- Email: [email protected]
Role of autologous blood transfusion in sacral tumor resection: patient selection and recovery after surgery and blood donation
J Orthop Sci (2000) 5:321–327
Role of autologous blood transfusion in sacral tumor resection: patient selection and recovery after surgery and blood donation Sadaaki Nakai, Hidezo Yoshizawa, Shigeru Kobayashi, Kumi Naga, and Hirofusa Ichinose Department of Orthopedic Surgery, Fujita Health University School of Medicine, Toyoake, Aichi 470-1192, Japan
Abstract We carried out sacral en-bloc resection in six patients (three with chordoma; one with pheochromocytoma; one with malignant schwannoma; and one with giant cell tumor) using preoperatively collected autologous blood, to avoid homologous blood transfusion. An average of 3200 ml was collected preoperatively, with patients receiving recombinant human erythropoietin (r-HuEPO), at a total dose of 130 000 units on average. In four patients, we were able to accomplish the surgery without homologous blood transfusion. Postoperatively, the hemoglobin level in these four patients recovered to the pre-collective level in 4.5 weeks, on average. These clinical results indicate that en-bloc sacrectomy, which requires a large volume of blood transfusion, can be accomplished with preoperatively collected autologous blood alone. Key words Sacral tumor · Malignant tumor · Sacrectomy · Autologous blood transfusion · Recombinant human erythropoietin
Introduction Tumors of the sacrum tend to expand anteriorly into the pelvic cavity and are often discovered late, so the lesions are usually quite large at diagnosis. Because reoperation for local recurrence of these tumors is extremely difficult, the initial procedure should be radical and curative. In addition, reconstruction of the sacroiliac joint with an implant is necessary when the first and second sacral vertebrae are removed. For such surgeries, the transfusion of a large volume of blood is necessary. Where possible, we perform this surgery using preoperatively donated autologous blood, especially because homologous transfusion generally has an
Offprint requests to: S. Nakai Received: July 2, 1999 / Accepted: December 14, 1999
immunosuppressive effect.3,11 In addition, we wish to avoid other risks associated with blood transfusion, including hepatitis, acquired immunodeficiency syndrome, and graft-versus-host disease. While reducing blood loss is the best way to avoid or minimize the need for homologous transfusion during or after surgery, the use of autologous blood is another alternative. Autologous blood transfusion is the safest form of transfusion in patients undergoing elective surgery, including oncologic procedures.20 A large volume of autologous blood can be obtained by collecting and storing blood preoperatively, but this may take a considerable length of time. In general, surgery should be performed without delay in patients who have bone tumors for which resection is indicated. In patients with slow-growing tumors,21 however, surgery can be delayed to allow for autologous blood donation. For patients with fast-growing tumors who require prompt surgery, as much autologous blood as possible must be collected during the time required to analyze the results of diagnostic imaging and biopsies prior to the operation. We studied six patients whose surgery was performed under hypotensive anesthesia to reduce intraoperative blood loss.
Patients, methods, and results Sacral resection has been performed on six patients at our institution to date (Table 1 and Table 2). Three patients were treated by single-stage, combined anteroposterior surgery (cases 1, 5, and 6). In one patient, surgery was carried out as a two-stage procedure, comprising tumor resection in the first stage and instrumentation with bone grafting in the second (case 4). Two other patients were managed via the posterior approach alone (cases 2 and 3). For storage of autologous blood, three different methods were used. Citratephosphate-dextrose (CPD) was used for storage of
55/F 72/M
55/M
29/F
39/F
2 3
4
5
6
Malignant schwannoma GCT
Pheochromocytoma
Chordoma Chordoma
Chordoma
Diagnosis
4 5
5 6
2000 2400
3200 3600 4000 4000
Quantity of autologous blood collected (ml)
Approach
48 000 90 000
144 000 144 000 192 000 162 000
r-HuEPO (units)
13.1 11.2
12.5 12.6 12.8 13.3
Surgery
10.6 10.3
11.4 10.9 10.8 10.7
Lowest preoperative Hb level (g/dl)
First Resection left sacrum; medial part of left ilium Second, Internal fixation Resection left sacrum; medial part of left ilium Resection S1–S5; medial parts of both ilia; internal fixation
Resection S2–S5 Resection S3–S5
Resection S2–S5
Hb level before initial blood collection (g/dl)
Anteroposterior
Anteroposterior
Two-stage anteroposterior
Posterior Posterior
Anteroposterior
rHuEPO, Recombinant human erythropoietin; Hb, hemoglobin
8 10 11 11
1 2 3 4
Case no.
Period of blood collection (weeks)
Table 2. Summary of patient data on blood transfusion
GCT, Giant cell tumor of bone
53/M
Age (years)/ sex
1
Case no.
Table 1. Summary of patient data on diagnosis, surgery, and postoperative course Follow-up
3883 4564 2466 3703 (First surgery) 1095 (Second surgery) 4212 6841
Deep infection
Superficial infection
Complication
7 Homologous transfusion Homologous transfusion
6 4 1
Period required for postoperative recovery of Hb level (weeks)
Died (7 months after surgery) 4 Years
2 Years Died (1 month after surgery) 4 Years
4 Years
Perioperative blood loss (ml)
21 h 5 min
2 h 39 min 16 h 17 min
13 h 8 min
14 h 3 h 59 min
12 h 23 min
Operative time
322 S. Nakai et al.: Autologous blood transfusion in sacral tumor resection
S. Nakai et al.: Autologous blood transfusion in sacral tumor resection
blood for less than 3 weeks, mannitol-adeninephosphate (MAP) for storage of blood for 3 to 6 weeks, and freezing was used for storage of blood for longer than 6 weeks. Case 1 Case 1 was a 53-year-old man diagnosed with a sacral chordoma. Sacral resection was indicated, because bone scans showed no abnormal tracer accumulation, suggesting the absence of metastasis. It was decided that autologous blood would be sufficient to meet the needs of intraoperative and postoperative transfusion. A volume of 400 ml of blood was collected at each session, based on the patient’s weight of 61 kg. Blood collection was started 8 weeks before surgery. A total of 3200 ml of autologous blood was collected, of which 2000 ml was frozen and 1200 ml was stored using MAP. During the 8-week period in which blood was collected, a total of 24 doses, of 6000 units each, of recombinant human erythropoietin (r-HuEPO) were administered by intravenous injection three times per week (total dose, 144 000 units). The hemoglobin (Hb) level was 12.5 g/dl before the initiation of collection and decreased to 11.4 g/dl during the collection period. However, on the day before surgery, the Hb level was 12.4 g/dl (Fig. 1). After embolization of the internal iliac artery on the day before surgery, the sacrum was resected en bloc from S2 to S5 via both the anterior and posterior approaches in a single operation. The operating time
323
was 12 h and 23 min. Intraoperative and postoperative blood loss totaled 3883 ml, comprising 3521 ml during surgery and 362 ml via a drain over the 2-day period after surgery. A volume of 2400 ml of autologous blood was transfused intraoperatively, and 800 ml was transfused over the 2-day period after surgery. The Hb decreased to 7.7 g/dl on the day after surgery, but recovered to 12.5 g/dl by 6 weeks after the operation. The use of homologous blood was avoided. The patient developed a superficial wound infection postoperatively, but this healed with conservative treatment. Four years after the operation, there has been no recurrence of tumor. Case 2 Case 2 was a 55-year-old woman with a sacral chordoma. Surgery was scheduled to allow time for autologous blood donation. Blood collection was started 10 weeks before surgery, with 400 ml being obtained at each session, based on the patient’s body weight of 56 kg. A total of 3600 ml of blood was collected preoperatively, of which 1600 ml was frozen, 400 ml was stored using MAP, and 1600 ml was stored using CPD. During the blood collection period, six doses, each of 24 000 units, of r-HuEPO were administered subcutaneously (total dose, 144 000 units). The Hb level was 12.6 g/dl before blood collection was started, decreased to 10.9 g/dl during the collection period, and increased again to 11.7 g/dl before the final collection on the day before surgery (Fig. 2).
