
Reasons of selecting TNFs
Choice of the referral physicians
During the study period, 93 patients with CD fulfilling the inclusion criteria were treated with anti-TNFs. Forty-four patients were treated with infliximab and 49 patients with adalimumab. Table 1 details reasons for choosing each anti-TNF treatment at enrolment.
Demographic and clinical characteristics of patients treated with anti-TNFs are summarised in Table 1. As shown, most of clinical characteristics (including age, gender, previous surgery, smoking habits, concomitant medications) were comparable between patients treated with infliximab vs. adalimumab. Differently, the two study populations differed in terms of CD duration (P = 0.02) and in the infliximab group, a higher percentage of patients were treated for perianal disease and a lower percentage of patients for refractory CD (P = 0.006) (Table 1). At enrolment, 7/44 patients (16%) in the infliximab group and 9/49 patients (18%) in the adalimumab group were on dual immunosuppression with thiopurine, initiated by >6 months, while no patients were on concomitant methotrexate treatment. Furthermore, in the infliximab group, 30/44 patients (75%) were on steroids (prednisolone 1 mg/kg tapered 5 mg/week in 14 patients or budesonide 9 mg). Accordingly, in the adalimumab group, 36/49 patients (73%) received steroids (prednisolone 1 mg/kg tapered 5 mg/week in 15 patients or budesonide 9 mg in 21 patients). In the infliximab group, 13 patients (30%) had a previous exposure to anti-TNFs (10 with infliximab, 3 adalimumab) (Table 1). In these patients, anti-TNFs were discontinued due to clinical remission (n = 11), no response (n = 1) or loss of response (n = 1). In the adalimumab group, 17 patients (35%) had previous exposure to infliximab (Table 1). Reasons for infliximab discontinuation was clinical remission (n = 13), infusion reaction (n = 2) or no response (n = 2). The flow chart in Figure 1 summarises the outcome of patients treated with either infliximab or adalimumab. During the study period, none of the patients had doses of infliximab or adalimumab increased or dosing intervals reduced.
Indication for infliximab included active CD in 15 (34%) patients, refractory CD in 13 (30%) patients and perianal disease in 16 (36%) patients (Table 1).
Among the 44 patients treated with infliximab, only 34 (77.3%) completed the induction treatment (Figure 1). Reasons for drop out before the end of the induction included no response requiring urgent surgery after week 2 in 2 (4.5%) patients and severe adverse events (SAE) in 8 (18.2%) patients. When considering the outcome of both the 34 patients completing the induction regimen and the 2 patients requiring surgery due to no response (n = 36), at week 6, 25 (70%) were in remission, 2 (5%) showed clinical improvement and 9 (25%) were no responders. In particular, two of these nine patients showing no response required urgent surgery due to refractory CD.
In the per-protocol analysis, when considering the clinical outcome at the end of the induction treatment in the 34 patients completing the induction regimen, the CDAI value significantly reduced when comparing week 0 vs. both week 2 and week 6 (P < 0.0001 for both) (Figure 2). Among the 23 patients on steroids at enrolment, 10 (43%) were able to wean off steroids within week 6. Among the 14 patients showing active perianal disease at baseline and completing the induction regimen in the per-protocol analysis, none showed fistula closure at week 6. All these 14 patients had seton drainage from enrolment to week 6.
The IBDQ was performed in 16 compliant patients, showing a significant improvement at week 0 vs. week 6 (P < 0.001) (Figure 2).
Among the 44 patients enrolled, 33 received maintenance treatment. As shown in Figure 1, 10 (30%) patients dropped out before week 6, while in one patient remission was maintained by using azathioprine with no infliximab. The maintenance treatment at week 54 was completed by 23 of these 33 (60%) patients, as 10 patients dropped out. Reasons for drop out included clinical worsening or loss of response (n = 5), SAE (n = 5). Three of the five patients with loss of response required surgical resection due to refractory CD at weeks 8, 10 and 30 respectively. None of these three patients showed early post-operative complications.
When considering the clinical outcome of all patients starting the maintenance treatment (n = 33), including not only the 23 patients completing the 54 weeks treatment, but also the 5 patients dropped out due loss of efficacy, the analysis included 28 patients. Among these 28 patients, 20 (74%) patients were in clinical remission, 1 (3%) patient showed clinical improvement, 4 (13%) patients clinical worsening and 3 (10%) patients loss of response.
