Date of Award

January 2011

Document Type

Open Access Thesis

Degree Name

Medical Doctor (MD)



First Advisor

Shamsuddin Akhtar

Subject Area(s)



A Comparison of Blood Transfusion Practice Guidelines: What Quality of Evidence is Being Utilized to Develop Transfusion Guideline Recommendations?

Janice Y. Man, BA and Shamsuddin Akhtar, MD. Department of Anesthesiology, Yale University, School of Medicine, New Haven, CT.


Transfusion of blood components is widely utilized in the management of medical and surgical conditions. Though transfusion is a life-saving intervention, there has been debate about the standardization of blood transfusion practices. There has been a tremendous response in literature generated from multiple medical specialties regarding appropriate use of blood products to guide clinicians in their transfusion decisions. However, the consequence of numerous guidelines from multiple specialties results in varying recommendations for transfusion practices. This study was designed to compare and analyze current guidelines to determine if the recommendations generated to guide clinicians in transfusion decisions are truly supported by quality evidence.

We performed a literature search on clinical transfusion practice guidelines from January 2005 to October 2010 with the following computer databases: PubMed/Medline, Cochrane Central, Scopus and the National Guideline Clearinghouse. Additional websites and publications, such as the Australian and New Zealand Society of Blood Transfusion, were also searched for guidelines missed from the computer database search. Key words that were used for the search include the combination of the following keywords: blood, blood component, blood product, transfusion, guidelines. The resulting eleven guidelines were analyzed for the following areas: characteristics and composition of the guideline working group panel, literature and evidence utilized for the systematic review, databases utilized to retrieve evidence and literature for the systematic review, methodologies employed by guideline committees to grade strength and quality of evidence and recommendations, quantity of recommendations suggested, and specific transfusion thresholds and/or clinical settings for transfusion of blood products. We developed a three-tiered classification system in order to compare the level of evidence and strength of recommendations generated by each guideline even with the utilization of seven difference grading systems.

A total of 107 recommendations were generated about packed red blood cells, fresh frozen plasma, platelets, and cryoprecipitate transfusion. Of the 107 recommendations, 48 (48.86%) of the recommendations were specific to the use of packed red blood cells, 31 (28.97%) of the recommendations were specific to the use of fresh frozen plasma, 15 (12.02%) of the recommendations were specific for the use of platelets, and only 13 (12.15% recommendations were specific to the use of cryoprecipitate. Only 12 (11.21%) recommendations were generated from "strong" level evidence, 25 (23.36%) recommendations were generated from "intermediate" level evidence, and 70 (65.42%) recommendations were generated from "low" level evidence. Of the total recommendations, 36 (33.64%) recommendations were classified as a "strong" recommendation to perform the intervention, 46 (42.99%) recommendations were classified as an "intermediate" recommendation to perform the intervention, and 25 (23.36%) recommendations were classified as a "weak" recommendation to perform the intervention.

Analysis of these eleven international guidelines suggests that currently a large body of recommendations concerning blood component therapy is based solely on "low" quality evidence. Clearly there is a significant scarcity of strong evidence as well as clearly explicit recommendations to guide clinician practice of transfusion of blood products. Future research should thus be stimulated and directed at providing more abundant and high quality evidence regarding the use and safety of blood components in the perioperative setting.