Fig. 1. Hemoglobin (Hb) values from the start of blood collection until after surgery in case 1. Blood collection was started 8 weeks before surgery, and 3200 ml of autologous blood was obtained. Hb recovered to the level before collection by 6 weeks (ws.) after surgery. Large arrow, Day of surgery; small arrows, days on which 400 ml of blood was collected
Fig. 2. Hemoglobin (Hb) values from the start of blood collection until after surgery in case 2. Blood collection was started 10 weeks before surgery, and 3600 ml of autologous blood was obtained. Hb recovered to the level before collection by 6 weeks (ws.) after surgery. Large arrow, Day of surgery; small arrows, days on which 400 ml of blood was collected
324
S. Nakai et al.: Autologous blood transfusion in sacral tumor resection
Fig. 3. Hemoglobin (Hb) values from the start of blood collection until after surgery in case 3. Blood collection was started 10 weeks before surgery, and 4000 ml of autologous blood was obtained. Hb recovered to the level before
collection by 1 week (w.) after surgery, because the patient lost only 2466 ml of blood (1931 ml intraoperatively and 535 ml postoperatively.) Large arrow, Day of surgery; small arrows, days on which 400 ml of blood was collected
After embolization of the internal iliac artery on the day before surgery, the patient underwent en-bloc resection of the sacrum from S2 through S5, via the posterior approach. The operating time was 14 h. Intraoperative and postoperative blood loss totaled 4564 ml, comprising 3454 ml during surgery and 1110 ml over the 2-day period after surgery. A volume of 3200 ml of autologous blood was transfused intraoperatively, and 400 ml was transfused on the day after surgery. The Hb in this patient decreased to 7.1 g/dl on the day after surgery, but after 4 weeks, it had recovered to 12.8 g/dl. The use of homologous blood was avoided. Two years after the operation, there has been no recurrence of the tumor.
A total of 2000 ml of autologous blood was transfused during surgery, 800 ml was transfused on the day after surgery, and 400 ml was transfused on each of postoperative days 2, 4, and 7. The Hb level in this patient was 10.5 g/dl on the day after surgery, the lowest value observed postoperatively. By 1 week after surgery, the Hb had increased to 13.4 g/dl (Fig. 3). Transfusion of homologous blood was not required. The patient died 1 month after surgery because of respiratory distress syndrome and diffuse intravascular coagulopathy.
Case 3 Case 3 was a 72-year-old man who was hospitalized with a sacral chordoma. Surgery was scheduled to allow time for autologous blood donation. Blood collection was started 11 weeks before the operation, and 400 ml of blood was obtained at each session, based on the patient’s body weight of 66 kg. A total of 4000 ml of blood was collected, of which 2400 ml was frozen, 1200 ml was stored using MAP, and 400 ml was stored using CPD. During the blood collection period, 24 000 units of r-HuEPO was administered subcutaneously once a week for 10 weeks (total dose, 192 000 units, because, on two occasions, blood was collected without using r-HuEPO). The Hb level was 12.8 g/dl before blood collection was started, fell to 10.8 g/dl during the collection period, and recovered to 13.3 g/dl on the day before the operation. After embolization of the internal iliac artery, the patient underwent en-bloc resection of the sacrum from S2 through S5 via the posterior approach. The operating time was 3 h 59 min. Intraoperative and postoperative blood loss totaled 2466 ml, comprising 1931 ml during surgery and 535 ml over the 2-day period after surgery.
