Early Serum Infliximab Levels in Pediatric Ulcerative Colitis

Background: Data on serum infliximab concentrations during induction in pediatric ulcerative colitis are limited. The study aim is to evaluate the relationship between serum infliximab concentrations during induction and short-term clinical remission in children with ulcerative colitis. Methods: We carried out a prospective, multi-center cohort study in pediatric patients with ulcerative colitis. Serum infliximab concentrations were collected at peak dose #1, week 1, trough pre-dose #2, and trough pre-dose #3. Infliximab dosing was left to investigator discretion. Clinical remission was defined by pediatric ulcerative colitis activity index <10 at week 8. Results: Twenty-four of thirty-four subjects (71%) achieved clinical remission at week 8. The median infliximab concentrations were 33.0 μg/mL (interquartile range: 26.5–52.1 μg/mL) pre-dose #2 and 22.5 μg/mL (interquartile range:15.9–32.3 μg/mL) pre-dose #3. Trough pre-dose #2 infliximab concentration yielded area under receiver operator characteristic curve 0.7, 95% CI: 0.5–0.9 in predicting week 8 clinical remission; a cut-off of 33.0 μg/mL yielded 62.5% sensitivity, 66.7% specificity. Trough pre-dose #3 infliximab concentrations were lower for subjects <10 years compared to ≥ 10 years [median 15.9 μg/mL, interquartile range (IQR) 8.5–21.8 μg/mL vs. 27.7 μg/mL, IQR 17.2–46.7 μg/mL, p = 0.01] and correlated with baseline weight (Spearman's rank correlation coefficient 0.45, p = 0.01). The median half-life following first IFX dose was 6.04 days (IQR 5.3–7.9 days). Conclusions: Infliximab concentrations ≥33 μg/mL prior to the second dose were associated with week 8 clinical remission. As young age and low body weight impact infliximab concentration, prospective studies with proactive adjustment in pediatric patients with ulcerative colitis should be carried out. Clinicians caring for children with UC should diligently adjust and monitor infliximab to optimize response.


INTRODUCTION
Infliximab (IFX) is a monoclonal antibody against tumor necrosis factor α (TNF-α) used in adult and pediatric ulcerative colitis (UC) that is initiated in up to 31% of newly diagnosed children with UC by 52 weeks of follow-up (1)(2)(3)(4)(5). However, in children with moderate to severely active UC who commence IFX, 26.7% fail to respond to induction and up to 39% have inadequate initial response or lose response and require colectomy by 24 months (2,6). Therefore, strategies to optimize IFX response are still desired.
Numerous factors affect pharmacokinetics of IFX. In adults with UC, higher body weight, lower serum albumin, antibodies against IFX (ATI), and female sex accelerated clearance (7). In children with UC who may have differing weights and disease burden compared to adults, limited data are available. In a single study, age and baseline use of concomitant immunomodulators did not influence IFX pharmacokinetics and the effects of weight, serum albumin, ATI, and sex were not evaluated (8). A recent study of children with IBD suggested that young age may further affect IFX pharmacokinetics; the majority of young children (age <10 years) had inadequate IFX trough levels at the start of maintenance and were more likely to have an intensive treatment regimen for the first year of IFX treatment compared to older children (9).
Serum IFX concentrations correlate with clinical and endoscopic outcomes in UC (10). Limited studies have evaluated IFX levels during induction of active UC, a distinctive disease state with high inflammatory burden. In adults with UC, IFX concentration thresholds during induction are associated with endoscopic response and short-term mucosal healing (11,12). In children with UC, there are limited data on IFX concentrations during induction. In this study, we evaluated the relationship between serum IFX concentrations during induction and short-term clinical remission in children with UC. We also examined potential factors associated with IFX concentrations, pharmacokinetics, and achievement of shortterm clinical remission.
Subjects received IFX (Remicade R ; Janssen Biotech, Inc., Horsham, Pennsylvania, USA) with a typical induction of IFX 5 mg/kg at weeks 0, 2, and 6; this could be adjusted at clinician's discretion. All doses and dates were recorded. Intensified IFX schedule was defined as IFX dose #3 ≥ 7 mg/kg and/or time from dose #1 to #3 ≤35 days. This stringent definition acknowledged practice patterns of rounding up IFX dose to nearest hundred and accounted for scheduling issues.

