 |
 |
 |
|
| home | CME | help | feedback | subscriptions | archive |
|
© 1997 The Journal of Bone and Joint Surgery, Inc.
Coagulopathy Complicating Intraoperative Blood Salvage in a Patient Who Had Idiopathic Scoliosis. A Case Report*  , BALTIMORE, MARYLANDInvestigation performed at the University of Maryland Medical School, Baltimore
Techniques that have been designed to avoid the use of homologous blood, including preoperative donation of autologous blood and intraoperative blood salvage, are now commonly employed in major orthopaedic procedures9,11,12,14,15,17. Coagulopathy is a rare complication of spinal arthrodesis; the reasons for its occurrence may be variable and may include activation of platelets and white blood cells4, consumption of clotting factors18, dilutional coagulopathy in association with massive transfusion20, and defibrination triggered by injury secondary to decortication or chipping at bone23. We present here the case of a patient who had life-threatening disseminated intravascular coagulation. We believe that this complication was caused by a combination of factors, including dilutional coagulopathy in conjunction with intraoperative blood salvage and the use of absorbable gelatin sponges (Gelfoam; Upjohn, Kalamazoo, Michigan) soaked in topical thrombin (Thrombostat; Parke Davis, Morris Plains, New Jersey). Excessive hemodilution can precipitate the activation of platelets and leukocytes during the spin cycle in the Cell Saver (Haemonetics, Braintree, Massachusetts)3-6. This initiates the cascade that has been called the salvaged-blood syndrome or disseminated intravascular inflammation, also known as hemodilution-induced platelet-and-leukocyte activation syndrome. Thrombin and Gelfoam may have been aspirated along with the salvaged blood and, on reinfusion, may have contributed to the disseminated intravascular coagulation.
A seventeen-year-old girl who had adolescent idiopathic scoliosis had a posterior spinal arthrodesis for the treatment of a double major curve (58 degrees to the right from the fourth to the tenth thoracic vertebra, with correction to 40 degrees on side-bending, and 56 degrees to the left from the eleventh thoracic to the third lumbar vertebra, with correction to 24 degrees on side-bending). Previous treatment had consisted of two years of surface electrical stimulation with an electrospinal orthosis (Medtronic, Minneapolis, Minnesota), which was worn while the patient was sleeping. The curve progressed despite this treatment, and the patient did not have additional follow-up until she was seen by us. The patient was healthy but of slight build: she was 145 centimeters tall and weighed thirty-eight kilograms. She donated three units of autologous blood before the procedure. Preoperatively, the hemoglobin level was 108 grams per liter (normal, 115 to 165 grams per liter), the platelet count was 400 x 109 per liter (normal, 150 to 400 x 109 per liter), the activated partial thromboplastin time was twenty-nine seconds (normal, twenty-one to thirty-two seconds), and the prothrombin time was 11.6 seconds (normal, 10.5 to 12.7 seconds). General anesthesia was maintained with nitrous oxide and isoflurane, and the systolic blood pressure was kept between eighty and ninety millimeters of mercury (10.66 and 12.00 kilopascals) with nitroprusside. The patient was under anesthesia for a total of nine hours; the duration of the operation was six hours and fifteen minutes. A routine posterior approach was used, and an arthrodesis was performed from the third thoracic to the second lumbar vertebra with use of Cotrel-Dubousset instrumentation. There was slightly more bleeding than usual during the dissection, so thrombin-soaked Gelfoam sponges were used to aid hemostasis. Shed blood was salvaged with the Cell Saver 3 (Haemonetics), and heparin (30,000 units per liter of normal saline solution) was used as the anticoagulant. The heparinized saline solution was dripped in at fifteen milliliters per 100 milliliters of whole-blood loss. In this instance, three units of salvaged cells were infused, with the anticoagulant volume totaling 650 milliliters (a total of 650/1000 x 30,000 x 19,500 units of heparin). A standard wake-up test to determine the function of the spinal cord was performed four and one-half hours after the induction of anesthesia, when the instrumentation was in place and correction had been achieved. The patient moved the feet well and was put back under anesthesia. At that time, the estimated blood loss was 1400 milliliters, the pulse rate was eighty beats per minute, the blood pressure was 70/40 millimeters of mercury (9.33/5.33 kilopascals), and the urinary output was 400 milliliters (average, eighty milliliters per hour). By that time, the patient had received 7000 milliliters of normal saline solution, the hemoglobin level was fifty grams per liter, and the hematocrit was 15 per cent. Bone graft was taken from