Infliximab (IFX, Remicade®
) is a chimeric monoclonal IgG1
antibody against tumor necrosis factor (TNF), a central cytokine in inflammatory bowel disease (IBD). This drug is effective in inducing and maintaining remission in Crohn’s disease (CD) and ulcerative colitis (UC), the two principal entities of IBD [1
Baert et al. identified the potential immunogenicity of IFX in IBD patients, suggesting that patients may develop antibodies toward IFX (ATI), especially with episodic administration of the drug [4
]. However, 6%–17% of the patients develop ATI even with scheduled treatment [5
]. The association of ATI and infusion reactions is clear, and several studies demonstrate significantly lower trough serum IFX (s-IFX just before the next infusion) in patients with ATI, explained by increased elimination of IFX [4
]. Low s-IFX and ATI formation have been associated with loss of response (LOR), but it has been difficult to establish any significant correlation between ATI and clinical parameters other than LOR as shown in a meta-analysis including 10 studies and 668 patients [4
]. Treatment response seems to be more related to drug levels rather than to ATI, and a clear correlation has been found in several reports between trough s-IFX and clinical remission, C-reactive protein (CRP), and endoscopic improvement in adult patients [8
The role of monitoring s-IFX and ATI, therapeutic drug monitoring (TDM), in the clinic is unclear, and the level of evidence is low given a limited number of studies with small cohorts, the use of retrospective designs, and different methodological approaches [10
]. During induction of IFX, low trough levels have been suggested to support dose escalation in case of poor response [16
]. In patients with loss of response, TDM may support dose escalation or switch to another drug [7
]. A majority of the TDM studies have been performed in adult patients, and even though there are findings suggesting different pharmacodynamics and kinetics in children, very few studies have been undertaken to assess trough levels and ATI in pediatric cohorts [20
]. The current study includes 45 children with CD or UC on maintenance IFX treatment. The study aims at correlating clinical activity and response to treatment with s-IFX levels and ATI in children.
Forty-five children receiving IFX maintenance treatment were included and patients contributed with one sample for each visit with a total of 93 specimens; 15 patients contributed with one sample, 19 patients with two samples, four patients with three samples, and seven patients with four samples. The median age of this pediatric cohort was 16.0 years (range 7–18). CD was diagnosed in 32 patients (71%) and 13 patients (29%) had UC. In the anti-TNF-treated CD cohort, 27 patients (84%) had colonic or ileocolonic inflammation and 15 children (47%) also had involvement of upper gastrointestinal tract (mucosal inflammation found proximal to the ligamentum Treitz) at diagnosis. A proportion of 84% showed an inflammatory phenotype without stricturing or penetrating disease (B1). Only four patients (13%) had penetrating phenotype (B3) and seven children (22%) presented with perianal disease (P). Among the UC patients, 11/13 (85%) had extensive colitis or pancolitis (Table 1
The duration of the IFX treatment ranged from 3 to 60 months, and the children had received a mean number of 13 IFX infusions (range 4–48). The mean dose of IFX ± standard deviation (SD) was 6.4 ± 1.7 mg/kg (median 6.2 mg/kg, range 3.44–10.5) with a mean interval of 44.8 ± 11.2 days. The mean s-IFX trough level was 5.2 µg/mL (median 4.5 µg/mL; range from <0.2 to 21), showing a right-shifted Gaussian distribution, as seen in Figure 1
. One CD patient in remission with s-IFX 40 µg/mL was excluded from the analysis.
The assessment of disease activity was based on the validated scoring indices Pediatric CD Activity Index (PCDAI) and Pediatric UC Activity Index (PUCAI). The children were in clinical remission at 44 out of 93 visits (47%). With a stricter definition of remission using a combination of low clinical scoring and normalized C-Reactive Protein (CRP, mg/L) and Erytrocyte Sedimentation Rate (ESR, mm/h), the patients were in remission at 26 of the 93 test occasions (28%). Nine children were in remission at all visits, while 28 children were not in remission at any visit (10 of these non-remitters had only one visit). The clinical indices and biochemistry are summarized in Table 2
. As shown in Figure 2
, s-IFX was significantly higher in samples taken during remission (mean 7.2) as compared with sera collected during active disease (mean 4.5 µg/mL, p
< 0.05). No significant difference was observed in dose-interval (days) between patients in active disease and those in remission (mean 43.0 days in active disease vs. mean 42.7 days in remission, p
= 0.88) or in mean dose of IFX between the children in active disease (6.4 mg/kg) and those in remission (6.5 mg/kg, p
The trough levels indicated a statistically significant correlation with clinical indices, as well as with CRP, ESR, and albumin levels, as illustrated in Figure 3
a–d. No correlation was detected between trough levels and Fecal Calprotectin (FCP, mg/kg), and no difference was noted in s-IFX trough levels between CD and UC (not shown).
Interestingly, intra-individual variations in dosing (mg/kg/infusion interval in days) did not show any clear correlation with changes in trough levels (µg/mL), as seen in Figure 4
. Dose changes were not planned within the framework of the study, and were prescribed by the treating physician at his or her own discretion and not necessarily based on trough levels of s-IFX. The dose variations were also due to changes in weight. In the 30 patients who supplied two to four tests (total 48 samples) we identified 18/48 (38%) tests with decreased dose (mean −22%, SD ±13%), and in 11 of these 18 samples s-IFX decreased. Of these 18 samples, seven represented children in remission. At 22/48 (46%) test occasions, there was a dose increase (mean +44%, SD ±32%), and 15/22 tests showed subsequent increased s-IFX. Of these 22 samples, 17 tests represented patients with active disease.