Serum Infliximab and Antibody to Infliximab Concentrations
Serum IFX and ATI concentrations were collected at peak dose #1 (1-hour post-infusion), week 1, trough pre-dose #2, and trough pre-dose #3. Sera were processed at Prometheus Laboratories Inc., San Diego, California using the drug-tolerant PROMETHEUS R Anser R IFX liquid-phase mobility shift assay. The lower limit of detection was 0.98 µg/mL for serum IFX and 6.6 U/mL for ATI with no upper limit of quantification. Physicians were blinded to results.

Outcome
At week 8, subjects underwent PUCAI assessment. The primary outcome was clinical remission (PUCAI < 10 at week 8, in absence of colectomy) (16). A secondary outcome was clinical response at week 8 (at least 50% reduction in PUCAI from baseline).

Statistical Analysis
Data are presented as medians (interquartile range [IQR]). Categorical data were compared using χ 2 or Fisher's exact test and continuous data using the Student t-test, Wilcoxon rank sum test, or Wilcoxon signed-rank test, as appropriate.
Receiver operating characteristic (ROC) curves were used to assess the diagnostic ability of models of serum IFX concentrations at week 1, pre-dose #2, and pre-dose #3 in predicting clinical remission at week 8. Area under ROC (AUROC) values were interpreted as AUC <0.70 corresponding to low diagnostic accuracy; 0.70-0.90 = moderate diagnostic accuracy; and AUC ≥ 0.90 = high diagnostic accuracy. The clinical utility of a range of pre-dose #2 and pre-dose #3 IFX levels are shown reporting sensitivity, specificity, positive predictive value, and negative predictive value. The Youden Index was used to select an optimal threshold IFX concentration in models that showed moderate diagnostic ability. ROC curves were also used to assess the diagnostic ability of models of serum IFX concentrations at week 1, pre-dose #2, and pre-dose #3 in predicting CRP remission (<10.0 mg/L) and ESR remission (<10.0 mm/h) week 8.
Univariate logistic regression analysis was performed to evaluate association of clinical and laboratory variables with week 8 clinical remission. The variables evaluated included: male gender, young age (<10 years, based on Paris Classification) (15), high CRP (>10.0 mg/L), high ESR (>10.0 mm/h), low albumin (<33 g/L), concomitant immunomodulator, body mass index (BMI), and intensified IFX schedule. This analysis was repeated for outcome of steroid-free clinical remission, and achievement of clinical response and/or remission. The CRP, ESR, and albumin interpretation were based on the upper and lower limits of normal from participating centers' laboratories.
If an AUROC demonstrated moderate diagnostic ability, then univariate logistic regression analysis was performed to evaluate association between the above variables and achievement of the cut-off serum IFX concentration, according to the above analysis.
The Kruskal-Wallis test was performed to determine if there were statistically significant differences between serum IFX concentrations pre-dose #2 among categories differentiated by: male gender, young age (<10 years, based on Paris Classification) (14), high CRP (>10.0 mg/L), high ESR (>10.0 mm/h), low albumin (<33 g/L), concomitant immunomodulator, body mass index (BMI), intensified IFX schedule, and IFX dose #1 ≥ 7 mg/kg. A similar analysis was performed for serum IFX concentrations pre-dose #3 using the above variables with the addition of dose #3 ≤ 35 days from dose #1. By determining Spearman's rank correlation coefficient, we evaluated for association between serum pre-dose #2 and pre-dose #3 IFX concentrations and age, baseline laboratory parameters (CRP, ESR, albumin), body mass index, and baseline infliximab dose (mg/kg).
We calculated the half-life associated for dose #1. We compared IFX dose, regimen, concentrations, and half-life by age [categorical variable (young age <10 years) (15), Wilcoxon ranksum test] and by anthropometric measures and age [continuous variables, Spearman's rank correlation coefficient].
Results were considered statistically significant when P < 0.05. We aimed to enroll 32 children for this pilot study to evaluate early induction IFX concentrations to guide future prospective studies implementing strategies to optimize IFX response in pediatric UC. Subjects lost to follow-up were excluded from overall analysis.
Data were entered into Research Electronic Data Capture (REDCap). Statistical analysis was performed using Intercooled Stata 13 software (Stata Corporation, College Station, Texas).

