ORIGINAL ARTICLE

Serum Iniximab, Antidrug Antibodies, and Tumor Necrosis Factor

Predict Sustained Response in Pediatric Crohns Disease
Ronen Stein, MD,* Dale Lee, MD, MSCE,* , Mary B. Leonard, MD, MSCE, Meena Thayu, MD, MSCE,* ,
Lee A. Denson, MD,k Emil Chuang, MD, Rita Herskovitz, MS, RD,# Theresa Kerbowski, BA,*
and Robert N. Baldassano, MD* ,**

Background: Serum iniximab (s-IFX) levels, antibodies to IFX (ATI), and inammatory markers are important in predicting clinical outcomes in
adults, but their roles in pediatric Crohns disease (CD) require further study. The primary aim of this study was to determine the association between
serologic parameters during induction and ongoing IFX therapy at 12 months in pediatric CD.

Methods: S-IFX, ATI, serum tumor necrosis factor alpha (s-TNF-a), and C-reactive protein were measured at IFX initiation, 10 weeks, 6 months, and
12 months in a prospective cohort study of children with CD at a single tertiary care center.

Results: At 12 months, 60 of 77 participants (78%) remained on IFX. Participants who completed 12 months of IFX had higher 10-week median s-IFX
levels (20.40 mg/mL; interquartile range [IQR], 11.2035.00] versus 8.70 mg/mL; IQR 0.9016.90; P 0.01), a greater proportion with undetectable 10week ATI (P 0.008), and a greater median change in s-TNF-a between baseline and week 10 (25.96 pg/mL; IQR, 28.73 to 24.17 versus 21.76 pg/
mL; IQR, 25.60 to 0.30; P 0.006). Receiver operating characteristic analysis to predict ongoing IFX at 12 months showed area under the curve (95%
condence interval) for 10-week s-IFX and change in s-TNF-a from baseline to 10 weeks to be 0.71 (0.540.88) and 0.74 (0.580.91), respectively.
C-reactive protein was not associated with ongoing therapy.

Conclusions: ATI, s-IFX, and s-TNF-a during IFX induction are associated with 12-month clinical outcomes in pediatric CD. Future studies are
needed to further dene the clinical role of s-TNF-a measurement and to compare the clinical utility of 10 and 14-week ATI and s-IFX levels.
(Inamm Bowel Dis 2016;22:13701377)
Key Words: therapeutic drug monitoring, loss-of-response, biologic therapies

rohns disease (CD) is an idiopathic chronic inammatory

condition of the gastrointestinal tract with an increasing incidence in children.1 Iniximab (IFX), a monoclonal antibody against

Supplemental digital content is available for this article. Direct URL citations
appear in the printed text and are provided in the HTML and PDF versions of this
article on the journals Web site (www.ibdjournal.org).
Received for publication January 4, 2016; Accepted February 1, 2016.
From the *Department of Pediatrics, Division of Gastroenterology, Hepatology and
Nutrition, The Childrens Hospital of Philadelphia, Philadelphia, Pennsylvania; Department of Gastroenterology, Seattle Childrens Hospital, Seattle, Washington; Departments of Pediatrics and Medicine, Division of Nephrology, Stanford University School
of Medicine, Stanford, California; Janssen Pharmaceuticals, Horsham, Pennsylvania;
k
Department of Pediatrics, Division of Gastroenterology, Hepatology, and Nutrition,
Cincinnati Childrens Hospital Medical Center, Cincinnati, Ohio; Nestl Health Science, Vevey, Switzerland; #Department of Pediatrics, Division of Nephrology, The
Childrens Hospital of Philadelphia, Philadelphia, Pennsylvania; and **Perelman School
of Medicine at the University of Pennsylvania, Philadelphia, Pennsylvania.
Supported by National Institutes of Health Grants K23 DK082012 (M.T.) and
K24 DK076808 (M.B.L.); the Clinical and Translational Science Award
UL1RR024134 and UL1TR000003; the Penn Joint Center for Inammatory Bowel
Diseases; and by Prometheus Laboratories Inc.
R. N. Baldassano is a consultant for Janssen Pharmaceuticals. The remaining
authors have no conict of interest to disclose.
Reprints: Ronen Stein, MD, Division of Gastroenterology, Hepatology, and
Nutrition, The Childrens Hospital of Philadelphia, Philadelphia, PA 19104 (e-mail:
steinr2@email.chop.edu).
Copyright 2016 Crohns & Colitis Foundation of America, Inc.
DOI 10.1097/MIB.0000000000000769
Published online 8 April 2016.

