- Email: [email protected]
Prevention of Antidrug Antibody Formation to Infliximab in Crohn's Patients With Prior Failure of Thiopurines
Accepted Manuscript Prevention of anti-drug antibody formation to infliximab in Crohn's patients with prior failure of thiopurines Haggai Bar-Yoseph, Matti Waterman, Ronit Almog, Thomas Billiet, Séverine Vermeire, Bella Ungar, Henit Yanai, Iris Dotan, Shomron Ben-Horin, Yehuda Chowers PII: DOI: Reference:
S1542-3565(16)30375-5 10.1016/j.cgh.2016.06.028 YJCGH 54819
To appear in: Clinical Gastroenterology and Hepatology Accepted Date: 21 June 2016 Please cite this article as: Bar-Yoseph H, Waterman M, Almog R, Billiet T, Vermeire S, Ungar B, Yanai H, Dotan I, Ben-Horin S, Chowers Y, Prevention of anti-drug antibody formation to infliximab in Crohn's patients with prior failure of thiopurines, Clinical Gastroenterology and Hepatology (2016), doi: 10.1016/ j.cgh.2016.06.028. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT 1
Crohn's patients with prior failure of thiopurines
2
Haggai Bar-Yoseph,1 Matti Waterman,2* Ronit Almog,3 Thomas Billiet,4 Séverine Vermeire,4
3
Bella Ungar,5 Henit Yanai,6 Iris Dotan,6 Shomron Ben-Horin,5 Yehuda Chowers2
4
RI PT
Prevention of anti-drug antibody formation to infliximab in
5
School of Medicine, Technion Israel Institute of Technology, Haifa, Israel
6
2 Department of Gastroenterology, Rambam Health Care Campus & Bruce Rappaport School
7
of Medicine, Technion Israel Institute of Technology, Haifa, Israel
8
M AN U
SC
1 Department of Internal Medicine H, Rambam Health Care Campus & Bruce Rappaport
3 Department of Epidemiology, Rambam Health Care Campus & School of Public Health,
9
University of Haifa, Haifa, Israel
10
11
5 Department of Gastroenterology, Chaim Sheba Medical Center, Ramat-gan & Sackler
12
TE D
4 Department of Gastroenterology, University Hospitals Leuven, Leuven, Belgium
13
6 IBD Center, Department of Gastroenterology, Tel Aviv Sourasky Medical Center &
14
Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel
15
* First and second authors had equal contribution to this study
16
AC C
EP
School of Medicine, Tel Aviv University, Tel Aviv, Israel
Short title: Thiopurine response & infliximab immunogenicity
17
This study was presented at the ECCO congress, February 2015, Barcelona, Spain & at the
18
DDW congress, May 2015, Washington DC, USA
19
This study was partially supported by a generous grant from the Leona M. and Harry B.
20
Helmsley Charitable Trust
21
Disclosures – all authors have nothing to disclose with respect to the current study
22 1
ACCEPTED MANUSCRIPT 23
Study concept and design - HBY, MW, RA, SBH, YC
24
Acquisition of data– HBY, MW, TB, BU, HY
25
Analysis and interpretation of data - HBY, MW, RA,SBH, YC
26
Drafting of the manuscript - HBY, MW, RA, YC
27
RI PT
Authors contribution:
Critical revision of the manuscript for important intellectual content - HBY, MW, RA, TB,
28
SV, ID, SBH, YC
29 30
SC
Statistical analysis - RA
Haggai Bar-Yoseph, MD Department of Internal Medicine H,
Haifa, 3109601 Israel
32 33 34 35
TE D
Rambam Health Care Campus
M AN U
Corresponding author details:
31
36 37
Cellular: +972-50-2064878
38
EP
E-mail: [email protected]
Fax: +972-4-7771620
AC C
39
2
ACCEPTED MANUSCRIPT 40
Background
41
Combination thiopurine-infliximab (IFX) therapy is associated with reduced generation of
42
anti-drug antibodies (ADA) compared with IFX monotherapy. Whether past clinical response
43
RI PT
ABSTRACT
to thiopurine therapy bears an effect on ADA prevention is unknown.
44
Methods
45
SC
This was a retrospective observational multicenter study of Crohn's disease (CD) patients
46 47
into past thiopurine response or its lack of, de-novo combination or IFX monotherapy. The
48
primary end-point was risk of ADA appearance.
49
Results
50
M AN U
treated by IFX and thiopurines who had serial ADA measurements. Therapy was classified
51
final analysis. The 1 year cumulative risk of ADA development was similar in past thiopurine
52
responders (19.3%) compared to past thiopurine failures (16.1%) (Log-Rank p=0.54).
53
ADA were found in 46.6% of the monotherapy group and was significantly different
54
compared to past thiopurine responders (p=0.007) and past thiopurine failures (p=0.007). The
55
adjusted hazards for ADA development were significantly lower in past responders and past
56
AC C
EP
TE D
Out of 494 patients with serial ADA measurements 207 eligible patients were included in the
failures compared with the monotherapy group ((HR 0.47 (95% CI 0.22-1.00) and HR 0.32
57
(95% CI 0.11-0.93), respectively).
58
Conclusion
59
Thiopurines-IFX co-therapy in CD patients is associated with reduced ADA formation
60
compared with IFX monotherapy. This is probably regardless of initial thiopurine therapeutic
61
effect.
62
3
ACCEPTED MANUSCRIPT Key words: Crohn's disease, infliximab, thiopurines, anti-drug antibodies
AC C
EP
TE D
M AN U
SC
RI PT
63
4
64
Loss of response (LOR) to anti-TNF agents occurs in up to 13% of patients annualy1 and
65
significantly impacts long-term efficacy of these agents. Anti-drug immunogenicity is one of
66
the main causes for LOR. Several studies demonstrated that IFX-thiopurines co-therapy
67
(combination) reduces anti-drug antibody (ADA) The mechanism by which thiopurines exert
68
their effect is still not fully elucidated, suggesting that the ADA-protective effect may differ
69
from their clinical anti-inflammatory effect. However, combination therapy is associated with
70
enhanced immune suppression and increased risks for infection and malignancy2–4. Thus,
71
identifying the sub-groups of patients who benefit most from combination therapy is a
72
pertinent issue, which may help avoiding combination therapy in patients unlikely to benefit
73
from such intervention. Specifically, it is hitherto unknown whether patients who clinically
74
failed prior thiopurine therapy still benefit from combination therapy with respect to
75
prevention of anti-drug immunogenicity, or whether this practice exposes them to
76
unnecessary risks, which are not associated with a clinical benefit.
77
Therefore, we aimed to examine the impact of past thiopurine response on ADA appearance
78
in patients with Crohn's disease (CD) co-treated with IFX. A secondary objective was to
79
EP
TE D
SC
RI PT
INTRODUCTION
M AN U
ACCEPTED MANUSCRIPT
80
groups to the monotherapy group.
81
AC C
compare the risk of ADA appearance between the three IFX-thiopurine combination sub-
5
ACCEPTED MANUSCRIPT 82
Patients
83
This was a retrospective observational cohort study. Inclusion criteria were patients with CD
84
treated by ≥ 4 scheduled IFX infusions (0, 2, 6 and 14 weeks) who had a documentation of
85
RI PT
METHODS
86
trough ADA measurements. Patients who had early (before week 14) appearance of ADA,
87
and did not reach week 14 infusion, were also included in this study. Patients treated with
88
methotrexate-IFX co-therapy, patients treated by episodic IFX therapy, those who did not
89
receive induction therapy, or those who had a long 'drug holiday' (more than 10 weeks
90
M AN U
SC
response/non-response to past/ongoing thiopurine treatment (see below), and had serial
91
permitted. All data were collected from patients' medical records. Patients were classified
92
into four groups: IFX monotherapy and three combined IFX-thiopurine groups according to
93
classification of past outcome of thiopurine therapy: primary failure, past response and de-
94
novo combination (no thiopurine exposure before IFX initiation). The internal review board
95
of each institution authorized the study.