Ethics
This study was approved by the institutional review board at each institution. All participants provided written consent.
The IFX induction regimen is presented in Table 2 and Supplementary Figure 1. Eleven subjects started with dose #1 ≥ 7 mg/kg (32.4%). During induction, five subjects had dose escalation for breakthrough symptoms (two at dose # 2; three at dose #3) and five subjects had frequency shortening (one received dose #2 at < 12 days from dose #1; four received dose #3 at < 21 days from dose #2). In total, 13 subject were classified as intensified IFX induction (nine based on IFX dose #3 ≥ 7 mg/kg alone, one based on time from dose #1 to #3 ≤35 days alone, and three based on both dose and time).
Thirty subjects were on corticosteroids at baseline. Two subjects started or increased prednisone during the study period. At week 8, ten patients remained were on corticosteroids, of which nine were on a weaning regimen with dose ≤ 20 mg/d; eight of these nine patients on weaning corticosteroids were in clinical remission.

IFX Concentrations and Week 8 Outcomes
Serum IFX concentrations are shown in Table 2 and Figure 1.

Anti-drug Antibodies
One subject who was on concomitant azathioprine during the entire study period developed ATI (9.2 U/mL) with undetectable serum IFX concentrations (<1.0 µg/mL) at pre-dose #3. This subject developed clinical symptoms prior to dose #3, received systemic corticosteroids and dose #3 escalation, and ultimately achieved clinical remission by week 8.

Evaluation of Predictors of Week 8 Outcomes
In a univariate analysis, achievement of clinical remission at week 8 was not associated with male gender, young age, high baseline CRP, low baseline albumin, BMI, or intensified IFX schedule (Supplementary

Association of Young Age and Low Baseline Weight With Lower IFX Concentration at Week 6
Pre-dose #3 IFX concentration differed by age (Figure 3) [median 15.9 µg/mL, IQR 8. Pre-dose #3 IFX concentration correlated with baseline weight (Spearman's rank correlation coefficient 0.45, p = 0.01) but did not correlate with baseline BMI (Figure 4). IFX concentrations at week 1 and pre-dose #2 did not correlate with baseline weight or BMI.

Infliximab Half-Life
Twenty subjects provided sufficient measurements to calculate IFX dose #1 half-life. The median half-life was 6.04 days (IQR 5.3-7.9 days); there was no difference between subjects in clinical remission at week 8 (median 6.68 days, IQR 5.35-9.03 days) vs. not in clinical remission (5.50 days, IQR 4.97-5.92 days) (p = 0.2). IFX half-life did not correlate with age; however only three of these 20 subjects were <10 years of age. IFX half-life did not correlate with baseline weight, baseline BMI, or dose #1 amount (mg/kg).

