Anesth Analg 2003;97:958-963
© 2003 International Anesthesia Research Society
CARDIOVASCULAR ANESTHESIA
Blood Use in Patients Undergoing Coronary Artery Bypass Surgery: Impact Of Cardiopulmonary Bypass Pump, Hematocrit, Gender, Age, and Body Weight
Bharathi H. Scott, MD*,
Frank C. Seifert, MD
,
Peter S. A. Glass, MBChB*, and
Roger Grimson, PhD*
Departments of *Anesthesiology and
Surgery, State University of New York at Stony Brook, Stony Brook, New York
Address correspondence and reprint requests to Bharathi H. Scott, MD, Department of Anesthesiology, SUNY at Stony Brook, Health Sciences Center, L4–060, Stony Brook, NY 11794–8480. Address email to bscott{at}anesthes.sunysb.edu
 |
Abstract
|
|---|
We investigated the impact of cardiopulmonary bypass pump (CPB), hematocrit, gender, age, and body weight on blood use in patients undergoing coronary artery bypass graft surgery at a major university hospital. Participants were 1235 consecutive patients undergoing primary coronary artery surgery over a period of 2 yr (1999 and 2000); 681 patients underwent coronary surgery with use of CPB, and 554 patients underwent off-pump coronary artery bypass surgery using a median sternotomy incision. There were 881 males and 354 females. Average packed red blood cells (PRBC) transfusion for patients on CPB was 3.4 U compared with 1.6 U for the off-pump group (P = <0.001). Patients on CPB received more frequent PRBC transfusion (72.5%) compared with 45.7% of off-pump patients (P = <0.001). Average PRBC transfusion for males was 2.2 U compared with 3.6 U for females (P = <0.001). A lower percentage of males (52.6%) than females (79.4%) received transfusion (P = <0.001). The impact of CPB, off-pump status, preoperative hematocrit <35%, gender, age 
65 yr, and weight 
83 kilograms using median values as cut points, on blood use was examined using logistic regression models. Use of CPB, preoperative hematocrit, (<35%) female gender, increasing age, and decreased body weight were significant predictors of transfusion (P = <0.001). Preoperative hematocrit <35% and use of CPB were the strongest predictors of PRBC transfusion.
IMPLICATIONS: We examined the impact of cardiopulmonary bypass, preoperative hematocrit, gender, age, and body weight on blood use in patients undergoing primary coronary artery bypass surgery at a tertiary care institution. We found that all five of these variables are significant predictors of blood use in patients undergoing coronary artery bypass surgery.
|
Introduction
|
|---|
Approximately 571,000 coronary artery bypass procedures are performed annually in the United States (1). Blood transfusion is common in patients undergoing coronary artery bypass (CABG) surgery with the use of a cardiopulmonary bypass pump (CBP). Despite this, there is relatively scant information regarding blood and blood component use in patients undergoing coronary artery surgery. Until a few years ago, almost all CABG surgery was performed with the use of a CBP. In the last few years there is a resurgence of interest in performing coronary artery surgery, popularly referred to as "beating heart surgery," without the use of CBP. In our institution we began performing off-pump coronary artery bypass (OPCAB) surgery in 1997. A database was created to capture data for all cardiac surgical procedures. Caretakers involved in managing these patients had an impression that there were differences in blood use among patients undergoing cardiac surgery with and without the use of a CPB pump. The goal of this study was to query our database to determine the impact of CPB, preoperative hematocrit, gender, age, and body weight on blood use in patients undergoing CABG surgery over a period of 2 yr (1999, 2000).