96
Past thiopurine response definitions
97
EP
TE D
between treatments) were excluded from analysis. Past exposure to adalimumab was
98
adequate dosing (>1 mg/kg for mercaptopurine (6-MP) and >2 mg/kg for azathioprine (AZA)
99
if no adverse effects were reported) and good adherence to therapy as documented by the
100
treating physician. Active disease was defined in the study protocol as symptomatic patients
101
who failed to discontinue concurrent steroids, or had to switch/add biological therapy or
102
another immunomodulator, or needed disease related surgery. Past response group included
103
patients with secondary thiopurine failure, defined as patients who had disease exacerbation
104
after a good initial response (>6 months steroid independent) to thiopurines. Physicians and
105
AC C
Primary thiopurine failure was defined as active disease despite more than 3 months of
6
ACCEPTED MANUSCRIPT 106
thiopurine exposure were excluded from the study.
107
ADA & drug level assays
108
ADA and drug levels were performed on trough serum samples at each center laboratory: Tel
109
RI PT
patients were contacted if no data was found in the medical record. Patients with no data on
110
systems (Frankfurt, Germany), Leuven University Hospital (LUH) utilized a self-developed
111
bridging ELISA as previously described5. Sheba Medical Center (SMC) and Rambam Health
112
SC
Aviv Sourasky Medical Center (TASMC) utilized kits manufactured by Immunodiagnostik
113
assay, as previously described6. Since no significant differences in drug serum levels between
114
M AN U
Care Campus (RHCC) determined ADA level using a self-developed anti-lambda chain
different kits were reported in the past7, drug level analysis compiled drug concentration data
115
from all centers.
116
117
The primary end-point was ADA appearance during the first treatment year, defined as a
118
TE D
Primary end-point and end of follow-up
119
Patients were followed and included in the analysis up to the primary end point or one of the
120
end of follow up points: last follow up visit or first change in therapy (shortening IFX
121
EP
positive ADA result according to the positivity cutoff level as defined by each center.
122
oral steroids) or change in thiopurine therapy).
123
Additional potential determinants of ADA appearance
124
Collected information included: age, gender, ethnicity, smoking status, BMI, age at
125
diagnosis, disease behavior and anatomical location according to the Montreal classification,
126
extra-intestinal manifestations, perianal disease, past surgery, concomitant IV steroid
127
treatment (200 mg IV hydrocortisone) and time from diagnosis to start of IFX.
128
Clinical status at the end of follow up
129
AC C
intervals, IFX dose escalation, IFX discontinuation, immunomodulator drug added (including
7
ACCEPTED MANUSCRIPT 130
possible. Therefore, response and remission were defined according to the treating
131
physician’s comments in the patient’s chart. To objectively substantiate the clinical
132
observation, improvement from pre-treatment baseline of biomarkers of inflammation was
133
also required. These included serum CRP, fecal calprotectin and/or endoscopic activity
134
assessment before and after IFX/thiopurine initiation.
135
136
SC
Statistical analysis
RI PT
Given the retrospective nature of this study a standard definition of response was not
137
categorical baseline demographic and clinical variables between the 4 compared treatment
138
M AN U
The χ2 test and Fisher exact test were performed to determine significant differences in
139
variables according to data distribution. Kaplan-Meier analysis was used to compare ADA
140
free survival between the 4 groups. The differences in survival between the groups were
141
evaluated with log-rank test. Because ADA mostly appear during the first year of therapy8
142
and due to the small number of patients in each of the combined treatment groups we limited
143
the survival analysis to 1 year follow up. In order to detect a difference of 20% in the 1 year
144
cumulative incidence of ADA with power of 0.8 and an α of 0.05 we needed 60 patients in
145
EP
TE D
groups. ANOVA or the non-parametric Kruskal-Wallis tests were used for continuous
146
allowed detection of such difference with power of 0.6.
147
AC C
each thiopurine response group. The actual sample size that we reached of 34 and 52 patients
Cox proportional hazards model was constructed in order to determine the independent
148
hazard ratios (HRs) of 1 year ADA appearance risk by past thiopurine response and between
149
1 year ADA appearance risk in combination treatments compared to monotherapy, adjusting
150
for other potential determinants of ADA appearance. HRs for ADA appearance between the
151
compared groups and 95% confidence intervals (CIs) were calculated. Covariates at the
152
model included demographic and clinical characteristics that were significantly different
153
between the 4 treatment groups (Table 1). Study center was entered to the multivariate model
154
8
ACCEPTED MANUSCRIPT 155
the use of different ADA assays in different centers. Statistical significance was defined as a
156
two-tailed P<0.05. All analyses were performed using SPSS software, version 19.
157
AC C
EP
TE D
M AN U
SC
RI PT
because of differences (although not statistically significant) in treatment groups and due to
9
ACCEPTED MANUSCRIPT 158
Study groups
159
Four hundred ninety four patients with serial ADA measurements collected between 2008
160
and 2014 were reviewed. Two hundred and eighty seven patients were excluded from
161
RI PT
RESULTS
162
analysis: 93 received IFX monotherapy and 114 received IFX-thiopurines co-therapy (34
163
were defined as previous primary thiopurines failures, 52 defined as previous primary
164
thiopurines responders and 28 received de-novo combination treatment). The median follow
165
up duration (as defined by study protocol) across all study groups was 7.4 months (IQR 4.9-
166
M AN U
SC
analysis as shown in Figure 1. Two hundred and seven patients were included in the final
17.8).
167
168
differences were found between study groups in most demographic and clinical parameters.
169
The age at CD diagnosis was older in the de novo group albeit not clinically significant. Time
170
TE D
Table 1 presents the baseline characteristics of patients by study group. No significant
171
groups. Penetrating disease behavior was significantly less frequent in patients with past
172
thiopurine response. The proportion of patients who received IV steroids prior to IFX dose
173
EP
from diagnosis to IFX use was shorter in the de-novo combination group compared with other
174
thiopurine responders groups.
175
Patients reached end of follow up (as stated above) due to the following: ADA appearance in
176
36.7%, 31.4% stopped/changed IFX regimen, 18.4% reached their last follow up, 12.6%
177
changed their thiopurine status and only 1% was lost to follow up.
178
AC C
was more common in the primary failure group compared with IFX monotherapy and past
179
Figure 1 – Flow chart of patient exclusion and study groups
180
ADA appearance
181 10
ACCEPTED MANUSCRIPT 182
monotherapy and all patients treated with thiopurines regardless to their past response (Figure
183
2a). No difference was observed in the 1 year cumulative risk for ADA development between
184
past thiopurine responders (19.3%) and past thiopurine failures (16.1%) (Log-Rank p=0.54).
185
The highest risk for ADA was found in the monotherapy group (46.6%) and was significantly
186
different compared to past thiopurine treatment groups - 19.3% for past responders (p=0.007)
187
and 16.1% for past failures (p=0.007). A trend for difference between monotherapy and de-
188
novo combination group was observed (21.9%), but under-powered for statistical significance
189
SC
RI PT
Kaplan-Meier analysis revealed a significant difference between patients treated by IFX
190
at diagnosis, time to IFX onset, study center and IV steroid pre-treatment, no significant
191
difference in ADA development was observed between past thiopurine failures compared to
192
responders (HR 0.7 (95% CI 0.2-2.3, p=0.56)) (Table 2). The hazards for ADA development
193
were significantly lower in past responders and past failures compared with the monotherapy
194
group (HR 0.47 (95% CI 0.22-1.00) and HR 0.32 (95% CI 0.11-0.93), respectively), while a
195
TE D
M AN U
(p=0.102) (Figure 2b). In a multivariate analysis, adjusting for Montreal B classification, age
196
1.12)). Age at diagnosis significantly affected ADA formation. The type of thiopurine in the
197
combination groups did not significantly impact ADA appearance: 24.6% developed ADA on
198
AZA and 17.9% developed ADA on 6-MP (p=NS).
