| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
***,
#
*Canadian Blood Services; Departments of Medicine, Pediatrics, Pathology and Molecular Medicine, and #Medicine & Epidemiology and Biostatistics, McMaster University, Hamilton, Ontario; ¶Centre for Transfusion Research, University of Ottawa, Ottawa, Ontario; **Department of Statistics and Actuarial Science, University of Waterloo, Waterloo, Ontario; and ***Juravinski Cancer Centre, Hamilton, Ontario, Canada
Address correspondence to D. J. Cook, MSc, MD, Department of Medicine & Epidemiology and Biostatistics, McMaster University Health Sciences Center, Room 2C11, 1200 Main Street West, Hamilton, Ontario, Canada, L8N 3Z5. Address e-mail to debcook{at}mcmaster.ca.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Aprotinin, a serine protease inhibitor, is a potent antifibrinolytic medication that rapidly inhibits human plasmin, trypsin, and kallikrein. It is indicated for the prevention and treatment of the bleeding diatheses associated with profibrinolytic states and mitigates CPB-induced platelet dysfunction by preserving glycoprotein Ib and glycoprotein IIb/IIIa function on the platelet surface (3). Side effects of aprotinin include hypersensitivity reactions in 0.3% to 0.6% of patients upon re-exposure to the drug (4–6), an increased risk of perioperative myocardial infarction (7) and venous thrombosis (8).
A meta-analysis of 61 trials (n = 7027) of adults undergoing elective surgery demonstrated that aprotinin reduced the proportion of patients exposed to at least one unit of allogeneic red blood cells (RBC) by 30% (relative risk [RR] = 0.70; 95% confidence interval [CI], 0.64–0.76) compared with controls (9). Similarly, a meta-analysis of 35 trials (n = 3879) of adult patients undergoing coronary artery bypass grafting showed that aprotinin reduced the proportion of patients transfused by 39% (RR = 0.61; 95% CI, 0.58–0.66) (10). The effectiveness of aprotinin in children undergoing cardiac surgery is unclear, as randomized clinical trials (RCTs) in this population have reported conflicting results. The objective of this systematic review was to determine the effect of IV aprotinin administered perioperatively to children undergoing cardiac surgery with CPB on the proportion of children requiring allogeneic RBC or whole blood transfusions, the volume of blood transfused, and the amount of chest tube drainage in the immediate postoperative period.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
After initial screening, two independent reviewers examined the abstracts of potentially eligible RCT and selected those which met all of the following predefined inclusion criteria: 1) random allocation of all study treatment arms; 2) enrollment of children <18 yr of age; 3) primary or redo open-heart surgery with CPB for repair or palliation of CHD; 4) preoperative or intraoperative administration of IV aprotinin in any dose; 5) use of placebos, no aprotinin or other antifibrinolytic drugs as controls; 6) clinical outcomes that included the proportion of children requiring blood transfusion, the amount of transfused blood and/or the amount of chest tube drainage. RBC or whole blood transfusions were counted as the outcome unless the type of "blood transfusion" was not specified. The only exclusion criterion was non-English language publications.
We abstracted the following data: study design and source of funding; patient inclusion and exclusion criteria; number of patients screened and enrolled; average age, weight, and body mass index; number of patients with cyanotic heart disease; types of corrective surgical procedures; CPB variables; details of transfusion protocols; and doses of aprotinin. Outcome data we abstracted were the number of children requiring transfusion of any blood product (in excess of pump prime); volume of blood transfused; volume of postoperative chest tube drainage; therapy-related complications; and mortality. We contacted the corresponding authors to obtain missing data when possible.
Methodologic quality of the included trials were judged by two independent reviewers blinded as to the authors, affiliated institutions, sponsors, journal name, date of publication, and study results. We used the Jadad quality assessment scale (14), which assigns 1 point for each of the following criteria: 1) randomized treatment allocation; 2) appropriate methods of randomization; 3) the use of a double-blind study maneuver; 4) appropriate methods for double-blinding; and 5) a description of all withdrawals and dropouts. An overall score of 2 or lower was considered "poor" methodological quality (14). Furthermore, reviewers were asked to judge the adequacy of the method of allocation concealment and the use of an objective, predefined transfusion protocol.
