Cohort members were followed from first endoscopy until HGD or EAC was diagnosed, the most recent endoscopy before withdrawal from further follow-up, end of study period, or death whichever came first. Incidence rates (IR) and incidence rate ratios (IRR) for developing HGD or EAC were calculated with 95% confidence intervals (CI). Trends were assessed by linear regression. An exploratory analysis, high throughput screening utilizing risk stratification, was performed
to identify patients that could be excluded from further surveillance. Cost-utility analysis used a Markov simulation model based on surveillance data and international guidelines to calculate US$ cost per quality-adjusted life year (QALY) ratios. Results: 826 patients were followed for a median 2.7 years (range 0–8.9) and 2,067 person years of follow-up. 17 cases of HGD or EAC were identified – IR of 0.8% annually. There was a statistically significant association between the length of Columnar lined oesophagus (CLE) and the development of HGD or EAC (r = 0.79, p = 0.02). Individuals with CLE of ≥2 cm an annual IR of 1.2% and a close to 8-fold increased relative risk of HGD or EAC, compared to individuals
with CLE < 2 cm at an IR of 0.1% (IRR 7.9, 95% CI 4.5–12.8). Only individuals with intestinal metaplasia progressed to HGD or EAC. Compared with no surveillance, surveillance of the entire cohort had an incremental cost per QALY of $60,858. Modeling showed that limiting the surveillance cohort, after the first endoscopy to individuals with a CLE segment of at least 2 cm, or 5-Fluoracil ic50 a CLE segment with dysplasia (any length) resulted in a reduction of 316 (38%) persons and 681 (33%) person-years under surveillance.
Application of a further restriction after the second endoscopy – exclusion of all patients without intestinal metaplasia at both the first or second endoscopy – removed a further 61 (12%) patients and 86 (4%) person-years from surveillance. Combining these strategies reduced the number under surveillance by 377 (46%) patients, and 767 Suplatast tosilate (37%) person-years, which translated to an estimated incremental cost per QALY of approximately $40,000 for surveillance of the remainder. Conclusions: Using a two-step limitation process, based on stratification of risk for HGD or EAC, the number of patients with Barrett’s oesophagus requiring surveillance can be reduced by at least a third. This needs to be validated in other populations, but suggests endoscopic surveillance for Barrett’s oesophagus can be tailored to achieve cost-effectiveness in Australia. NS DING, E FLANAGAN, T NGUYEN, DM ISER, T HONG, L LUIZ, M RYAN, SJ BELL, PV DESMOND, AJ THOMPSON St Vincent’s Hospital, Victoria, Australia Aims: Endoscopic screening for oesophageal varices is currently recommended for all cirrhotic patients, but <50% of patients with compensated cirrhosis show clinically significant portal hypertension (CSPH).