the posterior iliac crest while decortication of the site of the arthrodesis was completed. The first unit of predonated autologous blood was to be transfused approximately fifteen minutes before the wake-up test was performed, but transfusion was delayed by half an hour because the identification number on the blood did not match that on the patient's wristband. While the issue of blood identification was being resolved at the blood bank, the patient received 500 milliliters of 5 per cent albumin followed by the first unit from the Cell Saver. After matching had been confirmed, the first unit of predonated autologous blood was transfused with another 500 milliliters of 5 per cent albumin. During this delay, the patient remained stable, with a systolic blood pressure of sixty-five to seventy millimeters of mercury (8.66 to 9.33 kilopascals) and a pulse rate of ninety-five to 100 beats per minute. Thirty minutes after initiation of the transfusion of the first unit of washed salvaged cells, both wounds began to fill rapidly with blood while the incisions were being closed. There was no indication of a transfusion reaction, so the second unit of predonated autologous blood, the second unit from the Cell Saver, 500 milliliters of 5 per cent albumin, and 1500 milliliters of normal saline solution were infused. Fifteen minutes later, the systolic blood pressure dropped to fifty millimeters of mercury (6.67 kilopascals), and the pulse rate, to forty beats per minute; ST-segment depression was seen on the electrocardiogram. Atropine, epinephrine, and bicarbonate were administered along with the third unit of predonated autologous blood and the third unit from the Cell Saver while wound closure was completed expeditiously. Despite the administration of another 1500 milliliters of normal saline solution (which brought the total fluid replacement to 10,000 milliliters of normal saline solution, 1500 milliliters of 5 per cent albumin, three units of predonated autologous blood, and three units of washed salvaged blood), the patient remained hypotensive with a systolic blood pressure of fifty millimeters of mercury (6.67 kilopascals) for another twenty minutes until the systolic blood pressure returned to 120 millimeters of mercury (16.00 kilopascals). The patient was rolled to the supine position and was noted to have a grossly distended abdomen, generalized edema (especially about the face), and hematuria. Fifteen minutes after wound closure and during the following two hours, she received two units of homologous blood, three units of fresh-frozen plasma, another 4000 milliliters of normal saline solution, and ten units of platelets. Thus, the total fluid replacement included 14,000 milliliters of normal saline solution, 1500 milliliters of 5 per cent albumin, three units of predonated autologous blood, three units of washed salvaged blood, two units of banked blood, ten units of platelets, and three units of fresh-frozen plasma. A radiograph of the chest showed bilateral pulmonary infiltrates consistent with either adult respiratory distress syndrome or transfusion-related acute lung injury. A computed tomography scan, made on an emergency basis, showed marked amounts of ascites (with a density ranging from approximately one to twenty Hounsfield units) within the abdomen but no evidence of retroperitoneal hematoma. Forty minutes after the hemodynamic collapse, the prothrombin time was 17.4 seconds, the activated partial thromboplastin time was 150 seconds, the fibrinogen level was 0.8 gram per liter (normal, 1.48 to 4.22 grams per liter), the level of fibrin degradation products was more than forty milligrams per liter (normal, less than ten milligrams per liter), and the hemoglobin level was ninety-seven grams per liter. There had been no rash or evidence of a transfusion reaction intraoperatively, and subsequent Coombs' tests and repeat blood-typing of returned blood-unit bags showed no incompatibility. The patient required five days of mechanical ventilation in the intensive-care unit and received four additional units of packed red blood cells, eight units of fresh-frozen plasma, and one unit of cryoprecipitate. The coagulation profile returned to almost normal on the first postoperative day, but the clinical course was further complicated by acute pancreatitis that necessitated parenteral nutrition for ten days. She was discharged to home in good condition thirteen days after the operation with satisfactory correction of the scoliosis. Six months after the operation, the patient had a thoracic curve of 20 degrees, a lumbar curve of 32 degrees, and intact neurological function. The hardware had not migrated but it was somewhat prominent and was palpable under the skin, with the stabilization rod on the right causing minor irritation. The patient was able to begin physical therapy and physical-education class but was not permitted to participate in contact sports.