|
Methods
|
|---|
This study included all consecutive patients undergoing primary coronary artery surgery at our institution over a period of 2 yr (1999 and 2000). IRB approval was obtained for this study. There were 1235 patients. Six-hundred-eighty-one patients underwent CABG surgery with the use of CPB. Five-hundred-fifty-four patients underwent OPCAB. Both groups had a median sternotomy incision. Reoperations and coronary artery surgery via anterior and lateral thoracotomy approach were excluded. Aspirin use was discontinued in elective patients 7 days before surgery. Approximately 20% of coronary surgery is elective in our institution. Patients who were in the hospital the night before surgery were premedicated with combinations of morphine and scopolamine as deemed appropriate by the anesthesiologist. Intraoperative monitors included standard ASA monitors along with our standard monitors for patients undergoing coronary artery surgery. This included 5-lead electrocardiograms (ECG) with continuous automated ST segment analysis, a radial artery catheter for continuous arterial pressure determination, and a pulmonary artery catheter. All patients received our standard fast-track anesthetic induction using a combination of fentanyl 5–10 µg/kg and midazolam 0.05 to 0.1 mg/kg; rocuronium 1 mg/kg or vecuronium 0.1 mg/kg was used for neuromuscular blockade. Anesthesia was maintained with isoflurane, with supplements of propofol as indicated. Patients undergoing on-pump CABG received heparin 300 U/kg before institution of cardiopulmonary bypass and activated coagulation time (ACT) was maintained above 400 s. Patients undergoing OPCAB were given heparin 100–200 U/kg to maintain ACT above 350 s. Both groups received additional heparin as deemed necessary to maintain ACT at these values. Patients undergoing on-pump CABG were cannulated with an aortic cannula and a single two-stage venous cannula. Both retrograde and antegrade intermittent cold (6°C) blood cardioplegia solution was administered for myocardial protection. Systemic temperature was allowed to drift to 30–32°C during CBP. Hollow fiber membrane oxygenators (Dideco, Modena, Italy) were used in all patients. Our pump prime consisted of 1500 mL of Isolyte solution that was reduced to 900 mL after aortic cannulation and retro-priming in all patients. All distal and proximal anastomoses were constructed using a single cross-clamp technique. Cardiotomy suction was used after adequate anticoagulation for blood salvage. The protocol used for blood transfusion during the period of this study was as follows. During CPB, packed red blood cells (PRBC) were infused for hematocrits less than 22%. Blood was also administered if hematocrits were <25% in patients with good ventricles (ejection fraction>35%) and between 25–30% in compromised ventricles (ejection fraction<35%) at the termination of CBP. The end-points of PRBC transfusion was to maintain hematocrits between 25% and 30% in "healthy" patients with good ventricular function and stable hemodynamics and more than 30% in patients with compromised hemodynamics and ventricular function in both on- and off-pump groups. After weaning from CPB all the blood in the bypass circuit was salvaged and reinfused as tolerated by patients based on the hemodynamics.
In patients undergoing OPCAB, special off-pump coronary retractors and stabilizers (Guidant, Indianapolis, IN) were used during coronary artery anastomosis. After anticoagulation, suction was used to salvage blood into a chest tube drainage system (Pleur-evac plus, Deknatel); this blood was reinfused in the operating room at the end of the procedure as appropriate based on hemodynamics. All on-pump patients received 
aminocaproic acid during surgery. No antifibrinolytics were used in OPCAB patients.
All demographic, clinical and surgical data were prospectively collected on standardized data collection forms as required by New York State cardiac surgical reporting system. This included age, gender, weight, body surface area (BSA), use of pump, number of CABGs, comorbid conditions such as diabetes, hypertension, and renal failure (creatinine >2.5 mg/dL), use of intraaortic balloon pump, and left ventricular ejection fraction and postoperative bleeding requiring reoperation, for predicting New York state risk adjusted mortality rate (NYSRAMR) (2). We also collected data on pre- and postoperative hematocrits. Data pertaining to total units of PRBC, fresh-frozen plasma (FFP) and platelets transfused were obtained from our blood bank database. Preoperative hematocrit is defined as the first hematocrit obtained in the operating room before surgery. Postoperative hematocrit is defined as the last hematocrit obtained before discharge from the intensive care unit after surgery. These were obtained from anesthesia and intensive care records respectively. All the data were entered into our computerized CABG database.
