Therapeutic Monitoring of Adalimumab at Non-Trough Levels to Gauge Its Ecacy in Patients with Inammatory Bowel Disease in Clinical Practice

Background: Adalimumab (ADA) trough level and anti-ADA antibody (AAA) positivity could inuence mucosal healing and loss of response in patients with inammatory bowel disease (IBD). This study aimed to clarify the correlation between ADA monitoring, including non-trough level and real-world IBD clinical outcomes. Methods: This retrospective, observational, single-center study included 19 patients with ulcerative colitis (UC) and 33 patients with Crohn’s disease (CD) treated with ADA from January 2007 to August 2018. Serum ADA and AAA levels were measured 4 ‒ 14 days after ADA administration. Results: The AAA positivity rate was 23.1% (12/52). The ADA continuity was higher in the AAA-negative group than in the AAA-positive group (P = 0.223). Receiver operating characteristic (ROC) analysis revealed that a serum AAA cut-off value of 9.2 µg/mL was associated with ADA continuity (area under the curve [AUC]: 0.767, 95% condence interval [CI]: 0.636–0.899). The ADA level was signicantly higher in the endoscopic remission group than in the non-remission group (12.4 vs. 6.4 µg/mL, P = 0.02). Based on the ROC curve analysis results of serum ADA level and endoscopic remission, the cut-off value of serum ADA level was set to 11.1 µg/mL (AUC: 0.716, 95% CI: 0.533–0.900). Conclusions: Regardless of iniximab administration history, under combined use of ADA with immunomodulators and AAA positivity, ADA continuity was signicantly higher when the serum AAA level 4–14 days after ADA administration was ≥ 9.2 µg/mL. Furthermore, endoscopic remission can be expected with a serum ADA level of ≥ 11.1 µg/mL.


Background
Although the etiology of refractory in ammatory lesions in the intestinal tract of patients with in ammatory bowel diseases (IBDs) such as ulcerative colitis (UC) and Crohn's disease (CD) is unknown, in ammation caused by immune system dysregulation in the intestinal mucosa has been postulated to be the underlying cause of IBD pathogenesis [1]. The mainstay treatment for IBD is the suppression of immunocompetent cells by the systemic administration of steroids [2]. However, during recent years, biologics targeting cytokines produced by immunocompetent cells have been increasingly used in IBD treatment [3]. Among the biologics, anti-TNFα agents are widely used for IBD. In Japan, in iximab (IFX) and adalimumab (ADA) are approved for CD treatment and golimumab is approved for UC treatment.
Biologics such as IFX and ADA not only induce clinical remission in patients with IBD but also cause mucosal healing [4][5][6]. However, loss of response (LOR) is a major concern in IBD management, as it occurs within 1 year in an estimated 13% of patients treated with IFX [7] and 20% of patients treated with ADA [8]. Although the mechanism of LOR is not clear, immunogenicity against anti-TNFα inhibitors, manifested as a decrease in serum level (trough level) of anti-TNFα owing to the production of anti-drug antibodies or undetectability of anti-TNFα, is considered to be involved in LOR [5].
The importance of maintaining a high trough level of anti-TNFα agents has been shown in previous studies, which demonstrated that a low IFX trough level makes it di cult to achieve mucosal healing [4,5,9]. As reported, the risk of LOR in patients with antibodies against IFX (ATI) is three times that of patients without ATI [10]; it is important to suppress the production of anti-drug antibodies. Both trough level and anti-ADA antibody (AAA) have been associated with mucosal healing and LOR [11]. During recent years, therapeutic drug monitoring (TDM), a strategy to optimize treatment by measuring the serum trough level and anti-drug antibody level, has been recommended to maximize the effects of these biological agents [12]. Therapeutic drug monitoring is of two types: reactive TDM, which promotes decision making for patients who show treatment failure, and proactive TDM, which optimizes treatment and potentially avoids are and LOR in patients who have achieved some therapeutic effect [13]. However, the treat-to-target strategy, which has been advocated during recent years, emphasizes the importance of proactive TDM in the treatment of patients with IBD in clinical remission [14]. A review by Mitrev et al. recommends steady-state trough levels between 3 and 8 µg/mL for IFX and between 5 and 12 µg/mL for ADA [12]. The upper limit of the therapeutic window of ADA trough level to achieve clinical remission in patients with IBD has not been de ned.
Another problem in real-world clinical practice of TDM is the regular visit dates of patients do not always coincide with the trough measurement date, i.e. 14 days after ADA administration. It is not always possible to schedule another consultation visit only for the measurement of trough or to change the time of injection to measure the trough level during the consultation visit. Interestingly, Ward et al. reported that the trough level could be estimated from serum ADA level from days 3 to 9 after ADA injection, suggesting the feasibility of TDM even without trough level measurement of ADA [15]. In addition, authors reported that 8 weeks after the initiation of ADA therapy, the average uctuation in ADA level during the 2 weeks between successive administrations of ADA was approximately 3 µg/mL [16]. Therefore, the purpose of our study was to analyze the effects of serum ADA levels, including its nontrough levels, and AAA presence on subsequent treatment outcomes in patients with IBD receiving ADA. We also aimed to clarify the correlation between TDM for ADA and clinical outcomes in the real-world setting.

