| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2000;69:1070-1075
© 2000 The Society of Thoracic Surgeons
a Department of Anesthesiology, Duke University Medical Center, Durham, North Carolina, USA
b Department of Surgery, Duke University Medical Center, Durham, North Carolina, USA
Address reprint requests to Dr Hill, Department of Anesthesiology, Duke University Medical Center, Box 3094, Durham, NC 27710
e-mail: hill0012{at}mc.duke.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and methodsResultsCommentReferences |
|---|
Methods. Outcome data from 2,862 CABG patients were merged with intraoperative physiologic data. A preoperative mortality risk index was calculated for each patient. Variables found significant (p < 0.05) by univariate logistic regression were tested in a multiple variable model to determine risk-adjusted association with mortality.
Results. Overall mortality rate was 1.85%. The preoperative risk index was significantly associated with mortality (p = 0.0001). No significant association was present between mortality and intraoperative variables. Preexisting hypertension was an independent predictor of mortality after controlling for risk index and bypass duration.
Conclusions. Preexisting hypertension proved to be an independent predictor of mortality in our patient population. This study found no evidence to support the hypothesis that mean arterial pressure less than 50 mm Hg, lower hematocrit, or elevated glucose while on bypass increases in-hospital mortality.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentReferences |
|---|
While most of the risk stratification variables are based on unalterable preoperative characteristics, there are a number of intraoperative physiologic variables that might influence operative morbidity and mortality. Hematocrit, mean arterial pressure (MAP), and glucose concentration are all controlled and easily manipulated during cardiopulmonary bypass (CPB). However, evidence regarding the optimal range of these variables based on outcome models is absent. Although a recent retrospective study suggested lower hematocrit is associated with greater mortality [4], the majority of evidence from anecdotal reports suggests hematocrit levels as low as 10% are acceptable [5, 6]. Similar to hematocrit, the role that MAP plays in the determination of perioperative morbidity and mortality is controversial and has been debated for years, creating uncertainty regarding the regulation of perfusion pressure [7–9]. Although animal data show correlation between higher glucose levels and extent of damage in neurologic injury [10, 11], contribution of hyperglycemia to cardiac surgical mortality in humans has never been documented.
Intraoperative physiologic data acquisition systems permit automated recording of the many physiologic variables measured during cardiac surgery [12]. These data, when merged with clinical outcome data, allow us to determine the effect of physiologic variables on major outcome variables, such as mortality and stroke. The purpose of our two-part study was to determine if hematocrit, MAP, glucose concentration, or CPB duration influence risk-adjusted in-hospital mortality (Part I) or adverse neurologic outcome (Part II).
|
Material and methods |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentReferences |
|---|
|
Determination of intraoperative variables
Anesthesia personnel at the Duke Heart Center have used an automated intraoperative information system (Synergy, Inc, San Diego, CA) during all cardiac surgical procedures performed since 1988 [12]. Data from this automated intraoperative record were downloaded to an anesthesia information network [13]. A query of this database identified the intraoperative variables of interest for the patients in the study. Specifically, we searched for the following five variables: (1) MINHCT, defined as the lowest recorded hematocrit of those sampled at 15-minute intervals during CPB; (2) MAXGLC, defined as the highest recorded glucose concentration during the surgical procedure; (3) MAP <50, defined as the integrated area below a MAP of 50 mm Hg at each minute during CPB; (4) MAP >50, defined as the integrated area greater than or equal to a MAP of 50 mm Hg at each minute during CPB; and (5) CPBTIME, defined as the CPB duration in minutes.
A MAP less than 50 mm Hg during CPB has historically been considered to represent hypotension [14] and falls below the low MAP group investigated in the paper by Gold and associates [7]. The MAP <50 was computed as the area between an upper boundary of 50 mm Hg and the measured MAP over time. This calculation thus represents a measure of hypotension that takes into account how low the MAP was and how long it remained below 50 mm Hg during CPB. If MAP never fell below 50 mm Hg for a particular patient, the MAP <50 was zero. The MAP >50 was computed as the area greater than or equal to a lower boundary of 50 mm Hg and the measured MAP over time. This calculation represents a measure of normal to high pressure, which takes into account how high the MAP was and how long it remained above 50 mm Hg during CPB.
Determination of outcome
Physiologic data from the intraoperative database inquiry were merged with outcome data of the cardiac surgical patients maintained in the Duke Cardiovascular Database [15] using hospital record number and surgery date as unique identifiers. From this database, patients were identified who died before hospital discharge. All causes of death were identified.
