Interventions for reducing red blood cell transfusion in adults undergoing hip fracture surgery: an overview of systematic reviews
Sharon R Lewis,Michael W Pritchard,Lise J Estcourt,Simon J Stanworth,Xavier L Griffin
DOI: https://doi.org/10.1002/14651858.CD013737.pub2
2023-06-08
Abstract:Background: Following hip fracture, people sustain an acute blood loss caused by the injury and subsequent surgery. Because the majority of hip fractures occur in older adults, blood loss may be compounded by pre-existing anaemia. Allogenic blood transfusions (ABT) may be given before, during, and after surgery to correct chronic anaemia or acute blood loss. However, there is uncertainty about the benefit-risk ratio for ABT. This is a potentially scarce resource, with availability of blood products sometimes uncertain. Other strategies from Patient Blood Management may prevent or minimise blood loss and avoid administration of ABT. Objectives: To summarise the evidence from Cochrane Reviews and other systematic reviews of randomised or quasi-randomised trials evaluating the effects of pharmacological and non-pharmacological interventions, administered perioperatively, on reducing blood loss, anaemia, and the need for ABT in adults undergoing hip fracture surgery. Methods: In January 2022, we searched the Cochrane Library, MEDLINE, Embase, and five other databases for systematic reviews of randomised controlled trials (RCTs) of interventions given to prevent or minimise blood loss, treat the effects of anaemia, and reduce the need for ABT, in adults undergoing hip fracture surgery. We searched for pharmacological interventions (fibrinogen, factor VIIa and factor XIII, desmopressin, antifibrinolytics, fibrin and non-fibrin sealants and glue, agents to reverse the effects of anticoagulants, erythropoiesis agents, iron, vitamin B12, and folate replacement therapy) and non-pharmacological interventions (surgical approaches to reduce or manage blood loss, intraoperative cell salvage and autologous blood transfusion, temperature management, and oxygen therapy). We used Cochrane methodology, and assessed the methodological quality of included reviews using AMSTAR 2. We assessed the degree of overlap of RCTs between reviews. Because overlap was very high, we used a hierarchical approach to select reviews from which to report data; we compared the findings of selected reviews with findings from the other reviews. Outcomes were: number of people requiring ABT, volume of transfused blood (measured as units of packed red blood cells (PRC)), postoperative delirium, adverse events, activities of daily living (ADL), health-related quality of life (HRQoL), and mortality. Main results: We found 26 systematic reviews including 36 RCTs (3923 participants), which only evaluated tranexamic acid and iron. We found no reviews of other pharmacological interventions or any non-pharmacological interventions. Tranexamic acid (17 reviews, 29 eligible RCTs) We selected reviews with the most recent search date, and which included data for the most outcomes. The methodological quality of these reviews was low. However, the findings were largely consistent across reviews. One review included 24 RCTs, with participants who had internal fixation or arthroplasty for different types of hip fracture. Tranexamic acid was given intravenously or topically during the perioperative period. In this review, based on a control group risk of 451 people per 1000, 194 fewer people per 1000 probably require ABT after receiving tranexamic acid (risk ratio (RR) 0.56, 95% confidence interval (CI) 0.46 to 0.68; 21 studies, 2148 participants; moderate-certainty evidence). We downgraded the certainty for possible publication bias. Review authors found that there was probably little or no difference in the risks of adverse events, reported as deep vein thrombosis (RR 1.16, 95% CI 0.74 to 1.81; 22 studies), pulmonary embolism (RR 1.01, 95% CI 0.36 to 2.86; 9 studies), myocardial infarction (RR 1.00, 95% CI 0.23 to 4.33; 8 studies), cerebrovascular accident (RR 1.45, 95% CI 0.56 to 3.70; 8 studies), or death (RR 1.01, 95% CI 0.70 to 1.46; 10 studies). We judged evidence from these outcomes to be moderate certainty, downgraded for imprecision. Another review, with a similarly broad inclusion criteria, included 10 studies, and found that tranexamic acid probably reduces the volume of transfused PRC (0.53 fewer units, 95% CI 0.27 to 0.80; 7 studies, 813 participants; moderate-certainty evidence). We downgraded the certainty because of unexplained high levels of statistical heterogeneity. No reviews reported outcomes of postoperative delirium, ADL, or HRQoL. Iron (9 reviews, 7 eligible RCTs) Whilst all reviews included studies in hip fracture populations, most also included other surgical populations. The most current, direct evidence was reported in two RCTs, with 403 participants with hip fracture; iron was given intravenously, starting preoperatively. This review did not include evidence for iron with erythropoietin. The methodological quality of this review was low. In this review, there was low-certainty evidence from two studies (403 participants) that there may be little or no difference according to whether intravenous iron was given in: the number of people who required ABT (RR 0.90, 95% CI 0.73 to 1.11), the volume of transfused blood (MD -0.07 units of PRC, 95% CI -0.31 to 0.17), infection (RR 0.99, 95% CI 0.55 to 1.80), or mortality within 30 days (RR 1.06, 95% CI 0.53 to 2.13). There may be little or no difference in delirium (25 events in the iron group compared to 26 events in control group; 1 study, 303 participants; low-certainty evidence). We are very unsure whether there was any difference in HRQoL, since it was reported without an effect estimate. The findings were largely consistent across reviews. We downgraded the evidence for imprecision, because studies included few participants, and the wide CIs indicated possible benefit and harm. No reviews reported outcomes of cognitive dysfunction, ADL, or HRQoL. Authors' conclusions: Tranexamic acid probably reduces the need for ABT in adults undergoing hip fracture surgery, and there is probably little or no difference in adverse events. For iron, there may be little or no difference in overall clinical effects, but this finding is limited by evidence from only a few small studies. Reviews of these treatments did not adequately include patient-reported outcome measures (PROMS), and evidence for their effectiveness remains incomplete. We were unable to effectively explore the impact of timing and route of administration between reviews. A lack of systematic reviews for other types of pharmacological or any non-pharmacological interventions to reduce the need for ABT indicates a need for further evidence syntheses to explore this. Methodologically sound evidence syntheses should include PROMS within four months of surgery.