1370

| www.ibdjournal.org

tumor necrosis factor alpha (TNF-a), is effective at inducing and

maintaining remission in moderate-to-severe pediatric CD.2 However, the long-term efcacy of IFX is complicated by loss of
response, which occurs at a rate of 13% per patient-year.3
The role of therapeutic drug monitoring in preventing loss of
response, previously investigated in adult studies, suggests that it
may be due to the immunogenicity of IFX and the development of
antibodies to IFX (ATI).4,5 The presence of ATI is also a risk factor
for hypersensitivity reactions to IFXanother common cause of
premature treatment cessation. However, ATI may be transient, and
ATI alone may not predict poor outcomes.5 Several studies have
indicated that higher serum IFX (s-IFX) trough levels, particularly
early in the treatment course, are strongly associated with long-term
treatment response.68 These studies have used a variety of measures for clinical response including continuation of IFX therapy, Creactive protein (CRP), Crohns Disease Activity Index, need for
surgery, and the HarveyBradshaw index.48
The inammatory burden early in the treatment course may
also be predictive of long-term efcacy. Cornillie et al6 showed
that larger decreases in CRP from baseline to week 14 of therapy
were associated with better clinical outcomes. In contrast to CRP,
the role of serum TNF-a (s-TNF-a) concentration as a predictor of
clinical outcomes is less clear. Louis et al9 reported that higher
baseline s-TNF-a levels were associated with primary response to
Inamm Bowel Dis  Volume 22, Number 6, June 2016

Copyright 2016 Crohns & Colitis Foundation of America, Inc. Unauthorized reproduction of this article is prohibited.

Inamm Bowel Dis  Volume 22, Number 6, June 2016

IFX, whereas others have found higher baseline levels associated

with nonresponse.10
Pediatric CD often has a more severe presentation and
aggressive disease course with more frequent colonic involvement
as compared with adult CD.11,12 Body mass index, which has been
shown to be associated with s-IFX levels and loss of response to
IFX, is lower in children.13,14 Therefore, the impact of IFX pharmacokinetics and immunogenicity may be different in children
compared with adults. Few pediatric studies have investigated
the association between ATI, s-IFX level, serum inammatory
markers, and long-term treatment outcomes in CD. A recent,
single-center, cross-sectional study found that ATI was a predictor
of lower s-IFX levels and higher risk of surgery among pediatric
inammatory bowel disease patients.15 Additionally, Singh et al16
performed a prospective study among pediatric patients with both
ulcerative colitis and CD initiating IFX therapy and found that
CRP at 14 weeks of treatment in conjunction with the s-IFX level
at that time predicted sustained clinical response at 54 weeks.
The primary aim of this study was to determine the
association between s-IFX, ATI, CRP, and s-TNF-a during induction and ongoing IFX therapy at 12 months among pediatric CD
patients initiated on IFX therapy. A second aim was to compare
the predictive values of s-TNF-a and CRP on continuation of IFX
therapy at 12 months in this sample.

Pediatric Crohns Disease

obtained just before an IFX infusion, whereas the 10-week samples were midcycle levels and obtained 4 weeks from the previous
IFX infusion and 4 weeks before the subsequent IFX infusion.
Pediatric Crohns Disease Activity Index (PCDAI) scores19 were
calculated at each visit.

Measurement of S-IFX Levels and ATI

A portion of the collected serum at each visit was stored at
2708C for future use. This serum was sent to Prometheus Laboratories, Inc. (San Diego, CA) for measurement of s-IFX and
ATI levels using a mobility shift assay. The mobility shift assay
detects both drug-free and drug-bound ATI and therefore allows
for the measurement of ATI in the presence of circulating IFX
(up to 60.00 mg/mL of s-IFX).20 The reporting range of s-IFX
was 1.00 to 34.00 mg/mL. The lower reporting limit of ATI was
3.10 mg/mL.

Measurement of CRP
Analysis of CRP was performed at the laboratory of The
Childrens Hospital of Philadelphia. The lower limit of detection
was 0.30 mg/dL. According to the reference range for this laboratory, CRP values of ,1.00 mg/dL are considered to be normal.
As such, participants with CRP values of ,1.00 mg/dL were
considered to be in CRP-dened remission.

Measurement of S-TNF-a
MATERIALS AND METHODS
Patient Population and Study Design
This study is ancillary to a prospective study that collected
blood samples from 77 children aged 5 to 21 years with active CD
starting IFX at the Inammatory Bowel Disease Center at The
Childrens Hospital of Philadelphia from 2006 to 2011. The decision to start IFX was made by the treating gastroenterologist, and
induction dosing was given at 0, 2, and 6 weeks at a dose of 5
mg/kg followed by standard maintenance dosing. Participants
were originally enrolled in a study investigating changes in bone
and mineral metabolism after the initiation of antiTNF-a therapy,17,18 and they had provided written informed consent and
assent for future use of their data and samples for other studies.
The Institutional Review Board at The Childrens Hospital of
Philadelphia approved the study protocol.
Exclusion criteria included previous antiTNF-a therapy,
pregnancy, and other medical conditions that could affect bone
health, including nonambulatory status, kidney disease, liver disease, and seizure disorders. Participants who transitioned to an
outside provider before study completion were also excluded.
Blood samples for measurement of hematocrit, erythrocyte
sedimentation rate, CRP, and albumin were obtained before the
rst IFX infusion and also 10 weeks, 6 months, and 12 months
after IFX initiation, based on the protocol of the original study on
antiTNF-a therapy and bone and mineral metabolism.17
S-TNF-a was measured at baseline, 10 weeks, and 12 months.
The samples from 6 and 12 months were trough levels and

Total s-TNF-a (both bound and unbound to IFX) was measured by the Luminex platform (Luminex Corporation, Austin,
TX) using nonneutralizing antibodies with an interassay variation
of 8.3%. The lower limit of detection was 0.18 pg/mL.

