SummaryBackgroundOesophageal adenocarcinoma is the sixth most common cause of cancer death worldwide and Barrett's oesophagus is the biggest risk factor. We aimed to evaluate the efficacy of high-dose esomeprazole proton-pump inhibitor (PPI) and aspirin for improving outcomes in patients with Barrett's oesophagus.MethodsThe Aspirin and Esomeprazole Chemoprevention in Barrett's metaplasia Trial had a 2 × 2 factorial design and was done at 84 centres in the UK and one in Canada. Patients with Barrett's oesophagus of 1 cm or more were randomised 1:1:1:1 using a computer-generated schedule held in a central trials unit to receive high-dose (40 mg twice-daily) or low-dose (20 mg once-daily) PPI, with or without aspirin (300 mg per day in the UK, 325 mg per day in Canada) for at least 8 years, in an unblinded manner. Reporting pathologists were masked to treatment allocation. The primary composite endpoint was time to all-cause mortality, oesophageal adenocarcinoma, or high-grade dysplasia, which was analysed with accelerated failure time modelling adjusted for minimisation factors (age, Barrett's oesophagus length, intestinal metaplasia) in all patients in the intention-to-treat population. This trial is registered with EudraCT, number 2004-003836-77.FindingsBetween March 10, 2005, and March 1, 2009, 2557 patients were recruited. 705 patients were assigned to low-dose PPI and no aspirin, 704 to high-dose PPI and no aspirin, 571 to low-dose PPI and aspirin, and 577 to high-dose PPI and aspirin. Median follow-up and treatment duration was 8·9 years (IQR 8·2–9·8), and we collected 20 095 follow-up years and 99·9% of planned data. 313 primary events occurred. High-dose PPI (139 events in 1270 patients) was superior to low-dose PPI (174 events in 1265 patients; time ratio [TR] 1·27, 95% CI 1·01–1·58, p=0·038). Aspirin (127 events in 1138 patients) was not significantly better than no aspirin (154 events in 1142 patients; TR 1·24, 0·98–1·57, p=0·068). If patients using non-steroidal anti-inflammatory drugs were censored at the time of first use, aspirin was significantly better than no aspirin (TR 1·29, 1·01–1·66, p=0·043; n=2236). Combining high-dose PPI with aspirin had the strongest effect compared with low-dose PPI without aspirin (TR 1·59, 1·14–2·23, p=0·0068). The numbers needed to treat were 34 for PPI and 43 for aspirin. Only 28 (1%) participants reported study-treatment-related serious adverse events.InterpretationHigh-dose PPI and aspirin chemoprevention therapy, especially in combination, significantly and safely improved outcomes in patients with Barrett's oesophagus.FundingCancer Research UK, AstraZeneca, Wellcome Trust, and Health Technology Assessment.

115-4 (21-9), n=15) were not different to control results. Eight patients had carcinomas (rectal (2), sigmoid (4), and caecal (2)) all of which were adenocarcinomas. The cancers (193.6 (40.2), n=8) synthesised more PGE2 than control specimens (p<0001), but were not different to polyps. Cancerassociated mucosa (140.3 (27.7) n=8) synthesised more PGE2 than control and polypassociated mucosa. Colorectal neoplasia is associated with a progressive increase in PGE2 synthesis which may have a role in tumourigenesis and be a pathophysiological explanation for the beneficial effects of NSAIDs in animal models and human disease.

No abstract

Background Endoscopic full-thickness resection (eFTR) of the colon using the full-thickness resection device (FTRD) is a novel method for removing lesions involving, or tethered to, deeper layers of the colonic wall. The UK FTRD Registry collected data from multiple centres performing this procedure. We describe the technical feasibility, safety and early outcomes of this technique in the UK. Methods Data were collected and analysed on 68 patients who underwent eFTR at 11 UK centres from April 2015 to June 2019. Outcome measures were technical success, procedural time, specimen size, R0 resection, endoscopic clearance, and adverse events. Reported technical difficulties were collated. Results Indications for eFTR included non-lifting polyps (29 cases), T1 tumour resection (13), subepithelial tumour (9), and polyps at the appendix base or diverticulum (17). Target lesion resection was achieved in 60/68 (88.2%). Median specimen size was 21.7 mm (10–35 mm). Histologically confirmed R0 resection was achieved in 43/56 (76.8%) with full-thickness resection in 52/56 (92.9%). Technical difficulties occurred in 17/68 (25%) and complications in 3/68 (5.9%) patients. Conclusion eFTR is a useful technique with a high success rate in treating lesions not previously amenable to endoscopic therapy. Whilst technical difficulties may arise, complication rates are low and outcomes are acceptable, making eFTR a viable alternative to surgery for some specific lesions.