Although the patient had bled more than was expected during the initial dissection, the replacement of the estimated blood loss of 1400 milliliters with either crystalloid or colloid at that time would not normally have been expected to cause dilutional coagulopathy. Dilutional coagulopathy is rare unless transfusion requirements exceed seventy milliliters per kilogram20. At this point in the operation, the patient had received 7000 milliliters of normal saline solution, the hemoglobin level was fifty grams per liter, and the hematocrit was 15 per cent; hence, the blood being salvaged was extremely dilute. The sudden increased hemorrhage at the time of wound closure, with rapid pooling of blood in both wounds from previously dry margins, is consistent with disseminated intravascular coagulation7,19,21. However, salvage of extremely dilute blood can itself lead to a coagulopathy. This mechanism was reported by Bull et al.3-6, who described a platelet-and-leukocyte activation syndrome in which the reinfusion of such blood resulted in intravascular damage and inflammation, giving a disseminated intravascular coagulation-like picture. Those authors postulated that this mechanism is related to centrifugal separation of extremely dilute blood. Dilution may be related to the aspiration of excessive irrigation fluid with the salvaged blood or to the salvage of extremely dilute blood. The temporal relationship between the administration of salvaged blood and the sudden cardiovascular collapse is additional strong supporting evidence of disseminated intravascular coagulation. Transfusion incompatibility was excluded as a possible cause with use of Coombs' tests and repeat blood-typing. Experimental studies4-6 have shown that salvaged blood that has been diluted with large quantities of saline solution (either during irrigation or during the infusion of crystalloid) creates deposits of cellular aggregates on the polycarbonate bowl of the centrifuge; when this blood is reinfused, the increased vascular permeability leads to coagulopathy and adult respiratory distress syndrome. These cellular deposits do not form in the centrifuge when undiluted blood is used4. Although we did not use irrigation until after the wake-up test was performed, the patient had received a large volume (7000 milliliters) of normal saline solution by that time; this had resulted in severe hemodilution, as evidenced by a hemoglobin level of fifty grams per liter. Thus, the anasarca in our patient may have been more consistent with disseminated intravascular inflammation than disseminated intravascular coagulation. During the centrifugation of salvaged blood, a coating of red blood cells normally forms on the surface of the plastic bowl; this coating functions as a cushion to prevent leukocytes and platelets from touching the bowl. With severe hemodilution, the diluted blood enters the centrifuge bowl and arcs in and up in a different, longer path than when either undiluted blood or blood from a patient who has a high hematocrit is aspirated. This longer arc allows the blood components to strike the side of the bowl farther up before rolling downward. It is the journey down the unprotected surface of the plastic bowl that triggers the leukocytes and platelets. Thus, activated clotting factors contaminate the blood and, on reinfusion, cause a clinical disseminated intravascular coagulation syndrome3-6. To our knowledge, all previously reported cases of salvaged-blood syndrome have been associated with the use of heparin rather than citrate as the anticoagulant in the Cell Saver4,5. The reason for this is unknown, and the choice of heparin or citrate currently is based on the preference of the physician. In light of these reported cases, however, it may be preferable to use citrate in conjunction with the Cell Saver even though it is more expensive than heparin. Cellular debris from salvaged blood is another possible cause of coagulation abnormalities, especially if the suction vacuum is set too high or an inadequate wash cycle is used. In this instance, the suction vacuum was not adjusted to more than 100 millimeters of mercury (13.33 kilopascals), and each unit was washed with 1200 to 1500 milliliters of saline solution to give a clear effluent13,20. Another factor that may have contributed to the disseminated intravascular coagulation in