Outcome Measures
Our primary study outcome was ongoing IFX therapy at 12
months. Chart review was used to retrospectively document the
reason for IFX withdrawal.

Statistical Analysis
All analyses were conducted using STATA 13.1 (Stata,
Corp., College Station, TX). Signicance was set at P , 0.05 for
all analyses. Continuous variables were assessed for normality,
and skewed data were reported as median and interquartile range
(IQR). Because the reporting range of s-IFX was 1.00 to 34.00
mg/mL, s-IFX levels above and below the reporting range were
assigned values of 0.90 and 35.00 mg/mL, respectively. For evaluating medication exposure and changes over time in PCDAI and
CRP, a last-observation-carried-forward approach was used for
missing data. Differences in nonparametric continuous variables
were assessed using the Wilcoxon rank-sum test, and differences
in categorical variables were assessed with chi-square testing or
Fishers exact testing. Spearmans rank correlation was used to
test strength of association between ranked variables. Multivariate
models were constructed using logistic regression. Receiver operating characteristic (ROC) curves were constructed, and areaunder-the-curve (AUC) analysis was performed for s-IFX levels
and CRP at week 10, as well as change in s-TNF-a and CRP from
www.ibdjournal.org |

1371

Copyright 2016 Crohns & Colitis Foundation of America, Inc. Unauthorized reproduction of this article is prohibited.

Inamm Bowel Dis  Volume 22, Number 6, June 2016

Stein et al

baseline to week 10. Potential cutoff values for s-IFX, CRP, and
change in s-TNF-a were analyzed using sensitivity, specicity,
positive likelihood ratios (PLR), and negative likelihood ratios
(NLR). PLR was calculated as sensitivity divided by (1 2 specicity), and NLR was calculated as (1 2 sensitivity) divided by
specicity. Best cutoff values for predicting continued IFX therapy at 12 months were selected based on the best combination of
lowest NLR and highest PLR. Preference was given to lowest
NLR to best identify subjects with the highest risk of not achieving sustained response to IFX with the highest sensitivity.

RESULTS
ATI and s-IFX were measured from 77 participants (48
male; 62%) over 12 months (Table 1). The median age at initiation of IFX was 14.79 years (IQR, 12.2016.81 yr) with the
median disease duration of 1.66 years (IQR, 0.823.31 yr). The
median IFX dose at baseline was 5.03 mg/kg (IQR, 4.945.32
mg/kg). At baseline, clinical disease activity as measured by
PCDAI was normal in 15 of 77 subjects (20%), and CRP was
normal in 28 of 77 subjects (36%). Among the 7 of 77 subjects
(9%) with normal values for both PCDAI and CRP, the

TABLE 1. Baseline Characteristics of Participants

Initiating IFX (n 77)
Characteristics
Male
Caucasian race
Age at rst infusion, yr
,6
69
1013
1417
1821
Age at diagnosis, yr
,6
69
1013
1417
Disease duration, yr
,1
12
34
56
79
Lower gastrointestinal disease site
Ileal only
Colonic only
Ileocolonic
Upper gastrointestinal tract involvement
Perianal involvement

1372

n (%)
48 (62)
65 (84)
2
11
21
38
5

(3)
(14)
(27)
(49)
(7)

4
17
36
20

(5)
(22)
(47)
(26)

30
28
8
6
5

(39)
(36)
(10)
(8)
(7)

4
20
53
65
26

(5)
(26)
(69)
(84)
(34)

indications for starting IFX included steroid dependence, growth

failure, and refractory perianal disease. Over the 12 months, clinical disease activity as measured by PCDAI score improved with
35% participants having moderate-to-severe disease at baseline
and decreasing to 5% by 12 months (P , 0.001; Table 2). The
usage of systemic corticosteroids decreased from 32% to 6% over
12 months (P , 0.001).
At 6 months after the initiation of IFX, 69 of 77 participants
(90%) continued on therapy, whereas at 12 months 60 of 77 of
participants (78%) remained on IFX (Table 2). Of the participants
continuing on IFX therapy at 12 months, 53 of 60 (88%) were in
CRP-dened remission and 46 of 60 (73%) were in PCDAIdened remission (PCDAI , 10). The most common reason for
stopping IFX therapy was poor clinical response (12 of 17 participants), whereas anaphylaxis and other adverse reactions contributed to IFX discontinuation in the remainder of participants.
Subsequent treatment was most commonly with adalimumab
(Table 3).
Dose escalations were made in 20 of 77 participants (26%),
including in 5 subjects before the 10-week s-IFX level. There was
no signicant difference in sustained response with 16 of 60
participants (27%) remaining on IFX at 12 months having a dose
adjustment, compared with 4 of 17 participants (24%) off of IFX
at 12 months (P 1.00). Compared with the other study participants, the 5 participants who had an IFX dose escalation before
week 10 had a signicantly higher baseline s-TNF-a (24.01 pg/
mL; IQR, 9.2429.76 pg/mL versus 7.36 pg/mL; IQR, 4.93
10.49 pg/mL; P 0.03).