As shown in Figure 2, in the per-protocol analysis when considering the clinical outcome in these 23 patients, the CDAI value significantly reduced at each visit during the maintenance treatment vs. baseline (week 6 vs. week 0 P < 0.04; week 12 vs. week 0 P < 0.01; week 30 and week 54 vs. week 0 P < 0.001 for both), as also at different observations within the 54 weeks follow-up. Among the 23 patients completing the maintenance treatment, 20 were on steroids at enrolment and 16 patients (80%) were able to wean off steroids within week 54.
The IBDQ was performed in a small subgroup of compliant patients (n = 10) before treatment, at week6 and at week 54. A slight improvement at week 0 vs. both week 6 and week 54 was observed, not reaching statistical significance due to the small number of patients (Figure 2). At baseline, active perianal fistulae were observed in 13 out of the 23 patients completing the maintenance treatment. At week 54, fistula closure was achieved in 11 out of these 13 patients (84%).
Among the 49 patients treated with adalimumab, indications for treatment included active CD in 17 (35%), refractory CD in 26 (53%) and active perianal disease in 6 (12%) patients (Table 1).
As shown in Figure 1, all the 49 patients enrolled completed the induction treatment as no SAE were observed during the 4 weeks treatment.
When considering the outcome of all the 49 patients, at week 4, 32 (66%) were in clinical remission, 7 (14%) showed clinical improvement and 10 (20%) patients were no responders. The CDAI value significantly reduced when comparing week 0 vs. both week 2 and week 4 (P < 0.0001 for both) (Figure 3). Among the 36 patients on steroid treatment at enrolment, 13 (36%) were able to wean off steroids within week 4. Among the 8 patients showing active perianal fistula at week 0, all patients had seton drainage up to the end of the induction regimen, and therefore no patients showed fistula closure at week 4.
The IBDQ performed in 34 patients before and after treatment showed a significant improvement at week 0 vs. week 4 (P < 0.01) (Figure 3).
Maintenance treatment was completed by 33 out of 49 patients (67%), completing the induction regimen, as 16 patients dropped out before the scheduled 54 weeks treatment (Figure 1). Reasons for drop out included: clinical worsening or loss of response (n = 8), delayed hypersensitivity reactions (DHR) (n = 1), Herpes Zoster viral infection (n = 1), dysplasia of the uterine cervix (n = 1) or other causes (n = 5, including: lost to follow-up, ectopic pregnancy, remission while on combined azathioprine treatment, low compliance or pregnancy, on the basis of patient's decision after detailed information regarding the possibility to continue treatment).
Among the 8 patients with loss of response, 4 required elective surgery at week 6, 8, 20 and 28, respectively. None of these patients showed early post-operative complications.
The clinical outcome was first considered in the 41 patients including both the 33 patients completing the 54 week maintenance treatment and the 8 patients who dropped out due to loss of response. At week 54, 30 of these 41 patients (73%) were in clinical remission, 1 (2%) showed clinical improvement, 7 (17%) have clinical worsening and 3 (8%) showed loss of response.
In the per-protocol analysis, when considering the clinical outcome in the 33 patients completing the maintenance treatment, the CDAI value significantly reduced when comparing week 0 vs. each scheduled visit and during the maintenance treatment (week 54, week 30, week 12, week 4 vs. week 0 P < 0.001 for all) (Figure 3). Among the 24 patients on steroids at baseline, 17 (71%) were able to wean off steroids within week 54.
The IBDQ, performed in 10 compliant patients, showed a significant improvement at week 0 vs. each scheduled visit (P < 0.002) (Figure 3, panel d).
At week 54, fistula closure was observed in five of the six patients showing active perianal disease at baseline.