199
AC C
EP
non-significant trend was noted in the de-novo combination group (HR 0.47 (95% CI 0.19-
We next looked at differences in the dynamics of ADA formation between the groups. The
200
cumulative risk of ADA development at 6 months was also significantly higher in the
201
monotherapy group (35.1%) vs. past responders (13.9%, p=0.014) and past failures (9.6%,
202
p=0.012). However, the frequency of ADA formation in the monotherapy and de-novo
203
combination group curves diverges only after 5 months of therapy. Cumulative risk of
204
ADA development at 5 months from IFX initiation was (24.7%) in the monotherapy group
205
and (21.9%) in the de-novo combination group (p=0.728). In the following 7 months up to 1
206
11
ACCEPTED MANUSCRIPT 207
in the monotherapy group (p=0.024). Of note, the vast majority of patients with positive
208
ADA had persistently positive ADA measurements (not transiently detected).
209
Forty patients on IFX monotherapy received IV steroids pre-treatment. Interestingly,
210
corticosteroid pre-medication did not alter ADA appearance in the IFX monotherapy group
211
(41.0% with pre-treatment vs. 44.2% without, p=0.760).
212
RI PT
year, no new ADA events were recorded in the de-novo combination group (0%) vs. (29.9%)
SC
Study center was found to significantly impact ADA positivity (Table 2): it was significantly
213 214
all model variables study center still had overall significant effect on ADA appearance (p=
215
M AN U
more common at RHCC and SMC asP<0.001 compared with LUA and TASMC. After adjustment to
0.028). When monotherapy was compared to all combination groups within each center’s
216
cohort, ADA positivity was more common during monotherapy in every center, though this
217
did not reach statistical significance due to the small sample size in each center.
218
TE D
219
Figure 2a - ADA cumulative incidence in infliximab monotherapy versus all three
220
combination therapy groups compiled
221
EP
222
Figure 2b – Cumulative incidence of ADA in the four study groups
AC C
223
Drug levels
224
Significantly lower IFX serum concentrations were observed in ADA positive compared with
225
ADA negative patients at the end of follow up:P=0.102 the median was 0.3 mcg/ml (IQR 0.00-1.66)
226
P=0.007
P=0.007 for ADA negative, p<0.001. This effect for ADA positive and 2.33 mcg/ml (IQR 0.76-5.67)
227
occurred in all study groups, regardless of the therapy or treating center. The proportion of
228
patients with IFX drug levels less than 3 mcg/ml at the end of follow up was higher in the
229
12
ACCEPTED MANUSCRIPT 230
respectively), but was not significant (p=0.146).
231
Clinical status at end of follow up
232
Due to small sample size all thiopurine combination groups were compiled and compared to
233
RI PT
monotherapy group compared with all combination groups compiled (70.0% and 60.2%,
234
group were more likely to have active disease (21.5% vs. 8.8%, respectively), and
235
subsequently less likely to be in remission (40.9% compared to 58.8% of the combination
236
SC
monotherapy group. As compared to the combination groups, patients in the monotherapy
237
patients and ADA negative patient (though not significant). No significant differences in
238
M AN U
group)(p=0.009) (Figure 3). This trend was consistent after stratifying to ADA positive
clinical response were noted between the different combination groups.
239
240
241
combination
242
AC C
EP
TE D
Figure 3 – Clinical status at end of follow up in IFX monotherapy and IFX-thiopurine
13
ACCEPTED MANUSCRIPT 243
The findings from our retrospective cohort study suggest that adding a thiopurine to IFX
244
therapy prevents ADA formation independently of the past clinical response to thiopurines.
245
This effect was consistent after adjustment to clinical and treatment related factors.
246
RI PT
DISCUSSION
247
impact, since more patients in the combination groups (compiled) were in clinical remission
248
at the end of the study compared to the monotherapy group. However, an additional
249
explanation for the better clinical status of the combination groups compared to monotherapy
250
may be direct pharmacological effects of thiopurines. This was possibly supported by the
251
M AN U
SC
The results support the contention that prevention of ADA formation probably has a clinical
252
significant due to small numbers of patients. Not surprisingly, patients with ADA had lower
253
serum drug levels than those without ADA. A non-statistically significant trend was observed
254
for a higher proportion of patients with IFX drug levels less than 3 mcg/ml in the
255
monotherapy group compared with the combination groups. We had an unusually high
256
proportion (60-70%) of patients with IFX concentrations <3 mcg/ml at the study endpoint.
257
Likely, this stemmed from a bias introduced by the study's end-point definition of LOR or
258
EP
TE D
observation that the difference remained after stratification to ADA formation, although not
259
The present results need to be interpreted with caution. Because the study was retrospective it
260
may have been underpowered to detect a possible difference between past responders and
261
past non-responders to thiopurine. Notwithstanding, the present observations suggest that
262
even if such difference does exist, it is probably of small effect size, especially when
263
projected against the significant difference in ADA formation rate between each of these
264
groups and the monotherapy group.
265
AC C
ADA development.
14
ACCEPTED MANUSCRIPT 266
monotherapy groups during the first 5 months of therapy. However, patients in the
267
monotherapy group continued to form ADA throughout the study follow up period, while no
268
new ADA developed in the de-novo combination group thereafter, converging with the other
269
combination groups at 1 year. It appears therefore that the lag time between initiation of
270
thiopurines and their effect on ADA prevention is similar to the 3-6 months lag between
271
thiopurine initiation and their effect on clinical response as previously observed9. To avoid
272
such lag phenomenon it might be recommended to continue thiopurines when the decision to
273
SC
RI PT
We observed that ADA formation was similar in the de-novo combination group and
274
difference in cumulative incidence of ADA between patients receiving immunomodulator
275
before IFX or in de-novo combination therapy, both groups having significantly lower ADA
276
levels compared with monotherapy10. However, most of these patients were treated
277
episodically. Thus, prevention of "early" ADA might not be clinically relevant.
278
Our study had some limitations. The first is its retrospective nature of clinical data collection
279
and methodological limitations associated with this approach. However, because the study
280
was based on real life cohorts from four different centers in two countries, the data presented
281
EP
TE D
M AN U
start IFX is made in order to prevent "early" ADA appearance. Vermeire at al. reported no
282
exclusion of 287/494 of patients reviewed may limit the relevance of our finding to other
283
AC C
herein have a lower propensity for specific treatment bias. A potential bias related to the
patient populations, though exclusion criteria were explicit and led to a homogenic and
284
specific study cohort, relevant to study's question. Second, patients’ adherence and actual 6-
285
TGN blood levels of thiopurines metabolites could not be assessed in this study and may
286
impact the results. However, as shown in Table 1, no difference in reported weight based
287
thiopurine dose (mg/kg) was found between past thiopurine exposure groups. The planned
288
sample size in the combined treatment groups was not reached and thus a type II mistake
289
should be considered as one of the explanations for detecting no difference. Lastly, ADA
290
15
ACCEPTED MANUSCRIPT 291
more common in the two centers using the anti-lambda assay. In an attempt to address this
292
difference, adjustment to medical center as a co-variate in the multivariate analysis was done
293
(though no significant difference in distribution between study groups was noted between
294
study centers) but was not found to alter the results. To avoid interference of serum ADA by
295
serum IFX we looked at a harder endpoint of ADA positivity and serum IFX<3 mcg/ml. The
296
difference in ADA formation between study groups was not changed in univariate analysis,
297
and the same trend was noted in multivariate analysis but was not significant (data not
298
SC
RI PT
assays and ADA positivity rates differed between study's centers, with ADA positivity being
299
reported in past studies10–14. The validity of ADA positivity is also supported by the
300
consistency of our results with other published studies regarding the correlation between
301
ADA and IFX drug levels and clinical status.