S-IFX Levels and ATI

As shown in Table 4, participants who remained on IFX
at 12 months had higher median week 10 s-IFX levels as compared with those who stopped IFX (20.40 mg/mL; IQR, 11.20
35.00 mg/mL versus 8.70 mg/mL; IQR 0.9016.90 mg/mL; P
0.01). The median 6-month s-IFX trough levels were higher
among participants remaining on therapy at 12 months,
although this was not statistically signicant (4.00 mg/mL;
IQR 1.207.90 mg/mL versus 1.30 mg/mL; IQR 0.904.90
mg/mL; P 0.28). However, there was a signicant correlation
between 10-week s-IFX levels and s-IFX trough levels at 6
months (rs 0.28; P 0.03).
During the study period, 18 of 77 participants (23%) had
detectable ATI levels, with transient, detectable ATI found in only
1 subject. Undetectable ATI at 10 weeks was associated with
ongoing IFX therapy at 12 months (P 0.008). All 3 subjects
who developed ATI at 10 weeks were off of IFX therapy by 12
months. Of the 69 subjects remaining on IFX at 6 months, detectable ATI were found in 8 subjects (12%) with no signicant
difference between those continuing IFX at 12 months (6 of 60;
10%) and those off of IFX (2 of 9; 22%).
At week 10, there were 5 subjects whose ATI/IFX levels
were missing. These subjects had similar baseline PCDAI, CRP,
immunomodulator use, and steroid exposure compared with the
remainder of the cohort.

| www.ibdjournal.org

Copyright 2016 Crohns & Colitis Foundation of America, Inc. Unauthorized reproduction of this article is prohibited.

Inamm Bowel Dis  Volume 22, Number 6, June 2016

Pediatric Crohns Disease

TABLE 2. Disease Activity, Laboratory Results, and Medications Over the Study Interval (n 77)

PCDAI, n (%)
Not active (#10)
Mild (1130)
Moderatesevere (.30)
Laboratory evaluation, median (IQR)
s-IFX, mg/mL
s-TNF-a, pg/mL
CRP, mg/dL
CRP , 1 mg/dL, n (%)
Detectable ATI, n (%)
Medications, n (%)
5-aminosalicylic acid
6-mercaptopurine/azathioprine
Adalimumab
IFX
Methotrexate
Corticosteroids

Baseline

10 Weeks

6 Months

12 Months

15 (20)
34 (45)
26 (35)

47 (61)
26 (34)
4 (5)

52 (67)
23 (30)
2 (3)

54 (70)
19 (25)
4 (5)

7.58 (5.0610.61)
1.30 (0.602.90)
28 (36)

16.80 (9.0035.00)
1.61 (1.032.62)
0.50 (0.300.50)
62 (81)
3 (4)

3.80 (0.907.90)a

0.50 (0.300.70)
66 (86)
8 (12)a

5.85 (1.0511.05)b
3.32 (1.1710.05)
0.50 (0.300.60)
65 (84)
10 (17)b

60 (78)
28 (36)
0 (0)
77 (100)
14 (18)
25 (32)

59 (77)
18 (23)
0 (0)
77 (100)
17 (22)
12 (16)

56
11
2
69
17
5

(73)
(14)
(3)
(90)
(22)
(6)

50 (65)
7 (9)
8 (10)
60 (78)
17 (22)
5 (6)

Excludes 8 subjects off of IFX at the time of 6-month measurement.

Excludes 17 subjects off of IFX at the time of 12-month measurement.

To evaluate the predictive value of week 10 s-IFX levels,

ROC curves were constructed with continuation of IFX therapy at 12
months as the binary classier (Fig. 1). The AUC (95% condence
interval) for 10-week s-IFX was 0.71 (0.540.88), suggesting that
10-week s-IFX may be a fair predictor of ongoing IFX therapy at 12
months. Based on the estimated PLR and NLR, the best cutoff value
for week 10 s-IFX was $9.10 mg/mL (Table 5 and see Table 1,
Supplemental Digital Content 1, http://links.lww.com/IBD/B251).

C-Reactive Protein
At study baseline, 28 of 77 participants (36%) were in
CRP-dened remission. There was no difference in CRP at

TABLE 3. Reasons for IFX Discontinuation and Choice

of Post-IFX Therapy (n 17)
n (%)
Reason for discontinuation
Poor response
Other adverse effects
Anaphylaxis
Post-IFX therapy
Adalimumab
Surgical resection
Other medical therapy
Systemic corticosteroids
6-Mercaptopurine

baseline between participants who continued IFX therapy for 12

months and those who stopped IFX therapy before 12 months of
therapy (1.40 mg/dL; IQR 0.602.90 mg/dL versus 1.20 mg/dL;
IQR 0.505.00 mg/dL; P 0.92). Additionally, 10-week CRP
and change in CPR from baseline to 10 weeks between participants who remained on IFX at 12 months and those who stopped IFX before 12 months of therapy were not signicantly
different (0.50 mg/dL; IQR 0.300.50 mg/dL versus 0.50
mg/dL; IQR 0.351.35 mg/dL; P 0.24 for 10-week CRP
and 20.70 mg/dL; IQR 22.20 to 20.10 mg/dL versus 20.70
mg/dL; IQR 22.40 to 0.00 mg/dL; P 0.61 for change in CRP
from baseline to 10 wk). No association was seen between these
variables and sustained response even after exclusion of subjects
in CRP-dened remission from the analysis. Although further
analysis of CRP was planned, the rate and degree of CRP elevation at baseline was too small to allow for its investigation
with an ROC curve or estimation cutoff point values.