Case 4 Case 4 was a 55-year-old man in whom a solitary metastatic lesion was detected in the sacrum prior to resection of the left kidney and adrenal gland for pheochromocytoma. Because the lesion was confined to the left half of the sacrum, it was decided to resect this half of the sacrum and the medial portion of the left ilium, followed by reconstruction with instrumentation and bone grafting. Autologous blood donation was started 11 weeks before surgery. A volume of 400 ml of blood was collected at each session, based on the patient’s weight of 65 kg. A total of 4000 ml of blood was collected before surgery, of which 2400 ml was frozen and 1600 ml was stored using MAP. This patient received a total of 27 doses of 6000 units of rHuEPO given by intravenous injection three times per week (total dose, 162 000 units). The Hb level was 13.3 g/dl before the start of blood collection and decreased to 10.7 g/dl on the day before surgery, the minimum value observed during the blood collection period. Embolization of the internal iliac artery was performed on the day before surgery. In the first operation, the left half of the sacrum and the medial portion of the left ilium were resected en bloc via both the anterior and posterior approaches. The operating time was 13 h and 8 min. Intraoperative and posto-
S. Nakai et al.: Autologous blood transfusion in sacral tumor resection
325
Fig 4. Hemoglobin (Hb) values from the start of blood collection until after surgery in case 4. Blood collection was started 11 weeks before surgery, and 4000 ml of autologous blood was obtained. Hb recovered to virtually the pre-
collection level by 7 weeks (ws.) after the first operation (1st op.). Large arrows, Days of surgery; small arrows, days on which 400 ml of blood was collected
perative blood loss totaled 3703 ml, comprising 3393 ml during surgery and 310 ml over the 3-day period after surgery. A total of 3200 ml of autologous blood was transfused during surgery, and an additional 400 ml was transfused postoperatively when the Hb value fell to 6.8 g/dl 2 days after the procedure (the lowest level seen postoperatively). The second operation was performed 2 weeks after the first operation, when the Hb had recovered to 11.3 g/dl. In the second operation, internal fixation, using an implant, and bone grafting were performed to obtain bone union between L5 and the left ilium. The operating time was 2 h and 39 min. During surgery, 510 ml of blood was lost, and 400 ml of autologous blood was transfused. A further 585 ml of blood drained postoperatively, for a total blood loss of 1095 ml. The nadir was 8.2 g/dl 2 days after the second operation, and homologous blood transfusion was successfully avoided. Seven weeks after the first operation, the Hb had increased to 12.9 g/dl (Fig. 4). Four years after the two operations, there has been no recurrence of tumor. This patient’s preoperative blood adrenaline level was 621 pg/ml (normal range, less than 100 pg/ml), while the noradrenaline level was 2882 pg/ml (normal range, 100 to 450 pg/ml), and dopamine level was 20 pg/ml (normal range, less than 20 pg/ml). Therefore, the autologous blood transfused during surgery may have contained high levels of adrenaline, noradrenaline, and dopamine. However, the intraoperative levels ranged from 143 to 272 pg/ml for adrenaline, 531 to 979 pg/ml for noradrenaline, and 38 to less than 5 pg/ml for dopamine. In addition, there were no sudden increases in blood pressure intraoperatively.
resonance imaging. Autologous blood donation was started immediately after admission. Although her body weight was 42 kg, 400 ml of blood was successfully obtained at each session. A diagnosis of malignant schwannoma was made, based on the results of biopsy performed under local anesthesia. During the blood collection period, eight doses of 6000 units of r-HuEPO were administered intravenously over the 4-week period before surgery (total dose, 48 000 units). A total of 2000 ml of blood was collected, of which 800 ml was stored using MAP, and 1200 ml was stored using CPD. The Hb level was 13.1 g/dl at the start of blood collection and fell to a minimum of 10.6 g/dl during the blood collection period. After embolization of the internal iliac artery, the left side of the sacrum and the medial portion of the left ilium were resected en bloc, with internal fixation, using an implant and bone grafting to obtain union between L4, L5, and the left ilium. The operating time was 16 h and 17 min. A total of 4212 ml of blood was lost during surgery. A total of 1600 ml of homologous blood was transfused, in addition to the 2000 ml of autologous blood collected preoperatively. There was no recurrence at the sacral site, but the patient died of a massive abdominal tumor 7 months after surgery.
Case 5 Case 5 was a 29-year-old woman with hereditary neurofibromatosis type I,18 characterized by generalized cafe-au-lait spots and neurofibromas. She was referred after the detection of a sacral tumor on magnetic
Case 6 Case 6 was a 39-year-old woman referred for resection of a sacral tumor. A giant cell tumor of bone was diagnosed, based on the results of a biopsy performed under local anesthesia. Autologous blood was collected for 5 weeks before surgery, during which time 15 doses of 6000 units of r-HuEPO were administered intravenously (total dose, 90 000 units). A total of 2400 ml of autologous blood was collected, of which 1200 ml was stored using MAP, and 1200 ml was stored using CPD. The Hb level was 11.2 g/dl before blood donation, and fell to a minimum of 10.3 g/dl during the blood collection period.
326
S. Nakai et al.: Autologous blood transfusion in sacral tumor resection
After embolization of the internal iliac artery, the entire sacrum and the medial segments of both ilia were resected en bloc in one operation via both the anterior and posterior approaches. Internal fixation with an implant and bone grafting were performed at the same time to obtain union between L4, L5, and the ilia. The operating time was 21 h and 5 min. A total of 5241 ml of blood was lost during surgery, requiring the transfusion of 2000 ml of homologous blood, in addition to the 2400 ml of autologous blood deposited preoperatively. A total of 1600 ml of blood was lost over the 8-day period after surgery. Her Hb decreased to 7.5 g/dl 3 days postoperatively but later recovered. She developed a deep wound infection 11 days after the surgery, which was treated by curettage and continuous irrigation. However, the infection failed to resolve, and a second operation was required. The patient was subsequently followed for 4 years, without recurrence of the tumor.
Discussion Homologous blood transfusion during renal transplantation has been reported to have an immunosuppressive effect,8 and homologous transfusion during resection of malignant tumors also suppresses immunity13,16 and may facilitate tumor recurrence or growth.4,6,10,12,22,24 So although one cannot draw firm conclusions concerning the contribution of homologous transfusion to outcome in patients with malignancy, the risk that homologous blood transfusion may have a deleterious effect on immune function makes it advisable to avoid homologous transfusion as much as possible. The indications for the preoperative use of donated autologous blood transfusion in patients with metastatic tumor remain poorly defined. In preoperatively collected autologous blood, tumor cells are potential contaminants. Still, the level of contamination in collected blood does not seem to exceed the blood level of tumor cells at the time of surgery. Clinical studies of patients who received autologous transfusion of preoperatively donated blood for hepatic resection of metastasis colorectal cancer did not show worse postoperative mortality than patients who received homologous transfusions.5 Autologous blood transfusion can be justified in those patients with metastatic tumor in whom the primary tumor has been removed and a good prognosis is expected. Extensive sacral resection is indicated for benign tumors with marked invasion of the sacrum and for malignant tumors confined to the sacrum. If resection is minimized for fear of causing neurologic deficits, further surgery may be required because of tumor recurrence, subsequently resulting in a worse neurologic outcome. In addition, reoperation is extremely difficult and more damaging to the patient.