Operative management
During the study interval, the intraoperative management of patients undergoing cardiac surgery did not change significantly. After oral methadone or diazepam premedication, general anesthesia was induced with midazolam, fentanyl, and/or thiopental. Anesthesia was maintained with midazolam, fentanyl, and isoflurane. Most patients were monitored with electrocardiogram, ST-segment trending, radial arterial pressure, and Swan-Ganz catheter measurements. Patients underwent standard nonpulsatile hypothermic (28 to 34°C) cardiopulmonary bypass with a Cobe CML oxygenator (Cobe, Inc, Arvada, CO) and a crystalloid prime. Arterial line filters were used. Porcine heparin was administered as a bolus of 300 U/kg and supplemented as necessary during CPB to maintain a celite-activated clotting time of more than 450 seconds (Hemochron 801; International Technidyne Corp, Edison, NJ). Alpha stat blood gas management was used exclusively in regulating pO2, pCO2, and pH within normal limits. Insulin was administered for hyperglycemia at the discretion of the anesthesiologist in an attempt to keep the serum glucose less than 300 mg/dL (One-Touch Glucometer; Lifescan, Inc, Milpitas, CA). Hematocrits were determined from blood gas analysis (IL 1400 Series Blood Gas Analyzer; Instrumentation Laboratories, Lexington, MA). Low hematocrits (
20%) were confirmed by capillary tube centrifugation. Hematocrit and glucose were measured before anesthesia induction, before CPB, every 30 minutes during CPB, and after separation from CPB. These values were entered into the anesthesia information system immediately upon determination. Decisions to transfuse were based upon clinical evidence of inadequate oxygen delivery. No institutional transfusion trigger was in use during the period of the study.
MAP was automatically entered at 1-minute intervals. MAP was manipulated with phenylephrine, isoflurane, and sodium nitroprusside at the discretion of the anesthesiologist. Frequently, anesthesia personnel attempted to maintain MAP between 50 and 90 mm Hg. However, no rigid protocol exists as to a target MAP during CPB, and some individual practitioners elected not to use pressor medication at all while on bypass. A bladder temperature of greater than or equal to 36°C was required for separation from CPB. After CPB, packed red blood cells, crystalloid, and colloid solutions were administered to optimize intravascular volume and hematocrit. Before transport to the intensive care unit, temperature was maintained greater than 35°C by convective warming and arterial blood pressures were maintained less than 140/90 mm Hg using vasoactive drugs.
Statistical methodology
Statistical analysis was performed using SAS system, version 6.12 (SAS Inc, Cary, NC). Relationships between continuous variables were investigated using correlations. Analysis of mortality data was done with logistic regression. Hannan scores were determined for the entire study population. Univariate logistic regressions were done to test the association between mortality and the preoperative predictor variables of Hannan score and HTN as well as the intraoperative predictor variables of MINHCT, MAXGLC, CPBTIME, and MAP during CPB. MAP during bypass was considered in two ways: area less than 50 mm Hg (MAP <50) and area greater than or equal to 50 mm Hg (MAP >50). Variables found significant by univariate analysis were tested in a model controlling for Hannan score. Finally, potential covariates were identified and added to the model in order to identify important predictors of mortality.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentReferences |
|---|
Hannan scores ranged from 0.001 to 0.476, with a mean of 0.028 and standard deviation (SD) of 0.03. These scores represent a predicted mortality rate as a decimal (0.01 = 1% mortality risk). The incidence of HTN in the study population was 68.7%. MAXGLC ranged from 79 to 653 mg/dL, with a mean of 230 mg/dL and SD of 69.0 mg/dL. MAP <50 ranged from 0 to -2,629 mm Hg/min with a mean of -131 mm Hg/min and SD of 154. MAP >50 ranged from 8 to 6,966 mm Hg/min, with a mean of 1,185 mm Hg/min and SD of 762. MINHCT ranged from 10% to 39% with a mean of 19.0% and SD of 3.7%. CPBTIME ranged from 19 to 313 minutes with a mean of 107 minutes and SD of 34.3.
Univariate analysis did not show a significant association between mortality and MAXGLC or MAP <50. MINHCT did show a relationship with mortality but the trend did not reach statistical significance (p = 0.056, c-index = 0.565). The association between Hannan score and mortality was highly significant in the univariate analysis (p = 0.0001, c-index = 0.696). MAP >50, CPBTIME, and HTN also displayed a significant association with mortality in the univariate model (Table 2).
|
In the multivariate model predicting mortality, the potential predictor variables of Hannan score, HTN, CPBTIME, and MAP >50 were included. While Hannan score, HTN, and CPBTIME retained significance, MAP >50 became completely insignificant (p = 0.9679) in this model (Table 3).
