| ||||||||||||||
|
|
|||||||||||||
Department of Anesthesia, Nippon Medical School Chiba Hokusoh Hospital, Chiba, Japan; Department of Anesthesiology, Nippon Medical School Hospital, Tokyo, Japan
Address correspondence and reprint requests to Toshiya Shiga, MD, PhD, Department of Anesthesia, Nippon Medical School Chiba Hokusoh Hospital, Kamagari 1715, Inba-mura, Inba-gun, Chiba 2701694, Japan. Address e-mail to qzx02115{at}nifty.com.
| Abstract |
|---|
|
|
|---|
| Introduction |
|---|
|
|
|---|
Antifibrinolytic therapy is a promising strategy to reduce blood loss and blood transfusion requirements. Aprotinin, a serine protease inhibitor, is already used in cardiac surgery, and its efficacy in reducing blood loss and transfusion requirements is widely recognized (35). Aprotinin reduces the risk of stroke in patients undergoing coronary artery bypass surgery (4) and decreases (3) or does not affect (4) mortality associated with cardiac surgery. However, its efficacy and safety in patients undergoing noncardiac surgery remain controversial, and conflicting results have been published even recently (58). In orthopedic surgery, such as hip or knee replacement when patients are at the most intense risk of venous thromboembolism, the safety of using aprotinin is debatable; it may predispose the patient to a procoagulable state and result in an adverse event such as myocardial infarction, stroke, or atrial fibrillation (58). Furthermore, aprotinin is not currently approved for orthopedic surgery by the United States Food and Drug Administration or in most other developed countries. This may explain orthopedists reluctance to use aprotinin. Therefore, we performed a meta-analysis of randomized controlled trials on the efficacy and safety of aprotinin in patients undergoing major orthopedic surgery. We focused on blood loss and the transfusion requirement as primary end-points to clarify the efficacy of aprotinin and on the incidence of deep vein thrombosis (DVT) as a secondary end-point to clarify the safety of aprotinin.
| Methods |
|---|
|
|
|---|
Assessment of the methodological quality of the included studies was performed by two independent investigators (TS and ZW). Disagreements were resolved by consensus. Each study was assessed according to the 5-point scale introduced by Jadad et al. (10), which scores randomization, double-blinding, withdrawals, and dropouts. Briefly, if the study was described as randomized, 1 point was assigned. If the randomization process was appropriate, an additional point was assigned. If the randomization was inappropriate (e.g., allocation by date of birth), the original point was lost. One point was assigned if a study was described as blinded. If the blinding method was appropriate, an additional point was assigned. Finally, 1 point was assigned if the number of and reasons for withdrawals and dropouts were described. The maximum possible score was 5.
The information extracted included the following: patient characteristics, type of surgery, dose of aprotinin, blood infusion criteria, intraoperative and perioperative (intraoperative + postoperative) blood loss, and red blood cell (RBC) requirements, incidence of DVT, method of DVT diagnosis, use or nonuse of DVT prophylaxis. Data were extracted by two independent investigators (TS and ZW). Disagreements or uncertainties were resolved by consensus. When results were not presented in the original paper in a usable form, attempts were made to obtain additional data from the authors. Where these data were not forthcoming, the trial was excluded from the meta-analysis.
Treatment effects for dichotomous and continuous outcomes were expressed as odds ratios and WMDs. The effect sizes were calculated with the use of a random-effects model (11). When zero outcome occurred in one or both groups, 0.5 was added to each cell of the respective contingency table. Homogeneity of effect size across trials was tested by the Cochran Q statistic.
|
| Results |
|---|
|
|
|---|
|
Ten trials (18,20,2329) evaluated the use of aprotinin for reduction of intraoperative blood loss, with 214 aprotinin-treated patients and 213 control patients. The pooled intraoperative blood loss was significantly less in the aprotinin-treated patients than in the control patients (WMD = 228.9 mL, 95% confidence interval [CI] = 366.8 to 91.0 mL; P = 0.0011, with significant heterogeneity among trials (P < 0.0001) (Fig. 2A). Eight trials (18,20,2429) with 163 aprotinin-treated patients and 160 control patients evaluated the use of aprotinin for reduction of perioperative blood loss. The pooled perioperative blood loss was significantly less in the aprotinin-treated patients than in the control patients (WMD = 556.8 mL, 95% CI = 860.1 to 253.5 mL; P < 0.0001), with significant heterogeneity among trials (P < 0.0001) (Fig. 2B).
