Red blood cell transfusion for people undergoing hip fracture surgery.
Brunskill SJ., Millette SL., Shokoohi A., Pulford EC., Doree C., Murphy MF., Stanworth S.
BACKGROUND: The incidence of hip fracture is increasing and it is more common with increasing age. Surgery is used for almost all hip fractures. Blood loss occurs as a consequence of both the fracture and the surgery and thus red blood cell transfusion is frequently used. However, red blood cell transfusion is not without risks. Therefore, it is important to identify the evidence for the effective and safe use of red blood cell transfusion in people with hip fracture. OBJECTIVES: To assess the effects (benefits and harms) of red blood cell transfusion in people undergoing surgery for hip fracture. SEARCH METHODS: We searched the Cochrane Bone, Joint and Muscle Trauma Group Specialised Register (31 October 2014), the Cochrane Central Register of Controlled Trials (The Cochrane Library, 2014, Issue 10), MEDLINE (January 1946 to 20 November 2014), EMBASE (January 1974 to 20 November 2014), CINAHL (January 1982 to 20 November 2014), British Nursing Index Database (January 1992 to 20 November 2014), the Systematic Review Initiative's Transfusion Evidence Library, PubMed for e-publications, various other databases and ongoing trial registers. SELECTION CRITERIA: Randomised controlled trials comparing red blood cell transfusion versus no transfusion or an alternative to transfusion, different transfusion protocols or different transfusion thresholds in people undergoing surgery for hip fracture. DATA COLLECTION AND ANALYSIS: Three review authors independently assessed each study's risk of bias and extracted data using a study-specific form. We pooled data where there was homogeneity in the trial comparisons and the timing of outcome measurement. We used GRADE criteria to assess the quality (low, moderate or high) of the evidence for each outcome. MAIN RESULTS: We included six trials (2722 participants): all compared two thresholds for red blood cell transfusion: a 'liberal' strategy to maintain a haemoglobin concentration of usually 10 g/dL versus a more 'restrictive' strategy based on symptoms of anaemia or a lower haemoglobin concentration, usually 8 g/dL. The exact nature of the transfusion interventions, types of surgery and participants varied between trials. The mean age of participants ranged from 81 to 87 years and approximately 24% of participants were men. The largest trial enrolled 2016 participants, over 60% of whom had a history of cardiovascular disease. The percentage of participants receiving a red blood cell transfusion ranged from 74% to 100% in the liberal transfusion threshold group and from 11% to 45% in the restrictive transfusion threshold group. There were no results available for the smallest trial (18 participants). All studies were at some risk of bias, in particular performance bias relating to the absence of blinding of personnel. We judged the evidence for all outcomes, except myocardial infarction, was low quality reflecting risk of bias primarily from imbalances in protocol violations in the largest trial and imprecision, often because of insufficient events. Thus, further research is likely to have an important impact on these results.There was no evidence of a difference between a liberal versus restricted threshold transfusion in mortality, at 30 days post hip fracture surgery (risk ratio (RR) 0.92, 95% confidence interval (CI) 0.67 to 1.26; five trials; 2683 participants; low quality evidence) or at 60 days post surgery (RR 1.08, 95% CI 0.80 to 1.44; three trials; 2283 participants; low quality evidence). Assuming an illustrative baseline risk of 50 deaths per 1000 participants in the restricted threshold group at 30 days, these data equate to four fewer (95% CI 17 fewer to 14 more) deaths per 1000 in the liberal threshold group at 30 days.There was no evidence of a difference between a liberal versus restricted threshold transfusion in functional recovery at 60 days, assessed in terms of the inability to walk 10 feet (3 m) without human assistance (RR 1.00, 95% CI 0.87 to 1.15; two trials; 2083 participants; low quality evidence).There was low quality evidence of no difference between the transfusion thresholds in postoperative morbidity for the following complications: thromboembolism (RR 1.15 favouring a restrictive threshold, 95% CI 0.56 to 2.37; four trials; 2416 participants), stroke (RR 2.40 favouring a restrictive threshold, 95% CI 0.85 to 6.79; four trials; 2416 participants), wound infection (RR 1.61 favouring a restrictive threshold, 95% CI 0.77 to 3.35; three trials; 2332 participants), respiratory infection (pneumonia) (RR 1.35 favouring a restrictive threshold, 95% CI 0.95 to 1.92; four trials; 2416 participants) and new diagnosis of congestive heart failure (RR 0.77 favouring a liberal threshold, 95% CI 0.48 to 1.23; three trials; 2332 participants). There was very low quality evidence of a lower risk of myocardial infarction in the liberal compared with the restrictive transfusion threshold group (RR 0.59, 95% CI 0.36 to 0.96; three trials; 2217 participants). Assuming an illustrative baseline risk of myocardial infarction of 24 per 1000 participants in the restricted threshold group, this result was compatible with between one and 15 fewer myocardial infarctions in the liberal threshold group. AUTHORS' CONCLUSIONS: We found low quality evidence of no difference in mortality, functional recovery or postoperative morbidity between 'liberal' versus 'restrictive' thresholds for red blood cell transfusion in people undergoing surgery for hip fracture. Although further research may change the estimates of effect, the currently available evidence does not support the use of liberal red blood cell transfusion thresholds based on a 10 g/dL haemoglobin trigger in preference to more restrictive transfusion thresholds based on lower haemoglobin levels or symptoms of anaemia in these people. Future research needs to address the effectiveness of red blood cell transfusions at different time points in the surgical pathway, whether pre-operative, peri-operative or postoperative. In particular, such research would need to consider people who are symptomatic or haemodynamically unstable who were excluded from most of these trials.