In a randomized study, the administration of rHuEPO has been shown to increase the amount of autologous blood that could be collected before surgery.9 Although the collection of a large volume of autologous blood without the use of r-HuEPO could be possible over time, we could not find any reports on extensive short-term blood collection without the use of r-HuEPO. Little is known about the interaction between r-HuEPO and nonhematopoietic malignant cells. R-HuEPO showed no stimulation of clonal growth in solid tumor cell lines.2 There have been reports of autologous blood transfusion using r-HuEPO in the surgical treatment of solid malignant tumors, including hepatocellular, colon, and uterine cervical cancers. Neither enhancement of tumor growth by r-HuEPO nor any interaction between tumor and r-HuEPO has been documented in these reports.7,14,15,25 There have been reports that it is possible to deposit a large amount of blood within a short period by giving patients r-HuEPO,19 and this method appeared useful for patients who required the storage of a large blood volume in a brief time. In our patients, 400-ml volumes of blood were repeatedly collected at 1-week intervals, but the recovery of their Hb levels suggested that blood could have been collected more frequently. Resolution of postoperative anemia may be delayed by the preoperative administration of high doses of r-HuEPO, because endogenous EPO secretion is suppressed by exogenous r-HuEPO.23 Our experience showed that the period required for the Hb level to recover to the precollection level was 6 weeks (case 1), 4 weeks (case 2), 1 week (case 3), and 7 weeks (case 4). These are relatively short recovery periods, if we consider the large amount of blood collected, the scale of the surgery, and the considerable preoperative blood loss. Although the preoperative administration of rHuEPO may have a negative influence on postoperative anemia,1 no conclusions can be drawn from the results of our study, because we did not use a control group in whom large amounts of autologous blood were collected without r-HuEPO. There have been no previous reports, other than our own,17 on the transfusion of preoperatively deposited blood for sacral resection of pheochromocytoma. In case 4 in the present study, there were no sudden increases in blood pressure intraoperatively.
Conclusion We performed sacrectomy in six patients, using preoperatively collected autologous blood. In four of these patients, the sacrectomy was performed without homologous blood, despite the massive hemorrhage.
S. Nakai et al.: Autologous blood transfusion in sacral tumor resection
References 14. 1. Adamson JW, Spivak JL. Physiologic basis for the pharmacologic use of recombinant human erythropoietin in surgery and cancer treatment. Surgery 1994;115:7–15. 2. Berdel WE, Oberberg D, Reufi B, et al. Studies on the role of recombinant human erythropoietin in the growth regulation of human nonhematopoietic tumor cells in vitro. Ann Hematol 1991;63:5–8. 3. Blumberg N, Heal JM. Transfusion and recipient immune function. Arch Pathol Lab Med 1989;113:246–53. 4. Burrows L, Tartter P. Effect of blood transfusions on colonic malignancy recurrence rate. Lancet 1982;II:662. 5. Chan AC, Blumgart LH, Wuest DL, et al. Use of preoperative autologous blood donation in liver resections for colorectal metastases. Am J Surg 1998;175:461–5. 6. Foster RS Jr, Costanza MC, Foster JC, et al. Adverse relationship between blood transfusions and survival after colectomy for colon cancer. Cancer 1985;55:1195–201. 7. Fujimoto J, Okamoto E, Yamanaka N, et al. Efficacy of autotransfusion in hepatectomy for hepatocellular carcinoma. Arch Surg 1993;128:1065–9. 8. Gantt CL. Red blood cells for cancer patients. Lancet 1981;II: 363. 9. Goodnough LT, Rudnick S, Price TH, et al. Increased preoperative collection of autologous blood with recombinant human erythropoietin therapy. N Engl J Med 1989;321:1163–8. 10. Hyman NH, Foster RS Jr, DeMeules JE, et al. Blood transfusion and survival after lung cancer resection. Am J Surg 1985;149:502– 7. 11. Jensen LS, Hokland M, Nielsen HJ. A randomized controlled study of the effect of bedside leucocyte depression on the immunosuppressive effect of whole blood transfusion in patients undergoing elective colorectal surgery. Br J Surg 1996;83:973–7. 12. Kaneda M, Horimi T, Ninomiya M, et al. Adverse effect of blood transfusions on survival of patients with gastric cancer. Transfusion 1987;27:375–7. 13. Kase H. Host immunosuppression after blood transfusion at the operation for gastric cancer and an attempt of autogenous blood
15.
16.
17. 18. 19.
20.
21.
22.
23.
24.
25.
327
transfusion. Nihon Rinshougeka Igakkai Zasshi (J Jap Surg Assoc) 1993;54:2211–8 (in Japanese). Madura JA. Use of erythropoietin and parenteral iron dextran in a severely anemic Jehovah’s Witness with colon cancer. Arch Surg 1993;128:1168–70. Matsumoto K, Hukuda S, Ishizawa M, et al. Use of preoperative autologous blood donations and erythropoietin for treatment of giant cell tumor of the ischium. Clin Orthop 326 1996;246–9. Mezrow CK, Bergstein I, Tartter PI. Postoperative infections following autologous and homologous blood transfusions. Transfusion 1992;32:27–30. Nakai S, Yoshizawa H. Two cases of sacrectomy using autogenous blood transfusion. J Jpn Scoliosis Soc 1998;13:114–9. Riccardi VM. Neurofibromatosis: clinical heterogeneity. Curr Probl Cancer 1982;7:2–34. Sakamaki T, Yanagimoto S, Homma T, Seki M. Short term preoperative deposit of autologous blood by subcutaneous administration of recombinant erythropoietin. Hip Joint 1993;19:376–9 (in Japanese). Shaheen SP II, Lichtiger B. Autologous deposit of blood by oncology patients: an evaluation of actual use and wastage. Ann Surg Oncol 1995;2:457–61. Sundaresan N, Rosenthal DI, Schiller AL, et al. Chordomas. In: Sundaresan N, Schmidek HH, Schiller AL, Rosenthal DI, editors. Tumors of the spine: diagnosis and clinical management. Philadelphia: WB Saunders; 1990. p. 192–213. Tartter PI, Burrows L, Papatestas AE, et al. Perioperative blood transfusion has prognostic significance for breast cancer. Surgery 1985;97:225–9. Tasaki T, Tohyama Y, Hashimoto C, et al. Autologous blood donation in terms of serum erythropoietic concentration. Jpn J Transfusion Med 1992;38:40–5 (in Japanese). Voogt PJ, van de Velde CJH, Brand A, et al. Perioperative blood transfusion and cancer prognosis. Different effects of blood transfusion on prognosis of colon and breast cancer patients. Cancer 1987;59:836–43. Yamawaki T, Tanaka H, Takeuchi S, et al. Autologous blood transfusion using recombinant human erythropoietin in radical hysterectomy. Asia-Oceania J Obstet Gynaecol 1994;20:147– 53.
