Discussion
This meta-analysis is the first to compare efficacy of combined IFX and IS therapy with IFX alone in moderate-to-severe active UC. In this meta-analysis, the clinical remission rate is higher in the combination therapy group, with significant results at 4–6 months (P < 0.001). According to Orwin's method, four medium-sized nonsignificant studies would need to be included to cancel out significance, which underlines the robustness of this result.
In UC, the benefit of the combination of IFX with an IS is debated. Several trials have been undertaken to resolve this issue; however, the results are contradictory. Results of post hoc analyses of the ACT 1 and ACT 2 trials, did not find combined therapy (IFX-AZA) to be superior to IFX alone. On the other hand, Armuzzi et al. in their recently published prospective cohort found that combined IFX and AZA therapy was a predictor of steroid-free clinical remission at 6 and 12 months. A recent study has compared the combination of IFX with Methotrexate (MTX) to IFX alone and MTX alone to a control group (2 mg/kg prednisone) in patients with severe corticosteroid dependent UC. This study showed that the combination of IFX with MTX did not provide any benefit concerning mucosal healing at 16 weeks. However, the results of the SUCCESS trial have provided evidence of the superiority of the combination of IFX with AZA compared to IFX alone at 16 weeks for IS naïve patients with moderate-to-severe UC.
This meta-analysis did not demonstrate any significant difference between the two treatment groups at 12 months, which may be due to the fact that the two studies included for the 12 month analysis had contradictory results. The ACT 1 post hoc analysis did not show difference between combined IFX-IS vs. IFX alone, contrary to the Armuzzi et al. cohort in favour of combined IFX and AZA. These contradictory results could be due to differences concerning the patient population. In the Armuzzi et al. cohort, patients had a lower mean duration of disease (4 years vs. 6.6 years in the ACT 1 trial), and these patients had significantly higher steroid-free clinical remission rates at 6 and 12 months. Moreover, in the ACT 1 trial, the majority of patients had previously been treated with thiopurines over the last 5 years with no response, unlike the Armuzzi et al. cohort where an important proportion of patients were IS naive (45.2%). Moreover, the lack of a significant difference between the two treatment groups at 12 months should be considered with caution since only two studies were included for this analysis, and the results of a meta-analysis containing only two studies may be questionable. A subsequent meta-analysis with additional studies would be necessary in order to demonstrate a significant difference between the two treatment groups.
Data from the SUCCESS trials demonstrated the superiority of combined IFX-AZA therapy for IS naïve UC patients, but the question remains open for all UC patients regardless of their prior IS status. This study population differed somewhat from other trials as patients were IS naive and had a relatively short history of UC. Concerning prior IS use, data were clearly available for the two groups (IFX-IS and IFX alone) only in the SUCCESS trial with 8 patients in each group. However, in the Armuzzi et al. cohort, 54.8% of patients had prior IS therapy, whereas in the ACT1 and ACT2 trials, 50.4% and 42.3% of patients were on IS treatment at the time of induction of IFX therapy. In addition, patients having had no response to azathioprine or mercaptopurine within the preceding 5 years were eligible for inclusion the ACT1 and ACT 2 trials. We can therefore presume that a considerable number of patients included had prior IS therapy. Therefore, it is likely that our results could be extrapolated to all UC patients regardless of their prior IS use.
Heterogeneity is present in the patient population of our meta-analysis, since we included both trials with IS naïve patients and patients previously treated with IS, for which disease duration was variable. In real life practice, both these types of patients are regularly encountered; however it should be noted that in a hospital setting patients with a long disease duration having already received several treatments are more frequent, and few of these patients are IS naïve. We therefore decided not to limit our analysis to patients with a short disease duration naïve of IS.
Several studies have shown that concomitant immunosuppressant therapy decreases formation of antibodies to infliximab (ATI) in patients receiving IFX. The clinical remission rate would be correlated with the trough serum IFX levels and to the absence of ATI formation. However, despite the general belief that concomitant IS therapy decreases formation of ATI and leads to a higher IFX trough level in patients with IBD, data are inconsistent.
We decided to exclude studies using adalimumab or golimumab since there is no trial available comparing ADA or GLM vs. combination therapy in UC patients which would allow us to perform a meta-analysis for each anti-TNFα agent. Data concerning clinical remission rates on golimumab or adalimumab therapy vs. combination therapy (ADA-IS or GLM-IS) are not available neither in manuscript form nor in supplementary appendices of trials evaluating these two anti-TNFα agents' induction and maintenance efficacy in UC.
We chose to include both RCTs randomised according to the combination of IFX with an IS or not, and nRCSs. There is no denying that RCTs are studies which are of superior methodological quality with a lower risk of bias, and that post hoc analyses of RCTs are of poorer methodological quality. However, the nRCSs included in our meta-analysis are either studies randomised according to different criteria, or cohorts with prospective follow-up. All these studies are therefore of high methodological quality. On a practical level, prospective cohorts have the advantage of corresponding more to real life medical practice than RCTs. Moreover, patients enrolled in RCTs do not always accurately represent the IBD population.
The frequency of adverse events, including frequency of all and serious infections, all and serious infusion reactions and adverse events leading to discontinuation of study drugs were available in only two studies with separate results at 30 weeks for the ACT 1 trial and at 54 weeks for the ACT 2 trial. It was therefore not possible to perform a meta-analysis on these outcomes. However, in both post hoc analyses there was no significant difference between combination therapy and monotherapy groups concerning infections, serious infections, infusions reactions and adverse events leading to discontinuation of study drugs.
In conclusion, our meta-analysis reveals significant superiority for combination IFX and IS therapy (AZA or MP) compared to IFX alone for the achievement and maintenance of clinical remission of UC flares at 4–6 months of treatment. At 12 months of therapy, results should be considered with caution as the difference was not significant, possibly due to a lack of studies.