DISCUSSION
In this prospective multicenter cohort study of children with UC, we evaluated early IFX concentrations during IFX induction. Our results demonstrate that the majority (70%) of children with UC who undergo IFX induction achieve clinical remission by week 8. We demonstrate that pre-dose #2 IFX concentrations have moderate diagnostic ability to predict week 8 clinical clinical remission, with an IFX concentration ≥33.0 µg/mL yielding best test characteristics, including PPV 83%. Though there was a trend toward higher pre-dose #2 and pre-dose #3 serum IFX concentrations for subjects in clinical remission at week 8 compared to subjects not in clinical remission, this did not reach statistical significance, potentially due to larger required sample size to demonstrate this difference.
IFX induction concentrations have been described in adults with UC. Papamichael et al. demonstrated in a retrospective, single-center study of adults with UC similar serum IFX concentration thresholds associated with short-term mucosal healing: 28.3 µg/mL (AUROC 0.64, 95% CI 0.53-0.75) at week 2, and 15.0 µg/mL (AUROC 0.69, 95% CI 0.59-0.79) at week 6 (11). Post-hoc analysis of ACT-1 and ACT-2 data in adults with UC demonstrated that achievement of serum IFX concentration threshold 22 µg/mL at week 6 was associated with clinical response at week 8 (17). In both our study and the aforementioned study by Papamichael et al. (11) the AUROC yields only borderline moderate diagnostic ability, possibly suggesting that IFX exposure is overall generally adequate by using an induction weight-based strategy with doses provided in frequent intervals. However, assembling a larger sample size of patients with variability in severity of disease, age, weight, dose, and timing may provide an improved understanding on how these variables contribute to IFX exposure and therapeutic response.
The currently approved induction regimen for IFX is 5 mg/kg at weeks 0, 2, and 6. However, concerns have risen for higher inflammatory burden and fecal IFX loss in active UC, resulting in lower serum IFX concentrations and therefore requiring an intensified schedule to achieve adequate concentrations. Seow et al. demonstrate that 56% of adults with UC had a low serum IFX concentration (median 2.1 µg/ml) at 4 weeks and undetectable serum IFX concentration at 6 weeks (10). Ungar et al. demonstrate lower IFX trough serum concentrations at 2 weeks (mean 7.15 ± 5.3 µg/mL) for acute severe UC compared to moderately severe UC (mean 14.4 ± 11.2 µg/mL) (19). Church et al. retrospectively evaluated an intensified induction (mean induction dose ≥7 mg/kg or interval between dose #1 and dose #3 ≤5 weeks) in steroid-refractory pediatric UC patients; an intensified induction was associated with higher chance of clinical remission (hazard ratio 3.2, p = 0.02) and lower chance of colectomy (hazard ratio 0.4, p = 0.05) compared to standard regimen (20). A systematic review by Hindryckx et al. identified uncontrolled studies that suggest a benefit for intensified induction for adults with acute severe UC, primarily in reducing early (3-month) colectomy rates (21). At the time of our study conduct, intensified IFX induction was intermittently performed. Nine subjects received dose #3 ≥ 7 mg/kg, 1 subject received dose #3 ≤ 35 days, and 3 received both [intensified IFX=13 subjects (38%)]. Neither interval shortening nor increased IFX dose were associated with reaching pre-dose #3 IFX threshold or week 8 clinical outcomes. The incremental benefit of IFX intensification on short and long term clinical, biomarker, and endoscopic outcomes requires further evaluation in a prospective study.
Evaluating factors impacting serum IFX concentrations, subjects <10 years and subjects with lower baseline weight had lower pre-dose #3 serum IFX concentrations. Adedokun et al. evaluated the pharmacokinetics of IFX in children with UC and demonstrated that median serum IFX concentrations at several time points during induction and maintenance were ∼20% lower in children than adults (8). Using integrated data from pediatric patients with UC, Crohn's disease, juvenile rheumatoid arthritis, and Kawasaki disease, Xu et al. showed through population pharmacokinetic analysis that IFX clearance showed non-linear dependence on body weight and that systemic IFX exposure in very young children (aged 2 to 6 years) was predicted to be ∼40% lower compared to adults (22). Therefore, the discrepancy between a linear dosing regimen of 5 mg/kg and a nonlinear relationship between body weight and IFX clearance may predispose younger children to insufficient IFX exposure. This is affirmed by a recent study by Jongsma et al. which evaluated IFX trough levels in 110 young children (age < 10 years) and 105 older children (age ≥ 10 years) (9). The authors showed that 72% of young children had inadequate IFX trough levels at the