Role of autologous blood transfusion in sacral tumor resection: patient selection and recovery after surgery and blood donation Sadaaki Nakai, Hidezo Yoshizawa, Shigeru Kobayashi, Kumi Naga, and Hirofusa Ichinose Department of Orthopedic Surgery, Fujita Health University School of Medicine, Toyoake, Aichi 470-1192, Japan
Abstract We carried out sacral en-bloc resection in six patients (three with chordoma; one with pheochromocytoma; one with malignant schwannoma; and one with giant cell tumor) using preoperatively collected autologous blood, to avoid homologous blood transfusion. An average of 3200 ml was collected preoperatively, with patients receiving recombinant human erythropoietin (r-HuEPO), at a total dose of 130 000 units on average. In four patients, we were able to accomplish the surgery without homologous blood transfusion. Postoperatively, the hemoglobin level in these four patients recovered to the pre-collective level in 4.5 weeks, on average. These clinical results indicate that en-bloc sacrectomy, which requires a large volume of blood transfusion, can be accomplished with preoperatively collected autologous blood alone. Key words Sacral tumor · Malignant tumor · Sacrectomy · Autologous blood transfusion · Recombinant human erythropoietin
Introduction Tumors of the sacrum tend to expand anteriorly into the pelvic cavity and are often discovered late, so the lesions are usually quite large at diagnosis. Because reoperation for local recurrence of these tumors is extremely difficult, the initial procedure should be radical and curative. In addition, reconstruction of the sacroiliac joint with an implant is necessary when the first and second sacral vertebrae are removed. For such surgeries, the transfusion of a large volume of blood is necessary. Where possible, we perform this surgery using preoperatively donated autologous blood, especially because homologous transfusion generally has an
Offprint requests to: S. Nakai Received: July 2, 1999 / Accepted: December 14, 1999
immunosuppressive effect.3,11 In addition, we wish to avoid other risks associated with blood transfusion, including hepatitis, acquired immunodeficiency syndrome, and graft-versus-host disease. While reducing blood loss is the best way to avoid or minimize the need for homologous transfusion during or after surgery, the use of autologous blood is another alternative. Autologous blood transfusion is the safest form of transfusion in patients undergoing elective surgery, including oncologic procedures.20 A large volume of autologous blood can be obtained by collecting and storing blood preoperatively, but this may take a considerable length of time. In general, surgery should be performed without delay in patients who have bone tumors for which resection is indicated. In patients with slow-growing tumors,21 however, surgery can be delayed to allow for autologous blood donation. For patients with fast-growing tumors who require prompt surgery, as much autologous blood as possible must be collected during the time required to analyze the results of diagnostic imaging and biopsies prior to the operation. We studied six patients whose surgery was performed under hypotensive anesthesia to reduce intraoperative blood loss.
Patients, methods, and results Sacral resection has been performed on six patients at our institution to date (Table 1 and Table 2). Three patients were treated by single-stage, combined anteroposterior surgery (cases 1, 5, and 6). In one patient, surgery was carried out as a two-stage procedure, comprising tumor resection in the first stage and instrumentation with bone grafting in the second (case 4). Two other patients were managed via the posterior approach alone (cases 2 and 3). For storage of autologous blood, three different methods were used. Citratephosphate-dextrose (CPD) was used for storage of
55/F 72/M
55/M
29/F
39/F
2 3
4
5
6
Malignant schwannoma GCT
Pheochromocytoma
Chordoma Chordoma
Chordoma
Diagnosis
4 5
5 6
2000 2400
3200 3600 4000 4000
Quantity of autologous blood collected (ml)
Approach
48 000 90 000
144 000 144 000 192 000 162 000
r-HuEPO (units)
13.1 11.2
12.5 12.6 12.8 13.3
Surgery
10.6 10.3
11.4 10.9 10.8 10.7
Lowest preoperative Hb level (g/dl)
First Resection left sacrum; medial part of left ilium Second, Internal fixation Resection left sacrum; medial part of left ilium Resection S1–S5; medial parts of both ilia; internal fixation
Resection S2–S5 Resection S3–S5
Resection S2–S5
Hb level before initial blood collection (g/dl)
Anteroposterior
Anteroposterior
Two-stage anteroposterior
Posterior Posterior
Anteroposterior
rHuEPO, Recombinant human erythropoietin; Hb, hemoglobin
8 10 11 11
1 2 3 4
Case no.
Period of blood collection (weeks)
Table 2. Summary of patient data on blood transfusion
GCT, Giant cell tumor of bone
53/M
Age (years)/ sex
1
Case no.
Table 1. Summary of patient data on diagnosis, surgery, and postoperative course Follow-up
3883 4564 2466 3703 (First surgery) 1095 (Second surgery) 4212 6841
Deep infection
Superficial infection
Complication
7 Homologous transfusion Homologous transfusion
6 4 1
Period required for postoperative recovery of Hb level (weeks)
Died (7 months after surgery) 4 Years
2 Years Died (1 month after surgery) 4 Years
4 Years
Perioperative blood loss (ml)
21 h 5 min
2 h 39 min 16 h 17 min
13 h 8 min
14 h 3 h 59 min
12 h 23 min
Operative time
322 S. Nakai et al.: Autologous blood transfusion in sacral tumor resection
S. Nakai et al.: Autologous blood transfusion in sacral tumor resection
blood for less than 3 weeks, mannitol-adeninephosphate (MAP) for storage of blood for 3 to 6 weeks, and freezing was used for storage of blood for longer than 6 weeks. Case 1 Case 1 was a 53-year-old man diagnosed with a sacral chordoma. Sacral resection was indicated, because bone scans showed no abnormal tracer accumulation, suggesting the absence of metastasis. It was decided that autologous blood would be sufficient to meet the needs of intraoperative and postoperative transfusion. A volume of 400 ml of blood was collected at each session, based on the patient’s weight of 61 kg. Blood collection was started 8 weeks before surgery. A total of 3200 ml of autologous blood was collected, of which 2000 ml was frozen and 1200 ml was stored using MAP. During the 8-week period in which blood was collected, a total of 24 doses, of 6000 units each, of recombinant human erythropoietin (r-HuEPO) were administered by intravenous injection three times per week (total dose, 144 000 units). The hemoglobin (Hb) level was 12.5 g/dl before the initiation of collection and decreased to 11.4 g/dl during the collection period. However, on the day before surgery, the Hb level was 12.4 g/dl (Fig. 1). After embolization of the internal iliac artery on the day before surgery, the sacrum was resected en bloc from S2 to S5 via both the anterior and posterior approaches in a single operation. The operating time
323
was 12 h and 23 min. Intraoperative and postoperative blood loss totaled 3883 ml, comprising 3521 ml during surgery and 362 ml via a drain over the 2-day period after surgery. A volume of 2400 ml of autologous blood was transfused intraoperatively, and 800 ml was transfused over the 2-day period after surgery. The Hb decreased to 7.7 g/dl on the day after surgery, but recovered to 12.5 g/dl by 6 weeks after the operation. The use of homologous blood was avoided. The patient developed a superficial wound infection postoperatively, but this healed with conservative treatment. Four years after the operation, there has been no recurrence of tumor. Case 2 Case 2 was a 55-year-old woman with a sacral chordoma. Surgery was scheduled to allow time for autologous blood donation. Blood collection was started 10 weeks before surgery, with 400 ml being obtained at each session, based on the patient’s body weight of 56 kg. A total of 3600 ml of blood was collected preoperatively, of which 1600 ml was frozen, 400 ml was stored using MAP, and 1600 ml was stored using CPD. During the blood collection period, six doses, each of 24 000 units, of r-HuEPO were administered subcutaneously (total dose, 144 000 units). The Hb level was 12.6 g/dl before blood collection was started, decreased to 10.9 g/dl during the collection period, and increased again to 11.7 g/dl before the final collection on the day before surgery (Fig. 2).
