Pooled analysis of the outcome. *Mean based on a random-effects model (method of DerSimonian and Laird7); relative risk (RR) less than 1 favors intraoperative autotransfusion, whereas RR greater than 1 favors control. Solid diamonds indicate risk ratio of the individual study; open diamond, the pooled risk ratio; 1, equals equal risk.
Takagi H, Sekino S, Kato T, Matsuno Y, Umemoto T. Intraoperative Autotransfusion in Abdominal Aortic Aneurysm SurgeryMeta-analysis of Randomized Controlled Trials. Arch Surg. 2007;142(11):1098-1101. doi:10.1001/archsurg.142.11.1098
To determine whether intraoperative autotransfusion reduces the percentage of patients undergoing allogeneic blood transfusion.
Data Sources and Study Selection
Using a public domain database (MEDLINE) and a Web-based search engine (PubMed), all intraoperative autotransfusion vs control prospective randomized controlled trials that enrolled patients undergoing elective infrarenal abdominal aortic aneurysm surgery, published between January 1, 1966, and November 30, 2005, were searched. Relevant studies were identified through a manual search of secondary sources including references of initially identified articles.
Data on detailed inclusion criteria, autotransfusion system type, and incidence of allogeneic blood transfusion were abstracted from each study. Sensitivity analyses were performed by excluding individual trials one at a time and recalculating the pooled risk ratio estimates for the remaining studies.
Our search identified 4 randomized controlled trials including data for 292 patients. Pooled analysis demonstrated a statistically significant 37% reduction in risk of allogeneic blood transfusion with intraoperative autotransfusion compared with control (risk ratio, 0.63; 95% confidence interval, 0.41-0.95; P = .03) in a random-effects model. There was statistically significant trial heterogeneity of results (P = .02) but no evidence of statistically significant publication bias (P = .497). Two of 4 sensitivity analyses demonstrated statistically nonsignificant results favoring intraoperative autotransfusion.
Based on a meta-analysis of available randomized controlled trials, intraoperative autotransfusion reduces risk of allogeneic blood transfusion in elective infrarenal abdominal aortic aneurysm surgery.
Intraoperative autotransfusion (IAT) is a method of blood conservation. The technique is effective in reducing the requirement for allogeneic blood transfusion (ABT) (receipt of at least 1 U of allogeneic red blood cells) in abdominal aortic aneurysm (AAA) surgery.1,2 However, subanalysis of 2 small randomized controlled trials3,4 in a systematic review by Alvarez et al5 did not find sufficient evidence that IAT decreases exposure to ABT in infrarenal AAA surgery. Since the systematic review5 was undertaken, results of 2 larger randomized controlled trials2,6 have been recently published. To determine whether IAT reduces the percentage of patients undergoing ABT in elective infrarenal AAA surgery, we conducted a meta-analysis of available randomized controlled trials.
All prospective randomized controlled trials of IAT vs control (no IAT) that enrolled patients undergoing elective infrarenal AAA surgery were identified using a 2-level search strategy. First, a public domain database (MEDLINE) was searched using a Web-based search engine (PubMed). Second, relevant studies were identified through a manual search of secondary sources including references of initially identified articles. The MEDLINE database was searched from January 1, 1966, to November 30, 2005. Keywords included transfusion, abdominal aortic aneurysm, and randomized controlled trials.
Studies considered for inclusion met the following criteria: the design was a prospective randomized controlled clinical trial, the study population included patients undergoing elective infrarenal AAA surgery, patients were randomly assigned to IAT vs control groups, and main outcomes included incidence of ABT. All qualifying studies were assessed for adequate blinding of randomization and the intervention (IAT) and objectivity of the outcome assessment. Data on detailed inclusion criteria, autotransfusion system type, and incidence of ABT were abstracted, as available, from each study.