Univariate analysis of the data was conducted using Pearson 
2 analysis, Wilcoxon’s ranked sum test, and Spearman’s correlation methods. Means and confidence intervals are used to describe demographic and clinical data. The data are summarized by means and 95% confidence intervals. P values <0.05 are considered significant. Multivariate logistic models and the Mantel-Haenszel categorical data methodology are used in multivariate analysis involving multiple factors hypothesized to be associated with occurrence of transfusion. Interaction terms were also examined. Goodness-of-fit statistics were applied. The variables preoperative hematocrit, pump, gender, age, weight, and BSA are used in the analyses, but binary counterparts were used in some of the reported analysis for reasons of comparability and meaningfulness in the computation of odds ratios. The "cut points" for the binary variables are the median points. The median age of 65 yr is the cutoff point. The "older patients" are 65 yr of age and are considered the risk group. Median weight is 83 kilograms. Weight 83 kilograms is the risk group. Median initial hematocrit is 35% and the patients with hematocrits <35 are the risk group. The data were analyzed using SPSS 11.0 software package (SPSS, Chicago, IL).
|
Results
|
|---|
There were 1235 patients included from the database. Six-hundred-eighty-one patients underwent conventional coronary surgery with use of CPB and 554 patients underwent OPCAB surgery. All the demographic and transfusion data for on- and off-pump patients are displayed in Table 1. No significant differences were noted in demographic variables, with the exception of postoperative hematocrit. Average blood transfusion in the on-pump group was significantly more compared with the OPCAB group (P 0.001). Blood use data for on- and off-pump patients are presented in Figure 1.
There were 881 males and 354 female patients. Demographic and transfusion data for male and female patients are presented in Table 2. The females were significantly older and weighed less than males (P 0.001). The incidence of diabetes, hypertension, and NYSRAMR was significantly more frequent for females (P 0.001). The average preoperative hematocrits were significantly lower in females (P 0.001). There was no significant difference between males and females in the incidence of bleeding requiring reexploration. Average blood transfusion for females was significantly more than males (P 0.001). A greater percentage of females were transfused (79.4%) compared with males (52.6%) (P 0.001), and a greater percentage of females in the on-pump group (93.6%) received transfusion compared to the off-pump group (63.3%).
Table 3 displays the risk factors for transfusion. The model fit is good according to the Hosmer-Lemeshow goodness of fit test; P = 0.721 with a 79% correct classification. The greatest risk factors for transfusion are being on pump and having a low (<35%) preoperative hematocrit. The adjusted odds ratio for hematocrit is 15.5 (P 0.001), indicating that the odds of receiving a transfusion is several times larger for patients with low hematocrit readings than for those with larger readings. This odds ratio is essentially reduced by the pump- hematocrit interaction term, which has a negative regression coefficient. The percentage of patients with high hematocrit readings who received transfusion is 40.5% and the percent with low readings who received transfusion is 87.3%. Therefore the unadjusted rate ratio is 2.16. In this case the ratio of odds is much larger than ratio of rates; the two would necessarily be close if the dependent variable (in this case, transfusion) were a rare event. The adjusted odds ratio for on-pump is 7.91 (P 0.001), indicating an increased risk of receiving a transfusion for the on-pump patients than for the off-pump patients. The unadjusted transfusion rates are 72.5% for the on-pump group and 45.7% for the off-pump patients. Other risk factors are female gender, low body weight, and older age, with odds ratios of 2.37, 2.18, and 1.99, respectively. The unadjusted odds ratios are displayed in Table 3. The rate ratios are 1.50, 1.53, and 1.42, respectively.