When measures of variance were unavailable, they were imputed using mathematical formulae assuming a normal distribution of the data. The reported volumes of blood transfused and chest tube drainage were standardized by converting to mL/kg using the mean body weight or average body surface area of the study population where necessary. For multiarmed studies comparing different doses of aprotinin, the proportion of children transfused was determined by the total number of cases in all aprotinin arms divided by the total number of children; the amount of blood transfused and chest tube drainage were estimated by calculating the mean of all aprotinin arms. The random effects model of DerSimonian and Laird (15) was used to calculate the pooled RR for the proportion of children transfused, and the weighted mean difference (WMD) was calculated by pooling results of continuous variables (volume of transfused blood and volume of chest tube drainage), weighted by the inverse of the variance. We used Review Manager Version 4.2.3 (The Nordic Cochrane Centre, The Cochrane Collaboration 2003, Copenhagen, Denmark) for the analyses. Funnel plots were inspected for evidence of publication bias. We quantified the percentage of total variation across studies using the I2 test for heterogeneity and defined a low, moderate, and high I2 as 25%, 50%, and 75%, respectively (16). The a priori sources of heterogeneity we proposed were 1) study quality; 2) type of procedure (primary or redo sternotomies); 3) age or weight criteria; 4) cyanotic morphologies; and 5) aprotinin dose.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
= 0.52). Disagreement was often attributable to unspecified age criteria and/or surgical procedures in the abstracts of potentially eligible studies; however, disagreement was resolved by consensus in all cases. Initial agreement between reviewers on quality assessment was poor ( = 0.21); however, this scale has been associated with considerable interrater variability (17). Disagreement was resolved by third-party adjudication in all cases.
We identified 548 citations representing 541 published articles and 7 abstracts by the comprehensive literature search. Titles were screened for relevance, leaving 244 citations, of which 11 full publications (12,13,18–26) and one abstract (27) were included in this review, enrolling a total of 626 children between the ages of <1 and 16 yr (Fig. 1). Of the 12 studies, 7 were 2-arm trials comparing aprotinin to either placebo or no therapy (12,13,19,22,23,25,27) and 4 included 3 intervention groups (large-dose aprotinin, small-dose aprotinin, and either placebo or no treatment) (5,18,21,24). There was no uniform definition of large- or small-dose aprotinin regimens among studies. All treatments were randomly allocated except for the large-dose aprotinin arm in the study by Miller et al. (24); therefore, this arm was not considered in the analysis. One study included 2 groups with active controls, 
-aminocarproic acid and the combination of -aminocarproic acid plus aprotinin, in addition to a standard control group (no aprotinin) (20). The active control arms that used other antifibrinolytic drugs were not included in this analysis. The number of adverse events in the treatment arms was reported in 6 trials (12,21–24,26); none were considered to be attributable to aprotinin.
|
Of the 578 surgical procedures reported in the 12 studies, the most frequent was Tetralogy of Fallot repair (n = 140), followed by atrial septal defect and and/or ventricular septal defect repair including complete atrio-ventricular septal defects (n = 123); repair of transposition of the great arteries (n = 74); repair of hypoplastic left heart syndrome including Fontan and modified Fontan procedures (n = 71); and valvular replacements or repairs (n = 46). The details of CPB were similar among studies; in most studies, core body temperature was decreased to 24.0°C–30.1°C, blood flow rates were maintained 2.4 L/m2/min and a cardiac membrane oxygenator was used. The cardiac bypass pump was generally primed with colloid and crystalloid solutions in addition to allogeneic RBCs or whole blood. Further characteristics of each study, including aprotinin dosages, patient age, and weight are summarized in Table 1.
|
Four of 12 studies were judged to be of good methodological quality (12,21–23). Two studies adequately described the method of randomization (20,22), and in only one (22) were those methods appropriate. Four studies used placebo as the control group (12,21–23), 2 of which were described as identical (12,23), and none of the other 8 studies were described as double-blind (13,18–20,24–27). One study described the method of allocation concealment (20), and none of the studies adequately reported the number or reasons for withdrawals and dropouts (Table 2).
|
Six studies representing 362 children reported the proportion of children requiring at least one allogeneic RBC or whole blood transfusion after surgery with and without aprotinin (12,19,21–24). The mean age of children in these studies was 3.3 yr (range, 3.6 mo to 14.5 yr) and the mean weight was 12.6 kg (range, <3.5 to 42.5 kg). Aprotinin reduced the proportion of children transfused by 33% (RR = 0.67; 95% CI, 0.51 –0.89) (Fig. 2). The percentage of total variation across studies attributable to heterogeneity was low (I2 = 15%). When only the 4 studies of good methodological quality (12,21–23) were pooled, the effect of aprotinin on the proportion of children transfused remained statistically significant (RR = 0.60; 95% CI, 0.38-0.95). Similarly, when only the 3 studies that used an objective transfusion protocol were pooled (21–23), the effect of aprotinin was significant (RR = 0.72; 95% CI, 0.58-0.89). Three (12,19,23) of the 6 studies enrolled patients undergoing primary sternotomy only; the proportion of children transfused was reduced by 56% in this group (RR = 0.44; 95% CI, 0.26–0.76). The average weight of children in 5 of the 6 studies was more than 10 kg (19,21–24) and the effect of aprotinin remained significant in these studies (RR = 0.73; 95% CI, 0.59–0.89). Even in the single trial (12) that exclusively enrolled children less than 10 kg, the effect of aprotinin was significant. None of the studies enrolled children with cyanotic or noncyanotic CHD exclusively and each study used a different aprotinin dose regimen precluding analysis of these subgroups.
|
Seven studies enrolling 404 children reported the volume of blood transfused (13,20–23,25,27) and 10 studies enrolling 571 children reported the volume of chest tube drainage postoperatively with and without aprotinin (12,13,18–22,24,26,27). The effect of aprotinin on the volume of blood transfused (Fig. 3) and on the volume of chest tube drainage (Fig. 4) was not statistically significant (WMD = –8.42 mL/kg, 95% CI, –19.86 to 3.02; WMD = –0.97 mL/kg, 95%CI, –4.94 to 2.99, respectively). Heterogeneity across studies was high for these outcomes (I2 = 96% for volume of blood transfused and 77% for volume of chest tube bleeding).
