a study investigating the association of dermatological and infusion reactions to infliximab and infliximab trough levels

1Division of Gastroenterology, Department of Medicine; 2Division of Medical Biochemistry, Department of Laboratory Medicine and Pathology, University of Alberta, Edmonton, Alberta Correspondence: Dr Vivian Wai-Mei Huang, 2-14A Zeidler Building, University of Alberta, Edmonton, Alberta T6G 2X8. Telephone 780-492-6941, fax 780-492-8121, e-mail vwhuang@ualberta.ca Received for publication August 18, 2014. Accepted November 24, 2014 Anti-tumour necrosis factor (TNF) agents have transformed the management of inflammatory bowel diseases (IBDs) and other autoimmune diseases. Despite its efficacy as a therapeutic agent, there have been increasing reports of infliximab-associated adverse events, some of which appear to be paradoxical immune-mediated inflammatory disorders (1,2). Theses adverse events include dermatological conditions, such as psoriasis, eczema, erythema multiforme, lupus-like syndrome and nonmelanoma skin cancer (3-7), and infusion reactions, both acute and delayed (8,9). The management of these adverse events include conservative therapy, use of corticosteroids and adjustment of infliximab therapy. These are all attempts to overcome hypothesized mechanisms of action, and are often unsuccessful. In severe cases of adverse reactions to infliximab, patients often must discontinue the medication (9-14). Pathophysiological mechanisms that have been postulated to contribute to infliximab-induced adverse events differ depending on the adverse event; however, these reactions are believed to be mediated through the immune system. Pathogenesis may include the development of antibodies to infliximab (ATI), either immunoglobulin (Ig) E resulting in a hypersensitivity type reaction (15,16) or IgG resulting in the formation of antibody/antigen complexes (1,8,17-20), or induction of autoantibodies such as antinuclear, antidouble-stranded DNA and antihistone (1,10,13,19). Dysregulation of the immune system and cytokine responses may also be involved (2,4-6,14,21-24). Given the importance of the pharmacodynamics of infliximab, the level of infliximab in the blood may be central to the generation of infliximabinduced side effects. However, to date, there have been no published VWM Huang, N Dhami, D Fedorak, et al. A study investigating the association of dermatological and infusion reactions to infliximab and infliximab trough levels. Can J Gastroenterol Hepatol 2015;29(1):35-40.

The aims of the present study were to report the prevalence of infliximab-associated dermatological and infusion reaction events among IBD outpatients receiving stable maintenance infliximab therapy, and to correlate ITLs with infliximab-associated adverse events.

Study design and setting
The present cross-sectional study was conducted at the University of Alberta IBD Infliximab Infusion Clinic (Edmonton, Alberta) in 2013.

Participants
Consecutive IBD outpatients receiving infliximab maintenance therapy were systematically and prospectively identified at the infliximab infusion clinic and invited to participate in the present study.Patients were included in the study if they met the following criteria: a known endoscopic and histological and/or radiological diagnosis of Crohn disease (CD) or ulcerative colitis (UC); primary responders to the initial induction regimen of infliximab 5 mg/kg at weeks 0, 2 and 6 (defined as a decrease in their Harvey-Bradshaw index after induction therapy by 3 points, or a decrease in their partial Mayo score to 0 or 1); receiving stable maintenance infliximab infusions (defined as infusions subsequent to the third induction dose or subsequent to the first infusion after any dose or interval change); and having complete records of infliximab infusions since initial induction.There were no specific exclusion criteria.All consented patients had blood for the determination of ITLs drawn immediately before their infliximab infusion, and underwent a detailed history and chart review to confirm infliximabassociated adverse events.

Outcomes
The primary outcome of the present study was the prevalence of infliximab-associated dermatological and infusion reaction adverse events among IBD outpatients receiving stable maintenance infliximab therapy.The secondary outcome was the correlation between ITLs and infliximab-associated adverse events.

