Steven C. Fong

Optimizing the use of thiopurines in inflammatory bowel disease

Immunomodulator drugs, of which thiopurines can be considered the backbone, are widely used in the treatment of inflammatory bowel disease. They have been shown to be highly effective and safe; however, a significant proportion of patients are deemed to have a poor response or suffer adverse reactions. Knowing how to monitor and optimize thiopurine therapy in these scenarios is crucial to effective management. We discuss the metabolism of thiopurines, the use of enzyme/metabolite testing to guide treatment, as well as strategies to circumvent toxicity and side effects, such as allopurinol coprescription. The indications, use in pregnancy, safety profile and duration of thiopurine therapy are also discussed.

Getting the best out of thiopurine therapy: thiopurine S-methyltransferase and beyond.

Thiopurines are the cornerstone of treatment for a wide variety of medical disorders, ranging from pediatric leukemia to inflammatory bowel disease. Because of their complex metabolism and potential toxicities, the use of biomarkers to predict risk and response is paramount. Thiopurine S-methyltransferase and thiopurine metabolite levels have emerged as companion diagnostics with crucial roles in facilitating safe and effective treatment. This review serves to update the reader on how these tools are being developed and implemented in clinical practice. A useful paradigm in thiopurine therapeutic strategy is presented, along with fresh insights into the mechanisms underlying these approaches. We elaborate on potential future developments in the optimization of thiopurine therapy.

Sa1255 Higher Red Blood Cell Methotrexate Polyglutamates Correlate With Increased Disease Activity, and Are Useful in Assessing Adherence

Introduction Methotrexate (MTX) is commonly used in patients with inflammatory bowel disease (IBD). Within red blood cells (RBC), MTX is activated by sequential addition of glutamic acid residues to form polyglutamates (MTXPG 1–5 ). In rheumatoid arthritis, low [MTXPG] has been associated with active disease, whereas other studies have demonstrated an inverse relationship, including the only published data in IBD. The aim of this study was to determine if RBC [MTXPG] reflect clinical response in IBD patients and whether they are useful in assessing adherence. Methods This was a single-centre, retrospective pilot study of 21 IBD patients treated with weekly MTX. RBC MTXPG 1–5 was measured using high-performance liquid chromatography. Clinical status (active disease or remission) was assessed by 2 IBD physicians blinded to [MTXPG], using a combination of prospectively recorded clinical activity indices (Simple Colitis Activity Index, Harvey Bradshaw Index), endoscopy, faecal calprotectin and C reactive protein (CRP). Pearson correlation coefficient, r was calculated to assess the relationship between MTX dose and [MTXPG]. Association between [MTXPG] and clinical response was analysed with unpaired t-test. Results 4/21(22%) patients (3 of whom admitted non-adherence) had undetectable MTXPGs and were excluded from further analysis. MTXPG 2–4 were detected in all adherent patients. PG 3 was the predominant polyglutamate accounting for a mean of 43% of total MTXPG. A linear relationship between dose of MTX and PG 1–5 was observed. 12/21(57%) patients were assessed as having active disease. No significant difference in mean [MTXPG n ] was observed between those with active disease and remission. For each MTXPG n , a non-significant trend towards a higher concentration was observed in patients with active disease. Conclusion In this study, the largest to date in IBD, measuring RBC MTXPG was useful in assessing adherence to MTX. A trend towards higher PG concentrations was associated with active disease confirming the findings in the only other study in IBD. Whether this is confounded by higher doses being used in patients with more active disease warrants further study in larger, prospective trials. Reference Disclosure of Interest None Declared.

