Background Colonoscopic surveillance in inflammatory bowel disease (IBD) patients leads to earlier detection of colorectal cancer (CRC) and reduces CRC-associated deaths. However, adherence to international surveillance guidelines in practice is poor. Consequently, we hypothesised that patients with IBD-associated CRC had suboptimal endoscopic surveillance and conducted a root cause analysis of IBD-associated CRC diagnoses to identify the scale of missed surveillance opportunities. Methods We calculated the incidence of CRC in our IBD population between 1998 and 2019 by cross-referencing the hospital’s CRC and IBD databases. All cases were adjudicated by IBD specialists to determine eligibility for surveillance using guidelines contemporaneous to the time of CRC diagnosis. For all eligible patients, a root cause analysis was conducted to determine whether there were missed opportunities to detect CRC. Results Our search identified 94 patients with IBD and CRC. We excluded 16 patients diagnosed with IBD after the diagnosis of CRC. The overall incidence of IBD-associated CRC in East Devon from 1998 to 2019 was 0.17% per year (95% CI 0.14 %–0.18%) with no statistically significant change seen over time (R2= 0.15, p = 0.11). Fifty-one patients (65%) were male, the median age at diagnosis of CRC was 69 years (range 21–88) and the median duration of IBD prior to CRC diagnosis was 21 years (range 0–57). After adjudication, 42 (54%) patients were eligible for surveillance at the time of CRC diagnosis. Correctly timed surveillance colonoscopy identified 5/42 (12%) CRC cases, but failed to identify CRC in 6 (14%) patients with so-called interval cancers. Overdue surveillance colonoscopy identified a further 4/42 (10%) cases. Overall, 27/42 (64%) IBD-associated CRC cases were classified as having had missed opportunities for surveillance. Root cause analyses found that 10 (37%) patients had not been offered surveillance despite on-going secondary care follow-up. Four (15%) patients had a delayed diagnosis of CRC because of a failure to account for the adequacy of previous colonoscopic and histological findings. Thirteen patients were managed exclusively in primary care including 7 (26%) patients who had been discharged back to primary care without a plan for surveillance and 6 (22%) patients who were never known to secondary care. Conclusion Two-thirds of patients who were eligible for surveillance had a missed opportunity to diagnose CRC. In most cases, the patient was known to the secondary care IBD service but no recommendation for surveillance was made. There is a need to implement processes to facilitate identification and recall of patients eligible for surveillance across primary and secondary care.

Background & Aims Because of COVID-19 public health restrictions, telemedicine has replaced conventional outpatient follow up for most patients with chronic immune-mediated inflammatory disorders treated with biologic drugs. Innovative solutions to facilitate remote therapeutic drug monitoring are therefore required. Low-volume intracapillary blood sampling can be undertaken by patients at home and samples returned by post to central laboratories. We sought to report the effect of the COVID-19 pandemic on requests for therapeutic drug monitoring and the equivalence, acceptability and effectiveness of low volume Patient-led Remote IntraCapillary pharmacoKinetic Sampling (fingerPRICKS) compared to conventional venepuncture. Methods We undertook a cross-sectional blood sampling methods comparison study and compared sample types using linear regression models. Drug and antidrug antibody levels were measured using standard ELISAs. Acceptability was assessed using a purpose-designed questionnaire. Results Therapeutic drug monitoring requests for adalimumab (96.5 [70.5 - 106] per week to 52 [33.5 - 57.0], p < 0.001) but not infliximab (184.5 [161.2 - 214.2] to 161 [135 – 197.5], p = 0.34) reduced during the first UK stay-at-home lockdown compared with the preceding six months. Fingerprick sampling was equivalent to conventional venepuncture for adalimumab, infliximab, vedolizumab, and ustekinumab drug, and anti-adalimumab and -infliximab antibody levels. The median (IQR) volume of serum obtained using intracapillary sampling was 195µL (130-210). More than 87% (90/103) patients agreed that intracapillary testing was easy and 69% (71/103) preferred it to conventional venepuncture. In routine care, 75.3% (58/77) patients returned two blood samples within 14 days to permit remote assessment of biologic therapeutic drug monitoring. Conclusions Therapeutic drug monitoring can be undertaken using patient-led remote intracapillary blood sampling and has the potential to be a key adjunct to telemedicine in patients with immune-mediated inflammatory diseases.

symptom response (adequate relief), gut microbiome (16S rRNA sequencing), short-chain fatty acids (SCFA, gas liquid chromatography), volatile organic compounds (VOC) (gas chromatography/mass spectrometry) and urine metabolomics (NMR spectroscopy) at baseline (BL) and 4 weeks. Data were analysed by Chi 2 for dichotomous outcomes and Kruskal-Wall-isMann-Whitney U tests for continuous variables. Predictive analysis was tested using receiver operator curves (ROC) and orthoganol partial least squared discriminant analysis (OPLS-DA). Results Response rate (adequate relief) at 4 weeks was greater in the LFD+GOS (67%) and LFD (50%) than the Control group (30%) (p=0.046). The phylum Actinobacteria was lower (p<0.001), specifically Bifidobacterium (p=0.002) and Collinsella (p=0.004), in the LFD and LFD+B GOS groups compared to Control at 4 weeks. Prediction of response to the LFD was seen in the LFD group only, with significant separation of BL faecal VOC profiles (AUC=0.854, p=0.045), faecal propionate (AUC=0.848, p=0.009), and urine metabolite profiles (Q 2 =0.296 vs randomised-0.175). Roseburia, Blautia, Lachnospira and Faecalibacterium, were positively but not significantly associated with VOC classes that predicted response when all samples were correlated. Microbiome and VOC showed a significant negative correlation between propionic acid derivatives and Phascolarctobacterium (r=-0.85 and-0.81, p<0.05) at BL in the LFD group only although the difference in Phascolarctobacterium between responders (1.5%) and nonresponders (9.3%) at BL did not reach significance (p=0.055). Conclusions The LFD +B GOS improves symptoms in IBS compared to Control but the LFD significantly reduces Actinobacteria and addition of 1.4 g/d of prebiotic B-GOS does not overcome this effect. Faecal VOCs and SCFAs and urine metabolomes may predict clinical response to the LFD, and specific bacterial groups correlate with predictive metabolites. Larger studies are required to validate these algorithms to develop personalised nutrition in IBS.