302
In conclusion, in this 'real life' study we observed that adding a thiopurine to IFX therapy is
303
associated with lower rate of ADA appearance, probably regardless of patient's past clinical
304
thiopurine response. Further studies are needed to increase our understanding of factors
305
governing immunogenicity of infliximab in some but not all patients and to elucidate the
306
optimal methods to prevent it.
307
AC C
EP
TE D
M AN U
shown). Of note, the overall rate of ADA formation in our study was in the range of that
16
ACCEPTED MANUSCRIPT References: 1.
Gisbert JP, Panés J. Loss of response and requirement of infliximab dose intensification in Crohn’s disease: a review. Am J Gastroenterol. 2009;104(3):760-767. Lichtenstein GR, Feagan BG, Cohen RD, et al. Drug therapies and the risk of
RI PT
2.
malignancy in Crohn’s disease: results from the TREATTM Registry. Am J Gastroenterol. 2014;109(2):212-223.
Beaugerie L, Brousse N, Bouvier AM, et al. Lymphoproliferative disorders in patients
SC
3.
receiving thiopurines for inflammatory bowel disease: a prospective observational
4.
M AN U
cohort study. Lancet (London, England). 2009;374(9701):1617-1625. Osterman MT, Haynes K, Delzell E, et al. Effectiveness and Safety of Immunomodulators with Anti-TNF Therapy in Crohn’s Disease. Clin Gastroenterol Hepatol. 2015;13(7):1293-1301.e5.
Vande Casteele N, Buurman DJ, Sturkenboom MGG, et al. Detection of infliximab
TE D
5.
levels and anti-infliximab antibodies: a comparison of three different assays. Aliment
6.
EP
Pharmacol Ther. 2012;36(8):765-771.
Bein-Horn S, Yavzori M, Katz L, et al. The immunogenic part of infliximab is the
AC C
F(ab’)2, but measuring antibodies to the intact infliximab molecule is more clinically useful. Gut. 2011;60(1):41-48.
7.
Steenholdt C, Bendtzen K, Brynskov J, et al. Cut-off levels and diagnostic accuracy of infliximab trough levels and anti-infliximab antibodies in Crohn’s disease. Scand J Gastroenterol. 2011;46(3):310-318.
8.
Ungar B, Chowers Y, Yavzori M, et al. The temporal evolution of antidrug antibodies in patients with inflammatory bowel disease treated with infliximab. Gut. 2014;63(8):1258-1264. 17
ACCEPTED MANUSCRIPT 9.
Su C, Lichtenstein GR. Treatment of inflammatory bowel disease with azathioprine and 6-mercaptopurine. Gastroenterol Clin North Am. 2004;33(2):209-234, viii.
10.
Vermeire S, Noman M, Van Assche G, et al. Effectiveness of concomitant
Crohn’s disease. Gut. 2007;56(9):1226-1231. 11.
RI PT
immunosuppressive therapy in suppressing the formation of antibodies to infliximab in
Baert F, Noman M, Vermeire S, et al. Influence of immunogenicity on the long-term
12.
SC
efficacy of infliximab in Crohn’s disease. N Engl J Med. 2003;348(7):601-608. Colombel JF, Sandborn WJ, Reinisch W, et al. Infliximab, Azathioprine, or
13.
M AN U
Combination Therapy for Crohn’s Disease. N Engl J Med. 2010;362(15):1383-1395. Fefferman DS, Farrell RJ. Immunogenicity of biological agents in inflammatory bowel disease. Inflamm Bowel Dis. 2005;11(5):497-503.
Hanauer SB, Feagan BG, Lichtenstein GR, et al. Maintenance Infliximab for Crohn’s
EP
TE D
Disease: The ACCENT I Randomised Trial. Vol 359.; 2002:1541-1549.
AC C
14.
18
ACCEPTED MANUSCRIPT Table 1 – clinical and demographic parameters
Monotherapy (n=93)
Primary failure (n=34)
Past responders (n=52)
De-novo combination (n=28)
P value
52.7
41.2
44.2
57.1
0.467
Gender (% Female) Country of residence % Israel Belgium Smoking % Never
72.0 28.0
70.6 29.4
RI PT
0.358 67.3 32.7
85.7 14.3
0.533
76.5
63.3
69.2
Past
13.8
5.9
6.1
11.5
Current
23
17.6
30.6
19.2
22.9 (4.6)
21.2 (5.5)
21.5 (3.7)
22.2 (3.5)
0.121
22.7 (17.9-33.7)
19.3 (13.323.2)
19.6 (15.322.5)
23.5 (18.928.1)
0.001
9.4 (4.5-17.0)
1.8 (0.6-9.5)
<0.001
Age at diagnosis (y), Median (IQR) Disease duration (y) till IFX initiation, Median (IQR) Disease location (Montreal) % L1
7.6 (1.6-16.4)
L2
AC C
B2
B3 Perianal disease % Extra-intestinal manifestations % Past surgery % Past thiopurine dose (mg/kg) Median (IQR)
26.9
35.7
22.6
17.6
13.5
17.9
49.5
47.1
59.6
46.4
10.8
5.9
15.4
3.6
0.358 0.053
39.8
55.9
28.8
57.1
25.8
20.6
34.6
7.1
34.4 50.5
23.5 58.8
36.5 40.4
35.7 39.3
0.273
41.9
44.1
41.2
33.3
0.843
29.0
32.4
38.5
25.0
0.574
43.5
n=27 2.0 (1.7-2.3) n=7 1.2 (1.1-1.5) 55.9
n=32 2.1 (1.8-2.4) n=20 1.3 (1.1-1.4) 25.0
Azathioprine (AZA) 6-MP IV steroids % IBD Center %
0.749
35.3
EP
L3 Upper GI involvement % Disease behavior (Montreal) % B1
4.0 (1.2-8.6)
TE D
28.0
M AN U
BMI, Mean (±SD)
SC
63.2
0.474 0.580 39.3
0.031 0.134 19
ACCEPTED MANUSCRIPT 23.6
11.8
32.7
39.3
32.2
41.2
15.4
32.1
16.2
17.6
19.2
14.2
28.0
29.4
32.7
14.3
AC C
EP
TE D
M AN U
SC
RI PT
Rambam Health Care Campus Sheba Medical Center Tel-Aviv Medical Center Leuven University Hospital
20
ACCEPTED MANUSCRIPT Table 2 – Multivariable analysis (N=204)*
N
P value
HR (CI 95%)
0.034 51
1.00
Monotherapy
91
2.14 (1.00-4.60)
Past thiopurine failure
34
0.70 (0.20-2.30)
De-novo combination
28
1.00 (0.34-2.90)
ref.
0.051 0.560 1.000
SC
Past thiopurine responders
RI PT
Treatment group
Institute
0.028
Sheba Medical Center Tel-Aviv Medical Center Leuven University Hospital
1.00
ref.
59
1.31 (0.63-2.71)
0.460
35
0.35 (0.11-1.09)
0.070
57
0.66 (0.31-1.39)
0.270
TE D
Montreal B
53
M AN U
Rambam Health Care Campus
B1
0.169
86
1.00
ref.