All possible two-way interactions were examined. The pump-gender interaction nearly entered the model (P = 0.054) but its presence significantly worsened the fit. The negative pump-hematocrit interaction term (P = 0.006), was reflected in the model by an adjusted odds ratio of <1.00. These results indicate that when hematocrit is low, the risk of transfusion from being on-pump decreases slightly. Alternatively, when a patient is on-pump, the risk of transfusion from a low hematocrit slightly drops. This seemingly contradictory finding was independently confirmed by a Mantel-Haenszel stratified categorical analysis (Table 4) and may be explained by the observation that in the high-risk hematocrit group 92.2% of the on-pump patients and 81.2% of the off-pump patients received transfusions. These two rates are large and close together, yielding a rate ratio of 1.14 and an odds ratio of 2.73. However in the low-risk hematocrit group (35%), 57.6% of the on-pump and 19.7% of the off-pump patients received blood transfusion. These rates are smaller and farther apart than the comparable on-pump rates, yielding larger rate ratios. The rate ratio of 2.92 and the odds ratio of 5.55 are larger than those of the high-risk hematocrit group. The stratified analysis depicting the two highest risk factors, pump and hematocrit, and their interaction is presented in Table 4. A test for homogeneity revealed that odds ratios within each pair in Table 4, are significantly different from each other (P 0.05).
The interaction term allows for the good fit and is properly interpreted as a small effect modifier. It does not detract from the main effects of the model, which indicate that being on-pump or having a low preoperative hematocrit impart relatively high risks for transfusion.
|
Discussion
|
|---|
Bleeding and transfusion of blood and blood components are frequent complications of cardiac surgery among patients undergoing CABG surgery with the use of a CPB pump (3). The incidence of blood transfusion among patients undergoing coronary artery surgery has been reported to vary between 27% and 92% (4). More recently it has been reported to be 29.4% (5). There is a significant variation among physicians and institutions despite national guidelines for blood transfusion, especially in patients with ischemic heart disease (6,7,8). The goal of our investigation was to examine the impact of five major variables (CPB use, preoperative hematocrit, gender, age, and body weight) on blood use in patients undergoing CABG surgery at our institution over a period of 2 years. We found that on-pump (versus off-pump), preoperative hematocrit <35%, female gender, weight 83 kilograms, and age 65 years were significant predictors of blood administration for patients undergoing coronary artery surgery. Preoperative hematocrit is one of the predictors of blood transfusion in previous reports (5,9,10). We found it to be the most significant predictor of blood transfusion. Use of a CPB pump was the next significant predictor of blood transfusion.
Most literature examining blood use in coronary artery surgery deals with patients undergoing CABG surgery with CPB (4,5,6,9). Early reports of decreased blood usage in "beating heart surgery" are encouraging but not completely resolved (11–15). Ours is the only study examining the impact of CPB on blood usage in such a large group of patients.
Our data show that on-pump CABG is associated with significantly more blood and blood component transfusion compared with OPCAB. Interestingly our results show that, when the hematocrit is low (<35%), the risk of transfusion from being on-pump is less than when the hematocrit is >35%.
The use of a CPB is a risk factor for PRBC transfusion. This is in agreement with early reports examining blood use in smaller groups of patients undergoing OPCAB surgery (11–14,16). As both groups in our report were comparable in demographic and clinical features, we believe that CPB and its associated side effects were responsible for the differences in blood transfusion between groups. These effects are primarily the hemodilution effect of extracorporeal circuit on blood volume and, second, the adverse effects of CPB on red blood cells and platelets and coagulation factors (17,18). A potential for excessive bleeding exists in all patients undergoing cardiac surgery with the use of a CPB. CPB is an unphysiologic state, with its associated adverse effects of hemolysis, fibrinolysis, decrease in number and function of circulating platelets, decrease in fibrinogen, plasminogen, and coagulation factors V and VIII. It is also well established that complement is activated during CPB (17,18). In this study, postoperative bleeding requiring reexploration was slightly more frequent in the on-pump group, although it did not reach statistical significance. The number of patients transfused combined with the hematocrits observed at the end of surgery would imply greater blood loss in patients undergoing CABG with CPB. Previous reports are contradictory regarding blood loss in on- and off-pump groups (11,15,16); FFP and platelet use was significantly more in on-pump patients. This is in agreement with an earlier report that increased plasma and platelet use follows the increased use of PRBC (19).