Fig. 1. Hemoglobin (Hb) values from the start of blood collection until after surgery in case 1. Blood collection was started 8 weeks before surgery, and 3200 ml of autologous blood was obtained. Hb recovered to the level before collection by 6 weeks (ws.) after surgery. Large arrow, Day of surgery; small arrows, days on which 400 ml of blood was collected
Fig. 2. Hemoglobin (Hb) values from the start of blood collection until after surgery in case 2. Blood collection was started 10 weeks before surgery, and 3600 ml of autologous blood was obtained. Hb recovered to the level before collection by 6 weeks (ws.) after surgery. Large arrow, Day of surgery; small arrows, days on which 400 ml of blood was collected
324
S. Nakai et al.: Autologous blood transfusion in sacral tumor resection
Fig. 3. Hemoglobin (Hb) values from the start of blood collection until after surgery in case 3. Blood collection was started 10 weeks before surgery, and 4000 ml of autologous blood was obtained. Hb recovered to the level before
collection by 1 week (w.) after surgery, because the patient lost only 2466 ml of blood (1931 ml intraoperatively and 535 ml postoperatively.) Large arrow, Day of surgery; small arrows, days on which 400 ml of blood was collected
After embolization of the internal iliac artery on the day before surgery, the patient underwent en-bloc resection of the sacrum from S2 through S5, via the posterior approach. The operating time was 14 h. Intraoperative and postoperative blood loss totaled 4564 ml, comprising 3454 ml during surgery and 1110 ml over the 2-day period after surgery. A volume of 3200 ml of autologous blood was transfused intraoperatively, and 400 ml was transfused on the day after surgery. The Hb in this patient decreased to 7.1 g/dl on the day after surgery, but after 4 weeks, it had recovered to 12.8 g/dl. The use of homologous blood was avoided. Two years after the operation, there has been no recurrence of the tumor.
A total of 2000 ml of autologous blood was transfused during surgery, 800 ml was transfused on the day after surgery, and 400 ml was transfused on each of postoperative days 2, 4, and 7. The Hb level in this patient was 10.5 g/dl on the day after surgery, the lowest value observed postoperatively. By 1 week after surgery, the Hb had increased to 13.4 g/dl (Fig. 3). Transfusion of homologous blood was not required. The patient died 1 month after surgery because of respiratory distress syndrome and diffuse intravascular coagulopathy.
Case 3 Case 3 was a 72-year-old man who was hospitalized with a sacral chordoma. Surgery was scheduled to allow time for autologous blood donation. Blood collection was started 11 weeks before the operation, and 400 ml of blood was obtained at each session, based on the patient’s body weight of 66 kg. A total of 4000 ml of blood was collected, of which 2400 ml was frozen, 1200 ml was stored using MAP, and 400 ml was stored using CPD. During the blood collection period, 24 000 units of r-HuEPO was administered subcutaneously once a week for 10 weeks (total dose, 192 000 units, because, on two occasions, blood was collected without using r-HuEPO). The Hb level was 12.8 g/dl before blood collection was started, fell to 10.8 g/dl during the collection period, and recovered to 13.3 g/dl on the day before the operation. After embolization of the internal iliac artery, the patient underwent en-bloc resection of the sacrum from S2 through S5 via the posterior approach. The operating time was 3 h 59 min. Intraoperative and postoperative blood loss totaled 2466 ml, comprising 1931 ml during surgery and 535 ml over the 2-day period after surgery.