50
0.96 (0.44-2.09)
0.910
68
1.66 (0.89-3.12)
0.110
IV steroids pre-treatment (Yes)
82
1.07 (0.52-2.21)
0.850
Age at diagnosis
204
1.02 (1.00-1.05)
0.030
Time from diagnosis to IFX
204
1.02 (0.98-1.05)
0.350
B2
AC C
EP
B3
*3 patients were not included due to missing data
21
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
S1542-3565(16)30375-5 10.1016/j.cgh.2016.06.028 YJCGH 54819
To appear in: Clinical Gastroenterology and Hepatology Accepted Date: 21 June 2016 Please cite this article as: Bar-Yoseph H, Waterman M, Almog R, Billiet T, Vermeire S, Ungar B, Yanai H, Dotan I, Ben-Horin S, Chowers Y, Prevention of anti-drug antibody formation to infliximab in Crohn's patients with prior failure of thiopurines, Clinical Gastroenterology and Hepatology (2016), doi: 10.1016/ j.cgh.2016.06.028. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT 1
Crohn's patients with prior failure of thiopurines
2
Haggai Bar-Yoseph,1 Matti Waterman,2* Ronit Almog,3 Thomas Billiet,4 Séverine Vermeire,4
3
Bella Ungar,5 Henit Yanai,6 Iris Dotan,6 Shomron Ben-Horin,5 Yehuda Chowers2
4
RI PT
Prevention of anti-drug antibody formation to infliximab in
5
School of Medicine, Technion Israel Institute of Technology, Haifa, Israel
6
2 Department of Gastroenterology, Rambam Health Care Campus & Bruce Rappaport School
7
of Medicine, Technion Israel Institute of Technology, Haifa, Israel
8
M AN U
SC
1 Department of Internal Medicine H, Rambam Health Care Campus & Bruce Rappaport
3 Department of Epidemiology, Rambam Health Care Campus & School of Public Health,
9
University of Haifa, Haifa, Israel
10
11
5 Department of Gastroenterology, Chaim Sheba Medical Center, Ramat-gan & Sackler
12
TE D
4 Department of Gastroenterology, University Hospitals Leuven, Leuven, Belgium
13
6 IBD Center, Department of Gastroenterology, Tel Aviv Sourasky Medical Center &
14
Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel
15
* First and second authors had equal contribution to this study
16
AC C
EP
School of Medicine, Tel Aviv University, Tel Aviv, Israel
Short title: Thiopurine response & infliximab immunogenicity
17
This study was presented at the ECCO congress, February 2015, Barcelona, Spain & at the
18
DDW congress, May 2015, Washington DC, USA
19
This study was partially supported by a generous grant from the Leona M. and Harry B.
20
Helmsley Charitable Trust
21
Disclosures – all authors have nothing to disclose with respect to the current study
22 1
ACCEPTED MANUSCRIPT 23
Study concept and design - HBY, MW, RA, SBH, YC
24
Acquisition of data– HBY, MW, TB, BU, HY
25
Analysis and interpretation of data - HBY, MW, RA,SBH, YC
26
Drafting of the manuscript - HBY, MW, RA, YC
27
RI PT
Authors contribution:
Critical revision of the manuscript for important intellectual content - HBY, MW, RA, TB,
28
SV, ID, SBH, YC
29 30
SC
Statistical analysis - RA
Haggai Bar-Yoseph, MD Department of Internal Medicine H,
Haifa, 3109601 Israel
32 33 34 35
TE D
Rambam Health Care Campus
M AN U
Corresponding author details:
31
36 37
Cellular: +972-50-2064878
38
EP
E-mail: [email protected]
Fax: +972-4-7771620
AC C
39
2
ACCEPTED MANUSCRIPT 40
Background
41
Combination thiopurine-infliximab (IFX) therapy is associated with reduced generation of
42
anti-drug antibodies (ADA) compared with IFX monotherapy. Whether past clinical response
43
RI PT
ABSTRACT
to thiopurine therapy bears an effect on ADA prevention is unknown.
44
Methods
45
SC
This was a retrospective observational multicenter study of Crohn's disease (CD) patients
46 47
into past thiopurine response or its lack of, de-novo combination or IFX monotherapy. The
48
primary end-point was risk of ADA appearance.
49
Results
50
M AN U
treated by IFX and thiopurines who had serial ADA measurements. Therapy was classified
51
final analysis. The 1 year cumulative risk of ADA development was similar in past thiopurine
52
responders (19.3%) compared to past thiopurine failures (16.1%) (Log-Rank p=0.54).
53
ADA were found in 46.6% of the monotherapy group and was significantly different
54
compared to past thiopurine responders (p=0.007) and past thiopurine failures (p=0.007). The
55
adjusted hazards for ADA development were significantly lower in past responders and past
56
AC C
EP
TE D
Out of 494 patients with serial ADA measurements 207 eligible patients were included in the
failures compared with the monotherapy group ((HR 0.47 (95% CI 0.22-1.00) and HR 0.32
57
(95% CI 0.11-0.93), respectively).
58
Conclusion
59
Thiopurines-IFX co-therapy in CD patients is associated with reduced ADA formation
60
compared with IFX monotherapy. This is probably regardless of initial thiopurine therapeutic
61
effect.
62
3
ACCEPTED MANUSCRIPT Key words: Crohn's disease, infliximab, thiopurines, anti-drug antibodies
AC C
EP
TE D
M AN U
SC
RI PT
63
4
64
Loss of response (LOR) to anti-TNF agents occurs in up to 13% of patients annualy1 and
65
significantly impacts long-term efficacy of these agents. Anti-drug immunogenicity is one of
66
the main causes for LOR. Several studies demonstrated that IFX-thiopurines co-therapy
67
(combination) reduces anti-drug antibody (ADA) The mechanism by which thiopurines exert
68
their effect is still not fully elucidated, suggesting that the ADA-protective effect may differ
69
from their clinical anti-inflammatory effect. However, combination therapy is associated with
70
enhanced immune suppression and increased risks for infection and malignancy2–4. Thus,
71
identifying the sub-groups of patients who benefit most from combination therapy is a
72
pertinent issue, which may help avoiding combination therapy in patients unlikely to benefit
73
from such intervention. Specifically, it is hitherto unknown whether patients who clinically
74
failed prior thiopurine therapy still benefit from combination therapy with respect to
75
prevention of anti-drug immunogenicity, or whether this practice exposes them to
76
unnecessary risks, which are not associated with a clinical benefit.
77
Therefore, we aimed to examine the impact of past thiopurine response on ADA appearance
78
in patients with Crohn's disease (CD) co-treated with IFX. A secondary objective was to
79
EP
TE D
SC
RI PT
INTRODUCTION
M AN U
ACCEPTED MANUSCRIPT
80
groups to the monotherapy group.
81
AC C
compare the risk of ADA appearance between the three IFX-thiopurine combination sub-
5
ACCEPTED MANUSCRIPT 82
Patients
83
This was a retrospective observational cohort study. Inclusion criteria were patients with CD
84
treated by ≥ 4 scheduled IFX infusions (0, 2, 6 and 14 weeks) who had a documentation of
85
RI PT
METHODS
86
trough ADA measurements. Patients who had early (before week 14) appearance of ADA,
87
and did not reach week 14 infusion, were also included in this study. Patients treated with
88
methotrexate-IFX co-therapy, patients treated by episodic IFX therapy, those who did not
89
receive induction therapy, or those who had a long 'drug holiday' (more than 10 weeks
90
M AN U
SC
response/non-response to past/ongoing thiopurine treatment (see below), and had serial
91
permitted. All data were collected from patients' medical records. Patients were classified
92
into four groups: IFX monotherapy and three combined IFX-thiopurine groups according to
93
classification of past outcome of thiopurine therapy: primary failure, past response and de-
94
novo combination (no thiopurine exposure before IFX initiation). The internal review board
95
of each institution authorized the study.