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
With current methods of blood donor testing, allogeneic blood transfusion is extremely safe. Thus, the requirement for blood transfusion may be less clinically important than other "hard" outcomes such as reoperation rates resulting from bleeding, in-hospital morbidity, and death. Moreover, although the proportion of patients transfused may be reduced with the use of aprotinin, the number of overall donor exposures may not be similarly reduced, as most CPB circuits were primed with allogeneic blood.
The methodological quality of most studies included in this review was poor. Although all 12 studies were randomized, none provided an adequate description of withdrawals or dropouts, only 1 adequately described allocation concealment, and only half used an objective transfusion protocol. Our comprehensive search of the literature uncovered a subgroup report of children enrolled in an industry-sponsored compassionate use trial (28). The methods used in this study were not described, and substantially different numbers of patients were allocated to each of the 4 study arms. Moreover, the author stated that "we did not do hands-on monitoring of the trial, so the data may not be quite as clean as data from a more formal trial" (28). Therefore, this study was excluded from our review. In addition, 2 of the 12 trials in this review were industry-sponsored, a feature that has been associated with inflated estimates of benefit (31). This review also uncovered certain inconsistencies in reporting between aprotinin trials; each study used a different dose regimen (Table 1), and transfusion outcomes were occasionally reported as "blood transfusion" or "blood product" transfusions without specifying whether whole blood, RBCs, platelets, or plasma was transfused.
Any transfusion-sparing effect of aprotinin must be weighed against its associated complications and cost. Venous thrombosis and stroke are major causes of early and late morbidity and mortality after Fontan surgery, the definitive palliative surgical treatment for most congenital univentricular heart lesions; however, these complications have also been reported with other cardiac procedures (32). Seven of the 12 studies included in this review reported the frequency of complications and/or adverse events. No thrombotic or allergic complications of aprotinin were observed, and the frequency of adverse events were similar between groups in all studies, except one (22) where 5 early and 9 delayed adverse events were observed in the aprotinin group, compared with 2 early and 6 delayed adverse events in the placebo group (no test of significance was provided). Rare events, such as the thrombotic complications of aprotinin, are poorly captured in RCTs; even meta-analyses are usually underpowered to detect important effects because so few events are included in the original studies. The additional risk of thrombosis attributable to aprotinin in children undergoing corrective or palliative cardiac surgery remains unclear but must be considered for high-risk procedures. In addition, the cost of aprotinin is substantial; the average cost of aprotinin for a 10-kg child undergoing a 3-hour procedure using the large-dose regimen outlined in the Boldt et al. study (18) would be approximately $470 US.
In summary, pooling results of RCTs, aprotinin reduced the proportion of children who received allogeneic blood transfusions during cardiac surgery with CPB. However, aprotinin had no significant effect on the volume of blood transfused or on the amount of chest tube drainage. Among trials examining the effect of aprotinin in children, there is a need for consistency in reporting dosing regimens and transfusion requirements using objective transfusion protocols (30). Before the routine use of aprotinin in children undergoing cardiac surgery can be recommended, further independent RCTs are needed to carefully examine clinically important outcomes including bleeding, reoperation rates, and death in addition to the need for perioperative transfusion.
We thank Ms. Nancy Heddle and the McMaster Transfusion Research Program for critically appraising this research.
|
Footnotes |
|---|
Donald M. Arnold is a Transfusion Medicine Fellow funded by the Canadian Blood Services. Anthony Chan is a Career Investigator of the Heart and Stroke Foundation of Canada. Richard J. Cook is a Canada Research Chair. Graeme A. Fraser is a recipient of the Edith Turner Foundation Fellowship, Centre for Gene Therapeutics, McMaster University. Wendy Lim holds a Graduate Scholarship from the Canadian Institutes of Health Research. Deborah J. Cook is a Canada Research Chair.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. P. Eaton Antifibrinolytic Therapy in Surgery for Congenital Heart Disease Anesth. Analg., April 1, 2008; 106(4): 1087 - 1100. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. L. Backer, A. M. Kelle, R. D. Stewart, S. C. Suresh, F. N. Ali, R. A. Cohn, R. Seshadri, and C. Mavroudis Aprotinin is safe in pediatric patients undergoing cardiac surgery. J. Thorac. Cardiovasc. Surg., December 1, 2007; 134(6): 1421 - 1428. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