Data sources and definitions ITL:
ITLs were determined on serum samples collected within 30 min before the beginning of the infusion.Serum was separated and frozen within 4 h of collection.An ELISA method (Immunodiagnostik, Germany) was used to quantify levels of free infliximab.Results are reported down to 0.4 µg/mL, with an interassay precision of 8% at 1.8 µg/mL, 9% at 8.6 µg/mL and 20% at 12.9 µg/mL.Demographics and infliximab-associated adverse events: At the authors' centre, each infliximab infusion is characterized with the following: clinical disease activity scores (Harvey-Bradshaw index for CD, partial Mayo for UC); health care provider documentation of infliximab-associated adverse events; and patient completion of a questionnaire of infliximab-associated adverse events.
Data were extracted from several sources by two of the authors (VH and ND) using a standardized case report form: electronic medical records; infusion questionnaires completed by the patient at the time of each infusion; nurse-generated infusion reports produced at the time of each infusion; and physician-generated clinic letters at the time of each infusion or subsequent clinic visits.
Infliximab-associated adverse events were classified as dermatological or infusion reactions.Dermatological adverse events were documented by a gastroenterologist and, when possible, confirmed by a dermatologist.Infusion reactions were documented by the infusion nurse and confirmed by a gastroenterologist.Acute infusion reactions were defined as any adverse reaction occurring during or within 24 h of an infusion, while delayed infusion reactions were defined as any adverse reaction occurring between 24 h and 14 days after an infusion.The data were reviewed by four of the authors (VH, ND, KK and RF).
Demographic patient data included sex, age, type of disease, premedication with solumedrol or diphenhydramine, and concomitant immunosuppression (azathioprine, 6-mercaptopurine or methotrexate).Infliximab treatment characteristics were obtained from the above listed sources and included: infliximab dose in mg/kg; infliximab dosing interval in weeks; cumulative infliximab dose in mg; and number of infliximab infusions.
Time to infliximab-associated adverse event was calculated by determining the number of weeks from the first stable maintenance infliximab infusion until the first report of the infliximab-associated adverse event.Censoring occurred if the patient reported no infliximabassociated adverse events; time to censoring was calculated by determining the number of weeks from the first stable maintenance infliximab infusion until the ITL measurement.

Study size
A total of 100 consecutive IBD patients receiving stable maintenance infliximab therapy were recruited and consented to participate.However, the final evaluable sample size (n=71) was determined by subsequent chart review, which excluded 29 patients due to incomplete infusion records.

Continuous variables were presented as median (interquartile range [IQR]
) due to the small sample size and nonparametric distributions.Nonparametric Mann-Whitney and Kruskall-Wallis tests were used to determine significant differences between median values.For categorical variables, proportions were calculated and comparison between subgroups was performed using Fisher's exact test; P<0.05 was considered to be statistically significant.

Ethics
The present study was approved by the Health Research Ethics Board of the University of Alberta.Patients consented to provide a blood sample and allow review of their medical records.

Patient characteristics
Table 1 summarizes the demographic and infliximab treatment characteristics of the 71 evaluable patients.The number of men (39 of 71 [45.1%]) and women (32 of 71 [54.9%]) were similar, and there were more patients with CD (46 of 71 [64.8%]) than patients with UC (25 of 71 [35.2%]).The median duration of time on stable maintenance infliximab therapy was 71.9 weeks (IQR 30.0 to 126.7 weeks), with a median time since last infusion of 7.0 weeks (IQR 4.1 to 8.0 weeks).

Infliximab-associated adverse events
More than one-quarter (18 of 71 [25.4%]) of study participants experienced either a confirmed infliximab-associated dermatological event (nine of 71 [12.7%]) or an infusion reaction (nine of 71 [12.7%]) while receiving maintenance infliximab therapy (Figure 1).Of the nine patients who experienced dermatological events, two (22.0%) had psoriasis and seven (77.8%) had persistent nonpsoriatic skin eruptions that had developed while on infliximab and were attributed to infliximab.Of the nine patients with infusion reactions, five (55.6%) experienced acute infusion reactions and four (44.4%) experienced delayed infusion reactions.
Comparing patients who experienced adverse events with those who did not, the median age was similar (Table 2).There were more women (12 of 18 [66.7%])than men (six of 18 [33.3%]) in the group with adverse events (P=0.046).The proportions of CD and UC were similar between groups, with 62.3% (33 of 53) of patients without adverse events and 72.2% (13 of 18) of patients with adverse events having a diagnosis of CD, and 37.7% (20 of 53) of patients without adverse events and 27.8% (five of 18) of patients with adverse events having a diagnosis of UC (P=0.445).