96
Past thiopurine response definitions
97
EP
TE D
between treatments) were excluded from analysis. Past exposure to adalimumab was
98
adequate dosing (>1 mg/kg for mercaptopurine (6-MP) and >2 mg/kg for azathioprine (AZA)
99
if no adverse effects were reported) and good adherence to therapy as documented by the
100
treating physician. Active disease was defined in the study protocol as symptomatic patients
101
who failed to discontinue concurrent steroids, or had to switch/add biological therapy or
102
another immunomodulator, or needed disease related surgery. Past response group included
103
patients with secondary thiopurine failure, defined as patients who had disease exacerbation
104
after a good initial response (>6 months steroid independent) to thiopurines. Physicians and
105
AC C
Primary thiopurine failure was defined as active disease despite more than 3 months of
6
ACCEPTED MANUSCRIPT 106
thiopurine exposure were excluded from the study.
107
ADA & drug level assays
108
ADA and drug levels were performed on trough serum samples at each center laboratory: Tel
109
RI PT
patients were contacted if no data was found in the medical record. Patients with no data on
110
systems (Frankfurt, Germany), Leuven University Hospital (LUH) utilized a self-developed
111
bridging ELISA as previously described5. Sheba Medical Center (SMC) and Rambam Health
112
SC
Aviv Sourasky Medical Center (TASMC) utilized kits manufactured by Immunodiagnostik
113
assay, as previously described6. Since no significant differences in drug serum levels between
114
M AN U
Care Campus (RHCC) determined ADA level using a self-developed anti-lambda chain
different kits were reported in the past7, drug level analysis compiled drug concentration data
115
from all centers.
116
117
The primary end-point was ADA appearance during the first treatment year, defined as a
118
TE D
Primary end-point and end of follow-up
119
Patients were followed and included in the analysis up to the primary end point or one of the
120
end of follow up points: last follow up visit or first change in therapy (shortening IFX
121
EP
positive ADA result according to the positivity cutoff level as defined by each center.
122
oral steroids) or change in thiopurine therapy).
123
Additional potential determinants of ADA appearance
124
Collected information included: age, gender, ethnicity, smoking status, BMI, age at
125
diagnosis, disease behavior and anatomical location according to the Montreal classification,
126
extra-intestinal manifestations, perianal disease, past surgery, concomitant IV steroid
127
treatment (200 mg IV hydrocortisone) and time from diagnosis to start of IFX.
128
Clinical status at the end of follow up
129
AC C
intervals, IFX dose escalation, IFX discontinuation, immunomodulator drug added (including
7
ACCEPTED MANUSCRIPT 130
possible. Therefore, response and remission were defined according to the treating
131
physician’s comments in the patient’s chart. To objectively substantiate the clinical
132
observation, improvement from pre-treatment baseline of biomarkers of inflammation was
133
also required. These included serum CRP, fecal calprotectin and/or endoscopic activity
134
assessment before and after IFX/thiopurine initiation.
135
136
SC
Statistical analysis
RI PT
Given the retrospective nature of this study a standard definition of response was not
137
categorical baseline demographic and clinical variables between the 4 compared treatment
138
M AN U
The χ2 test and Fisher exact test were performed to determine significant differences in
139
variables according to data distribution. Kaplan-Meier analysis was used to compare ADA
140
free survival between the 4 groups. The differences in survival between the groups were
141
evaluated with log-rank test. Because ADA mostly appear during the first year of therapy8
142
and due to the small number of patients in each of the combined treatment groups we limited
143
the survival analysis to 1 year follow up. In order to detect a difference of 20% in the 1 year
144
cumulative incidence of ADA with power of 0.8 and an α of 0.05 we needed 60 patients in
145
EP
TE D
groups. ANOVA or the non-parametric Kruskal-Wallis tests were used for continuous
146
allowed detection of such difference with power of 0.6.
147
AC C
each thiopurine response group. The actual sample size that we reached of 34 and 52 patients
Cox proportional hazards model was constructed in order to determine the independent
148
hazard ratios (HRs) of 1 year ADA appearance risk by past thiopurine response and between
149
1 year ADA appearance risk in combination treatments compared to monotherapy, adjusting
150
for other potential determinants of ADA appearance. HRs for ADA appearance between the
151
compared groups and 95% confidence intervals (CIs) were calculated. Covariates at the
152
model included demographic and clinical characteristics that were significantly different
153
between the 4 treatment groups (Table 1). Study center was entered to the multivariate model
154
8
ACCEPTED MANUSCRIPT 155
the use of different ADA assays in different centers. Statistical significance was defined as a
156
two-tailed P<0.05. All analyses were performed using SPSS software, version 19.
157
AC C
EP
TE D
M AN U
SC
RI PT
because of differences (although not statistically significant) in treatment groups and due to
9
ACCEPTED MANUSCRIPT 158
Study groups
159
Four hundred ninety four patients with serial ADA measurements collected between 2008
160
and 2014 were reviewed. Two hundred and eighty seven patients were excluded from
161
RI PT
RESULTS
162
analysis: 93 received IFX monotherapy and 114 received IFX-thiopurines co-therapy (34
163
were defined as previous primary thiopurines failures, 52 defined as previous primary
164
thiopurines responders and 28 received de-novo combination treatment). The median follow
165
up duration (as defined by study protocol) across all study groups was 7.4 months (IQR 4.9-
166
M AN U
SC
analysis as shown in Figure 1. Two hundred and seven patients were included in the final
17.8).
167
168
differences were found between study groups in most demographic and clinical parameters.
169
The age at CD diagnosis was older in the de novo group albeit not clinically significant. Time
170
TE D
Table 1 presents the baseline characteristics of patients by study group. No significant
171
groups. Penetrating disease behavior was significantly less frequent in patients with past
172
thiopurine response. The proportion of patients who received IV steroids prior to IFX dose
173
EP
from diagnosis to IFX use was shorter in the de-novo combination group compared with other
174
thiopurine responders groups.
175
Patients reached end of follow up (as stated above) due to the following: ADA appearance in
176
36.7%, 31.4% stopped/changed IFX regimen, 18.4% reached their last follow up, 12.6%
177
changed their thiopurine status and only 1% was lost to follow up.
178
AC C
was more common in the primary failure group compared with IFX monotherapy and past
179
Figure 1 – Flow chart of patient exclusion and study groups
180
ADA appearance
181 10
ACCEPTED MANUSCRIPT 182
monotherapy and all patients treated with thiopurines regardless to their past response (Figure
183
2a). No difference was observed in the 1 year cumulative risk for ADA development between
184
past thiopurine responders (19.3%) and past thiopurine failures (16.1%) (Log-Rank p=0.54).
185
The highest risk for ADA was found in the monotherapy group (46.6%) and was significantly
186
different compared to past thiopurine treatment groups - 19.3% for past responders (p=0.007)
187
and 16.1% for past failures (p=0.007). A trend for difference between monotherapy and de-
188
novo combination group was observed (21.9%), but under-powered for statistical significance
189
SC
RI PT
Kaplan-Meier analysis revealed a significant difference between patients treated by IFX
190
at diagnosis, time to IFX onset, study center and IV steroid pre-treatment, no significant
191
difference in ADA development was observed between past thiopurine failures compared to
192
responders (HR 0.7 (95% CI 0.2-2.3, p=0.56)) (Table 2). The hazards for ADA development
193
were significantly lower in past responders and past failures compared with the monotherapy
194
group (HR 0.47 (95% CI 0.22-1.00) and HR 0.32 (95% CI 0.11-0.93), respectively), while a
195
TE D
M AN U
(p=0.102) (Figure 2b). In a multivariate analysis, adjusting for Montreal B classification, age
196
1.12)). Age at diagnosis significantly affected ADA formation. The type of thiopurine in the
197
combination groups did not significantly impact ADA appearance: 24.6% developed ADA on
198
AZA and 17.9% developed ADA on 6-MP (p=NS).