Patients
Fifty-two patients with IBD (UC or CD) were enrolled in this study and treated with ADA at Hamamatsu South Hospital, Hamamatsu, Japan, from January 2007 to August 2018. Serum drug (ADA) and AAA levels were measured during TDM. All patients provided informed consent before enrolling in this study.
Patients with Behcet's disease, indeterminate colitis, and unclassi ed IBD were excluded from the study. Adalimumab was prescribed at standard doses of 160 mg at week 0, 80 mg at week 2, and 40 mg subsequently every other week for maintenance; the drug was administered subcutaneously. The observation period was from the start of ADA administration to ADA discontinuation, and observations were continued until August 2018 in patients who were still on ADA in August 2018. We measured the anti-drug antibody and drug levels at patients' visit dates; we excluded patients whose antibody and drug levels were measured within 3 days of ADA administration.

Study design
This was a retrospective, observational single-center study. The primary endpoints of this study were the comparison of treatment continuity between (i) the AAA-positive and -negative groups and (ii) the highand low-serum ADA groups. The secondary endpoints were (i) comparison of endoscopic remission rate between patients with high-and low-serum ADA levels and (ii) extraction of factors that may in uence serum ADA level. Treatment continuity implied that neither the dose nor the dosing interval of ADA was changed during the observation period. Accordingly, patients who discontinued ADA therapy or received treatment enhancements, such as double-dose administration or shortened ADA administration period, did not t this description. Discontinuation of ADA because of transferring patients to another hospital, the preference of patients, or discontinuation because of adverse events were not considered "censored." Endoscopic remission was de ned as a Mayo endoscopic score of ≤ 1 in patients with UC and a simple endoscopic score for CD of ≤ 3 in patients with CD. Endoscopic evaluation was conducted only for patients who underwent endoscopy after introducing ADA.

Laboratory methods
We measured the serum level of ADA and AAA (free and bound) using enzyme-linked immunosorbent assays (ELISAs) (IDKmonitor® Adalimumab drug level ELISA and IDKmonitor® Adalimumab total ADA ELISA, Immundiagnostik AG, Germany).