The next significant predictor of transfusion we examined for CABG was gender. Female gender has been identified as a risk factor for blood transfusion in patients undergoing CABG surgery. This is noted in both on- and off-pump surgery (5,10,19,20). Our results are in agreement with previous reports. Preoperative anemia and smaller body size have been identified as predictors of blood transfusion (5,9). Our female patients had lower preoperative hematocrits and they were smaller than their male cohorts. The dilutional effects of CPB circuit along with smaller body weight and the associated smaller blood volume have been implicated for this difference in blood transfusion (19). Our females had a significantly more frequent incidence of comorbidities, leading to a significantly increased adjusted mortality rate (NYSDOHRAM). This gender difference is noted in noncardiac surgery where the hemodilution effect of CPB is not present (21). We examined the association of preoperative hematocrit and age with gender and CPB on the probability of transfusion in both males and females. The probability of transfusion increases with a low preoperative hematocrit and the use of a CPB pump, older age, and less body weight in cardiac surgery. These data are in agreement with previous reports examining risk factors for transfusion (5,9,10) and are important because the majority of patients undergoing coronary artery surgery are older than 65 years (1).
The economic impact of blood and blood component transfusion in cardiac surgery is important because a significant portion of all blood transfused in the US is associated with cardiac surgery (3). Approximately 13.9 million units of blood are donated and 4.5 million patients receive transfusions annually (National Blood Data Resource Center). Approximately 34,000 U of red blood cells are needed everyday in the US. The charge for processing 1 U of red blood cells for administration is approximately $500 in our institution (2002 charges). This includes a fixed charge for typing and screening blood that is not incurred during transfusion of subsequent units. These subsequent units are charged at approximately $400. Hence, a patient undergoing CABG on-pump incurs a charge of approximately $1460 for PRBC, compared with $740 for OPCAB. A female undergoing CABG incurs a charge of $1540 for blood use as compared with $980 for a male. It is apparent that blood and blood component transfusion has a significant impact on health care costs.
The limitations of this report are that it is a retrospective analysis and the majority of our patients undergo CABG during the same hospital admission for cardiac catheterization. This is considered a risk factor for transfusion (19). We have not considered other variables such as the operating room time, surgeon, and anesthesiologist.
In summary, we examined the impact of CPB, preoperative hematocrit, gender, age and body weight on blood use in patients undergoing primary CABG surgery at a tertiary care institution. We found that all five of these variables are significant predictors of blood use in patients undergoing coronary artery bypass surgery. Preoperative hematocrit <35% and the use of a CPB pump were the strongest predictors. Female gender, less body weight, and older age were also predictors of blood use in patients undergoing cardiac surgery.
|
References
|
|---|
- US Department of Health Statistics. Discharge survey advance data. 2001; 319. Available at: http://www.cdc.gov/nchs/fastats/insurg.htm.
- Hannan EL, Kilburn H, Racz M, et al. Improving outcomes of coronary artery bypass graft surgery in New York state. JAMA 1994; 271: 761–6.[Abstract]
- Munoz JJ, Birkmayer NJO, Birkmayer JD, et al. Is E-Aminocaproic acid as effective as aprotinin in reducing bleeding with cardiac surgery? Circulation 1999; 99: 81–9.[Abstract/Free Full Text]
- Stover EP, Siegel IC, Parks R, et al. Variability in transfusion practice for coronary artery bypass surgery persists despite national consensus guidelines: a 24 hr institution study. Anesthesiology 1998; 88: 327–33.[ISI][Medline]