199
AC C
EP
non-significant trend was noted in the de-novo combination group (HR 0.47 (95% CI 0.19-
We next looked at differences in the dynamics of ADA formation between the groups. The
200
cumulative risk of ADA development at 6 months was also significantly higher in the
201
monotherapy group (35.1%) vs. past responders (13.9%, p=0.014) and past failures (9.6%,
202
p=0.012). However, the frequency of ADA formation in the monotherapy and de-novo
203
combination group curves diverges only after 5 months of therapy. Cumulative risk of
204
ADA development at 5 months from IFX initiation was (24.7%) in the monotherapy group
205
and (21.9%) in the de-novo combination group (p=0.728). In the following 7 months up to 1
206
11
ACCEPTED MANUSCRIPT 207
in the monotherapy group (p=0.024). Of note, the vast majority of patients with positive
208
ADA had persistently positive ADA measurements (not transiently detected).
209
Forty patients on IFX monotherapy received IV steroids pre-treatment. Interestingly,
210
corticosteroid pre-medication did not alter ADA appearance in the IFX monotherapy group
211
(41.0% with pre-treatment vs. 44.2% without, p=0.760).
212
RI PT
year, no new ADA events were recorded in the de-novo combination group (0%) vs. (29.9%)
SC
Study center was found to significantly impact ADA positivity (Table 2): it was significantly
213 214
all model variables study center still had overall significant effect on ADA appearance (p=
215
M AN U
more common at RHCC and SMC asP<0.001 compared with LUA and TASMC. After adjustment to
0.028). When monotherapy was compared to all combination groups within each center’s
216
cohort, ADA positivity was more common during monotherapy in every center, though this
217
did not reach statistical significance due to the small sample size in each center.
218
TE D
219
Figure 2a - ADA cumulative incidence in infliximab monotherapy versus all three
220
combination therapy groups compiled
221
EP
222
Figure 2b – Cumulative incidence of ADA in the four study groups
AC C
223
Drug levels
224
Significantly lower IFX serum concentrations were observed in ADA positive compared with
225
ADA negative patients at the end of follow up:P=0.102 the median was 0.3 mcg/ml (IQR 0.00-1.66)
226
P=0.007
P=0.007 for ADA negative, p<0.001. This effect for ADA positive and 2.33 mcg/ml (IQR 0.76-5.67)
227
occurred in all study groups, regardless of the therapy or treating center. The proportion of
228
patients with IFX drug levels less than 3 mcg/ml at the end of follow up was higher in the
229
12
ACCEPTED MANUSCRIPT 230
respectively), but was not significant (p=0.146).
231
Clinical status at end of follow up
232
Due to small sample size all thiopurine combination groups were compiled and compared to
233
RI PT
monotherapy group compared with all combination groups compiled (70.0% and 60.2%,
234
group were more likely to have active disease (21.5% vs. 8.8%, respectively), and
235
subsequently less likely to be in remission (40.9% compared to 58.8% of the combination
236
SC
monotherapy group. As compared to the combination groups, patients in the monotherapy
237
patients and ADA negative patient (though not significant). No significant differences in
238
M AN U
group)(p=0.009) (Figure 3). This trend was consistent after stratifying to ADA positive
clinical response were noted between the different combination groups.
239
240
241
combination
242
AC C
EP
TE D
Figure 3 – Clinical status at end of follow up in IFX monotherapy and IFX-thiopurine
13
ACCEPTED MANUSCRIPT 243
The findings from our retrospective cohort study suggest that adding a thiopurine to IFX
244
therapy prevents ADA formation independently of the past clinical response to thiopurines.
245
This effect was consistent after adjustment to clinical and treatment related factors.
246
RI PT
DISCUSSION
247
impact, since more patients in the combination groups (compiled) were in clinical remission
248
at the end of the study compared to the monotherapy group. However, an additional
249
explanation for the better clinical status of the combination groups compared to monotherapy
250
may be direct pharmacological effects of thiopurines. This was possibly supported by the
251
M AN U
SC
The results support the contention that prevention of ADA formation probably has a clinical
252
significant due to small numbers of patients. Not surprisingly, patients with ADA had lower
253
serum drug levels than those without ADA. A non-statistically significant trend was observed
254
for a higher proportion of patients with IFX drug levels less than 3 mcg/ml in the
255
monotherapy group compared with the combination groups. We had an unusually high
256
proportion (60-70%) of patients with IFX concentrations <3 mcg/ml at the study endpoint.
257
Likely, this stemmed from a bias introduced by the study's end-point definition of LOR or
258
EP
TE D
observation that the difference remained after stratification to ADA formation, although not
259
The present results need to be interpreted with caution. Because the study was retrospective it
260
may have been underpowered to detect a possible difference between past responders and
261
past non-responders to thiopurine. Notwithstanding, the present observations suggest that
262
even if such difference does exist, it is probably of small effect size, especially when
263
projected against the significant difference in ADA formation rate between each of these
264
groups and the monotherapy group.
265
AC C
ADA development.
14
ACCEPTED MANUSCRIPT 266
monotherapy groups during the first 5 months of therapy. However, patients in the
267
monotherapy group continued to form ADA throughout the study follow up period, while no
268
new ADA developed in the de-novo combination group thereafter, converging with the other
269
combination groups at 1 year. It appears therefore that the lag time between initiation of
270
thiopurines and their effect on ADA prevention is similar to the 3-6 months lag between
271
thiopurine initiation and their effect on clinical response as previously observed9. To avoid
272
such lag phenomenon it might be recommended to continue thiopurines when the decision to
273
SC
RI PT
We observed that ADA formation was similar in the de-novo combination group and
274
difference in cumulative incidence of ADA between patients receiving immunomodulator
275
before IFX or in de-novo combination therapy, both groups having significantly lower ADA
276
levels compared with monotherapy10. However, most of these patients were treated
277
episodically. Thus, prevention of "early" ADA might not be clinically relevant.
278
Our study had some limitations. The first is its retrospective nature of clinical data collection
279
and methodological limitations associated with this approach. However, because the study
280
was based on real life cohorts from four different centers in two countries, the data presented
281
EP
TE D
M AN U
start IFX is made in order to prevent "early" ADA appearance. Vermeire at al. reported no
282
exclusion of 287/494 of patients reviewed may limit the relevance of our finding to other
283
AC C
herein have a lower propensity for specific treatment bias. A potential bias related to the
patient populations, though exclusion criteria were explicit and led to a homogenic and
284
specific study cohort, relevant to study's question. Second, patients’ adherence and actual 6-
285
TGN blood levels of thiopurines metabolites could not be assessed in this study and may
286
impact the results. However, as shown in Table 1, no difference in reported weight based
287
thiopurine dose (mg/kg) was found between past thiopurine exposure groups. The planned
288
sample size in the combined treatment groups was not reached and thus a type II mistake
289
should be considered as one of the explanations for detecting no difference. Lastly, ADA
290
15
ACCEPTED MANUSCRIPT 291
more common in the two centers using the anti-lambda assay. In an attempt to address this
292
difference, adjustment to medical center as a co-variate in the multivariate analysis was done
293
(though no significant difference in distribution between study groups was noted between
294
study centers) but was not found to alter the results. To avoid interference of serum ADA by
295
serum IFX we looked at a harder endpoint of ADA positivity and serum IFX<3 mcg/ml. The
296
difference in ADA formation between study groups was not changed in univariate analysis,
297
and the same trend was noted in multivariate analysis but was not significant (data not
298
SC
RI PT
assays and ADA positivity rates differed between study's centers, with ADA positivity being
299
reported in past studies10–14. The validity of ADA positivity is also supported by the
300
consistency of our results with other published studies regarding the correlation between
301
ADA and IFX drug levels and clinical status.