Su1101 Indeterminate and Inconclusive Results Are Common When Using Interferon Gamma Release Assay As Screening for TB in Patients With IBD

Introduction Anti-TNF treatment is widely used in inflammatory bowel disease (IBD) but has been linked with reactivation of tuberculosis (TB). Screening for active and latent TB prior to initiation of anti-TNF therapy is therefore mandated. ECCO recommends interferon gamma release assays (IGRAs) as, unlike tuberculin skin test, positive tests are not caused by previous Bacillus Calmette–Guerin (BCG) vaccine. However, immunosuppressive agents can result in indeterminate or unreportable results[i] and there is no clear guidance on managing them. We quantified the prevalence of indeterminate or unreportable TB IGRA Elispot results in a large tertiary centre cohort of patients with IBD. Methods A single centre retrospective study of IGRA tests performed on IBD patients prior to commencement of anti-TNF therapy between Oct 2010 and Oct 2013. Results We included 140 patients (median age 34, range 24–86, 50% males). 92% had Crohn’s disease, 4% ulcerative colitis, and 4% IBD-unclassified. At the time of IGRA testing, 115 patients were on immunomodulators and 6 on prednisolone. 3 were positive for latent TB and were referred to infectious disease (ID) department prior to anti-TNF therapy. 3 had indeterminate results; all were on immunosuppressants (2=azathioprine, 1=methotrexate). 2 had a lymphocyte count 10 had unreportable results, 9 of whom were on azathioprine. On repeat testing, 4 were negative, and the remainder were still unreportable, one of whom had risk factors for TB and was treated with isoniazid chemoprophylaxis on the advice of the ID team. The remaining 5 patients started anti-TNF based on the absence of risk factors for TB. No patient had reactivation of latent TB at follow up (range 1–18 months). Lymphopaenia was found to be associated with non-reportable cases as compared to the reported cases (median lymphocyte count unreportable = 0.4, reportable = 1.2; p = 0.015). Conclusion Our results demonstrate TB IGRA is a useful test to screen for latent infection before initiating anti-TNF therapy. However, a minority of results are indeterminate or unreportable. In such cases repeat testing can produce definitive results. Low lymphocyte counts in association with immunosuppression may contribute to unreportable and indeterminate results; clinical risk stratification appears to be a safe way of managing such cases in this small cohort. Reference Papay P et al . Predictors of indeterminate IFN-γ release assay in screening for latent TB in inflammatory bowel diseases. Eur J Clin Invest. 2011 Disclosure of Interest None Declared.

PWE-013 Should The Target TGN Range Be Different in Those with Intermediate Compared with Normal TPMT Activity?

Introduction Thioguanine nucleotides (TGNs) are the biologically active metabolites of thiopurines. Methylated thiopurine metabolites (MeMPs) formed via thiopurine-S-methyltransferase (TPMT) may also have immunosuppressive properties through inhibition of de novo purine synthesis (DNPS).1 Individuals with intermediate TPMT activity do not produce significant levels of methylated metabolites, with predictably less inhibition of DNPS and therefore may tolerate higher levels of TGNs. If confirmed this suggests that the therapeutic range of TGNs (240–450 pmol/8x108) may differ between patients with intermediate Vs normal TPMT activity. Methods All IBD patients at Guy’s and St Thomas’ Hospitals treated with either azathioprine or mercaptopurine who had TGNs measured from 2002–2013 were included. Patients receiving low dose thiopurine with allopurinol were excluded. Data was collected using the electronic patient record system and statistical analysis was completed in GraphPad Prism version 5.0. Results 2193 TGN measurements were included in the analysis, 340 (15.5%) from patients with intermediate TPMT activity and the remaining 1853 had normal TPMT activity. The average normalised thiopurine doses between groups was 1.13 Vs 1.87 mg/kg/day respectively, reflecting standard dose adjustment in intermediate TPMT patients. The median TGNs levels were significantly higher in patients with intermediate Vs normal TPMT activity; 370 Vs 277 pmol/8x108 (P = <0.00001) and the median MeMP levels were lower in patients with intermediate Vs normal TPMT activity; 104 pmol/8x108 RBC Vs 566 (P = <0.0001). No significant difference was found in the median values of Hb, WBCs, neutrophil or platelet counts between the groups. A lower lymphocyte count (1.3 Vs 1.2 x109/9/ L, P = 0.0010) and trend towards a higher MCV (93 Vs 94, P = 0.0625) was noted in patients with normal TPMT activity. Conclusion Despite significantly higher TGN levels in patients with intermediate as compared to normal TPMT activity, there was no difference in levels of Hb, WBCs, neutrophil or platelets counts. This suggests a greater tolerance to TGNs in patients with intermediate TPMT activity and may indicate that a different TGN therapeutic range is needed in these patients. The likely reason is that methylated metabolites are biologically active and also play a role in overall thiopurine immunosuppression. Reference 1 Tay BS, Lilley RM, Murray AW, et al. Inhibition of phosphoribosyl pyrophosphate amidotransferase from Ehrlich ascites-tumour cells by thiopurine nucleotides. Biochemical Pharmacology. 1969;18(4):936–8. Disclosure of Interest None Declared

