Transfusion Strategies

A summary of the NEJM 2013-01-03 article 'Transfusion Strategies for Acute Upper Gastrointestinal Bleeding'. Presented at DRH Journal Club 2013-03-06

1. Transfusion Strategies Vinod Kurup, MD 2013-03-06 DRH Journal Club
2. "Damn - We gotta stop transfusing these suckas!" -- Jonathan Lovins, MD, FHM Transfusion Strategies
3. The Article Transfusion Strategies for Acute Upper Gastrointestinal Bleeding C. Villanueva, et. al. Barcelona Spain NEJM 368;1, January 3, 2013
4. Background NEJM 1999;340:409-17: "A restrictive strategy of red-cell transfusion is at least as effective as and possibly superior to a liberal transfusion strategy in critically ill patients, with the possible exception of patients with acute myocardial infarction and unstable angina."
5. Background Blood transfusion is independent predictor of worse outcomes in trauma patients J Trauma 2003: 54:898-905 J Trauma 2005: 58:437-44
6. Background Observational studies suggest transfusion might be overused in GI bleeding Br J Surg 1986;73:783-5 Dig Dis Sci 2010;55:3430-5 Gastroenterology 1986;90:1232-40 (in rats)
7. Purpose of this study Randomized controlled trial to test if transfusion worsens outcomes in Acute upper GI bleeding
8. Patients Adults Hematemesis and/or melena Confirmed by hospital staff 2372 patients admitted with GI bleed 1610 screened 921 pts randomized 889 pts in final analysis
9. Excluded Pt declined transfusion Massive exsanguinating bleed Within past 90 days: ACS, Symptomatic PVD Stroke or TIA Transfusion Recent trauma/surgery Lower GI bleeding Rockall score of 0 with Hgb > 12
10. Protocol All got EGD + therapy if needed in < 6 hours PPI / somatostatin / antibiotics if indicated Portal pressure measured pre/post (varices) H/H q8 for 2 days, then daily Repeat H/H anytime clinician wanted H/H after each unit of transfusion
11. Randomization Stratified according to presence of cirrhosis Restrictive strategy: TF if Hgb < 7 Liberal strategy: TF if Hgb < 9 No significant differences between groups: rockall score, source of bleeding, cirrhosis
12. Outcomes Primary: Death within 45 days Secondary: further bleeding in-hospital complications
13. Results Hgb lower in restrictive group First 24 hours daily until discharge BUT Hgb similar at 45 days
14. Results How many patients received NO transfusion? Liberal: 14% Restrictive: 51%
15. Results How many units were transfused? Liberal: 3.7 Restrictive 1.5
16. Results Length of stay Liberal: 11.5 days Restrictive: 9.6 days
17. Results Increase in hospitalist satisfaction since consent forms didn't have to be filled out: Liberal: 0 % Restrictive: 239 % (note: just kidding)
18. Results - Secondary Outcomes Further bleeding: Liberal: 16% NNH: 16.7 Restrictive: 10% Adverse events: Liberal: 48% NNH: 12.5 Restrictive: 40%
19. Results: Primary Outcome Death at 45 days Liberal: 9% Restrictive: 5%
20. NNH = 25
21. Results: Primary Outcome Death at 45 days Liberal: 9% Restrictive: 5% Difference was most significant for Child-Pugh Class A or B cirrhosis
22. Problems 2372 pts with GI bleed, 1610 screened. Why? Were the 2 groups similar? Why so many complications? 18-22% had AKI 27-30% had bacterial infections 11-12% had pulmonary complications 11-16% had cardiac complications
23. My bottom line Data doesn't get much better than this. We should follow restrictive transfusion strategy Next question: What about lower GI bleeds?
24. Other studies JAMA Intern Med 2013;173(2):132-139 A liberal blood transfusion strategy is associated with higher all-cause mortality rates in patients with acute MI (Metanalysis)
26. SHM Choosing Wisely http://choosingwisely.org