Infliximab Is Associated With an Increased Risk of Serious Infection in Patients With Psoriasis in the U.K. and Republic of Ireland

Results From the British Association of Dermatologists Biologic Interventions Register (BADBIR)

Z.Z.N. Yiu; ; D.M. Ashcroft; I. Evans; K. McElhone; M. Lunt; C.H. Smith; S. Walton; R. Murphy; N.J. Reynolds; A.D. Ormerod; C.E.M. Griffiths; R.B. Warren; BADBIR Study Group


The British Journal of Dermatology. 2019;180(2):329-337. 

In This Article


Infliximab was associated with a twofold increase in the risk of serious infections compared with nonbiologic systemic therapies, and a threefold increase in the risk of serious infections compared with methotrexate, in patients with psoriasis. This is in contrast to our results from other biologic therapies in BADBIR, where etanercept, adalimumab and ustekinumab were not associated with a higher risk of serious infections than nonbiologic systemic therapies. There was a time-stratified difference in the risk of serious infections in the comparison between infliximab and nonbiologic therapies, with a higher risk in the first 6 months of therapy and a lower risk after 1 year of therapy, but these estimates had overlapping CIs given the fact that we had a lower precision owing to a lack of power within the specific time strata.

We have shown that combination treatment of infliximab with methotrexate, and other immunosuppressive therapies, is associated with a threefold increase in the risk of serious infection compared with patients on infliximab monotherapy in the biologic-naïve cohort. This is not the case in the overall cohort (i.e. both biologic-naïve and experienced) (Table 3). However, the propensity score method balances the baseline characteristics and not time-varying factors, and hence it cannot deal adequately with confounding by indication for the use of concomitant immunosuppressive therapy. Therefore, this estimated result should be interpreted with caution.

The major strengths of this study are the prospective cohort study design, fully industry-independent data analysis and the participation of multiple dermatology centres (N = 153) in the U.K. and Republic of Ireland. This study reports on the risk of serious infection in the largest incident cohort of patients with psoriasis on infliximab to date. Owing to the capture of numerous important covariates, we were able to account for significant confounding through weighting by propensity score. The substantial reduction in the point estimate of the HR after adjustment (> 40% change) suggests that significant positive confounding was reduced after IPTW propensity score adjustment.

Limitations include recall bias, which may occur with patient-reported characteristics, and residual confounding through variables that were not measured and not known to be associated with either exposure or outcome. There is a possibility that clinicians may have a lower threshold and a heightened awareness for admitting patients on infliximab with suspected infections compared with nonbiologic systemic therapies, thereby introducing confounding by indication that cannot be adjusted for. As patients on infliximab are by indication (higher PASI and DLQI) different from all other patients on the registry, there may be a higher degree of selection bias in our comparisons. The potential confounding introduced by this selection bias may be partly adjusted for from the measured covariates, and the propensity score IPTW method reduced expected bias substantially (Table S3; see Supporting Information), but there may be unmeasured covariates that have introduced confounding that cannot be adjusted for. For example, we were not able to adjust for previous serious infection within the past year. Conversely, our results have high external validity for patients with psoriasis who are eligible for infliximab in the U.K. or any other country with similar prescription eligibility criteria for the drug. A total of 844 registered patients were classified as dropping out owing to lack of follow-up. Of these patients, 92·5% (785 patients) discontinued the study owing to withdrawal of consent and, because the reason for withdrawal is not provided, it is unclear how or whether this would have introduced systematic bias.

We performed an analysis with an incident cohort, which avoids left truncation. Left truncation occurs when there is a period of time during which the event could have occurred that cannot be observed. Involving a prevalent cohort would introduce left truncation by selecting the involvement of patients on infliximab who have not discontinued owing to a serious infection, and selecting out those patients who have discontinued infliximab owing to a serious infection, thereby underestimating the associated risk in infliximab. To maximize our sample size we included all lines of infliximab therapy (i.e. biologic-experienced individuals who progressed onto infliximab as the second-line, third-line and fourth-line biologic therapy); however, we also performed a sensitivity analysis restricted to first-line infliximab therapy in biologic-naïve participants. There is a possibility that patients with severe psoriasis who experienced a serious infection on the first-line biologic therapy are not subsequently prescribed infliximab. However, it is reassuring that the results are contrary to this hypothesis, as the adjusted results for infliximab involving all lines of therapy are higher than those of first-line therapy with tighter CIs, suggesting that the difference is due to sample size and power rather than left truncation.

Our crude incidence rate for serious infections for infliximab (47·8 per 1000 person-years) is higher than that published in the Psoriasis Longitudinal Assessment and Registry (PSOLAR), a large study based mainly in the U.S.A. and Europe sponsored by a single pharmaceutical company (infliximab serious infection rate of 24·9 per 1000 person-years),[7] and the Spanish Registry of Adverse Events from Biological Therapy in Psoriasis (BIOBADADERM) (infliximab serious infection rate of 18·9 per 1000 person-years).[6] However, the adjusted relative risks from these cohorts are broadly similar to our results. PSOLAR reported an adjusted HR of 2·51 (95% CI 1·45–4·33) in a mixed prevalent/incident population and an adjusted HR of 1·78 in the incident population (95% CI 0·64–4·98), where the chosen comparator was a cohort on acitretin and/or phototherapy. BIOBADADERM reported an adjusted rate ratio of 2·52 (95% CI 0·83–7·69) in an incident cohort where the chosen comparator was a cohort on methotrexate only. The use of prevalent cohorts should be avoided in the assessment of adverse events because of the risk of left truncation, especially given the finding of an early high risk of serious infections in the first 90 days of treatment for infliximab in rheumatoid arthritis.[11] Both PSOLAR and BIOBADADERM incident cohorts did not have the requisite power to achieve adequate estimate precision.

As discussed previously, infliximab is currently recommended for patients with severe psoriasis only.[2] A recent network meta-analysis of evidence from clinical trials has found infliximab to be of high efficacy but of poorer tolerability.[1] Given our findings of a higher risk of serious infection associated with infliximab, we provide real-world evidence to reinforce the position of infliximab in the psoriasis treatment hierarchy. However, it should be noted that there is no real-world evidence, as yet, for the risk of serious infection of alternative, newer, licensed biologic therapies for psoriasis, such as secukinumab or ixekizumab.

Infliximab is associated with an overall twofold increase in the risk of serious infections when compared with nonbiologic systemic therapies. Patients with severe psoriasis who fulfil the criteria for the prescription of infliximab should be counselled for the risk of serious infection. These results are relevant to patients in the U.K. and the Republic of Ireland, and also in countries that have similar eligibility criteria for the prescription of infliximab.