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentReferences |
|---|
Fang and associates examined the relationship between minimum bypass hematocrit and 30-day mortality [4]. While they found a significant relationship between the minimum hematocrit and 30-day mortality, the statistical methodology was very different from ours. For a logistic model to be accurate and stable, it has been recommended that the number of variables examined be no more than 10% of the number of outcomes [16]. For the 71 deaths in their study group, Fang and coauthors should have examined a maximum of seven variables. Instead, they included 31 variables in their analysis. The analysis also included multiple cut-point analyses with multiple samplings of the data set introducing the possibility of type I statistical error. Type I statistical error represents the likelihood of finding a statistically significant p value when no significant difference exists. If only one comparison is made, a p value of 0.05 indicates only a 5% chance of type I error. If more comparisons are made, type I error increases markedly.
In our attempt to minimize the problem of overfitting the model, the Hannan score (which contains 15 variables previously found to confer risk) was used as a single independent variable. The Hannan score accounts for multiple preoperative risk factors without adding each risk factor individually to the regression analysis. The Hannan score was initially tested in a univariate analysis to confirm its suitability for the multivariate model. When combined as the only two variables in the multivariate logistic regression model, minimum hematocrit and Hannan score were used as covariates for 52 outcomes, avoiding the statistical problems of earlier studies. Using this statistical methodology, minimum hematocrit lost any trend toward significant association with mortality.
In a recent study by DeFoe and associates [17] an inverse association between hematocrit during CPB and mortality was identified. While minimum hematocrit did not retain significance as an independent predictor of mortality in our study, Hannan score and hematocrit showed highly significant inverse correlation with each other. This suggests that low hematocrit may be a surrogate marker for other risk factors included in the Hannan score and could well explain the positive findings of hematocrit and mortality by Fang and DeFoe. Preoperative factors in the DeFoe study, which were significantly associated with lower hematocrit on CPB, were female gender, older age, smaller body surface area, increased number of comorbid conditions, and increased acuity at the time of surgery. Many of these factors are part of the Hannan score and would explain why the significant association between minimum hematocrit and mortality vanished in our study when controlling for Hannan score. The patient’s preoperative status likely impacts upon initial and subsequent hematocrits during the operative procedure. Whether low hematocrit on CPB contributes to increased perioperative mortality for CABG surgery or rather is merely a marker for other risk factors that increase mortality can only be answered by a prospective, randomized clinical trial with sufficient power to adequately limit both alpha and beta error.
Preoperative history of hypertension is not included as one of the variables in the Hannan score. However, because the baseline blood pressure is likely to influence intraoperative MAP or the clinician’s management of blood pressure, it was included as an independent predictor variable in this study. Interestingly, a preoperative history of hypertension proved to be a significant predictor of in-hospital mortality in this patient population, even when controlling for other predictor variables in the multivariate analysis. Not only did hypertension remain a significant predictor when controlling for Hannan score, the predictive value of the model, as measured by c-index, improved when hypertension was added to the model along with Hannan score. Both findings argue that in this study population, preoperative history of hypertension is an independent risk factor for increased mortality in coronary artery bypass patients. Why hypertension did not prove to be a significant predictor variable in the Hannan study [1] or a significant risk factor in the Society of Thoracic Surgeons Database [3] is unknown. We speculate that chronic hypertension may be a contributor to atherosclerotic plaque formation throughout the body, and therefore contribute to increased perioperative mortality.
Consistent with prior studies [18–20], we found a highly significant association between CPB duration and mortality. This finding is not surprising when considering that complicated cases with higher procedural risk such as repeat sternotomy or patients with poor target vessels require longer bypass times. Patients who fail to wean from bypass or must be returned to bypass for complications would also have longer bypass duration and would be expected to have a higher mortality.
Controversy exists over the influence intraoperative blood pressure has on perioperative mortality after coronary artery bypass grafting. To date, no studies have conclusively identified a relationship between MAP during CPB and mortality. Gold and associates have presented data suggesting that combined cardiac and neurologic morbidity is decreased by maintaining higher MAP than customarily done during CPB [7]. However, this study found no statistically significant difference in mortality between the treatment (high MAP) and the control (low MAP). Our study addresses a possible relationship between MAP and mortality in a large patient population with sufficient power to conclude MAP <50 mm Hg during CPB is not associated with risk-adjusted in-hospital mortality. Given our sample size, we had 70% power to detect an odds ratio of 1.5, or 99% power to determine an odds ratio of 2.0. An odds ratio of less than 1.5 would not be considered clinically significant.