Serum Tumor Necrosis Factor Alpha

12 (71)
3 (18)
2 (11)
8
3
3
2
1

(47)
(18)
(18)
(11)
(6)

There was no difference in median s-TNF-a levels at baseline between participants who continued on IFX for 12 months
and those who stopped IFX before 12 months of therapy. However, participants completing 12 months of IFX therapy were
found to have lower 10-week s-TNF-a levels (1.46 pg/mL; IQR
0.891.94 pg/mL versus 2.53 pg/mL; IQR 1.356.86 pg/mL; P
0.009) and a greater decrease in s-TNF-a level over the rst 10
weeks of therapy, adjusted for baseline TNF-a (25.96 pg/mL;
IQR 28.73 to 24.17 pg/mL versus 21.76 pg/mL; IQR 25.60 to
0.30 pg/mL; P 0.006).
www.ibdjournal.org |

1373

Copyright 2016 Crohns & Colitis Foundation of America, Inc. Unauthorized reproduction of this article is prohibited.

Inamm Bowel Dis  Volume 22, Number 6, June 2016

Stein et al

TABLE 4. Median (IQR) s-IFX, ATI, s-TNF-a, and Disease Activity by IFX Status at 12 Months

s-IFX, mg/mL
10 wk s-IFX
6 mo s-IFX trough
Detectable ATI
10 wk, n (%)
6 mo, n (%)
s-TNF-a, pg/mL
Baseline s-TNF-a
10 wk s-TNF-a
Ds-TNF-a 010 wk
PCDAI
Baseline PCDAI
10 wk PCDAI
DPCDAI 010 wk
6 mo PCDAI
12 mo PCDAI

On IFX (n 60)

Off IFX (n 17)

20.40 (11.20 to 35.00)

4.00 (1.20 to 7.90)

8.70 (0.90 to 16.90)

1.30 (0.90 to 4.90)a

0.01
0.28

0 (0)
6 (10)

3 (18)
2 (22)a

0.008
0.30

7.92 (5.67 to 11.10)

1.46 (0.89 to 1.94)
25.96 (28.73 to 24.17)

6.13 (4.18 to 9.15)

2.53 (1.35 to 6.86)
21.76 (25.60 to 0.30)

0.13
0.009
0.006b

27.50 (16.25 to 40.00)

10.00 (5.00 to 15.00)
215.00 (225.00 to 23.75)
5.00 (0.00 to 12.50)
5.00 (2.50 to 12.50)

30.00 (12.50 to 50.00)

15.00 (10.00 to 20.00)
210.00 (230.00 to 0.00)
12.50 (5.00 to 17.50)
10.00 (5.00 to 20.00)

0.71
0.07
0.23b
0.09
0.20

Excludes 8 subjects off of IFX at the time of 6-month measurement.

Adjusted for baseline.

There were 8 subjects who were missing values for change

in s-TNF-a over the rst 10 weeks of therapy. These subjects had
similar baseline PCDAI, CRP, immunomodulator use, and steroid
exposure compared with the remainder of the cohort.
An ROC curve was constructed for change in s-TNF-a over
the rst 10 weeks of therapy with an AUC (95% condence
interval) of 0.74 (0.580.91), suggesting that change in s-TNF-a
over the rst 10 weeks of therapy may be a fair predictor of
ongoing IFX therapy at 12 months. A best cutoff value of
#23.60 pg/mL was estimated based on PLR and NLR (see
Table 2, Supplemental Digital Content 2, http://links.lww.com/
IBD/B252).
There was no difference in median baseline s-TNF-a levels
between participants in CRP-dened remission at baseline and
those not in remission (8.27 pg/mL; IQR 4.4411.14 pg/mL versus 7.08 pg/mL; IQR 5.4110.31 pg/mL; P 0.54). Likewise,
correlation of s-TNF-a and CRP at 10 weeks was not signicant
(rs 0.23; P 0.05).

Pediatric Crohns Disease Activity Index

There was no difference in median baseline PCDAI
between participants who continued on IFX for 12 months and
those who stopped IFX before 12 months of therapy. Although
not statistically signicant, subjects who remained on IFX therapy
at 12 months had lower PCDAI scores at 10 weeks (10.00; IQR
5.0015.00 versus 15.00; IQR 10.0020.00; P 0.07) and 6
months (5.00; IQR 0.0013.00 versus 13.00; IQR 5.0018.00;
P 0.09) compared with those who discontinued IFX therapy
before 12 months.

1374

Immunomodulator Therapy
There were 42 participants on immunomodulator therapy at
baseline, 14 on methotrexate and 28 on azathioprine/6mercaptopurine (Table 2). Among these participants, 16 of 42
(38%) remained on immunomodulator therapy throughout the
12-month study period. Among the participants on immunomodulator therapy at baseline, 7 of 42 (17%) developed ATI during
the study period compared with 11 of 35 (31%) on IFX monotherapy (P 0.13). There was no association between immunomodulator therapy at baseline and the primary study outcome,
with 35 of 42 participants (83%) on immunomodulators remaining on IFX at 12 months compared with 25 of 35 participants
(71%) on IFX monotherapy (P 0.21).