For each study, data about the incidence of ABT in both the IAT and control groups were used to generate risk ratios (RRs) (RR <1, favors IAT; RR >1, favors control) and 95% confidence intervals (CIs). Between-study heterogeneity was analyzed using standard χ2 tests. P < .05 was deemed statistically significant. Where significant statistical heterogeneity was identified, a random-effects model (method of DerSimonian and Laird7) was used preferentially as the summary measure. Sensitivity analyses were performed to assess the contribution of each study to the pooled estimate by excluding individual trials one at a time and recalculating the pooled RR estimates for the remaining studies. Publication bias was assessed graphically using a funnel plot and mathematically using an adjusted rank correlation test (method of Begg and Mazumdar8).
Our search identified 5 prospective randomized controlled clinical trials of IAT vs control that enrolled patients undergoing elective infrarenal AAA surgery.1- 4,6 The trial by Thompson et al,1 however, was excluded because the number of patients exposed to ABT was not clearly stated. Although the trials by Mercer et al2 and Spark et al3 were undertaken by the same unit in the same hospital, they were assessed independently of each other because different autotransfusion systems were used.
The 4 trials analyzed were by Mercer et al,2 Spark et al,3 Clagett et al,4 and Wong et al.6 There was substantial qualitative heterogeneity in trial design. Only patients in the IAT group of the study by Wong et al6 underwent acute normovolemic hemodilution to reduce a hemoglobin concentration of 11 g/dL (to convert to grams per liter, multiply by 10.0). In the trial by Clagett et al,4 all patients were given ABT according to the guidelines advocated by Nelson et al,9 which include intraoperative transfusion for hemodynamic instability; hemoglobin concentration less than 10 g/dL (hematocrit <30%; to convert to proportion of 1.0, multiply by 0.01); and postoperative transfusion because of hemoglobin concentration less than 8 g/dL (hematocrit <25%) or 8 to 10 g/dL (hematocrit, 25%-30%) in those with compromised cardiopulmonary status. Patients with hemoglobin concentration of 10 g/dL or greater (hematocrit >30%) did not receive transfusions. In the trial by Mercer et al,2 patients received blood products to maintain a hemoglobin concentration of 8 g/dL during and after surgery. In the trial by Spark et al,3 patients received transfusions of allogeneic blood if the hematocrit value decreased to less than 25%. In the trial by Wong et al,6 allogeneic blood was transfused when the hemoglobin concentration decreased to less than 8 g/dL or when ischemic electrocardiographic changes persisted after correction of hypovolemia. Patients were younger in the trial by Clagett et al4 (mean age, 63 years in the IAT group and 65 years in the control group) than in the other 3 trials.
Among the 4 trials that were used for the present meta-analysis, those by Spark et al3 and Mercer et al2 included only patients undergoing AAA surgery. Although 100 patients undergoing AAA repair (n = 50) or aortofemoral bypass because of occlusive disease (n = 50) were randomized to IAT and control groups in the trial by Clagett et al,4 only data for those undergoing AAA repair were analyzed. The trial by Wong et al6 included 34 patients who underwent elective infrarenal aortic surgery because of occlusive disease of 145 randomized participants, but we were able to obtain data for the other 111 patients (F. Torella, FRCS, written communication, September 13, 2005). In total, our meta-analysis included data for 292 patients undergoing elective infrarenal AAA surgery randomized to the IAT vs control groups (Table).
Blinding of randomization and the intervention and objectivity of the outcome assessment were not stated in the trial by Spark et al.3 The trial by Clagett et al4 was unblinded. In the trial by Mercer et al,2 patients were blinded to transfusion group allocation. Members of the operating surgical team were responsible for the continuing care of patients, decision to use blood transfusion, and investigation of postoperative complications, and they were independent of the research team but were not blinded to the use of IAT. The trial of Wong et al6 was single-blinded, but the decision to give ABT was made using a rigid protocol and by a physician independent from the research team.
Three of 4 trials demonstrated a statistically significant benefit of IAT over control for risk of ABT (Figure). Only the trial by Clagett et al4 demonstrated a statistically nonsignificant ABT risk reduction with IAT over control.
