Blood Use in Cerebrovascular Neurosurgery
Background and Purpose— This study reviews the perioperative use of red blood cell transfusion in cerebrovascular neurosurgery. The current algorithm for preoperative ordering of red cells is historical and dated. More blood is ordered than is actually transfused, and considerable variability exists between different institutions. We determine the use of blood transfusion in cerebrovascular surgery to develop a rational blood ordering practice.
Methods— Records of 301 patients undergoing cerebrovascular neurosurgery at the University of Virginia were reviewed to quantitatively evaluate red blood cell transfusion practices. The amount and reason for transfusion were noted in each case.
Results— In 126 patients undergoing carotid endarterectomy, there were no preoperative or intraoperative transfusions and 5 postoperative transfusions (4.0%). In 71 ruptured aneurysm patients, there were 2 preoperative blood transfusions (2.8%), 4 intraoperative transfusions (5.6%), and 15 postoperative transfusions (21.1%). Forty-seven patients underwent surgery for unruptured aneurysms, with no preoperative transfusions, 2 intraoperative transfusions (4.3%), and 8 postoperative blood transfusions (17.0%). Of the 54 patients undergoing surgery for arteriovenous malformations, 5 patients (9.3%) were transfused preoperatively, 4 were transfused intraoperatively (7.4%), and 22 were transfused postoperatively (40.7%). None of the 3 patients undergoing surgery for concomitant arteriovenous malformations and aneurysms received intraoperative blood transfusions, but 1 received blood both preoperatively and postoperatively, and another received a transfusion postoperatively only. The overall ratio of perioperative cross-match to transfusion in this series is 41.4.
Conclusions— In vascular neurosurgery at our institution, blood has routinely been ordered excessively. We recommend an ABO-Rh type and antibody screen for aneurysm and arteriovenous malformation surgery and no screen for carotid endarterectomy to efficiently utilize transfusion therapy in cerebrovascular surgery.
The amount of blood transfused in the perioperative period has changed significantly over the last decade, largely because of increased awareness of transfusion-associated risks and increasing costs of transfusions.1 The number of units of red blood cells (RBCs) ordered before surgery has been based on the physician’s transfusion experiences with previous patients. At the University of Virginia, 4 U has been ordered before aneurysm and arteriovenous malformation (AVM) surgeries and 2 U before carotid endarterectomy (CEA). Studies examining the physiology and effects of anemia during and after surgery have led to new strategies to reduce the perioperative use of blood products. The number of RBC units ordered in the past may not be an appropriate standard of practice. This study analyzes the existing blood ordering policy for CEA, AVM, and ruptured and unruptured aneurysm surgeries at a single institution. New guidelines are proposed to direct future blood ordering in cerebrovascular surgery.
Records of 301 patients who underwent neurological surgery for cerebrovascular disease at the University of Virginia were retrospectively reviewed. One hundred twenty-six patients undergoing CEA, 118 patients with cerebral aneurysms (71 ruptured, 47 unruptured), 54 patients with AVMs, and 3 patients with both AVMs and aneurysms were reviewed. Determination of RBC use was undertaken in the preoperative, intraoperative, and postoperative periods (within 3 days of surgery), and the reason for use was noted in each case. This information was compared with the policies at other institutions.
Results of this review are depicted in Table 1. In 126 patients undergoing CEA, there were no preoperative or intraoperative transfusions and 5 postoperative transfusions for postoperative complications. One patient was transfused on postoperative day (POD) 1 after blood was noted in the oropharynx, most likely because of a traumatic nasal intubation, and the hematocrit (Hct) was slowly trending downward. One patient developed a non–Q-wave myocardial infarction postoperatively and received an RBC transfusion on POD 0 for an Hct of 30.3. One patient received RBCs on POD 1 for a coagulopathy, and another patient developed a postoperative wound hematoma requiring an emergency clot evacuation and received a transfusion of RBCs on POD 0 before his emergency surgery (Hct not checked). The last patient underwent a CEA without complication and with an estimated blood loss of 50 cm3, but on arrival at the postanesthesia care unit, she was noted to have hypotension and a prolonged QT interval, as well as a small amount of bleeding from the operative site. Her Hct was 28.5, and she was transfused with 2 U of RBCs. The hypotension was later determined to be due to an idiosyncratic response or hyperresponse to β-blockers received intraoperatively.
