This large retrospective database analysis of patients with non-variceal UGI bleed found, after adjusting for confounders, that those who received transfusions were significantly more likely to rebleed than those that did not receive transfusions (odds ratio 1.8, CI 1.2 to 2.8). As with other conditions, RBC transfusions should only be administered when absolutely necessary, but the exact threshold will need to be further delineated with randomized trials (abstract).
This large systematic review found rectal NSAIDs significantly reduced the risk of post-ERCP pancreatitis compared to pancreatic duct stents (abstract).
This large population-based cohort found the most common causes of drug induced liver injury to be augmentin and diclofenac, followed by herbal and nutritional supplements (abstract).
This large trial of patients with a relative contraindication for enteral feeds were randomized to early TPN or usual care. There were no differences in the groups in 60 day mortality or LOS, but those on TPN did have a shorter time ventilated and less muscle/fat loss. It is unclear based on this trial if […]