PWE-006 The Effects of Smoking on Thiopurine Metabolism: Abstract PWE-006 Table 1

Introduction There is an unpredictable relationship between thiopurine methyl transferase (TPMT) and hypermethylation. It is likely that cytochrome P450 (CYP450) enzymes play a role in the demethylation of methylated mercaptopurine (MMP), altering the MMP:Thioguanine nucleotide (TGN) ratio(Table 1). Smoking has been shown to induce CYP450 enzymes. One study on smoking and olanzapine demonstrated a 6 fold increase in CYP450 activity through smoking.1 This might suggest that smoking could upregulate the demethylation process, reducing MMP. Methods Patients on thiopurines had their smoking status recorded as they attended the inflammatory bowel disease (IBD) clinic. Retrospective data on demographics was collected (Table 1). Only measurable TGNs after 6 weeks of therapy were used for analysis. Metabolites measured whilst on allopurinol were excluded. Average TGNs over the treatment period, MMP and MMP:TGN ratios were compared between smokers and non-smokers. Hypermethylators were defined as any patient with an MMP:TGN ratio of ≥11 during their treatment Results 230 patients were analysed(117 males). Table 1 shows comparison in demographics between smokers and non-smokersAbstract PWE-006 Table 1 Smokers n = 106 Non-smokers n = 124 Male:female 51%:49% 47%:53% Age 37 33 Ethnicity – White:Black:Other 81%:19%:0% 80%:9%:11% Crohn’s:Ulcerative colitis 82%:18% 72%:27% Previous Surgery 45% 27% Mean TPMT 38.4 (median 35) 34.1 (median 35) Hepatotoxicity 15.1% 13.7% Hypermethylation 28.3% 42.7% Mean MMP for smokers was 1683(SD 2565, median 840) versus 2041 in non-smokers(SD 2401, median 1340)(P = 0.713). Mean MeMP:TGN ratio for smokers was 6.15 (SD 9.06, median 2.22) versus 7.81 in non-smokers (SD 8.01, median 4.70)(P = 0.593). Conclusion There were more hypermethylators in the non-smoking cohort than in the smoking cohort (42.7% versus 28.3%(P < 0.05)), however the difference between average MMP and MMP:TGN ratios were not significant although the median differences are compelling. There were significant differences between quantity smoked and the MMP:TGN ratio. Paradoxically, the reduction in methylation if smoking <5 cigarettes/day was reversed when smoking ≥15 cigarettes/day. This may suggest an alternative pathway affected by smoking. Smoking status and quantity should be taken into account when monitoring metabolites. Additionally, polymorphisms for CYP450 (CYP1A2*F and CYP1A2*1 C), known to affect drug metabolism between individuals, may be responsible for variation and require further research. Reference 1 Carrillo JA, Herráiz AG, Ramos SI, Gervasini G, Vizcaíno S, Benítez J. Role of the smoking-induced cytochrome P450 (CYP)1A2 and polymorphic CYP2D6 in steady-state concentration of olanzapine. J Clin Psychopharmacol. 2003 Apr;23(2):119–27. Disclosure of Interest None Declared

Sa1217 A Protocol to Avoid Corticosteroid Exposure Prior to Crohn's Disease Requiring Luminal Surgery