In 71 patients undergoing surgery for ruptured aneurysm clipping, 2 patients received RBC transfusions preoperatively to correct decreased Hct. Four patients received transfusions intraoperatively. One patient had an intraoperative rupture of a 14-mm ruptured left middle cerebral artery trifurcation aneurysm and lost ≈350 cm3 of blood. She received 2 U of RBCs intraoperatively. One patient received 2 U of RBCs intraoperatively after a 700-cm3 blood loss during emergent clipping of a ruptured middle cerebral artery aneurysm associated with intracerebral hemorrhage causing significant mass effect and clinical deterioration. Another patient with a ruptured basilar artery aneurysm lost 500 cm3 of blood during surgery and received 1 U of RBCs intraoperatively. The fourth patient had a 2-cm ruptured ophthalmic segment aneurysm and received 2 U of RBCs intraoperatively after intraoperative rupture of her aneurysm and an estimated blood loss of 500 cm3. There were 15 patients who received RBC transfusions in the postoperative period. Most of these patients received transfusions for dilutional decrease in Hct secondary to hypertension, hypervolemia, and hemodilution therapy to treat vasospasm except for the following exceptions. One underwent angioplasty postoperatively for vasospasm and received heparin during the procedure. Although the heparin was reversed afterward, there was noticeable hemorrhage on CT after the procedure, and she underwent a craniotomy for clot removal and received another 5 U for this procedure. Another received 2 U of RBCs for an Hct of 28.3 on POD 1 after clipping of a pericallosal artery aneurysm that had ruptured intraoperatively.
There were a total of 47 patients who underwent surgery for elective clipping of unruptured aneurysms. None of these patients received preoperative RBC transfusions, and 2 patients received intraoperative RBC transfusions. One patient experienced torrential hemorrhage as the frontal lobe was retracted because of rupture of the left carotid artery and received 2 U of blood intraoperatively. Another lost an estimated 1000 cm3 of blood and was transfused intraoperatively with 2 U. Eight patients were transfused postoperatively. Six of these patients received 2 U of RBCs for dilutional decrease in Hct. The other 2 patients experienced intraoperative bleeding with estimated blood loss of 500 cm3 each and were transfused on POD 0 with 1 U of RBCs each and then another 1 and 2 U, respectively, on POD 1.
Fifty-four patients underwent surgery for resection of AVMs. Five patients were transfused preoperatively (all had ruptured AVMs) for decreased Hct. Four patients were transfused intraoperatively. One of these patients developed an intraparenchymal hemorrhage during embolization and was taken emergently to the operating room. The blood loss was estimated to be 1800 cm3, and 4 U of RBCs was transfused intraoperatively. Another patient was embolized without apparent complication and was taken to the intensive care unit after embolization. She subsequently developed intracerebral hemorrhage with intraventricular extension and clinical deterioration and was taken to the operating room for decompression of the hematoma. Multiple bleeding vessels from the AVM were discovered, and a partial resection was performed. She received 3 U of RBCs for preoperative and intraoperative blood loss. One patient developed an occipital hematoma after surgical resection of an occipital AVM. Although a coagulopathy was diagnosed, the emergent circumstances necessitated surgery. She received 8 U of RBCs intraoperatively for intraoperative blood loss. One patient received 2 U of RBCs intraoperatively after an estimated blood loss of 800 cm3 caused by intraoperative bleeding from a large central parasagittal AVM that was extraordinarily difficult to resect because of unusual and unexpected anatomic features. Twenty-two patients were transfused postoperatively for decreased Hct after surgery.
Three additional patients were operated on for both aneurysms and AVMs, none of whom received intraoperative transfusions of RBCs. One of these patients received RBCs both preoperatively and postoperatively for decreased Hct. Another received RBCs postoperatively for decreased Hct from blood loss in surgery. Both patients who received blood had bled preoperatively.
Given the growing concerns over safety, cost, and adequacy of the national blood supply, much effort has focused on examining the blood ordering and transfusion practices in various types of surgery.2–9⇓⇓⇓⇓⇓⇓⇓ Although the blood supply is now safer than ever before as a result of screening, there are still risks of transmission of diseases such as HIV and hepatitis (about 1 of 1 000 000 for each). There is also the risk of a fatal hemolytic transfusion reaction (1of 300 000 to 1 of 700 000 transfusions) and transfusion-acquired lung injury (1 of 5000).1 Autologous blood transfusion has a decreased risk of infection but is more expensive10 and still has risks of anemia, administration of the wrong unit, and the potential need for more frequent blood transfusions. With the advent of healthcare reform, the cost of transfusions and the adequacy of the blood supply are becoming increasingly relevant. At many institutions, there is a set policy regarding the ordering of RBCs for vascular neurosurgical cases; however, these policies vary widely (Table 2).11 The policy is often based on tradition rather than a formal study or an audit of the transfusion practice of the facility. The number of units of RBCs cross-matched for surgery is always greater than the number of units actually transfused.2,7⇓
Several studies have examined the problem of excessive cross-matching through the ratio of cross-match to transfusion (C/T ratio).3,7,8⇓⇓ A high C/T ratio means that the blood bank must keep more blood in its inventory, which increases hospital costs, wastes personnel time, and increases the likelihood of outdated blood products. The perioperative C/T ratios for the various procedures in this study are listed in Table 3. Mintz and colleagues6 have recommended a C/T ratio of 2.0. In neurovascular surgery at the University of Virginia, the perioperative C/T value is much higher in every instance, with an overall C/T ratio of 41.4 (Table 3).