Statistical analysis
The Kaplan-Meier estimator was used to generate survival curves for the AAA-positive and -negative groups. The log-rank test was used to compare persistence rate (treatment continuity) with ADA (the drug) between strata. A receiver operating characteristic (ROC) curve analysis was performed for serum drug levels to identify the cut-off value associated with treatment continuity. The two groups of patients were divided into a low-drug level group below the cut-off value and a high-drug level group above the cutoff value. We used the Kaplan-Meier estimator to generate survival curves for the high-drug level and the low-drug level groups and compared persistence rate with ADA between strata using the log-rank test. Next, the ROC curve analysis was performed to determine the cut-off value of serum drug level for endoscopic remission. We divided the patients into two groups, the low-drug level group below the cut-off value and the high-drug level group above the cut-off value, and compared the endoscopic remission rate between the two groups using the χ 2 test. A univariate analysis was performed to identify factors affecting the serum ADA level via Mann-Whitney's U-test, Pearson's correlation coe cient test, and Spearman's correlation coe cient using the log-rank test. Factors included in the model were sex, age, body weight, IBD type (UC or CD), disease duration, the time from diagnosis to rst ADA administration, the period from the start of ADA therapy to TDM implementation, IFX treatment history, and concomitant use of immunomodulators. These data were obtained when TDM was performed on the aforementioned factors. All statistical analyses were performed using EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan), which is a graphical user interface for R (The R Foundation for Statistical Computing, Vienna, Austria). Results with P < 0.05 were considered statistically signi cant.

Ethical considerations
The study protocol was reviewed and approved by the Committee on the Ethics of Clinical Trials Involving Human Subjects at Hamamatsu South Hospital. In addition, the trial adhered to the principles of Good Clinical Practice and the ethical standards stipulated in the 1964 Declaration of Helsinki and its subsequent amendments. Informed consent was obtained from all patients after the purpose of the study and nature of the procedures were explained to them.

Patient characteristics
Nineteen patients with UC and 33 patients with CD were enrolled in this study (Table 1)

AAA and treatment continuity
Here, AAA positivity was observed in 12 of the 52 patients (23.1%): 6 of the 19 (31.6%) patients with UC and 6 of the 33 (18.2%) patients with CD were positive for AAA. We found that treatment continuity was slightly lower in the AAA-positive group than in the AAA-negative group (P = 0.223 by log-rank test; Fig. 1). Next, we compared the endoscopic remission rate between the AAA-positive and AAA-negative groups, and no signi cant difference was observed between the two groups (P = 0.46, data not shown).

Adalimumab level and treatment continuity
Based on the results of the ROC curve analysis of serum ADA level performed to identify the cut-off value associated with treatment continuity, we set the cut-off value of serum ADA level to 9. Patients with serum ADA level of < 9.2 µg/mL were assigned to the low-drug level group and those with ≥ 9.2 µg/mL to the high-drug level group. Treatment continuity was signi cantly higher in the high-drug level group than in the low-drug level group (P < 0.01, log-rank test; Fig. 2b).

Adalimumab level and endoscopic remission
We analyzed the remission rate in patients who underwent endoscopy (44/52) after ADA administration. Endoscopic remission was observed in 15 of the 44 patients (34.1%), and the serum ADA level was signi cantly higher in the endoscopic remission group than in the non-remission group (12.4 vs. 6.4 µg/mL, P = 0.02; Fig. 3a). An ROC curve analysis was performed to determine the optimal serum ADA level for achieving endoscopic remission, and a serum ADA level of 11.1 µg/mL was identi ed as the cutoff value (AUC = 0.716, 95% CI 0.533-0.900, PPV = 75%, NPV = 81%, sensitivity = 0.600, speci city = 0.897, accuracy = 80%; Fig. 3b). We then compared the endoscopic remission rate between the low-drug level group with a serum ADA level of < 11.1 µg/mL and the high-drug level group with a serum ADA level of ≥

Relationship between patient-and disease-related factors and drug levels
We performed a univariate analysis to identify the factors that may in uence the serum ADA level ( Table 2). None of the factors were found to be associated with ADA level. We also compared treatment continuity with and without immunomodulators (combination therapy vs. monotherapy) but found no signi cant difference between the groups (P = 0.40; Fig. 4).