The effects of Roter (compound bismuth subnitrate) on antacid activity and mucosal prostaglandin E2 (PGE2) synthesis were variously investigated in patients with duodenal ulcer disease and healthy volunteers. Roter had a significant but small antacid activity with a buffering capacity of 10.9 mmol per tablet. In healthy volunteers, this was assessed by 24 h gastric pH monitoring on matched days with and without Roter treatment. The percentage of time that gastric pH was above 3 and the time after a standard meal that the pH was above 3, were both significantly increased by treatment with Roter (II tds post-cibal) (P less than 0.01). Endogenous PGE2 synthesis was measured in endoscopic duodenal biopsies taken both before and after Roter treatment in patients with acute untreated duodenal ulceration. There was a significant deficiency of mucosal PGE2 synthesis in untreated patients compared with controls (P less than 0.005). However, following 4 weeks' treatment with Roter, there was a 90% rate of healing accompanied by a significant increase in PGE2 synthesis (P less than 0.05) up to control levels. These findings suggest that Roter heals by the combined effects of a modest antacid neutralizing capacity and the ability to restore mucosal prostaglandins to normal levels, thereby mediating prostaglandin-dependent defence mechanisms.

Group (cl) ulcer rim made 49-8 (22.7) and at all stages ulcer rim and scar made less than the control duodenal bulb (p<002). Uninvolved duodenal bulb from groups (cl) (63-4 (31i0)), (c2) (83-6 (38 5)), and (c3) (81-5 (31-1)), however, also made significantly less than controls (p<002) and a similar though non-significant trend was seen in group (c4). Biopsies from the second part of the duodenum did not synthesise significantly less than the control group but a similar trend was noticed at each stage of ulcer treatment. Biopsies of control antrum synthesised 124-5 (32-2) but only 93 7 (44 2) in group (ci) (p<0005). All stages of duodenal ulcer healing were associated with a decreased capacity to synthesise the major prostaglandin PGE2 at the ulcer site and the uninvolved duodenal bulb and, in acute untreated duodenal ulcer, the uninvolved antrum. This decreased capacity may be the consequence of the disease process itself and not secondary to the treatment, indicating a basic pathophysiological abnormality which may explain the characteristic tendency of the disease to relapse.Animal work suggests that endogenous prosta-gastric and duodenal ulcer."'3 The majority of studies glandins may be a component of mucosal protection have looked at relatively few patients and methodoand rapid repair. I-3 Attempts to estimate mucosal logical problems may have failed to reveal a conprostaglandin synthesis in patients with peptic ulcera-sistent pattern. 14 None have looked at synthesis tion, however, has produced conflicting results in during the healing of peptic ulcers. In this study,

We investigated whether impaired duodenal mucosal prostaglandin E2 (PGE2) production previously observed in duodenal ulcer (DU) was a primary pathophysiological abnormality or secondary to mucosal architectural changes that accompany ulceration. One hundred patients were studied: at endoscopy, paired duodenal biopsies were taken in patients with normal endoscopies and from the ulcer edge or scar and background mucosa in active or healed DU. One of the pair of biopsies was used to estimate PGE2 synthesis ability, the other was processed for histology and histochemistry. The following features graded: goblet cell numbers and staining with Periodic acid-Schiff reagent (PAS), epithelial staining with PAS, villous atrophy, columnar cell height, inflammatory cell infiltrate and micro-erosions and gastric metaplasia taken as a whole. Patients were found to have normal endoscopy (n = 31), active untreated DU (n = 20) active DU on treatment with either cimetidine or ranitidine (n = 13), healed DU on maintenance treatment (n = 27) and healed DU off treatment (n = 9). Active duodenal ulceration was found to be associated with decreased numbers of goblet cells, loss and blunting of villi, increased columnar cell height, increased epithelial cell PAS staining and with gastric metaplasia. After healing, only villous blunting remained. These changes were present, but less marked, at sites removed from the ulcer and were not apparent in the patient groups with healed ulcers. A strong correlation between overall gastric metaplasia and epithelial cell PAS staining and the reduced ability to synthesize PGE2 (P < 0.001) was only apparent when biopsies from all patients were grouped together, but not within individual patient subgroups. There was no consistent correlation between PGE2 generation and individual parameters of pathological change in duodenum. We conclude that, although inflammatory and mucosal changes may contribute, the evidence suggests that the impaired PGE2 generation in DU disease is, to a large extent, independent of histological and histochemical features.

No abstract