our patient may have been the use of topical thrombin and Gelfoam to aid hemostasis during the early part of the procedure, as some portion of these agents may have been aspirated with the blood. The injection of therapeutic doses of thrombin has been shown to cause fatal coagulopathy in laboratory animals. In one study28, rabbits that received an intravenous injection of a weight-adjusted dose of thrombin (the equivalent of what would be used during an operation on the eye) died within thirty seconds because of total intravascular coagulation. In another study26, dogs that received an infusion of thrombin had development of disseminated intravascular coagulation, with deposition of fibrin in the microvascular circulation, structural changes in the vessel walls, and extravasation of blood from the circulation. Allergic reactions to thrombin have been observed in isolated instances8,22,24,25. While such a reaction does not seem to be a likely explanation for the coagulopathy in our patient, the potential risk of this complication must be considered before a patient receives any future treatment in which the use of thrombin is an option. Disseminated intravascular coagulation has been reported after procedures involving the spine, but previous investigators have not discussed the potential interaction of thrombin and intraoperative blood salvage18,20,23. We wonder if the use of topical hemostatic agents may have been a factor in previously reported cases of disseminated intravascular coagulation involving the use of the Cell Saver. Both the manufacturer of the Cell Saver (Haemonetics) and the American Association of Blood Banks have recommended against the use of the Cell Saver in conjunction with blood-clotting agents, such as fibrin glue, microfibrillar collagen, and thrombin, which may stimulate coagulopathy or cause the intraoperative blood-salvage reservoir to clot1. They also have warned against contamination with amniotic or gastric fluid, which may stimulate clotting because of the high concentrations of proteolytic enzymes. Povidone-iodine, hydrogen peroxide, water, alcohol, and bone cement may cause hemolysis. We believe that the coagulopathy in our patient may have been caused by a combination of factors, including severe hemodilution in association with the use of the Cell Saver as well as the aspiration of topical thrombin and Gelfoam. Unfortunately, this contraindication is not mentioned either in the orthopaedic literature or in the package insert for thrombin. The potential for disseminated intravascular coagulation after intravascular injection of thrombin is recognized and is mentioned in the package insert. Therefore, thrombin should not be used in conjunction with intraoperative blood salvage. Disseminated intravascular coagulation also has been reported after embolization of esophageal varices with Gelfoam2, and this application also should be avoided. We hope to raise the awareness of this problem in the orthopaedic community so that other orthopaedic surgeons will be able to avoid it. Some surgeons have told us anecdotally that they occasionally use Gelfoam and thrombin during procedures involving the spine but that they are careful not to place the suction catheter of the Cell Saver near that area. We believe that this practice may be too risky. The salvaged-blood syndrome4, or disseminated intravascular inflammation, is not well known in orthopaedic circles. Orthopaedic surgeons who use the Cell Saver should be cautioned against salvaging blood that has been excessively diluted either during irrigation or during intravenous infusion of crystalloid. If the surgeon wishes to use copious amounts of saline solution for irrigation, then a separate suction device connected directly to waste should be used. Standard textbooks dealing with the operative management of patients who have scoliosis fail to mention this specific preventable problem that may be associated with intraoperative blood salvage10,16,27.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
  CiteULike  Connotea  Del.icio.us  Facebook  Technorati  Twitter What's this?
This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Introduction
Case Report
| home | CME | help | feedback | subscriptions | archive | site index |
|
Stanford University Libraries' HighWire Press (TM) assists in the publication of JBJS Online. |