of 50 patients (92% CD, 4% UC, 4% indeterminate colitis) were included in the study (Table 1). The median CTL was 18.85 ug/mL (0-77.6); it was 20.75 ug/mL in patients receiving CZP every 2 weeks vs 15.25 ug/mL in the every 4 weeks group (p=0.52). Fifteen patients (30%) had anti-CZP antibodies, with lower median CTL compared to patients with no antibodies (3.9 vs. 27.5; p=0.0001). Median CTL were similar in patients with and without IMM (23.15 vs. 17.9; p=0.44). MH was achieved in 33% of patients (Table 2). Median CTL was 15.05 ug/mL (4.7-59.2) and 12.8 ug/mL (0-77.6) in patients with and without MH, respectively (p=0.52). RR and RH were 54% and 29%, respectively. CTL >27.55 ug/ml were significantly associated with both RH (p=0.029) and RR (p=0.023). Median CTL levels were higher in patients with vs without RR (28 vs 16.25; p=0.014). Low CTL (<27.5 ug/ml) were significantly associated with abnormal CRP levels (>3 mg/dL) (p= 0.0009). Median CTL levels were 36.8 ug/mL and 10.5 ug/mL in those with normal and high CRP, respectively (p=0.005). SR was noted in 54% and was significantly associated with both CTL >27.55 ug/ml (p=0.001) and absence of antibodies (p=0.028). CTL resulted in change in clinical management in 60%: decrease in dose interval (n=11), addition of an IMM (n=4) and discontinuation of CZP (n=15). CTL <27.5 ug/ml were significantly associated with changes in clinical management (p=0.007). Conclusions: In this cohort, about a third of patients developed CZP antibodies. Lower CTL were significantly associated with abnormal CRP and changes in clinical management, whereas CTL >27.5 ug/ml were associated with RH, RR and SR. There was no association between median CTL and MH or concomitant IMM. However, our results are limited by a small sample size. Table 1. Baseline Characteristics Table 2. Outcomes S-827 AGA Abstracts Mo1856 CLINICAL FEASIBILITY OF DRIED BLOOD SAMPLING FOR INFLIXIMAB IN IBD-PATIENTS Sophie E. Berends, Geert R. D'Haens, Karien Bloem, J Schaap, Annick de Vries, Theo Rispens, Ron Mathot Background Therapeutic drug monitoring (TDM) is important to optimize outcome of infliximab (IFX) treatment in inflammatory bowel disease (IBD). Dried blood spot (DBS) sampling using capillary blood obtained via a finger prick could facilitate TDM, since patients can administer this finger prick themselves at any time and away from the hospital. We investigated the predictive performance and the feasibility of DBS for measuring IFX concen- trations in IBD patients. Methods We studied 40 adult IBD patients receiving IFX therapy according to standard guidelines. From each patient blood was obtained simultaneously via venepuncture and DBS via finger prick by a trained employee at 3 different timepoints (trough, peak, 3-5 weeks after infusion). One week before IFX infusion (timepoint 4), patients performed DBS at home and the sample was directly sent to Sanquin laboratories, Amsterdam, The Netherlands. The corresponding serum concentration for this time point was estimated using Bayesian pharmacokinetic analysis using the samples at trough and mid- infusion. Capillary blood was obtained by a microsampling device developed by Neoteryx™. Hematocrit (Hct) values of each individual patient were used to convert DBS eluate results to values which can be compared to (venous) serum concentrations. Spearman's correlation coefficient was used to assess correlation and bias was calculated. Results Forty IBD patients were included with median [interquartile range] age: 41 [32-50], albumin: 43 mg/L [41- 45], and CRP: 1.3 mg/L [0.4-4.4]. IFX concentrations obtained from the DBS method correlated strongly with serum results from the same patient for IFX trough- and mid- interval concentrations (Spearman correlation coefficient >0.88) and moderately for IFX peak concentrations (Spearman correlation coefficient = 0.69). IFX serum concentrations from the DBS performed at home showed strong correlation with the concentrations obtained by Bayesian analysis (Spearman correlation coefficient = 0.71). No structural bias was shown for DBS results compared to venous IFX serum concentrations(table 1). Conclusion DBS via finger prick can be used for the assessment of serum IFX concentrations. More importantly, we showed the feasibility of using DBS via finger prick at home. This method greatly facilitates the use of TDM in the treatment of IBD patients using IFX. Table 1: Concentration range and relative bias * Values are expressed as mean prediction error (95% confidence interval) Mo1857 VEDOLIZUMAB LEVELS DURING INDUCTION ARE ASSOCIATED WITH LONG-TERM CLINICAL AND ENDOSCOPIC REMISSION IN PATIENTS WITH INFLAMMATORY BOWEL DISEASE Andres J. Yarur, Alexandra Bruss, Brandon Berens, Caroline Fox, Poonam M. Beniwal- Patel, Amir Patel, Ryan C. Ungaro, Marla Dubinsky, Snehal Naik, Daniel J. Stein BACKGROUND: Cross-sectional serum vedolizumab (VDZ) levels (SVL) have been associ- ated with disease activity in Crohn's disease (CD) and ulcerative colitis (UC) but the role of therapeutic drug monitoring (TDM) early in the course of therapy of VDZ is unknown. The aim of this study was to assess the relationship of serum vedolizumab (VDZ) levels (SVL) during induction and disease remission in CD and UC after 52 weeks of therapy. We also sought to assess predictive variables associated with SVL through the first 22 weeks of therapy. METHODS: Prospective cohort study including patients with active UC and CD starting standard therapy with VDZ. Predictive variables included demographics, pre-infusion SVL measured using a validated drug-tolerant assay at weeks 2, 6, 14 and 22 (Anser® VDZ), C-reactive protein (CRP), albumin level, fecal calprotectin (FC), Harvey Bradshaw index (HBI) in CD and Mayo Clinical Score (MCS) in UC, Simple endoscopic score-CD (SES-CD) in CD and Mayo endoscopic score (MES) in UC. Primary outcome was deep remission at week 52, defined as HBI<5 (in CD)/MCS<3 (in UC), and SES-CD #2 (in CD) or MES #1 (in UC). Secondary outcome was discontinuation of VDZ due to non-response. RESULTS: Of the 53 patients in the study population, 28 (53%) had UC. Twenty-one (40%) achieved deep remission by week 52. These patients had higher SVL at weeks 2, 6, 14 and 22 of therapy versus those that did not, but only the differences at weeks 2 and 6 achieved statistical significance (Figure 1). Nineteen (36%) patients discontinued VDZ due to non- response (63% of those discontinued before week 30). Differences between those that were and were not in deep remission at week 52 are shown in Table 1. Patients that started VDZ on combination therapy with an immunomodulator did not have significantly higher SVL at any time-point when compared to those who did not (p>0.05 for all). Patients with UC had a significantly higher level SVL at week 2 when compared to CD (20 Vs. 25 mgr/mL, p=0.005), but no difference was found at other time points. There was a positive, significant correlation between baseline albumin levels and SVL at week 6 (rho: 0.4, p=0.003) and positive but not significant correlation and SVL at week 2 (rho: 0.2, p=0.09). Furthermore, there was a significant negative correlation between baseline FC and SVL at week 2 (rho: -0.5, p<0.001) and week 6 (rho: -0.33, p=0.03). CONCLUSIONS: Vedolizumab levels at induction are associated with achievement of remission at week 52 of therapy and drug discontinuation due to non-response. While interventional studies are needed, the results of this study suggest that the use of early TDM in patients on VDZ may improve outcomes. AGA Abstracts