start of maintenance (<5.4 µg/mL) and young children required higher dose per 8 weeks compared to older children for the first year of IFX treatment [young children; 9.0 mg/kg (IQR 5.0-12.9) vs. older children; 5.5 mg/kg (IQR 5.0-9.3); p < 0.001] while maintaining similar overall duration of response to IFX. To affirm the uniqueness of pediatric pharmacokinetics, the median half-life in our study following IFX dose #1 in our study was 6.04 days (IQR 5.3-7.9 days) in comparison to reported median half life for IFX ∼14 days (23). Further studies are required to explore optimal dosing regimen for children of young age and/or lower body weight to better define this patient population and ensure adequate IFX exposure to optimize response.
In our study, the majority of children achieved clinical remission at week 8 along with statistically significant improvement in serum biomarkers of inflammation. However, fecal calprotectin continued to remain elevated at week 6, likely suggesting that mucosal healing lags after clinical remission. This is consistent with results from a prospective multicenter cohort study of children with UC which demonstrated that PUCAI is more responsive to change in short time periods of re-evaluation than fecal calprotectin in severe UC (24). In a study of 36 children with Crohn's disease and UC, though 37% had normal fecal calprotectin <100 µg/g at 2 weeks following IFX start, only 8.3% had fecal calprotectin <100 µg/g at 6 weeks (25).
The strengths of our study center on the real world, multicenter, prospective nature of our study and the contribution of IFX trough levels during induction in pediatric UC to the very limited existing body of evidence. We acknowledge the limitations of our study. We selected a clinical outcome, PUCAI, as the endpoint of our study. Indeed, mucosal healing is a gold standard. However, in children, re-evaluation endoscopy is challenging as it is an invasive procedure often requiring anesthesia. Therefore, we selected a clinical outcome with strong validity to optimize feasibility; a PUCAI-defined remission has high concordance with complete mucosal healing at week 8 and PUCAI scores post-induction predict 1-year steroid-free remission (26). We recognize the importance of biomarker remission as a on outcome; however adherence to stool collection for fecal calprotectin was poor. Though this was included in our study protocol, only 7 subjects provided a complete series of three stool samples. Another limitation is the short duration of follow-up to 8 weeks. Further followup would enable improved understanding into how weight and age could impact achievement of medium and long-term serum IFX concentrations and clinical outcomes. Lastly, the size of our cohort limits the ability to make comparisons between groups.
In conclusion, in this multi-center prospective study of realworld clinical practice, the majority of children who undergo IFX induction for UC achieve short term clinical remission. IFX concentrations ≥33 µg/mL prior to the second dose are associated with week 8 clinical remission. As young age and low body weight impact IFX concentration, prospective studies with proactive adjustment in pediatric patients with UC should be carried out to optimize response to therapy. Clinicians caring for children with UC should adjust and monitor IFX dosing and interval to obtain therapeutic response.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by Health Research Ethics Boards of University of Calgary, University of Alberta, University of British Columbia, and University of Manitoba. Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin.

AUTHOR CONTRIBUTIONS
JD was involved in study concept and design, acquisition of data, analysis and interpretation of data, drafting of the manuscript, critical revision of the manuscript for important intellectual content, and final approval of submitted version. KJ, WE-M, EW, MC, IW, and HH were involved in study concept and design, acquisition of data, analysis and interpretation of data, critical revision of the manuscript for important intellectual content, and final approval of submitted version. CG was involved in analysis and interpretation of data, critical revision of the manuscript for important intellectual content, and final approval of submitted version. RP was involved in study concept and design, analysis and interpretation of data, critical revision of the manuscript for important intellectual content, and final approval of submitted version. All authors contributed to the article and approved the submitted version.

FUNDING
The authors declare that this study received funding from Future Leaders in IBD and in-kind support from Prometheus Laboratories. The funders were not involved in the study design, collection, analysis, interpretation of data, the writing of this article or the decision to submit it for publication.