We also investigated a possible relationship between a measure of normal to high MAP (
50 mm Hg) and mortality. Interestingly, MAP >50 did demonstrate a statistically significant association with increased mortality in the univariate analysis and after, controlling for Hannan score, although the association disappeared when we added duration of cardiopulmonary bypass and preoperative history of hypertension to the multivariate model. Unlike MAP >50, Hannan score and HTN remain significant predictors of mortality when controlling for length of cardiopulmonary bypass. The explanation for this finding involves the method of calculating MAP >50. Because MAP >50 is a function of the height of MAP elevation as well as the number of minutes during which the blood pressure remains 50 mm Hg, length of CPB will alter MAP >50 in a direct fashion. Therefore, MAP >50 becomes essentially a marker for bypass time in prolonged cases, which will have an elevated mortality. Only if MAP >50 had retained a significant association with mortality after controlling for CPB time could any conclusions be drawn from this study about elevated MAP on bypass and mortality.
Conversely, conclusions can be drawn about low MAP on bypass and mortality. Prolonged CPB would also tend to increase the absolute value of MAP <50 (negative number), which is a function of duration and extent of MAP depression less than 50 mm Hg. In prolonged and difficult cases, the absolute value of MAP less than 50 would tend to increase due to hemodynamic instability and time period available, during which the MAP could drop less than 50 mm Hg. Therefore, lack of a significant relationship between MAP less than 50 and outcome in this study argues strongly that perfusion pressure less than 50 mm Hg during CPB in the presence of adequate flow does not contribute to in-hospital mortality.
Human data conflict regarding the influence hyperglycemia has on the progression of cerebral infarction [21–24]. There have been no studies suggesting that intraoperative hyperglycemia affects mortality during CABG surgery. While the presence of long-standing hyperglycemia in diabetic patients is a component of the Hannan score and does confer risk of increased mortality during coronary artery bypass surgery, our data demonstrate no correlation between maximum glucose concentration during CPB and in-hospital mortality for patients undergoing this procedure. This result is likely due to the differing physiologic impacts of chronic hyperglycemia compared with transient intraoperative hyperglycemia. Chronic diabetes produces diffuse microvascular coronary artery disease, which is less amenable to coronary artery bypass grafting as well as diffuse vascular disease, which raises the risk of multiple organ dysfunction in the perioperative period. Transient hyperglycemia in a patient with adequate intravascular volume, adequate circulation, and carefully controlled pH during CPB is unlikely to have the same clinical impact as chronic hyperglycemia.
There are several limitations to this study. First, this study is observational and we did not randomize patients to high or low physiologic variables. However, our anesthesia information system receives physiologic data in real time without investigator bias. Outcome events were entered into the surgical database as they occurred. Second, the question posed by this study was conceived after data had been entered, which prevented tighter control of the intraoperative variables. While the data reflect actual practice behaviors without manipulation bias, the variables ranged widely increasing the possibility of missing correlations that are actually significant. Finally, the outcome of in-hospital mortality is an uncommon occurrence in this patient group. This increases the risk of error developing as a result of a small number of missed outcomes. Fortunately, mortality is strictly followed in the postoperative cardiac surgery database and accurately reflects the discrete outcome variable of in-hospital mortality at the Duke Heart Center.
In summary, we failed to demonstrate a significant relationship between controllable intraoperative variables and in-hospital mortality. Hannan score did accurately predict in-hospital mortality. Preoperative history of clinical hypertension also predicted mortality independent of Hannan score. An association between minimum hematocrit and mortality did not achieve statistical significance in the multivariate model, but did show statistically significant inverse correlation with Hannan score, which suggests that the association of low hematocrit with comorbid disease produced the trend toward significance in the univariate analysis. A prospective study examining minimum hematocrit while accounting for the interrelationship between preoperative patient status and hematocrit is necessary to determine if minimum hematocrit alone is a significant predictor of mortality. In spite of earlier studies, which suggest that high MAP during CPB improves outcome, this study finds no evidence to support the hypothesis that low MAP during CPB increases in-hospital mortality. Regrettably, we found no association between easily controlled physiologic variables and risk-adjusted in-hospital mortality.
|
References |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentReferences |
|---|
Related Article
Ann. Thorac. Surg. 2000 69: 1075-1076.