Pooled analysis of the 4 trials, representing 292 patients, demonstrated a statistically significant 37% reduction in risk of ABT with IAT compared with control (RR, 0.63; 95% CI, 0.41-0.95; P = .03) in the random-effects model (Figure). There was statistically significant trial heterogeneity of results (P = .02). To assess the effect of qualitative heterogeneity in trial design and patient selection on the pooled effect estimate, we performed several sensitivity analyses. First, we excluded the trial by Spark et al3 with the lowest RR (0.14); combining the remaining 3 trials generated an attenuated but still statistically significant result favoring IAT (RR, 0.76; 95% CI, 0.62-0.92; P = .006). Second, we excluded the trial by Clagett et al4 with the highest RR (0.89). Without the trial by Clagett and colleagues, there was still a statistically significant benefit for IAT in pooled analysis of the remaining 3 trials (RR, 0.51; 95% CI, 0.29-0.92; P = .02). Third, we excluded the trial by Wong et al,6 which enrolled the most patients (n = 111); combining the remaining 3 trials generated statistically nonsignificant results favoring IAT (RR, 0.55; 95% CI, 0.29-1.05; P = .07). Fourth, we excluded the trial by Mercer et al.2 Without the trial by Mercer and colleagues, there was a statistically nonsignificant benefit for IAT in pooled analysis of the remaining 3 trials (RR, 0.55; 95% CI, 0.28-1.06; P = .07).
To assess publication bias, we generated a funnel plot of the logarithm of effect size vs the standard error for each trial. There was no evidence of statistically significant publication bias (P = .497).
Alvarez et al5 conducted a systematic review of published studies to determine whether IAT reduced the exposure of ABT in abdominal vascular surgery. To our knowledge, their study is only a meta-analysis of randomized controlled trials. In the subgroup of 100 patients who underwent infrarenal AAA surgery in 2 small randomized controlled trials,3,4 the pooled RR for IAT was 0.37 (95% CI, 0.06-2.36). The present meta-analysis of 4 randomized controlled trials2- 4,6 including 292 patients, however, demonstrated a statistically significant 37% reduction in risk of ABT with IAT compared with control. The trial by Wong et al6 is the largest (n = 111) of the trials included in the present meta-analysis; the other 3 trials2- 4 included fewer than 100 patients each.
The present analysis must be viewed in the context of its limitations. First, only data from randomized controlled trials were used, but patients enrolled in randomized trials may not be representative of those typically seen in clinical practice. Second, the oldest trial (published in 1990 by Thompson et al1) among those retrieved was excluded because the appropriate information, not included in the article, was sought by the authors without success. Third, the present results may be influenced by a publication bias favoring IAT. Although the statistical tests did not indicate publication bias, there is clearly limited power to detect such bias, given the small number of studies examined. Fourth, 2 of 4 sensitivity analyses, without the trials by Mercer et al2 or Wong et al,6 demonstrated statistically nonsignificant results favoring IAT. In conclusion, based on a meta-analysis of available randomized controlled trials, IAT reduces risk of ABT in elective infrarenal AAA surgery.
Correspondence: Hisato Takagi, MD, PhD, Department of Cardiovascular Surgery, Shizuoka Medical Center, 762-1 Nagasawa, Shimizu-cho, Sunto-gun, Shizuoka 411-8611, Japan (firstname.lastname@example.org).
Accepted for Publication: August 2, 2006.
Author Contributions:Study concept and design: Takagi and Umemoto. Acquisition of data: Takagi, Sekino, Kato, and Matsuno. Analysis and interpretation of data: Takagi, Sekino, Kato, and Matsuno. Drafting of the manuscript: Takagi. Critical revision of the manuscript for important intellectual content: Takagi, Sekino, Kato, Matsuno, and Umemoto. Statistical analysis: Takagi and Kato. Administrative, technical, and material support: Takagi and Umemoto. Study supervision: Umemoto.
Financial Disclosure: None reported.
Additional Contributions: Francesco Torella, FRCS, Julian C. L. Wong, FRCS, Sarah L. Haynes, PhD, Kirsteen Dalrymple, BSc(Hons), Andrew J. Mortimer, FRCA, and Charles N. McCollum, FRCS, on behalf of the ATIS (Autologous vs Allogeneic Transfusion in Aortic Surgery) investigators, from the Academic Surgery Unit, Wythenshawe Hospital, Manchester, England, supplied us with data from their study.