Introduction Corticosteroids (CS) are prescribed to control inflammatory and obstructive symptoms in Crohn’s disease (CD). Preoperative CS are associated with higher risk of all complications post-operatively including sepsis, anastomotic breakdown and venous thromboembolic disease.High doses of CS also preclude a primary anastomosis. In our institution, we have implemented a protocol to prevent CS administration or permit a rapid wean in patients requiring ileal/ileocaecal resection, maximising the use of exclusive enteral nutrition (EEN) or parenteral nutrition (PN). This study aimed to determine the achievability of CS avoidance in this setting. Method In patients selected for ileal/ileocaecal resection, EEN (with either Modulen IBD Nestle or Fortisip Nutricia, as tolerated) was prescribed to a target dose of 30 kcal/kg. Acute sepsis was controlled with intravenous and/or oral antibiotics. If EEN was not tolerated because of obstructive symptoms, PN was instituted or surgery expedited. Patients were not offered this protocol if they had minimal symptoms, adequate nutritional status and no CS exposure. Results In total, 39 patients with CD had IC resection over a period of 20 months from January 2013. 9 of these underwent urgent surgery due to symptom severity (severe obstruction or suspected cancer where delay not possible [4 of 9 on CS (45%)]. 8 patients with stable symptoms secondary to fibrotic stricture (4 on biologic, 4 on immunomodulator), adequate nutritional status and CS free were not suitable for the protocol. 22 patients with acute symptoms were therefore included. All had penetrating and/or stricturing disease. 8 patients were receiving CS of whom 7 (88%) successfully weaned off CS for greater than 4 weeks using EEN. 1 did not tolerate EEN and proceeded to surgery on CS. A further 11 of 22 patients successfully treated via protocol to avoid CS (EEN n = 7, PN n = 4) 3 patients did not tolerate/declined EEN/PN and proceeded to expedited surgery. Hence, in all patients suitable for the protocol, CS were successfully avoided (14/14 patients, 100%) and wean >4 weeks successful in majority (7/8 patients, 87.5%). CS exposure was successfully avoided in 21 patients via protocol (95.5%). Conclusion Aggressive use of EEN or PN provides an effective approach to avoid CS exposure, optimise nutrition and control acute symptoms in patients selected for IC resection. The implementation of the protocol requires close liaison between gastroenterologists, colorectal surgeons and dieticians. The protocol does not delay emergency surgery. Disclosure of interest None Declared.

PWE-088 The Impact Of Thiopurine Hypermethylation On Clinical Outcomes In Patients With Ibd

Introduction 15% of patients with inflammatory bowel disease (IBD) prescribed azathioprine (AZA)/mercaptopurine (MP) demonstrate a skewed drug metabolism with low thioguanine nucleotides (TGN) and unexpectedly high methylmercaptopurine (MeMP) levels. This predicts a lack of treatment efficacy and excess adverse events; however studies on the effect of thiopurine hypermethylation (TH) in adult IBD patients are lacking. Importantly, where identified TH can be circumvented with the use of low dose AZA/MP and allopurinol. The aim of this study is to characterise the impact of TH in IBD patients and determine if thiopurine metabolite profiles at week 4 of treatment can predict its occurrence to allow early combination treatment. Methods 273 patients with IBD, who were anti-TNF-α therapy naïve and matched for red blood cell thiopurine-S-methyltransferase activity, were retrospectively identified. Of these 181 patients demonstrated average MeMP:TGN ratios <11 (at least 2 profiles between weeks 12 – 52) and were used as a control group that was compared to 92 patients with average MeMP:TGN ratios ≥11. Clinical outcomes were recorded for the first 12 months of therapy. A failure of AZA/MP as monotherapy was determined by 3 gastroenterologists with expertise in IBD. Thiopurine metabolite profiles were measured in 139 patients at week 4 of therapy and were compared to average metabolite profiles between 12 – 52 weeks of treatment. Results The average MeMP:TGN ratios in those with and without TH were 21.0 and 2.52 respectively (p = <0.0001). The normalised dose of thiopurine was higher in patients with TH (1.91 vs. 2.09 mg/ kg; p = < 0.0001). Patients with TH were more likely to fail AZA/MP monotherapy as first line treatment during the first 12 months of therapy, in comparison with patients with normal methylation profiles (p = 0.0088, Gehan-Breslow-Wilcoxon Test). The difference in the number of treatment failures at 12 months was 15.7%. There was no difference in the occurrence of gastrointestinal intolerance, myelotoxicity or pancreatitis between groups, however there was an excess of hepatotoxicity with TH (p = 0.0006; OR = 8.058; 95% CI: 2.188 – 29.670; Fisher’s exact). TH was demonstrated in 83.7% of cases by 12 weeks of therapy. A MeMP : TGN level > 6.17 at week 4 accurately predicted TH (ratio ≥11) after 12 weeks of therapy (AUC = 0.839; p = < 0.0001; sensitivity = 75.4%; specificity = 88.4%). Conclusion TH is associated with an excess of thiopurine treatment failures as monotherapy immunosuppression during the first 12 months of treatment. We also confirm its association with hepatotoxicity. A MeMP:TGN >6.17 may be useful in early identification of patients likely to benefit from combination treatment. Disclosure of Interest None Declared.