Case 4 Case 4 was a 55-year-old man in whom a solitary metastatic lesion was detected in the sacrum prior to resection of the left kidney and adrenal gland for pheochromocytoma. Because the lesion was confined to the left half of the sacrum, it was decided to resect this half of the sacrum and the medial portion of the left ilium, followed by reconstruction with instrumentation and bone grafting. Autologous blood donation was started 11 weeks before surgery. A volume of 400 ml of blood was collected at each session, based on the patient’s weight of 65 kg. A total of 4000 ml of blood was collected before surgery, of which 2400 ml was frozen and 1600 ml was stored using MAP. This patient received a total of 27 doses of 6000 units of rHuEPO given by intravenous injection three times per week (total dose, 162 000 units). The Hb level was 13.3 g/dl before the start of blood collection and decreased to 10.7 g/dl on the day before surgery, the minimum value observed during the blood collection period. Embolization of the internal iliac artery was performed on the day before surgery. In the first operation, the left half of the sacrum and the medial portion of the left ilium were resected en bloc via both the anterior and posterior approaches. The operating time was 13 h and 8 min. Intraoperative and posto-
S. Nakai et al.: Autologous blood transfusion in sacral tumor resection
325
Fig 4. Hemoglobin (Hb) values from the start of blood collection until after surgery in case 4. Blood collection was started 11 weeks before surgery, and 4000 ml of autologous blood was obtained. Hb recovered to virtually the pre-
collection level by 7 weeks (ws.) after the first operation (1st op.). Large arrows, Days of surgery; small arrows, days on which 400 ml of blood was collected
perative blood loss totaled 3703 ml, comprising 3393 ml during surgery and 310 ml over the 3-day period after surgery. A total of 3200 ml of autologous blood was transfused during surgery, and an additional 400 ml was transfused postoperatively when the Hb value fell to 6.8 g/dl 2 days after the procedure (the lowest level seen postoperatively). The second operation was performed 2 weeks after the first operation, when the Hb had recovered to 11.3 g/dl. In the second operation, internal fixation, using an implant, and bone grafting were performed to obtain bone union between L5 and the left ilium. The operating time was 2 h and 39 min. During surgery, 510 ml of blood was lost, and 400 ml of autologous blood was transfused. A further 585 ml of blood drained postoperatively, for a total blood loss of 1095 ml. The nadir was 8.2 g/dl 2 days after the second operation, and homologous blood transfusion was successfully avoided. Seven weeks after the first operation, the Hb had increased to 12.9 g/dl (Fig. 4). Four years after the two operations, there has been no recurrence of tumor. This patient’s preoperative blood adrenaline level was 621 pg/ml (normal range, less than 100 pg/ml), while the noradrenaline level was 2882 pg/ml (normal range, 100 to 450 pg/ml), and dopamine level was 20 pg/ml (normal range, less than 20 pg/ml). Therefore, the autologous blood transfused during surgery may have contained high levels of adrenaline, noradrenaline, and dopamine. However, the intraoperative levels ranged from 143 to 272 pg/ml for adrenaline, 531 to 979 pg/ml for noradrenaline, and 38 to less than 5 pg/ml for dopamine. In addition, there were no sudden increases in blood pressure intraoperatively.
resonance imaging. Autologous blood donation was started immediately after admission. Although her body weight was 42 kg, 400 ml of blood was successfully obtained at each session. A diagnosis of malignant schwannoma was made, based on the results of biopsy performed under local anesthesia. During the blood collection period, eight doses of 6000 units of r-HuEPO were administered intravenously over the 4-week period before surgery (total dose, 48 000 units). A total of 2000 ml of blood was collected, of which 800 ml was stored using MAP, and 1200 ml was stored using CPD. The Hb level was 13.1 g/dl at the start of blood collection and fell to a minimum of 10.6 g/dl during the blood collection period. After embolization of the internal iliac artery, the left side of the sacrum and the medial portion of the left ilium were resected en bloc, with internal fixation, using an implant and bone grafting to obtain union between L4, L5, and the left ilium. The operating time was 16 h and 17 min. A total of 4212 ml of blood was lost during surgery. A total of 1600 ml of homologous blood was transfused, in addition to the 2000 ml of autologous blood collected preoperatively. There was no recurrence at the sacral site, but the patient died of a massive abdominal tumor 7 months after surgery.
Case 5 Case 5 was a 29-year-old woman with hereditary neurofibromatosis type I,18 characterized by generalized cafe-au-lait spots and neurofibromas. She was referred after the detection of a sacral tumor on magnetic
Case 6 Case 6 was a 39-year-old woman referred for resection of a sacral tumor. A giant cell tumor of bone was diagnosed, based on the results of a biopsy performed under local anesthesia. Autologous blood was collected for 5 weeks before surgery, during which time 15 doses of 6000 units of r-HuEPO were administered intravenously (total dose, 90 000 units). A total of 2400 ml of autologous blood was collected, of which 1200 ml was stored using MAP, and 1200 ml was stored using CPD. The Hb level was 11.2 g/dl before blood donation, and fell to a minimum of 10.3 g/dl during the blood collection period.
326
S. Nakai et al.: Autologous blood transfusion in sacral tumor resection
After embolization of the internal iliac artery, the entire sacrum and the medial segments of both ilia were resected en bloc in one operation via both the anterior and posterior approaches. Internal fixation with an implant and bone grafting were performed at the same time to obtain union between L4, L5, and the ilia. The operating time was 21 h and 5 min. A total of 5241 ml of blood was lost during surgery, requiring the transfusion of 2000 ml of homologous blood, in addition to the 2400 ml of autologous blood deposited preoperatively. A total of 1600 ml of blood was lost over the 8-day period after surgery. Her Hb decreased to 7.5 g/dl 3 days postoperatively but later recovered. She developed a deep wound infection 11 days after the surgery, which was treated by curettage and continuous irrigation. However, the infection failed to resolve, and a second operation was required. The patient was subsequently followed for 4 years, without recurrence of the tumor.
Discussion Homologous blood transfusion during renal transplantation has been reported to have an immunosuppressive effect,8 and homologous transfusion during resection of malignant tumors also suppresses immunity13,16 and may facilitate tumor recurrence or growth.4,6,10,12,22,24 So although one cannot draw firm conclusions concerning the contribution of homologous transfusion to outcome in patients with malignancy, the risk that homologous blood transfusion may have a deleterious effect on immune function makes it advisable to avoid homologous transfusion as much as possible. The indications for the preoperative use of donated autologous blood transfusion in patients with metastatic tumor remain poorly defined. In preoperatively collected autologous blood, tumor cells are potential contaminants. Still, the level of contamination in collected blood does not seem to exceed the blood level of tumor cells at the time of surgery. Clinical studies of patients who received autologous transfusion of preoperatively donated blood for hepatic resection of metastasis colorectal cancer did not show worse postoperative mortality than patients who received homologous transfusions.5 Autologous blood transfusion can be justified in those patients with metastatic tumor in whom the primary tumor has been removed and a good prognosis is expected. Extensive sacral resection is indicated for benign tumors with marked invasion of the sacrum and for malignant tumors confined to the sacrum. If resection is minimized for fear of causing neurologic deficits, further surgery may be required because of tumor recurrence, subsequently resulting in a worse neurologic outcome. In addition, reoperation is extremely difficult and more damaging to the patient.