DISCUSSION
In this single-center study of 77 pediatric CD patients
starting IFX therapy, 60 patients remained on IFX at 12 months.
These patients had higher s-IFX levels and lower ATI during
induction and a greater change in s-TNF-a from baseline to 10
weeks as compared with patients not able to complete 12 months
of therapy.
These results provide new insights into the importance of
s-IFX and ATI during induction therapy for pediatric CD. They
conrm the ndings of previous adult studies that showed
associations between s-IFX and ATI shortly after induction and
long-term therapeutic outcomes, measured in our study by
continued IFX therapy at 12 months. A new nding not seen in
the adult or pediatric literature is that change in s-TNF-a during

| www.ibdjournal.org

Copyright 2016 Crohns & Colitis Foundation of America, Inc. Unauthorized reproduction of this article is prohibited.

Inamm Bowel Dis  Volume 22, Number 6, June 2016

Pediatric Crohns Disease

FIGURE 1. ROC curves for ongoing iniximab therapy at 12 months

evaluating s-IFX at 10 weeks (A) and change in s-TNF-a from baseline
to 10 weeks (B).

induction is also associated with long-term outcomes in IFX

therapy.
It is currently not the standard clinical practice to measure
s-TNF-a in patients receiving IFX therapy, yet our ndings suggest that s-TNF-a may have a clinical role. Akin to previous
studies that have shown change in CRP during induction or
week-14 CRP levels to be predictive of long-term outcomes,
our study shows a relationship among change in s-TNF-a during
induction, s-TNF-a at 10 weeks, and ongoing IFX therapy at 12

months. However, we found change in s-TNF-a from baseline to

10 weeks to be a better predictor for ongoing IFX therapy than
change in CRP. CRP and s-TNF-a were poorly correlated at
baseline and week 10, respectively. Week 10 and change from
baseline to 10-week values of CRP were not associated with
remaining on IFX therapy at 12 months. However, at these same
time points, s-TNF-a values were signicantly associated with
continuation of IFX therapy. It may be the case that s-TNF-a is
a more sensitive marker of inammatory burden in pediatric CD
than CRP. It is also possible that the patients with the greatest
change in s-TNF-a during induction are primary responders, as
opposed to those without a large change in s-TNF-a, who may be
primary nonresponders. In our study, early response was not associated with greater baseline s-TNF-a burden, and it remains
unclear if higher IFX dosing has the potential to overcome early
nonresponse. However, subjects who did receive IFX dose escalation before week 10 were found to have a greater baseline
inammatory burden as measured by s-TNF-a. Future studies will
need to further investigate the role of s-TNF-a measurement, as
well as higher IFX dosing, in IFX therapy.
We did not nd change in CRP during induction to be
predictor of long-term clinical response. Although Cornillie et al6
found that changes in CRP from baseline to week 14 to be indicative of sustained clinical response in adults, Singh et al16 did not
reproduce this nding in the pediatric population. Instead, week14 CRP levels were found to be predictor of sustained clinical
response in children with ulcerative colitis and CD. Our study did
not nd week-10 CRP or change in CRP from baseline to 10
weeks to be signicantly associated with remaining on IFX therapy at 12 months, although this nding may be limited by the fact
that baseline CRP levels in the cohort were only mildly elevated.
Additionally, measuring CRP during midcycle, at 10 weeks, may
have led to a lower heterogeneity of values for CRP.
Our ndings demonstrate that s-IFX, ATI, and s-TNF-a early
in the treatment course may be important predictors of long-term
outcomes. In contrast to previous studies, which used the s-IFX
trough levels and ATI at week 14 to investigate long-term outcomes,6,7,16 our study used levels from week 10. In our sample,
several patients had already formed ATI by week 10, raising the
question of whether therapeutic drug monitoring even earlier than
week 14 may improve long-term outcomes. Effective week-10
predictors could potentially allow for the modication of therapy
and induction of remission earlier in the treatment course. It has
been previously suggested that the inammatory burden early on in

TABLE 5. Predictive Values of s-IFX and s-TNF-a on Continued IFX Therapy at 12 Months

10 week s-IFX $ 9.10 mg/mL

Ds-TNF-a 010 wk #23.60 pg/mL

Sensitivity, %

Specicity, %

PLR

NLR

AUC (95% Condence Interval)

Pa

80.70
81.48

60.00
66.67

2.02
2.44

0.32
0.28

0.71 (0.540.88)
0.74 (0.580.91)

0.003
0.001

Fishers exact test used to test the association between each cutoff and ongoing therapy at 12 months.

www.ibdjournal.org |

1375

Copyright 2016 Crohns & Colitis Foundation of America, Inc. Unauthorized reproduction of this article is prohibited.