302
In conclusion, in this 'real life' study we observed that adding a thiopurine to IFX therapy is
303
associated with lower rate of ADA appearance, probably regardless of patient's past clinical
304
thiopurine response. Further studies are needed to increase our understanding of factors
305
governing immunogenicity of infliximab in some but not all patients and to elucidate the
306
optimal methods to prevent it.
307
AC C
EP
TE D
M AN U
shown). Of note, the overall rate of ADA formation in our study was in the range of that
16
ACCEPTED MANUSCRIPT References: 1.
Gisbert JP, Panés J. Loss of response and requirement of infliximab dose intensification in Crohn’s disease: a review. Am J Gastroenterol. 2009;104(3):760-767. Lichtenstein GR, Feagan BG, Cohen RD, et al. Drug therapies and the risk of
RI PT
2.
malignancy in Crohn’s disease: results from the TREATTM Registry. Am J Gastroenterol. 2014;109(2):212-223.
Beaugerie L, Brousse N, Bouvier AM, et al. Lymphoproliferative disorders in patients
SC
3.
receiving thiopurines for inflammatory bowel disease: a prospective observational
4.
M AN U
cohort study. Lancet (London, England). 2009;374(9701):1617-1625. Osterman MT, Haynes K, Delzell E, et al. Effectiveness and Safety of Immunomodulators with Anti-TNF Therapy in Crohn’s Disease. Clin Gastroenterol Hepatol. 2015;13(7):1293-1301.e5.
Vande Casteele N, Buurman DJ, Sturkenboom MGG, et al. Detection of infliximab
TE D
5.
levels and anti-infliximab antibodies: a comparison of three different assays. Aliment
6.
EP
Pharmacol Ther. 2012;36(8):765-771.
Bein-Horn S, Yavzori M, Katz L, et al. The immunogenic part of infliximab is the
AC C
F(ab’)2, but measuring antibodies to the intact infliximab molecule is more clinically useful. Gut. 2011;60(1):41-48.
7.
Steenholdt C, Bendtzen K, Brynskov J, et al. Cut-off levels and diagnostic accuracy of infliximab trough levels and anti-infliximab antibodies in Crohn’s disease. Scand J Gastroenterol. 2011;46(3):310-318.
8.
Ungar B, Chowers Y, Yavzori M, et al. The temporal evolution of antidrug antibodies in patients with inflammatory bowel disease treated with infliximab. Gut. 2014;63(8):1258-1264. 17
ACCEPTED MANUSCRIPT 9.
Su C, Lichtenstein GR. Treatment of inflammatory bowel disease with azathioprine and 6-mercaptopurine. Gastroenterol Clin North Am. 2004;33(2):209-234, viii.
10.
Vermeire S, Noman M, Van Assche G, et al. Effectiveness of concomitant
Crohn’s disease. Gut. 2007;56(9):1226-1231. 11.
RI PT
immunosuppressive therapy in suppressing the formation of antibodies to infliximab in
Baert F, Noman M, Vermeire S, et al. Influence of immunogenicity on the long-term
12.
SC
efficacy of infliximab in Crohn’s disease. N Engl J Med. 2003;348(7):601-608. Colombel JF, Sandborn WJ, Reinisch W, et al. Infliximab, Azathioprine, or
13.
M AN U
Combination Therapy for Crohn’s Disease. N Engl J Med. 2010;362(15):1383-1395. Fefferman DS, Farrell RJ. Immunogenicity of biological agents in inflammatory bowel disease. Inflamm Bowel Dis. 2005;11(5):497-503.
Hanauer SB, Feagan BG, Lichtenstein GR, et al. Maintenance Infliximab for Crohn’s
EP
TE D
Disease: The ACCENT I Randomised Trial. Vol 359.; 2002:1541-1549.
AC C
14.
18
ACCEPTED MANUSCRIPT Table 1 – clinical and demographic parameters
Monotherapy (n=93)
Primary failure (n=34)
Past responders (n=52)
De-novo combination (n=28)
P value
52.7
41.2
44.2
57.1
0.467
Gender (% Female) Country of residence % Israel Belgium Smoking % Never
72.0 28.0
70.6 29.4
RI PT
0.358 67.3 32.7
85.7 14.3
0.533
76.5
63.3
69.2
Past
13.8
5.9
6.1
11.5
Current
23
17.6
30.6
19.2
22.9 (4.6)
21.2 (5.5)
21.5 (3.7)
22.2 (3.5)
0.121
22.7 (17.9-33.7)
19.3 (13.323.2)
19.6 (15.322.5)
23.5 (18.928.1)
0.001
9.4 (4.5-17.0)
1.8 (0.6-9.5)
<0.001
Age at diagnosis (y), Median (IQR) Disease duration (y) till IFX initiation, Median (IQR) Disease location (Montreal) % L1
7.6 (1.6-16.4)
L2
AC C
B2
B3 Perianal disease % Extra-intestinal manifestations % Past surgery % Past thiopurine dose (mg/kg) Median (IQR)
26.9
35.7
22.6
17.6
13.5
17.9
49.5
47.1
59.6
46.4
10.8
5.9
15.4
3.6
0.358 0.053
39.8
55.9
28.8
57.1
25.8
20.6
34.6
7.1
34.4 50.5
23.5 58.8
36.5 40.4
35.7 39.3
0.273
41.9
44.1
41.2
33.3
0.843
29.0
32.4
38.5
25.0
0.574
43.5
n=27 2.0 (1.7-2.3) n=7 1.2 (1.1-1.5) 55.9
n=32 2.1 (1.8-2.4) n=20 1.3 (1.1-1.4) 25.0
Azathioprine (AZA) 6-MP IV steroids % IBD Center %
0.749
35.3
EP
L3 Upper GI involvement % Disease behavior (Montreal) % B1
4.0 (1.2-8.6)
TE D
28.0
M AN U
BMI, Mean (±SD)
SC
63.2
0.474 0.580 39.3
0.031 0.134 19
ACCEPTED MANUSCRIPT 23.6
11.8
32.7
39.3
32.2
41.2
15.4
32.1
16.2
17.6
19.2
14.2
28.0
29.4
32.7
14.3
AC C
EP
TE D
M AN U
SC
RI PT
Rambam Health Care Campus Sheba Medical Center Tel-Aviv Medical Center Leuven University Hospital
20
ACCEPTED MANUSCRIPT Table 2 – Multivariable analysis (N=204)*
N
P value
HR (CI 95%)
0.034 51
1.00
Monotherapy
91
2.14 (1.00-4.60)
Past thiopurine failure
34
0.70 (0.20-2.30)
De-novo combination
28
1.00 (0.34-2.90)
ref.
0.051 0.560 1.000
SC
Past thiopurine responders
RI PT
Treatment group
Institute
0.028
Sheba Medical Center Tel-Aviv Medical Center Leuven University Hospital
1.00
ref.
59
1.31 (0.63-2.71)
0.460
35
0.35 (0.11-1.09)
0.070
57
0.66 (0.31-1.39)
0.270
TE D
Montreal B
53
M AN U
Rambam Health Care Campus
B1
0.169
86
1.00
ref.
50
0.96 (0.44-2.09)
0.910
68
1.66 (0.89-3.12)
0.110
IV steroids pre-treatment (Yes)
82
1.07 (0.52-2.21)
0.850
Age at diagnosis
204
1.02 (1.00-1.05)
0.030
Time from diagnosis to IFX
204
1.02 (0.98-1.05)
0.350
B2
AC C
EP
B3
*3 patients were not included due to missing data
21
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT