You are using an old and unsupported browser. Most core functionality will not work. Please upgrade to a supported browser - Google Chrome

ANZMOSS - ANZGOSA 2020 Virtual Conference
ANZMOSS – ANZGOSA 2020 Virtual Conference
ANZMOSS - ANZGOSA 2020 Virtual Conference

Initial radiological appearance rather than the initial management strategy predicts outcomes of sleeve gastrectomy leaks

Multiple modalities have been advocated for the treatment of sleeve gastrectomy leak and there remains no consensus on the best treatment paradigm. Over 10years we have variably attempted luminal occlusive therapies and repeated endoscopic debridement as treatment options. By evaluating the outcomes from thesetwo approaches we aimed to determine if one management strategy provided superior outcomes.

Patients were analysed by group (luminal occlusive versus repeated endoscopic debridement). Leaks were then stratified by radiological appearance on CT, defined as phlegmon, collection, contrast leak, or fistula. The primary outcome was length of stay (LOS). Secondary outcomes were comprehensive complication index (CCI) and the need for resection.

There were 54 patients with 22 in the luminal occlusive group and 32 in the repeated debridement group. There was no difference in LOS 59.8±41.6 versus 46.5±51.2 days (p=0.179), and no difference in the requirement for resection 4 versus 3 (p=0.425). Subset analysis suggested that patients who underwent an operative versus conservative management (p=0.006) had a longer stay. Excluding management strategy, radiological appearance on admission significantly predicted LOS (p=0.0053). Patients presenting with fistula (84±25.4 days) and contrast leak (64.1±40 days) had significantly longer LOS than those diagnosed with phlegmon (13.5±5.5 days).Radiological appearance was predictive of complication severity (p<0.0001) and salvage resection (p=0.008).

There was no significant difference in outcomes between patients treated with intraluminal occlusion or repeated debridement. Initial radiological appearance was predictive of LOS and complication severity. This highlights the need for routine use of a validated classification system in studies reporting outcomes and treatment of sleeve leaks.
Paul Burton - Alfred Health, Wendy Brown - Alfred Health, Samantha Ellis - Alfred Health, Peter Nottle - Alfred Health, Edward Forrest - Alfred Health, Yazmin Johari - Monash University, and William Catchlove - Alfred Health