Predictors of steroid-free remission and comparison between anti-TNFs
To assess whether clinical characteristics of patients at baseline may have influenced the likelihood of success of treatments, univariate analysis including patients with no steroid-free remission at week 0 was performed. This analysis therefore included the 80 patients with no steroid-free remission at week 0 (35 treated with infliximab and 45 with adalimumab). None of the following factors was predictive of steroid-free remission in univariate analysis (log-rank test): different anti-TNFs (infliximab or adalimumab) (P = 0.4), age (continuous variable, years) (P = 0.1), CD location (ileum, ileum-colon, colon) (P = 0.056), CD duration (continuous variable, years) (P = 0.5), previous appendectomy (yes, no) (P = 0.3), previous CD-related surgery (yes, no) (P = 0.06), perianal disease (active draining fistula or not) (P = 0.7), previous exposure to anti-TNFs (infliximab, adalimumab or naive patient) (P = 0.7), indication for anti-TNFs (active, refractory or perianal CD) (P = 0.7). Only two clinical characteristics were predictors of steroid-free remission. In particular, smoking was a negative predictor of steroids free remission when compared with no smoking [HR 0.49 (95% CI 0.26–0.92); log-rank test P = 0.02]. An additional parameter was CD behaviour, as B1 (non stricturing non penetrating) was predictive of steroid-free remission when compared with CD behaviour B3 (penetrating) and B2 (structuring) [HR 2.72 (95% CI 0.94–7.8); log-rank test P = 0.01].
The Kaplan–Meier survival analysis showed the risk of steroid-free remission in patients treated with infliximab vs. adalimumab for 54 weeks and subsequently followed up for 6 months (week 76) (Figure 4). As shown, log-rank test analysis detected no significant differences between the two groups (P = 0.4). The same analysis performed for patients treated with anti-TNFs for 54 weeks and subsequently followed up for 6 months (week 76), showed significant differences according to CD behaviour (log-rank test P = 0.01) (Figure 5) and smoking habits (log-rank test P = 0.02) (Figure 6).
Among the tested variables, only age, smoking habits, CD site, CD behaviour and previous CD-related surgery resulted in candidate variables. The variable comparing the two anti-TNFs (infliximab or adalimumab) was included in the final model, as they represent a key issue of the study. Absence of interaction and multicollinearity between variables was checked. In the final model, only no smokers and non stricturing non penetrating CD behaviour showed a significant association with an increased steroid-free remission risk [no smokers vs. smokers HR 2.94 (95% CI 1.52–5.70), P = 0.00142; non stricturing non penetrating vs. fistulising and penetrating CD behaviour HR 3.11 (95% CI 1.06–9.13); P = 0.038]. The use of adalimumab vs. infliximab was not significantly associated with the risk of steroid-free remission. Surgery for CD was associated with both the pattern (P = 0.0071) and the site (P = 0.0092), which were in turn associated (P = 0.0026).
Safety profile of anti-TNFs
In the infliximab group, SAE occurred in 13/44 patients (29.5%). During the induction regimen, DHR were observed in three patients and infusion reactions in five patients (subsiding after i.v. steroids and anti-histaminic drugs). All these eight patients had a previous exposure to infliximab (mean 7 years; range 6–9). During maintenance with infliximab, DHR occurred in four patients with no previous infliximab exposure, while infusion reaction was observed in one patient previously treated with adalimumab. In the adalimumab group, only 1 of 49 patients (2%) had a SAE, showing a DHR during the maintenance treatment. This patient was treated with infliximab in the previous 12 months. One additional AE was observed in one patient developing Herpes Zoster infection during the maintenance treatment, requiring drug discontinuation.
Several independent studies, including clinical trials, observational and cohort studies, support the marked efficacy of anti-TNFs in CD. Nevertheless, to our knowledge, no studies reported the efficacy and safety of infliximab and adalimumab, in a consecutive series of active CD patients referring to a tertiary IBD centre. We therefore aimed to assess, in an open, single centre study, the efficacy and safety of infliximab and adalimumab in active CD patients and to compare the efficacy of these anti-TNFs in clinical practice.
In our cohort, the use of infliximab in CD showed a remission rate comparable to the efficacy reported in controlled trials when considering both the induction and the maintenance regimens.[5, 6, 26] Differently, the induction regimen using adalimumab determined a higher remission rate than the rate reported in the CLASSIC I and GAIN I trials(66% vs. 36% and 21% respectively). The rate of remission when using adalimumab, as maintenance treatment was comparable to the rate reported in the CLASSIC II trial (73% vs. 79%). Several reasons may account for the observed differences between the present study and clinical trials. Among these, in our study protocol, the first clinical assessment was performed later (week 4) than in the reported trials (week 2).[12, 16] The different number and characteristics of the study populations and study designs (open vs. placebo-controlled trials) may also account for the observed discrepancies. The response and the remission rate in open-label studies may indeed differ from selected CD patients included in controlled trials, showing differences, including a higher compliance and adherence to treatment. According to this concept, the observed efficacy in our study was comparable to the efficacy reported when using adalimumab in clinical practice. In particular, Trinder et al. described their clinical experience using adalimumab in CD, showing at 8 and 12 weeks a response rate of 81.8% and 84.4%, and a remission rate of 54.5% and 63.6% respectively. The observed discrepancies may not be related to a different modality of assessment of CD activity, as most of the clinical trials use the CDAI score at this purpose. Despite the known limitations of this score for assessing CD activity (i.e. diarrhoea related to previous resections), the CDAI currently represents the gold standard for this purpose. Mucosal healing after anti-TNFs was not considered as an end point, due to the need of repeated colonoscopies, reducing the adherence of patients to the study protocol, requiring a long-term follow-up and several clinical assessments.