Compatibility testing in the blood bank consists of ABO and Rh typing of the patient’s RBCs, an antibody detection test on the patient’s serum (antibody screen), and a cross-match. The antibody screen involves reacting patient serum with commercially purchased RBCs that express all of the common, clinically significant antigens. This test provides almost complete assurance that the patient does not have any detectable RBC alloantibodies and that transfusing ABO-compatible RBCs will not result in a hemolytic transfusion reaction.12 The cross-match involves mixing RBCs from a donor unit with patient serum to look for evidence of direct incompatibility between the donor and prospective recipient. In most US hospitals, if the patient has a nonreacting antibody screen and no history of RBC alloantibodies, the blood bank performs only an “immediate-spin” cross-match to confirm ABO compatibility before issuing blood. For patients who have an RBC alloantibody (or a history of one), a complete cross-match using anti-human globulin is performed with units of blood negative for the antigen reactive with the patient’s antibody.
For each unit of RBCs ordered, a cross-match is performed. At the University of Virginia, the charge is $37.00 per unit cross-matched, whether the product is transfused or not. There is an additional charge (processing fees) for the blood product when administered to its ultimate recipient. The costs for any products that are outdated because of preparation for patients but not used must be absorbed by the hospital.
Problems with the blood supply occur when products that have a short shelf life are prepared and sent to the operating room but not used. When blood is cross-matched for surgery, it is unavailable for others for 24 to 48 hours, and the chance of outdating the RBCs is increased. RBCs can be stored for 35 days when stored in citrate phosphate dextrose adenine, 13 which can be extended to 42 days when AS-1 (Adsol) or AS-3 (Nutricel) is used.14,15⇓ This duration has been set by federal regulation and is determined by the requirement that ≥75% of the transfused cells remain in circulation after infusion. However, evidence suggests that the oxygen delivery capacity of the RBCs decreases with time in storage and that patients fare worse after receiving blood stored for increasing lengths of time.16
Several new strategies shown to effectively reduce the perioperative transfusion of blood products are being implemented to reduce the perioperative use of blood components.17 Many physicians use a hemoglobin concentration of 70 rather than 100 g/L as the level that prompts transfusion. There is no evidence that mild to moderate anemia contributes to perioperative morbidity. The timing of transfusion has also changed, so many patients who were previously transfused intraoperatively are now given blood in the postoperative period.9 Thus, cross-matches performed preoperatively may result in an increasing number of reserved units not being transfused, thereby unnecessarily elevating the preoperative C/T ratio.
Another method to increase efficiency of preoperative blood ordering is to develop a preoperative blood ordering schedule based on local experiences and implement a type and screen more regularly to avoid routine cross-matching for surgical procedures in which blood is rarely transfused. The efficacy of this practice has been demonstrated.6 A type and screen invokes a one-time charge of $30.50.
A blood transfusion performed after a type and screen is still very safe; the chance of missing a potentially dangerous antibody is estimated to be no more than 1 in 10 000.18 In neurovascular procedures, there is a possibility that the blood will be needed urgently. If this is the case, an immediate-spin cross-match can be performed before transfusion to eliminate reactions that may result from human errors in ABO-Rh typing. This test and the release of blood take only 10 to 15 minutes to perform.18 In fact, as noted, many hospitals perform only an immediate-spin cross-match for patients whose antibody screen demonstrates no RBC alloantibodies and have no history of any such antibodies. Blood given in this manner is very safe.19 In addition, crystalloids and colloids can be infused while a sample is being cross-matched to gain additional time.
Interestingly, in a recently published article reviewing blood transfusion practices in a large series of aneurysm operations only, the authors found a 24.5% intraoperative transfusion rate.20 This is much higher than our experience, and the reason for this apparent disparity is unclear.
From the results of our study, we feel an ABO-Rh type and antibody screen is sufficient preparation preoperatively for aneurysm and AVM surgery. In this study, if a transfusion was needed for CEA, the need was always postoperative and never emergent. Therefore, we believe even a type and screen is unnecessary for CEA. Implementing these changes not only would reduce costs but also would allow more efficient use of the nation’s blood supply.
- Received November 16, 2001.
- Accepted December 17, 2001.