In 12 samples from eight children (seven with CD and one with UC) collected at different sampling occasions, s-IFX trough levels were below detection and all of these samples were positive for ATI. None of the eight children were in remission at the time of the ATI positive samples. In six additional patients s-IFX was detectable, but below 1.0 µg/mL, giving a total of 14 patients with s-IFX of <1.0 µg/mL. All but one of these 14 patients showed active disease (CRP ≥ 5, ESR ≥ 10, and/or PCDAI ≥ 10 or PUCAI ≥ 10).
Of 14 patients with s-IFX of <1.0 µg/mL only two children had concomitant immunosuppressives during maintenance treatment. In the whole set of 93 s-IFX trough samples, 28 (30%) were collected at the time of concomitant immunosuppression. Mean trough IFX in these samples was 6.5 µg/L (0.2–21) compared with 4.8 µg/L (0.2–14) in samples from patients on monotherapy (n = 65, 70%) without reaching a significant difference between the two groups.
The present study investigated s-IFX trough levels and ATI in a cohort of 45 pediatric IBD patients comprising IFX-treated children in the counties of Stockholm and Västmanland. S-IFX trough concentrations showed a significant correlation with clinical response and inflammatory activity. The intra-individual variations in the trough levels between visits were evident, and there was no clear correlation with dose changes, as seen in Figure 4
. Low s-IFX trough levels were associated with the formation of ATI.
To our knowledge, there are only a few reports on IFX trough levels in children. In our study, we found a relatively high trough level (mean 5.2 µg/mL, median 4.5 µg/mL), even though within the proposed therapeutic interval 3–7 µg/mL for adults [18
]. Hämäläinen et al. and Hoekman et al. both reported a median s-IFX of 3.5 µg/mL in a mixed pediatric UC and CD cohort, whereas Adedokun et al. reported a median s-IFX of 1.9 µg/mL at week 30 and 2.6 µg/mL at week 46 in a UC population receiving 5 mg/kg IFX q8w [22
]. In the latter study, even a double dose of 10 mg/kg only gave a median trough of 2.9 µg/mL [24
In our study, the children who responded to IFX and who were in clinical remission based on combined clinical index activity and biomarkers presented a significantly higher mean trough concentration compared to non-remitters (7.2 vs. 4.5 µg/mL). Previous reports have not shown consistency with regard to the correlation between serum IFX levels and clinical activity indices/biomarkers [22
]. In adults, the correlation between pre-infusion s-IFX and clinical response is well established [11
Surprisingly, the children were in clinical remission, defined as a reduction in the clinical activity index as well as normalization of CRP/ESR, only at 26 of the 93 visits during maintenance treatment. This finding should be interpreted with some caution in the light of the restrictive definition of remission in this study, including clinical indices as well as normalization of biomarkers. Previous reports on IFX and clinical response in children have been divergent, and some studies suggest high rates of LOR up to 50% in children [20
]. The finding of only 28% remission rate based on the 93 recorded visits was surprising, especially in light of the relatively high mean trough level of 5.2 µg/mL in this report. A sub-study of the pivotal Crohn trial with IFX (the ACCENT I trial) revealed median week 14 trough levels of 4.0 µg/mL in patients with sustained response to IFX 5 mg/kg and 1.9 µg/mL in patients without sustained response [30
We could not detect any clear impact of dose changes on the s-IFX levels. In this study, the s-IFX is a mixture of tests taken in the clinic by the treating physician, as well as tests obtained within the current study. Therefore, not every decision to change dosing has been based on s-IFX levels. Moreover, since the dosing also depends on weight and infusion interval, the variation may not be a result of active dosing decisions.
ATI was found in all patients with undetectable s-IFX, and all these patients had active disease. This observation may suggest that ATI plays an important role in children with low trough levels and incomplete response to the treatment. Yet, a major limitation in the enzyme-linked immuno sorbent assay (ELISA) analysis of ATI is the inability to detect antibodies in the presence of IFX residue, which makes it difficult to speculate about the role of ATI in children exhibiting low but detectable trough levels. Vande Casteele et al. investigated the relationship between IFX concentrations, ATI, and disease activity in 1487 IFX trough serum samples from 483 adult CD patients [26
]. Their method allowed for the analysis of ATI in the presence of IFX, showing that ATI even at low as well as therapeutic concentrations of IFX increased the probability of active disease. It is conceivable that the same condition prevails in the pediatric population. The combination therapy with immunosuppressives is thought to reduce the frequency of ATI. We found a numerical trend toward a higher s-IFX concentration in children on azathioprine, but without reaching statistical significance, probably due to the small number [4
]. Children were on concomitant immunosuppressant only at 30% of the visits during maintenance treatment. This could probably reflect the fear for hepatosplenic T-cell lymphoma in young male patients [31
Whether TDM is a valuable tool to maintain remission in patients with IBD is not clear. In a randomized controlled study including 263 adult IBD patients dosing was optimized for IFX trough levels 3–7 µg/mL at the start and the patients were then randomly assigned to either continued TDM-based or clinically based dosing [32
]. The study showed no superiority of concentration-based dosing after one year with regard to remission rates. However, TDM-based dosing was associated with fewer flares during the course of treatment. Furthermore, a retrospective study examining the use of proactive TDM in 48 adult patients compared to 78 patients with standard care demonstrated a higher probability of remaining on IFX in patients with s-IFX trough levels >5 µg/mL [33
]. In a study by Minar et al., TDM was evaluated in IFX-treated children with CD who experienced LOR. The authors found that ESRs at the previous infusion were significantly associated with IFX concentrations [23
]. The current report is of observational and retrospective nature. Interventions were not implemented within the study and therefore conclusions with regard to the role of TDM-based IFX dosing are beyond the scope of this report. Nevertheless, in this cohort with children not subjected to active TDM, only one third was in clinical and biochemical remission.