Source: Carson JL, Carless PA, Hebert PC. Transfusion thresholds and other strategies for guiding allogeneic red blood cell transfusion. Cochrane Database Syst Rev. 2012;4:CD002042.
Holst LB, Petersen MW, Haase N, Perner A, Wetterslev J. Restrictive versus liberal transfusion strategy for red blood cell transfusion: systematic review of randomised trials with meta-analysis and trial sequential analysis. BMJ. 2015;350:h1354.
Rohde JM, Dimcheff DE, Blumberg N, et al. Health care-associated infection after red blood cell transfusion: a systematic review and meta-analysis. JAMA. 2014;311(13):1317-26.
Villanueva C, Colomo A, Bosch A, et al. Transfusion strategies for acute upper gastrointestinal bleeding. N Engl J Med. 2013;368(1):11-21.
De almeida JP, Vincent JL, Galas FR, et al. Transfusion requirements in surgical oncology patients: a prospective, randomized controlled trial. Anesthesiology. 2015;122(1):29-38.
Efficacy Endpoints: Mortality, functional status, ischemic events, hospital length of stay
Harm Endpoints: Mortality, serious infections, rebleeding
Narrative: Transfusion of red blood cells (rbc) occurred roughly 14 million times in the U.S. in 2011. [Whitaker BI, Hinkins S. The 2011 national blood collection and utilization survey.http://www.bloodsurvey.org. Accessed 4/10/2015]. While rbc transfusion is traditionally utilized in emergent settings of shock due to acute blood loss, its use outside of this context has been increasingly challenged based largely on an unclear ratio of benefits to harms. In recent years new research efforts have been undertaken, typically comparing Hb thresholds labeled ‘liberal’ (Hb=9 mg/dL or above) to those labeled ‘conservative’ (usually Hb=7-7.5 or below).
The Cochrane Collaboration review summarized here attempts to determine outcome differences between liberal and conservative thresholds in 19 clinical trials of more than 6000 subjects (nearly quadruple the number in the 2010 Cochrane review of the same topic) with acute or worsening anemia in association with a broad spectrum of medical and surgical conditions.(1)[Cochrane, 2012] A second, updated, larger and more inclusive review examined for this summary reports on nearly 10000 subjects from 31 studies.(2)[BMJ, 2015]
Ultimately, based on (necessarily) non-blinded trials of medium overall quality the investigators find no identifiable or measurable benefits to a liberal transfusion threshold. Most notably, the Cochrane authors do identify a clinically important increase in deaths among the groups assigned to a liberal strategy, although the statistical significance of this finding varies with the mortality endpoint used. While ‘hospital mortality’ is higher in the liberal threshold groups, 30-day- and overall mortality are both statistically borderline, with each outcome using a different denominator of subjects due to variable reporting of outcome measures. In the larger, more inclusive review mortality is unchanged, but again remains borderline.
In addition, the authors note a borderline statistically significant increase in overall infections among the six trials reporting this outcome. This finding agrees with a more recent and more inclusive systematic review specifically examining the topic of infections as a harmful outcome of transfusion.(3)[Rohde et al, JAMA 2014]
Caveats: The increasing numbers of individuals involved in trials of transfusion strategies is gradually helping to clarify benefits and harms, though the target continues to move. In the 2012 Cochrane review pulmonary edema, infections, and death appeared to be significantly more common in the liberal strategy group, while in updated reviews infections and death are variably more common while pulmonary edema is not. Of particular note a third meta-analysis, one examining exclusively trials using a lower threshold of <7g/dL as a restrictive trigger for transfusion, finds statistically significant benefits to this approach on deaths, pulmonary edema, infections, and rebleeding, with no indications of harms on cardiac function or outcome.(4)[Salpeter et al, AM J Med 2014]
It is also worth noting that a large 2013 study of patients with active acute gastrointestinal bleeding found a statistically significant increase in mortality with a liberal strategy, an important finding both because it is the first high quality, large single trial to show a clear mortality harm with a liberal strategy, and because it was performed in actively bleeding patients.(5)[Villanueva et al, NEJM 2013] This finding suggests that a group many would characterize as being the most likely to benefit from transfusion—those with active ongoing blood loss—fared considerably worse when transfused more liberally. On the flip side, a single, medium-size, relatively well done study of cancer patients undergoing abdominal surgery found a substantial mortality benefit with a liberal transfusion strategy, raising the possibility of either an anomalous finding or a specific context in which liberal strategies may be importantly beneficial.(6)[de Almeida et al, Anesthesiology 2015]
The quality and reliability of all of these data are variable but in a number of cases are reasonably high, although the lack of blinding is a necessary evil in such trials. In a number of more recent studies blinded outcome assessors and data analysis were undertaken to minimize bias. Despite this, there may be considerable residual bias, although direction of bias most often runs in favor of more intensive bedside management and treatment, suggesting that the bias here is unlikely to have caused the harms suggested, and more likely underestimates them. We have chosen, therefore, to classify this intervention as Black, indicating that data suggest to us that there is proven harm to use of a liberal transfusion strategy, without benefits in the great majority of study settings.
Author: David H. Newman, MD
Published/Updated: April 16, 2015
The title bar is color-coded with our overall recommendation.
If you have suggestions, requests, or questions about a particular NNT review, please send us a message and we’ll try to address it as soon as possible.