Tu1925 Novel Pathway-Centric Analysis Reveals Variants Associated With Toxicity and Response to Thiopurines in Patients With Inflammatory Bowel Disease

Introduction Thiopurines remain the first line immunosuppressants recommended in the management of inflammatory bowel disease (IBD). Unfortunately, 30–40% of patients prescribed these agents develop adverse drug reactions or fail to derive therapeutic benefit. Candidate gene studies have identified loci that explain some of these treatment failures; however a substantial fraction of the genetic contribution remains undefined. Using whole thiopurine pathway analysis the aim of this study was to identify novel loci associated with toxicity and response to azathioprine (AZA)/mercaptopurine (MP) in patients with IBD. Methods Genomic DNA was extracted from EDTA blood samples of 472 well-characterised IBD patients treated with AZA/MP. We examined exome array data using the Illumina HumanExome Beadchip and restricted the analysis to variants associated with the thiopurine pathway as defined by the KEGG database (100 genes, 639 single nucleotide polymorphisms). Using a case-control design we firstly tested for genetic associations between patients with ( n = 1 54) and without ( n = 2 58) adverse drug reactions, and secondly for polymorphisms differentiating patients with ( n = 1 88) and without ( n = 1 41) response to thiopurines after 12 months of treatment. One year intervention-free clinical response was defined by 3 investigators (PB, PI, JS). Results Following adjustment for principal components, the minor alleles at ADK rs946185 (p = 0.0078; OR 1.675), SLC28A1 rs2242046 (p = 0.0168; OR 1.600) and ABCA1 rs4149268 (p = 0.033; OR 1.487) were associated with the development of drug toxicity, whereas the minor alleles at ABCB5 rs2301641 (p = 0.0170; OR 0.608), ABCC4 rs4148549 (p = 0.027; OR 0.652) and AOX1 rs55754655 (p = 0.038; OR 0.549) protected against it. The minor allele at RRM2 rs1130609 (p = 3.80 × 10 –5 ; OR 0.461), which codes a subunit of ribonucleotide reductase involved in the conversion of thioguanine nucleotide to deoxy-thioguanine nucleotide, and a higher normalised dose of AZA/MP were associated with protection from non-response. Conversely, the minor allele at ABCA1 rs2230808 (p = 0.008; OR 2.585) and Crohn’s disease (p = Conclusion High-throughput sequencing using exome array technology has revealed new loci, other than thiopurine- S- methyltransferase, explaining toxicity and response to thiopurines. Validation of these markers in separate cohorts will allow the development of biomarker panels to predict outcomes prior to the start of treatment. Disclosure of Interest None Declared.