- Karkouti K, Cohen MM, Stuart A, et al. A multivariate model for predicting the need for blood transfusion in patients undergoing first time elective coronary artery bypass graft surgery. Transfusion 2001; 41: 1193–203.[ISI][Medline]
- Goodnough IT, Johnston MF, Toy PT. The variability of transfusion practice in coronary artery bypass surgery. JAMA 1991; 265: 86.[Abstract]
- Goodnough LT, Johnson MF, Ramsey G, et al. Guidelines for transfusion support in patients undergoing coronary artery bypass grafting. Ann Thorac Surg 1990; 50: 675–83.[Abstract]
- American Society of Anesthesiologists’ Task Force on Blood Component Therapy. Practice guidelines for blood component therapy. Anesthesiology. 1996; 84: 732–47.[ISI][Medline]
- Magovern JA, Sakert T, Benckart DH, et al. A model for predicting transfusion after coronary artery bypass grafting. Ann Thorac Surg 1996; 61: 27–32.[Abstract/Free Full Text]
- Shevde K, Pagala M, Kashikar A, et al. Gender is an essential determinant of blood transfusion in patients undergoing coronary artery bypass graft procedure. J Clin Anesth 2000; 12: 109–16.[Medline]
- Nader DN, Khadra WZ, Reich NT, et al. Blood product use in coronary revascularization: comparison of on and off-pump techniques. Ann Thorac Surg 1999; 68: 1640–43.[Abstract/Free Full Text]
- Cartier R. Systematic off pump coronary artery revascularization: experience of 275 cases. Ann Thorac Surg 1999; 68: 1494–7.[Abstract/Free Full Text]
- Pukas JD, Wright CE, Ronson RS, et al. Off pump multivessel coronary bypass via sternotomy is safe and effective. Ann Thorac Surg 1998; 66: 1068–72.[Abstract/Free Full Text]
- Nathoe HM, Dijk D, Jansen EWL, et al. A comparison of on-pump and off-pump coronary bypass surgery in low risk patients. N Eng J Med 2003; 348: 394–402.[Abstract/Free Full Text]
- Pearsall RA, Place DG, Cheng DCH, et al. Comparison of postoperative bleeding in off pump (OPCAB) versus traditional coronary artery bypass grafting surgery. Anesth Analg 2000; 90: SCA1-SCA95.
- Demaria RG, Carrier M, Fortier S, et al. Reduced mortality and strokes with off-pump coronary artery bypass grafting surgery in octogenarians. Circulation. 2002; 106: I-5–10.
- Kirklin JK, Westaby S, Blackstone EH, et al. Complement and damaging effects of cardiopulmonary bypass. J Thorac Cardiovasc Surg 1983; 86: 845–57.[Abstract]
- Ellison N, Jobes D. Effective hemostasis in cardiac surgery. Philadelphia: WB Saunders, 1988.
- Surgenor DM, Churchill WH, Wallace RJ, et al. Determinants of red cell, platelet, plasma, and cryoprecipitate transfusions during coronary artery bypass graft surgery: the collaborative hospital transfusion study. Transfusion 1996; 36: 521–32.[ISI][Medline]
- Scott B, Seifert FC, Glass PSA. Does gender influence resource utilization in patients undergoing off pump coronary artery bypass surgery? J Cardiothorac Vasc Anesth 2003; 17: 346–51.[Medline]
- Churchill WH, Chapman RH, Rutherford CJ, et al. Blood product utilization in hip and knee arthroplasty: effect of gender and autologous blood on transfusion practice. Vox Sang 1994; 66: 182–7.[Medline]
Accepted for publication May 30, 2003.
This article has been cited by other articles:
|
|
|
|
 
H. B. Bittner, J. Lemke, M. Lange, A. Rastan, and F. W. Mohr
The impact of aprotinin on blood loss and blood transfusion in off-pump coronary artery bypass grafting.
Ann. Thorac. Surg.,
May 1, 2008;
85(5):
1662 - 1668.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
M J ten Brinke, P W Weerwind, S Teerenstra, J C. Feron, W van der Meer, and M H. Brouwer
Leukocyte removal efficiency of cell-washed and unwashed whole blood: an in vitro study
Perfusion,
December 1, 2005;
20(6):
335 - 341.
[Abstract]
[PDF]
|
|
|
|
|
|
|
A. A. Fox and N. A. Nussmeier
Does Gender Influence the Likelihood or Types of Complications Following Cardiac Surgery?
Seminars in Cardiothoracic and Vascular Anesthesia,
December 1, 2004;
8(4):
283 - 295.
[Abstract]
[PDF]
|
|
|
Top
Abstract
Introduction