This article has been cited by other articles:
|
|
 |
|
  J. W. Hammon Extracorporeal Circulation: Perfusion System Card. Surg. Adult, January 1, 2008; 3(2008): 350 - 370. [Full Text]
|
|
|
|
 |
|
 L. A. Fleisher, J. A. Beckman, K. A. Brown, H. Calkins, E. L. Chaikof, K. E. Fleischmann, W. K. Freeman, J. B. Froehlich, E. K. Kasper, J. R. Kersten, et al. ACC/AHA 2007 Guidelines on Perioperative Cardiovascular Evaluation and Care for Noncardiac Surgery: A Report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Revise the 2002 Guidelines on Perioperative Cardiovascular Evaluation for Noncardiac Surgery) Developed in Collaboration With the American Society of Echocardiography, American Society of Nuclear Cardiology, Heart Rhythm Society, Society of Cardiovascular Anesthesiologists, Society for Cardiovascular Angiography and Interventions, Society for Vascular Medicine and Biology, and Society for Vascular Surgery J. Am. Coll. Cardiol., October 23, 2007; 50(17): e159 - e242. [Full Text] [PDF]
|
|
|
|
 |
|
 L. A. Fleisher, J. A. Beckman, K. A. Brown, H. Calkins, E. L. Chaikof, K. E. Fleischmann, W. K. Freeman, J. B. Froehlich, E. K. Kasper, J. R. Kersten, et al. ACC/AHA 2007 Guidelines on Perioperative Cardiovascular Evaluation and Care for Noncardiac Surgery: A Report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Revise the 2002 Guidelines on Perioperative Cardiovascular Evaluation for Noncardiac Surgery) Circulation, October 23, 2007; 116(17): e418 - e500. [Full Text] [PDF]
|
|
|
|
 |
|
 The Society of Thoracic Surgeons Blood Conservatio, V. A. Ferraris, S. P. Ferraris, S. P. Saha, E. A. Hessel II, C. K. Haan, B. D. Royston, C. R. Bridges, R. S.D. Higgins, G. Despotis, et al. Perioperative Blood Transfusion and Blood Conservation in Cardiac Surgery: The Society of Thoracic Surgeons and The Society of Cardiovascular Anesthesiologists Clinical Practice Guideline Ann. Thorac. Surg., May 1, 2007; 83(5_Supplement): S27 - S86. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. J. Murphy and G. D. Angelini Indications for Blood Transfusion in Cardiac Surgery Ann. Thorac. Surg., December 1, 2006; 82(6): 2323 - 2334. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. W. Hogue Jr, C. A. Palin, and J. E. Arrowsmith Cardiopulmonary bypass management and neurologic outcomes: an evidence-based appraisal of current practices. Anesth. Analg., July 1, 2006; 103(1): 21 - 37. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Swaminathan, B. G. Phillips-Bute, P. J. Conlon, P. K. Smith, M. F. Newman, and M. Stafford-Smith The association of lowest hematocrit during cardiopulmonary bypass with acute renal injury after coronary artery bypass surgery Ann. Thorac. Surg., September 1, 2003; 76(3): 784 - 791. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S Al-Ruzzeh, G Asimakopoulos, G Ambler, R Omar, R Hasan, B Fabri, A El-Gamel, A DeSouza, V Zamvar, S Griffin, et al. Validation of four different risk stratification systems in patients undergoing off-pump coronary artery bypass surgery: a UK multicentre analysis of 2223 patients Heart, April 1, 2003; 89(4): 432 - 435. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. A. Hessel II and L. H. Edmunds Jr. Extracorporeal Circulation: Perfusion Systems Card. Surg. Adult, January 1, 2003; 2(2003): 317 - 338. [Full Text]
|
|
|
|
|
|
I. J. Welsby, E. Bennett-Guerrero, D. Atwell, W. D. White, M. F. Newman, P. K. Smith, and M. G. Mythen The Association of Complication Type with Mortality and Prolonged Stay After Cardiac Surgery with Cardiopulmonary Bypass Anesth. Analg., May 1, 2002; 94(5): 1072 - 1078. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. K. van Wermeskerken, J.-W. H. Lardenoye, S. E. Hill, H. P. Grocott, B. Phillips-Bute, P. K. Smith, J. G. Reves, and M. F. Newman Intraoperative physiologic variables and outcome in cardiac surgery: part II. Neurologic outcome Ann. Thorac. Surg., April 1, 2000; 69(4): 1077 - 1083. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