Stein et al

inammatory bowel disease course is greater and that aggressive

immunosuppressive therapy to address this inammatory burden is
associated with better long-term outcomes.21,22 The potential predictive value of earlier therapeutic drug monitoring has already
been reported in the adult ulcerative colitis literature.23 Future studies will need to investigative the utility of earlier therapeutic drug
monitoring during IFX induction therapy.
We found no statistically signicant association between
combination therapy with an immunomodulator, ATI development, and the continuation of IFX therapy at 12 months.
However, interpretation of these results is limited given the small
sample size, the nonrandomized study design, and because the
majority of participants on immunomodulator therapy at baseline
were weaned off by 12 months. The high rate of medication
nonadherence has been previously well described in the pediatric
CD population.24,25 Because this was not an intervention study,
ascertainment of adherence with the prescribed immunomodulatory therapy was not performed and is another limitation to the
interpretation of these results. Several recent pediatric studies
have also shown no signicant effect of combination therapy on
s-IFX levels or ATI, although Church et al26 reported an association between combination therapy and decreased loss of response
in pediatric CD.15,16 Two, large, adult clinical trials have demonstrated lower ATI levels and higher s-IFX levels with combination
therapy.27,28 More recently, Grossi et al29 found that pediatric CD
patients on concomitant therapy for at least 6 months after the
initiation of IFX were more likely to remain on IFX after 5 years
compared with those on IFX monotherapy or those on combination therapy for less than 6 months. Additionally, male subjects on
combination therapy with methotrexate were found to have
greater sustained response compared with those on a thiopurine.
This study demonstrates an association between early aggressive
immunosuppression and more favorable long-term outcomes. The
data from our study likewise suggest that early control of inammatory burden, as measured by s-TNF-a, is associated with a 12month outcome. The importance of early induction of remission
in children seems to be important, but further studies will be
required in the future. Large randomized studies are needed to
investigate the effect of immunomodulator use on clinical
response and IFX immunogenicity in pediatric CD and to compare clinical response between combination therapy and optimized IFX monotherapy using therapeutic drug monitoring.
Our study has several limitations. First, all dosing adjustments were made by the treating gastroenterologist and not the
investigators, although clinicians were blinded to the s-IFX and
ATI results. As such, dose adjustment of IFX is a potential
confounder. However, dose adjustments were similar between the
2 groups as we have described, and the presented study focuses on
serologic laboratory markers as predictors on ongoing IFX
therapy at 12 months. We used continuation of IFX therapy at
12 months as the primary outcome as opposed to mucosal healing
on colonoscopy, which is the gold standard treatment outcome in
CD. Nonetheless, the majority of the patients able to continue IFX
therapy at 12 months were in CRP and PCDAI-dened remission,

1376

Inamm Bowel Dis  Volume 22, Number 6, June 2016

indicating good disease control. Another limitation is the missing

data in our study. However, the number of missing data values
was minimal. A last-observation-carried-forward approach was
used for missing medication exposure, CRP, and PCDAI, which
we believe is a conservative method to impute missing data in the
setting of initiation of a new medical therapy. Missing data for
s-TNF-a, s-IFX, and ATI were not differential by disease characteristics. Finally, our s-TNF-a assay measured TNF-a both bound
and unbound to IFX. An assay distinguishing unbound and bound
TNF-a would be a better representation of the systemic inammatory burden and could possibly be a better predictor of sustained clinical response. Nonetheless, our assay demonstrated an
inverse correlation between s-IFX and s-TNF-a at week 10 (rs
20.49; P , 0.001) and was able to measure changes in inammatory burden as evidenced by decline in s-TNF-a from baseline
to week 10. This may be explained by the rapid clearance of
s-TNF-a once bound by IFX.
In conclusion, we found an association between higher sIFX levels during induction and ongoing IFX therapy at 12
months. Furthermore, we showed that patients who remained on
IFX at 12 months had lower ATI levels during induction and had
a greater decrease in s-TNF-a from baseline to 10 weeks. Further
investigation is needed to (1) evaluate the utility of week 10
versus week 14 s-IFX and ATI levels, (2) determine the role of
s-TNF-a measurement in clinical practice, and (3) study the effects of combination therapy on the immunogenicity of IFX in
children.

REFERENCES
1. Rabizadeh S, Dubinsky M. Update in pediatric inammatory bowel disease. Rheum Dis Clin North Am. 2013;39:789799.
2. Hyams J, Crandall W, Kugathasan S, et al. Induction and maintenance iniximab therapy for the treatment of moderate-to-severe
Crohns disease in children. Gastroenterology. 2007;132:863873;
quiz 11651166.
3. Gisbert JP, Panes J. Loss of response and requirement of iniximab dose
intensication in Crohns disease: a review. Am J Gastroenterol. 2009;
104:760767.
4. Vande Casteele N, Gils A, Singh S, et al. Antibody response to iniximab
and its impact on pharmacokinetics can be transient. Am J Gastroenterol.
2013;108:962971.
5. Vande Casteele N, Khanna R, Levesque BG, et al. The relationship
between iniximab concentrations, antibodies to iniximab and disease
activity in Crohns disease. Gut. 2014;64:15391545.
6. Cornillie F, Hanauer SB, Diamond RH, et al. Postinduction serum iniximab trough level and decrease of C-reactive protein level are associated
with durable sustained response to iniximab: a retrospective analysis of
the ACCENT I trial. Gut. 2014;63:17211727.
7. Bortlik M, Duricova D, Malickova K, et al. Iniximab trough levels may
predict sustained response to iniximab in patients with Crohns disease.
J Crohns Colitis. 2013;7:736743.
8. Marits P, Landucci L, Sundin U, et al. Trough s-iniximab and antibodies
towards iniximab in a cohort of 79 IBD patients with maintenance iniximab treatment. J Crohns Colitis. 2014;8:881889.
9. Louis E, Vermeire S, Rutgeerts P, et al. A positive response to iniximab
in Crohn disease: association with a higher systemic inammation before
treatment but not with -308 TNF gene polymorphism. Scand J Gastroenterol. 2002;37:818824.
10. Martinez-Borra J, Lopez-Larrea C, Gonzalez S, et al. High serum tumor
necrosis factor-alpha levels are associated with lack of response to iniximab in stulizing Crohns disease. Am J Gastroenterol. 2002;97:
23502356.