ITLs and infliximab-associated adverse events
The median infliximab dose at the time of trough collection was 5.3 µg/mL (IQR 4.8 to 5.7 µg/mL), with a median time since last infusion of 7.0 weeks (IQR 4.1 to 8.0 weeks).The median time to infliximabassociated adverse event was 38.2 weeks (IQR 18.0 to 94.7 weeks) and the median time to censorship was 73.1 weeks (IQR 34.0 to 143.1 weeks).As shown in Table 2, the proportion of patients who experienced dermatological or infusion reactions was similar among those receiving infliximab <6 weeks interval (31.0%) and those receiving infliximab at ≥6 weeks interval (21.4%) (P=0.360).
As shown in Figure 2A, patients who experienced any infliximabassociated adverse event during stable maintenance infliximab therapy had a median ITL of 7.2 µg/mL (IQR 2.0 to 13.3 µg/mL), which was similar to the median ITL of patients who experienced no adverse events of 6.6 µg/mL (IQR 3.2 to 12.7 µg/mL) (P=0.648).Dermatological adverse events: As shown in Figure 2B, patients who experienced infliximab-associated dermatological adverse events had a median ITL of 13.3 µg/mL (IQR 8.8 to 17.4 µg/mL) while the patients who experienced no adverse events had a median ITL of 6.6 µg/mL (IQR 3.2 to 12.7 µg/mL) (P=0.058).The ITL of the two infliximabinduced psoriasis cases were 9.9 µg/mL and 13.3 µg/mL.Infusion reaction adverse events: In contrast, Figure 2C shows that patients who experienced infliximab-associated infusion reactions had a significantly lower median ITL of 2.0 µg/mL (IQR 0.1 to 5.7 µg/mL) compared with the median ITL in those who experienced no adverse events of 6.6 µg/mL (IQR 3.2 to 12.7 µg/mL) (P=0.008).Separating the infusion reactions into acute and delayed, the median ITL of those who experienced acute infusion reactions was lower (0.60 µg/mL [IQR 0.20 to 3.1 µg/mL]) compared with those who experienced delayed infusion reactions (3.5 µg/mL [IQR 1.0 to 6.9 µg/mL]), but the difference did not reach statistical significance (P=0.611).However, the ITL in those who experienced acute infusion reactions (0.60 µg/mL [IQR 0.20 to 3.1 µg/mL]) was significantly lower than the median ITL of those who experienced no adverse reactions (6.6 µg/mL [IQR 3.2 to 12.7 µg/mL]) (P=0.0167).The median ITL in those who had delayed infusion reactions (3.5 µg/mL [IQR 1.0 to 6.9 µg/mL]) was slightly lower than the median ITL of those who had no adverse reactions (6.6 µg/mL [IQR 3.2 to 12.7 µg/mL]), but the difference did not reach statistical significance (P=0.163).
Finally, as shown in Figure 2D, patients who experienced any infusion reactions had a significantly lower ITL, with a median of 2.0 µg/mL (IQR 0.1 to 5.7 µg/mL), compared with patients who experienced dermatological events, with a median of 13.3 µg/mL (IQR 8.8 to 17.4 µg/mL) (P<0.001).

DISCUSSION
Infliximab is one of the cornerstone therapies for refractory CD and UC.
In the present study, we demonstrated that more than one-quarter of IBD outpatients receiving stable maintenance infliximab therapy experience infliximab-associated dermatological or infusion reaction adverse events.In addition, we have shown a correlation between ITLs and the type of infliximab-associated adverse event: patients with dermatological adverse events had significantly higher ITLs, while patients with infusion reactions had significantly lower ITLs compared with patients who did not experience adverse events.

Dermatological adverse events
Paradoxical skin eruptions of eczema and psoriasis are common in patients receiving anti-TNF therapy, with up to 39% of reported cases of anti-TNF-induced psoriasis occurring in IBD patients (5,(25)(26)(27).Among studies specifically investigating IBD patients treated with anti-TNF therapies, the main focus has been on the induction or exacerbation of psoriasiform lesions, with reported incidence rates between 5% and 22% (1,4).The literature varies due to different inclusion criteria and definition of the skin lesions of interest.
In our study, we included both inflammatory skin lesions and any de novo skin eruptions that occurred during stable maintenance infliximab therapy, and estimated a prevalence of 12.7% for dermatological adverse events and 2.8% for anti-TNF-induced psoriasis.In a retrospective study conducted by Groupe d'Etude Thérapeutique des Affections Inflammatoires du tube Digestif (GETAID) centres in France, Rahier et al ( 12) estimated an incidence of 5% for inflammatory skin lesions, with 2% for psoriasiform lesions and 3% for eczematiform lesions.In a systematic analysis conducted in Madrid (Spain), Guerra et al (28) diagnosed 21 of 1294 IBD patients with anti-TNF-α-induced psoriasis, 14 (67%) of whom were being treated with infliximab; they reported a cumulative incidence of 1.62% (95% CI 1.06% to 2.47%).Of the 21 patients, 15 were female and 17 had CD.Eighteen patients developed new-onset psoriasis and, in virtually every case, the onset of psoriasis occurred during maintenance anti-TNF therapy.A systematic literature review of 69 published cases of IBD patients with infliximabinduced psoriasis, performed by Denadai et al (23), and a case series of 30 IBD patients with anti-TNF-related psoriasiform lesions conducted by Cullen et al (29) yielded similar findings with female sex and CD predominance.Similar to these studies, our study also showed a female sex and CD predominance among patients who experienced dermatological adverse events.These inflammatory skin lesions can be treated topically; however, up to 43% of patients require withdrawal of anti-TNF therapies due to uncontrolled skin lesions (12,24,29).
Although the underlying mechanism of infliximab-induced psoriasis is unknown, there is general consensus in the literature regarding a possible role for alterations in cytokine levels such as interferon (IFN)-α (1,2,24).It is proposed that anti-TNF-α agents, such as infliximab, decrease TNF-α inhibition of IFN-α production (23).Tillack et al (24) histologically examined psoriasiform skin lesions in anti-TNF treated IBD patients, and characterized them as having infiltrates with INF-α-expressing cells.Consequently, the abnormally high levels of IFN-α resulting from decreased TNF-α inhibition can initiate psoriatic lesions.Although we did not measure cytokine levels in our study, the ITLs of patients who expereinced dermatological adverse events were higher than the ITLs of patients who did not.
Regarding dermatological adverse events, a limitation of the present study was the inclusion of skin eruptions documented by patient self-completed questionnaires; however, these data were only included if reported at least twice by patients, and confirmed by the gastroenterologist and, when possible, by a dermatologist, to have developed during infliximab therapy and attributed to infliximab therapy.
Based on our study findings and the pathophysiological mechanisms proposed in the literature, a management strategy for patients presenting with difficult-to-control infliximab-associated dermatological adverse events is to first measure ITLs.If the patient is clinically in remission and, if the ITLs are high, the infliximab dose should be lowered accordingly.If this dose adjustment and conservative measures are ineffective, then switching to another therapy may be required.

