In this issue of Gut, two papers describe the vast experience of the group in Leuven, Belgium, with the use of infliximab in Crohn’s disease. The first report (see page 492) describes 614 patients, mostly with luminal disease and about 25% with fistulising disease treated over a period of 12 years. A sustained benefit was seen in 63%; in a third even after stopping the drug while in remission.1 Twenty percent of patients withdrew from treatment due to loss of response and 11% had no initial response. Of those patients with an initially elevated C-reactive protein (CRP), 77% showed a “biological response”, which had a positive predictive value for a better clinical outcome. Out of 10 deaths, one was directly related to therapy, three were possibly rèlated and six were not related.

The second report (see page 501) assessed medical records of 734 patients with inflammatory bowel disease (IBD), the majority with Crohn’s disease, treated with infliximab during 14 years, and compared them with 666 patients who did not receive infliximab followed during this period.2 Median follow-up was more than twice as long for controls than for infliximab patients (58 months; 33–88 months). A total of 113 severe adverse events occurred in 93 treated patients, and 157 in 126 control patients. Mortality, malignancies and infection rates did not differ between the two populations. However, the control group was older and received steroids much more often during the much longer follow-up.

To me there is an important message from these data to all patients with Crohn’s disease who do not have a benign course of the disease, which occurs in 40–50% of patients in population-based cohorts3: present yourself to a referral centre with experienced doctors using all available methods to assess and to treat your disease if you have proven inflammation and do not respond to your current treatment.

The data confirm and expand previous results4^–7 from different countries in patients with Crohn’s disease. They are, furthermore, in agreement with outcomes in patients with other diseases such as psoriasis,8 and ankylosing spondylitis.9

Why are the data so much better than those observed in the pivotal trials where neither the initial results nor long-term data were as good?10 11 Obviously the clinical trials included a large number of patients. To me a drug in a study against a placebo needing >200 patients to show a significant effect is not very promising. The results of the study Schnitzler et al1 finally convince me that in the clinical trials leading to registration of infliximab many patients were included who either did not have inflammation (this is indicated by the high frequency of normal CRP as compared with this cohort) or, for example, had an increased Crohn’s disease activity index (CDAI) due to irritable bowel syndrome (IBS) symptoms or due to prior resection.12 Schnitzler et al1 obviously selected their patients more carefully, as evident from the study of Fidder et al2 where about 45% of patients did not receive infliximab during the same period. This is also evident from the fact that about 60% of patients had other immunosuppression at the time of the first infusion.1 Obviously this was a typical tertiary referral centre population.

The observation that in a considerable number of patients remission was maintained even after stopping infliximab is very important. Eighty-four of those were continued on other immunosuppressants, indicating a bridging effect of infliximab. However, two-thirds of patients with sustained benefit were continued on long-term infliximab, which is in accordance with data from paediatric patients defining “infliximab dependency”.13 We certainly should rethink our definitions of drug dependency since probably all forms of immune suppression, beginning with steroids, but including azathioprine, 6-mercaptopurine, methotrexate, ciclosporin A and the tumour necrosis factor (TNF) blockers, often result in a “dependency” and will have some side effects due to the lack of an adequate immune response to microbes and tumour cells.14^–17 Thus, any immunosuppressant increases the risks, and we should aim for patients to achieve remission with little or no risk of infectious complications.

The same observation questions a recent paper from Belgium indicating that azathioprine can be withdrawn without reducing the effect of infliximab.18 If we want to use infliximab as bridging therapy, we need to maintain this slow-acting immunosuppressant during infliximab treatment. Interestingly, Schnitzler et al1 describe a switch from azathioprine to methotrexate in a considerable number of patients—methotrexate acts faster than azathioprine. When considering the as yet unpublished results of the SONIC trial, we might need to reconsider combination therapy and it is necessary to define which concomitant immunosuppressant, if any, is better.

This need is also highlighted by the recent paper on early combined immunosuppression in Crohn’s disease19 where no long-term difference between top-down and step-up treatment was shown, in particular when remission and not steroid-free remission was used as an early criterion. The use of the former seems advisable in this trial since low dose steroids could be maintained in this open study at the will of the treating doctor.

The use of all immunosuppressants (the wording immunomodulator is not appropriate and is only used to mislead possible users) has been disputed due to the innate potential for infections and neoplastic complications.17 When searching PubMed using the key words infliximab and risk, there are 135 hits, indicating that there may be a problem. The study of Fidder et al2 attempts to study these risks. It comes to the conclusion that long-term infliximab has a good safety profile overall. To me, this study is much less convincing than that of Schnitzler et al.1 Comparing two cohorts with very different features regarding sex, age, concomitant medications (steroids) and in particular follow-up (see table 1 in Fiddler et al2) cannot over-rule the impressive number of papers indicating that even in controlled trials,17 as well in several other cohort studies4 5 and in other indications,16 20 the risk of infections and neoplasia is increased. This is certainly the case with other immunosuppressants21 22 as is well known from the transplant literature. The comparison between two cohorts in a referral centre probably does not allow for a reasonable assessment of malignancy risk since most patients will have received immunosuppressants anyway. This is particularly true for skin cancers, as evidenced by studies in rheumatoid arthritis.20 In order to assess the risk of malignancies in IBD and in particular with immunosuppressants, population-based studies are needed.

Regarding infectious complications, a recent study15 indicated that combinations of immunosuppressants increased the risk significantly, and that this risk is additionly increased by TNFα blockade, as also evidenced by studies in rheumatoid arthritis.16 However, the effect of concomitant steroids is very important and they should be tapered as soon as possible.

It is currently not clear if the use of infliximab is associated with an increased risk of perioperative or postoperative complications. Some authors describe an increased risk in Crohn’s disease23 and in ulcerative colitis24; others do not.25 26 Interestingly, 23% of patients in the study of Schnitzler et al needed major abdominal surgery, but there was no indication of an increased risk for complications.1

In summary, the two papers from Leuven published in this issue of Gut indicate that the use of infliximab in appropriately selected patients will help to improve the long-term outcome. It obviously does not solve the problems in all patients forever, and it has some relevant side effects typical of strong immunosuppression, which can be limited by careful observation and follow-up in the hands of experienced clinicians.