|
Six trials (1820,24,26,29) evaluated the use of aprotinin for reduction of the amount of blood transfused intraoperatively, with 100 aprotinin-treated patients and 102 control patients. The pooled number of intraoperatively transfused RBC units (U) was significantly less in the aprotinin-treated patients than in the control patients (WMD = 1.1 U, 95% CI = 1.7 to 0.4 U; P = 0.0001), with no heterogeneity among trials (P = 0.11) (Fig. 3A). Nine trials (18,2022,24,2629) evaluated the use of aprotinin for reduction of perioperative blood transfusion, with 160 patients in the treatment group and 153 patients in the control group. The pooled number of perioperatively transfused RBC U was significantly less in the aprotinin-treated patients than in the control patients (WMD = 1.1 U, 95% CI = 1.74 to 0.45 U; P < 0.0001), with significant heterogeneity among trials (P < 0.0001) (Fig. 3B).
|
In 4 studies, all patients underwent routine ultrasonography (18,26,28) or venography (27). In 3 studies (21,25,29), routine clinical examinations were performed in all patients, and then, if positive, further venography was performed. Some studies (19,20,24,30) did not clarify whether examinations were performed for all or specific patients. For the remaining 128 patients ((2224), the diagnostic method was not specified. In 364 patients 18,20,21,2430), DVT prophylaxis was performed by means of heparin (25,26,28,30), low-molecular weight heparin (20,21,24,27,29), warfarin (28), or a mechanical device (18,27,29). For the remaining 142 patients (19,22,23), the method of prophylaxis was not specified. Diagnosis of DVT was confirmed by clinical signs (18,21,25), ultrasonography (18,19,26,28), or venography (21,25,27,29,30). DVT occurred in 1 patient and arteriovenous thrombosis occurred in 1 of 204 aprotinin-treated patients, and DVT occurred in 15 of 204 control patients. Aprotinin did not significantly affect the incidence of thrombotic events (odds ratio = 0.38; 95% CI = 0.14 to 1.05; P = 0.061), with no heterogeneity among trials (P = 0.75) (Fig. 4). The d-dimer test for the diagnosis of DVT, the length of hospital stay, and the incidence of pulmonary embolism were reported in only 2 (18,21), 3 (18,19,29) and 2 trials (18,29), respectively. We, therefore, did not include these data in the analysis.
|
During consecutive recruitment of patients in all trials, the investigators excluded patients with suspected or known allergy to aprotinin or with previous exposure to aprotinin. There were no reported anaphylactic or anaphylactoid reactions to aprotinin.
There was marked asymmetry of the funnel plot, confirmed by a significant Kendall correlation coefficient for intraoperative blood loss of 0.42 (P = 0.09), for intraoperative transfusion of 0.60 (P = 0.09), and for incidence of DVT of 0.60 (P = 0.01); this suggests a publication bias for these outcomes. The Kendall correlation coefficients for total blood loss and total transfusion were 0.36 (P = 0.22) and 0.39 (P = 0.14), respectively, indicating no publication bias in the analyses of total blood loss and transfusion.
| Discussion |
|---|
|
|
|---|
This is the first systematic review of the efficacy of aprotinin in patients undergoing major orthopedic surgery. Another pharmacologic strategy, use of tranexamic acid, a lysine analog, has already been shown in the meta-analysis by Ho and Ismail (31) to be safe and effective for reducing blood loss and the transfusion requirement in patients undergoing orthopedic surgery. Their results are comparable to ours: tranexamic acid reduces perioperative blood loss by 460 mL (mean) and the transfusion requirement by 0.85 U (mean). Ho and Ismail did not focus on the incidence of DVT but rather on the overall risk/incidence of thromboembolic complications, which do not increase with hip and knee arthroplasty. Like tranexamic acid, aprotinin moderately decreases blood loss and the allogeneic blood transfusion requirement. It may be argued that if the efficacy of both drugs in reducing blood loss and the transfusion requirement is similar, tranexamic acid is advantageous over aprotinin because of its lack of anaphylactic effects and its lower cost (aprotinin costs approximately $1000 per patient). However, Ho and Ismail did not clarify the presence or absence of pharmacologic DVT prophylaxis, which may have strongly influenced outcomes. Thus, we cannot directly compare the efficacies of the two drugs.