In a randomized study, the administration of rHuEPO has been shown to increase the amount of autologous blood that could be collected before surgery.9 Although the collection of a large volume of autologous blood without the use of r-HuEPO could be possible over time, we could not find any reports on extensive short-term blood collection without the use of r-HuEPO. Little is known about the interaction between r-HuEPO and nonhematopoietic malignant cells. R-HuEPO showed no stimulation of clonal growth in solid tumor cell lines.2 There have been reports of autologous blood transfusion using r-HuEPO in the surgical treatment of solid malignant tumors, including hepatocellular, colon, and uterine cervical cancers. Neither enhancement of tumor growth by r-HuEPO nor any interaction between tumor and r-HuEPO has been documented in these reports.7,14,15,25 There have been reports that it is possible to deposit a large amount of blood within a short period by giving patients r-HuEPO,19 and this method appeared useful for patients who required the storage of a large blood volume in a brief time. In our patients, 400-ml volumes of blood were repeatedly collected at 1-week intervals, but the recovery of their Hb levels suggested that blood could have been collected more frequently. Resolution of postoperative anemia may be delayed by the preoperative administration of high doses of r-HuEPO, because endogenous EPO secretion is suppressed by exogenous r-HuEPO.23 Our experience showed that the period required for the Hb level to recover to the precollection level was 6 weeks (case 1), 4 weeks (case 2), 1 week (case 3), and 7 weeks (case 4). These are relatively short recovery periods, if we consider the large amount of blood collected, the scale of the surgery, and the considerable preoperative blood loss. Although the preoperative administration of rHuEPO may have a negative influence on postoperative anemia,1 no conclusions can be drawn from the results of our study, because we did not use a control group in whom large amounts of autologous blood were collected without r-HuEPO. There have been no previous reports, other than our own,17 on the transfusion of preoperatively deposited blood for sacral resection of pheochromocytoma. In case 4 in the present study, there were no sudden increases in blood pressure intraoperatively.
Conclusion We performed sacrectomy in six patients, using preoperatively collected autologous blood. In four of these patients, the sacrectomy was performed without homologous blood, despite the massive hemorrhage.
S. Nakai et al.: Autologous blood transfusion in sacral tumor resection
References 14. 1. Adamson JW, Spivak JL. Physiologic basis for the pharmacologic use of recombinant human erythropoietin in surgery and cancer treatment. Surgery 1994;115:7–15. 2. Berdel WE, Oberberg D, Reufi B, et al. Studies on the role of recombinant human erythropoietin in the growth regulation of human nonhematopoietic tumor cells in vitro. Ann Hematol 1991;63:5–8. 3. Blumberg N, Heal JM. Transfusion and recipient immune function. Arch Pathol Lab Med 1989;113:246–53. 4. Burrows L, Tartter P. Effect of blood transfusions on colonic malignancy recurrence rate. Lancet 1982;II:662. 5. Chan AC, Blumgart LH, Wuest DL, et al. Use of preoperative autologous blood donation in liver resections for colorectal metastases. Am J Surg 1998;175:461–5. 6. Foster RS Jr, Costanza MC, Foster JC, et al. Adverse relationship between blood transfusions and survival after colectomy for colon cancer. Cancer 1985;55:1195–201. 7. Fujimoto J, Okamoto E, Yamanaka N, et al. Efficacy of autotransfusion in hepatectomy for hepatocellular carcinoma. Arch Surg 1993;128:1065–9. 8. Gantt CL. Red blood cells for cancer patients. Lancet 1981;II: 363. 9. Goodnough LT, Rudnick S, Price TH, et al. Increased preoperative collection of autologous blood with recombinant human erythropoietin therapy. N Engl J Med 1989;321:1163–8. 10. Hyman NH, Foster RS Jr, DeMeules JE, et al. Blood transfusion and survival after lung cancer resection. Am J Surg 1985;149:502– 7. 11. Jensen LS, Hokland M, Nielsen HJ. A randomized controlled study of the effect of bedside leucocyte depression on the immunosuppressive effect of whole blood transfusion in patients undergoing elective colorectal surgery. Br J Surg 1996;83:973–7. 12. Kaneda M, Horimi T, Ninomiya M, et al. Adverse effect of blood transfusions on survival of patients with gastric cancer. Transfusion 1987;27:375–7. 13. Kase H. Host immunosuppression after blood transfusion at the operation for gastric cancer and an attempt of autogenous blood
15.
16.
17. 18. 19.
20.
21.
22.
23.
24.
25.
327
transfusion. Nihon Rinshougeka Igakkai Zasshi (J Jap Surg Assoc) 1993;54:2211–8 (in Japanese). Madura JA. Use of erythropoietin and parenteral iron dextran in a severely anemic Jehovah’s Witness with colon cancer. Arch Surg 1993;128:1168–70. Matsumoto K, Hukuda S, Ishizawa M, et al. Use of preoperative autologous blood donations and erythropoietin for treatment of giant cell tumor of the ischium. Clin Orthop 326 1996;246–9. Mezrow CK, Bergstein I, Tartter PI. Postoperative infections following autologous and homologous blood transfusions. Transfusion 1992;32:27–30. Nakai S, Yoshizawa H. Two cases of sacrectomy using autogenous blood transfusion. J Jpn Scoliosis Soc 1998;13:114–9. Riccardi VM. Neurofibromatosis: clinical heterogeneity. Curr Probl Cancer 1982;7:2–34. Sakamaki T, Yanagimoto S, Homma T, Seki M. Short term preoperative deposit of autologous blood by subcutaneous administration of recombinant erythropoietin. Hip Joint 1993;19:376–9 (in Japanese). Shaheen SP II, Lichtiger B. Autologous deposit of blood by oncology patients: an evaluation of actual use and wastage. Ann Surg Oncol 1995;2:457–61. Sundaresan N, Rosenthal DI, Schiller AL, et al. Chordomas. In: Sundaresan N, Schmidek HH, Schiller AL, Rosenthal DI, editors. Tumors of the spine: diagnosis and clinical management. Philadelphia: WB Saunders; 1990. p. 192–213. Tartter PI, Burrows L, Papatestas AE, et al. Perioperative blood transfusion has prognostic significance for breast cancer. Surgery 1985;97:225–9. Tasaki T, Tohyama Y, Hashimoto C, et al. Autologous blood donation in terms of serum erythropoietic concentration. Jpn J Transfusion Med 1992;38:40–5 (in Japanese). Voogt PJ, van de Velde CJH, Brand A, et al. Perioperative blood transfusion and cancer prognosis. Different effects of blood transfusion on prognosis of colon and breast cancer patients. Cancer 1987;59:836–43. Yamawaki T, Tanaka H, Takeuchi S, et al. Autologous blood transfusion using recombinant human erythropoietin in radical hysterectomy. Asia-Oceania J Obstet Gynaecol 1994;20:147– 53.