Infusion reactions
Infusion reactions to infliximab have been reported to occur in 5% to 10% of all infusions among IBD patients (30,31).When Steenholdt et al (32) examined 25 (8%) of 315 IBD patients who experienced acute severe infusion reaction to infliximab, they found that IgG ATI levels were highly positive in 19 of 20 patients (95%) after the reactions, but that IgE ATI was negative in all patients with reactions (32).The importance of ATI is emphasized in the systematic review and meta-analysis by O'Meara et al (33), which estimated risk ratios for any acute infusion reaction (RR 2.4 [95% CI 1.5 to 3.8]; P<0.001) and severe infusion reactions (RR 5.8 [95% CI 1.7 to 19]; P=0.004) that are higher in ATI-positive compared with ATInegative patients.
Delayed infusion reactions are believed to be type III immune complex-mediated reactions with formation of antigen-antibody complexes that deposit in the tissues (8).Interestingly, ATI may not be as important in the mechanism of action for delayed infusion reactions because O'Meara et al (33) showed that the RR of delayed hypersensitivity reactions is not significantly different between ATI-positive and ATI-negative patients (RR 2.8 [95% CI 0.2 to 33]; P=0.4) (33).
In our study, patients who experienced delayed infusion reactions had slightly higher ITLs than patients who experienced acute infusion reactions, and slightly lower ITLs than patients who had no adverse events.Interestingly, patients who experienced acute infusion reactions had statistically significantly lower ITLs than patients who had no adverse events.
Regarding infusion reactions, a limitation of our study was the lack of antinuclear, anti-double-stranded DNA, antihistone, IgE levels, and of ATI levels in patients who developed infusion reactions.These measurements may have been beneficial to identify patients with drug-induced lupus, and also to confirm whether reactions occurred due to IgE-mediated versus ATI mechanisms.

Figure 2) Infliximab trough level was correlated with type of infliximabassociated adverse event. No adverse event (n=53), experience adverse event (n=18) (A); dermatological adverse event (n=9) (B); infusion reaction (n=9) (C); infusion reactions versus dermatological adverse events (D). Median, 25th percentile to 75th percentile, minimum and maximum values for infliximab trough levels (μg/mL) are plotted
A limitation of the present study was the small number of dermatological and infusion reactions to infliximab.However, the prevalence of these adverse events is similar to that reported in previous studies, as mentioned above.
Another limitation was that at the time of the present study, we only had access to ELISA-based ITLs, but not to ATI levels.However, given the finding that the patients who experienced infusion reactions had lower ITLs, it would be important to determine whether they have ATIs.Both the ITL and the ATI level would be important to know in patients who develop adverse events such as infusion reactions to infliximab.Therefore, for future research studies, and also for clinical practice, it would be important to measure ATIs in patients who experience infusion reactions or who have low ITLs.
A strength of the present study was that it prospectively analyzed ITLs in all patients receiving stable maintenance infliximab therapy who consented to participate during the study period, regardless of whether they experienced dermatological or infusion reactions.This is in comparison with study designs that analyze ITLs in patients who have levels drawn for specific reasons (eg, loss of response or adverse events) because the latter study design would introduce selection bias into the analysis.
Based on the results of the present study, we propose a strategy to manage infusion reactions by measuring ITLs in patients who develop infusion reactions; if the ITL is low, ATIs should be measured with consideration to switch drugs if the patient has a high level of ATI.

TABle 1 Demographics of inflammatory bowel disease outpatients receiving stable maintenance infliximab therapy at the Infliximab Infusion Clinic, University of Alberta, edmonton, Alberta (n=71)
Data presented as n (%) unless otherwise indicated.IQR Interquartile range