Among limitations of the present study there is the open-label study design, not allowing an objective assessment of the clinical outcome by using anti-TNFs. According to the study design, present findings indeed represent real-life clinical data, and issues such as previous anti-TNFs therapy limit the interpretation of the data.
The s.c. route of delivery of adalimumab may increase the patients' compliance when compared with the need of hospitalisation when using infliximab. The different route of administration may play a role when choosing anti-TNFs in CD. On the other hand, the self-administered adalimumab requires a good adherence to treatment, thus limiting its use in patients with a known low compliance. In this regard, our findings, showing a comparable efficacy of these 2 anti-TNFs, further support that a proper selection of patients is required before choosing the route of administration of anti-TNFs.
Regarding the steroid-sparing effect of anti-TNFs, Cohen et al. and Farrell et al. reported, when using infliximab in clinical practice, complete steroid withdrawal rates of 40% at week 12 and of 5% at week 7 respectively. When considering adalimumab, Colombel et al. reported that 30–35% of patients at week 26 and 23–29% of patients at week 56 achieved steroid-free remission during maintenance treatment. Swoger et al. reported results from the Mayo Clinic practice, showing that 70% of CD patients on systemic steroids were able to taper and discontinue steroids by using adalimumab. In agreement with this finding, we report that at week 54, 80% of patients treated with Infliximab and 71% of patients treated with adalimumab were able to wean off steroids.
To our knowledge, no previous studies have compared the steroid-free remission when using infliximab or adalimumab in a consecutive series of active CD patients under regular follow-up. The observed findings suggest no significant differences regarding this outcome, when using infliximab or adalimumab. Present results support the observations from clinical trials, suggesting that the choice of anti-TNFs should take into account the patients' preference to the route of delivery, local feasibility and costs,[10, 19] rather than possible differences in terms of efficacy.
Post hoc analysis of clinical trials using anti-TNFs suggests that CD patients with shorter disease duration show a higher response and remission rates. This observation may be related to changes of the inflammatory response over the course of the disease  or to the development of fibrotic lesions showing less responsiveness to anti-TNFs. Vermeire et al. reported that older age, small bowel location and smoking are associated with a lower response rate to infliximab. Likewise, in our CD population, a better response to anti-TNFs, in terms of steroid-free remission, was observed both in no smokers and in non penetrating non structuring CD. An inflammatory CD phenotype has recently been reported to predict sustained clinical benefit using infliximab.
The rate of infliximab discontinuation due to SAE in our study was higher than expected.[10, 32] The observation that all patients showing these SAE had previous infliximab exposure at least 1 year before enrolment may account for this finding. This condition has indeed been associated with a higher risk of AE, mostly related to the development of anti-infliximab antibodies,[11, 32] not tested in our study. In our series, all patients developing either infusion reactions or DHR after infliximab discontinued the drug. This decision was taken according to current guidelines, suggesting that shifting to a different anti-TNF is appropriate in case of development of these side effects, although preliminary observations suggest that steroids may prevent subsequent infusion reactions.
Although present findings derive from an open study including a relatively small number of patients, they suggest a comparable efficacy of infliximab and adalimumab, thus providing the first data comparing these anti-TNFs in CD. Future prospective double blind studies in patients matched for clinical variables are required to further assess the efficacy and safety of different anti-TNFs in CD.
Declaration of personal interests: Livia Biancone and Francesco Pallone have served as speakers and advisory board members for Abbott and MSD (former Shering-Plough). Emma Calabrese and Sara Onali have served as speakers for Abbott. Declaration of funding interests: The study received a grant from the Fondazione Umberto Di Mario, Roma, Italy.