| www.ibdjournal.org

Copyright 2016 Crohns & Colitis Foundation of America, Inc. Unauthorized reproduction of this article is prohibited.

Inamm Bowel Dis  Volume 22, Number 6, June 2016

11. de Bie CI, Paerregaard A, Kolacek S, et al. Disease phenotype at diagnosis

in pediatric Crohns disease: 5-year analyses of the EUROKIDS Registry.
Inamm Bowel Dis. 2013;19:378385.
12. Freeman HJ. Natural history and long-term clinical course of Crohns
disease. World J Gastroenterol. 2014;20:3136.
13. Dotan I, Ron Y, Yanai H, et al. Patient factors that increase iniximab
clearance and shorten half-life in inammatory bowel disease: a population
pharmacokinetic study. Inamm Bowel Dis. 2014;20:22472259.
14. Harper JW, Sinanan MN, Zisman TL. Increased body mass Index is
associated with earlier time to loss of response to iniximab in patients
with inammatory bowel disease. Inamm Bowel Dis. 2013;19:
21182124.
15. Zitomersky NL, Atkinson BJ, Fournier K, et al. Antibodies to iniximab
are associated with lower iniximab levels and increased likelihood of
surgery in pediatric IBD. Inamm Bowel Dis. 2015;21:307314.
16. Singh N, Rosenthal CJ, Melmed GY, et al. Early iniximab trough levels
are associated with persistent remission in pediatric patients with inammatory bowel disease. Inamm Bowel Dis. 2014;20:17081713.
17. Augustine MV, Leonard MB, Thayu M, et al. Changes in vitamin Drelated Mineral metabolism after induction with anti-tumor necrosis
factor-alpha therapy in Crohns disease. J Clin Endocrinol Metab. 2014;
99:E991E998.
18. Grifn LM, Thayu M, Baldassano RN, et al. Improvements in bone
density and structure during anti-TNF-alpha therapy in pediatric Crohns
disease. J Clin Endocrinol Metab. 2015;100:26302639.
19. Hyams JS, Ferry GD, Mandel FS, et al. Development and validation of
a pediatric Crohns disease activity Index. J Pediatr Gastroenterol Nutr.
1991;12:439447.
20. Wang SL, Ohrmund L, Hauenstein S, et al. Development and validation
of a homogeneous mobility shift assay for the measurement of iniximab

Pediatric Crohns Disease

21.
22.
23.

24.
25.
26.
27.
28.
29.

and antibodies-to-iniximab levels in patient serum. J Immunol Methods.

2012;382:177188.
Gibson DJ, Heetun ZS, Redmond CE, et al. An accelerated iniximab induction regimen reduces the need for early colectomy in patients with acute
severe ulcerative colitis. Clin Gastroenterol Hepatol. 2015;13:330335.e1.
Lam MC, Lee T, Atkinson K, et al. Time of iniximab therapy initiation
and dose escalation in Crohns disease. World J Gastroenterol. 2014;20:
214218.
Kobayashi T, Suzuki Y, Motoya S, et al. First trough level of iniximab at
week 2 predicts future outcomes of induction therapy in ulcerative colitisresults from a multicenter prospective randomized controlled trial and its
post hoc analysis. J Gastroenterol. 2016;51:241245.
Ingerski LM, Baldassano RN, Denson LA, et al. Barriers to oral medication adherence for adolescents with inammatory bowel disease. J Pediatr
Psychol. 2010;35:683691.
LeLeiko NS, Lobato D, Hagin S, et al. 6-Thioguanine levels in pediatric
IBD patients: adherence is more important than dose. Inamm Bowel Dis.
2013;19:26522658.
Church PC, Guan J, Walters TD, et al. Iniximab maintains durable
response and facilitates catch-up growth in luminal pediatric Crohns
disease. Inamm Bowel Dis. 2014;20:11771186.
Colombel JF, Sandborn WJ, Reinisch W, et al. Iniximab, azathioprine, or
combination therapy for Crohns disease. N Engl J Med. 2010;362:
13831395.
Feagan BG, McDonald JW, Panaccione R, et al. Methotrexate in combination with iniximab is no more effective than iniximab alone in patients with Crohns disease. Gastroenterology. 2014;146:681688.e1.
Grossi V, Lerer T, Grifths A, et al. Concomitant use of immunomodulators affects the durability of iniximab therapy in children with Crohns
disease. Clin Gastroenterol Hepatol. 2015;13:17481756.

www.ibdjournal.org |

1377

Copyright 2016 Crohns & Colitis Foundation of America, Inc. Unauthorized reproduction of this article is prohibited.