The effective dosage of aprotinin remains inconclusive. There are two basic regimens: full dose (
2 million KIU [kallikrein inhibitor units] as an initial dose) and half doses (32). In most of the trials included in our analysis, an initial dose of 12 million KIU was administered. We did not analyze the relation between dose and the amount of blood loss because of the small number of trials. Nevertheless, a dose-related reduction in blood loss has been suggested in some studies and reviews (16,17,32). Further large randomized trials are needed.
Without thromboprophylaxis, DVT occurs in 50% to 80% of patients after total hip or knee replacement or after surgery for hip fracture (33), and in 1% to 13% of these cases, it may lead to fatal pulmonary embolism (34). In the trials included in our analysis, nearly 90% of patients received antithrombotic or mechanical prophylaxis for DVT; therefore, the incidence of DVT was infrequent at 7.4% among control patients. It is controversial whether aprotinin may produce a procoagulable state resulting in increased risk of DVT. Our results suggest that aprotinin has no effect on the incidence of DVT in patients who receive thromboprophylaxis. Although statistically not significant, there was a trend toward fewer episodes of DVT in aprotinin-treated patients. It is unlikely that aprotinin has a direct antithrombotic effect. Instead, given the fact that transfusion itself poses a risk of hypercoagulability (35), we hypothesize that the decrease in transfusion resulting from aprotinin could reduce the occurrence of thromboembolism. One aprotinin-treated patient developed serious arteriovenous thrombosis requiring knee amputation, so the trial was discontinued. The use of aprotinin was not necessarily a cause of the complication in this case.
Several issues remain to be addressed. First, no anaphylactic reactions were reported in patients in the included trials, probably because of the strict exclusion of patients at risk for anaphylaxis and its infrequent incidence (<0.1%) (32). However, the risk of anaphylaxis increases by approximately 5% when re-exposure occurs within a few weeks (32,36). Levy (5) recommends avoiding re-exposure to aprotinin within 6 months after an initial exposure. Thus, despite the notable efficacy of aprotinin, routine clinical use in patients undergoing major orthopedic surgery cannot be recommended. Its use is particularly contraindicated when subsequent surgery is planned within several months (e.g., revision surgery). Patient eligibility must be carefully determined. Nevertheless, enthusiasm for use of aprotinin in noncardiac surgery is growing because the wide variety of complications of blood transfusion and increasing conservation costs are now better understood (2). Second, aprotinin is metabolized in the kidneys and is potentially nephrotoxic at large concentrations (37). Although clinical trials did not show that its use was significantly associated with serious renal dysfunction (37,38), aprotinin may have limited indications in patients with existing renal insufficiency or in combination with drugs such as certain antibiotics (e.g., gentamicin) that have potential detrimental effects on the kidneys (39). Third, there was heterogeneity in the results of our meta-analysis. This could be attributable to differences in the types of surgery, the backgrounds of patients, the criteria for blood transfusion, and other unidentified factors. Last, the possible publication bias suggests that small trials with negative results might not have been reported.
In summary, results of our meta-analysis indicate that aprotinin reduces intraoperative and perioperative blood loss and allogeneic blood transfusion requirements. Aprotinin may not be associated with an increased risk of DVT in the presence of pharmacologic or nonpharmacologic prophylaxis; however, definitive conclusions are precluded as the result of several study limitations.
| Footnotes |
|---|
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
N. Soundararajan and M. Cunliffe Anaesthesia for spinal surgery in children Br. J. Anaesth., July 1, 2007; 99(1): 86 - 94. [Abstract] [Full Text] [PDF] |
||||
![]() |
K. L. Skidmore and A. Azakie Aprotinin and Sevoflurane Do Not Affect Renal Function During Single-Ventricle Palliative Surgery Anesth. Analg., December 1, 2006; 103(6): 1614 - 1615. [Full Text] [PDF] |
||||
![]() |
R. Poston and O. Toshinaga Reply Ann. Thorac. Surg., November 1, 2006; 82(5): 1951 - 1952. [Full Text] [PDF] |
||||
![]() |
M. A. Camarasa, G. Olle, M. Serra-Prat, A. Martin, M. Sanchez, P. Ricos, A. Perez, and L. Opisso Efficacy of aminocaproic, tranexamic acids in the control of bleeding during total knee replacement: a randomized clinical trial Br. J. Anaesth., May 1, 2006; 96(5): 576 - 582. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|