Discussion
In this study, we showed that the serum ADA levels, including non-trough values, but excluding values obtained within 3 days of ADA administration, re ect the subsequent clinical course and endoscopic outcome. Our study demonstrated that ADA continuity can be expected if the non-trough serum ADA level (except within 3 days of ADA administration) is ≥ 9.2 µg/mL and that endoscopic remission can be expected if the level is ≥ 11.1 µg/mL. A previous study on trough level of ADA during maintenance therapy reported that 5-5.8 and 4.9-7.1 µg/mL were the lower limits of the therapeutic range when clinical remission and mucosal healing were the desired targets, respectively; both limits are approximately 4 µg/mL lower than our non-trough level. As these are the lower limits of the therapeutic range and the variation in the levels over 2 weeks between successive ADA administrations is approximately 3 µg/mL [16], our results are valid.
In other studies, AAA positivity was estimated to be 12.7%-30.9% [9,11,15,17], which is consistent with our results in this study (23.1%). Nakase et al. reported that AAA and clinical remission were signi cantly correlated in patients with CD receiving ADA [18]. In contrast, Robin et al. showed that AAA positivity was higher in patients with clinically active disease than in patients with clinical remission (13.5 vs. 9.8 ng/mL, respectively) [17]. In addition, AAA positivity was lower in patients who achieved mucosal healing than in those who did not (6.5 vs. 14.1 ng/mL; P = 0.06). Our study also showed that AAA did not correlate with treatment continuity or endoscopic outcomes; hence, the importance of measuring AAA level in clinical practice should be investigated further.
Although there was no association between the concomitant use of immunomodulators with ADA versus only ADA in this study (P = 0.252), we found that AAA positivity was signi cantly lower in the immunomodulator combination group (1/27 [3.7%] vs. 11/25 [44%], P < 0.05, Mann-Whitney's U-test). However, there was no signi cant difference in treatment continuity with or without immunomodulators (P = 0.40). Using immunomodulators suppressed AAA production, but it did not improve any clinical outcomes in our study. However, a recent study reported that treatment continuity was higher with azathioprine when a second anti-TNFα agent was used in cases where the use of the rst anti-TNFα agent resulted in immune-mediated LOR [25]. Further research is needed to determine whether it is necessary to use an immunomodulator with ADA in real-world clinical settings. In our study, because we found no other factors related to ADA level, we concluded that ADA is clinically effective if the serum ADA level is high, regardless of sex, weight, disease duration and IFX treatment history. Therefore, to predict the therapeutic e cacy of ADA, actual measurement of ADA level is important, and not the timing of blood collection to speci cally measure the trough level (unless collected within 3 days of ADA administration).
A limitation of this study was that the time of measurement of ADA level differed for each patient. There are a few studies on the pharmacokinetics of ADA in patients with IBD, but information about the changes in drug level after administration is limited. A post-hoc study on the pharmacokinetics of 65 patients with CD receiving ADA reported that the half-life of ADA was 22 days in AAA-negative patients [26]. A prospective observational study in which 19 patients with CD treated with ADA underwent TDM at multiple timepoints indicated that there was negligible change in the serum ADA level over 9 days after administration [15]. Another prospective study on the pharmacokinetics of seven patients with CD undergoing remission maintenance therapy with ADA reported a minimal difference between the peak and trough blood ADA levels during ADA therapy [27]. From these studies, we inferred that the difference in the timing of TDM administration would have negligible effect on the serum ADA level in patients repeatedly receiving ADA. This is because the ADA level in the blood increases and disappears relatively slowly because of the subcutaneous formulation, and the administration interval is 2 weeks, which is shorter than other drugs.

Conclusions
In this study, we evaluated non-trough drug levels, which may be useful in predicting the outcomes of ADA therapy, unlike with other biologics. Although this was a retrospective study based on actual clinical data, our study had a long observation period: median of 102 weeks and maximum of 593 weeks. We could not nd any longer-term observational study that examined the relationship between the ADA level and its therapeutic effect.
In patients with IBD receiving ADA, a higher serum ADA level was associated with higher treatment continuity and endoscopic remission rates; however, AAA positivity did not affect ADA continuity. None of the factors we examined (such as the history of IFX administration and concomitant use of immunomodulators) were associated with serum ADA level. Therefore, to predict clinical outcomes, it would be useful to measure the blood level of ADA regardless of the timing of trough. Availability of data and materials Not applicable.