| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2005;80:2213-2220
© 2005 The Society of Thoracic Surgeons
a Department of Cardiothoracic Anesthesia and Intensive Care, Policlinico San Donato, Milan, Italy
b Department of Cardiovascular Perfusion, Policlinico San Donato, Milan, Italy
Accepted for publication May 19, 2005.
* Address correspondence to Dr Ranucci, Cardiothoracic Anesthesia and Intensive Care Department, Policlinico S. Donato, Via Morandi 30, San Donato Milanese, Milan, 20097 Italy (Email: cardioanestesia{at}virgilio.it).
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsCommentReferences |
|---|
METHODS: One thousand forty-eight consecutive patients undergoing coronary operations have been studied. For each patient we have recorded the lowest hematocrit on cardiopulmonary bypass, the correspondent lowest oxygen delivery, and the pump flow around the time of these determinations. The three variables have been explored in a multivariable model as possible risk factors for acute renal failure and postoperative serum creatinine levels increase. The role of transfusions in determining acute renal failure was subsequently included in the model.
RESULTS: The best predictor for acute renal failure and peak postoperative serum creatinine levels was the lowest oxygen delivery, with a critical value at 272 mL·min–1 ·m–2. The lowest hematocrit was an independent risk factor with a lowest predictive value at a cutoff of 26%. When corrected for the need for transfusions, only the lowest oxygen delivery remained an independent risk factor.
CONCLUSIONS: A high degree of hemodilution during cardiopulmonary bypass is a risk factor for postoperative renal dysfunction; however, its detrimental effects may be reduced by increasing the oxygen delivery with an adequately increased pump flow.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentReferences |
|---|
Many independent risk factors have been identified in previously published articles [4–8]: age, diabetes, preexisting renal dysfunction, cardiopulmonary bypass (CPB) duration, hypotensive states during CPB or the whole perioperative period, low output state, and use of adrenergic drugs.
In recent years, a considerable amount of literature has pointed out the relationship between the lowest hematocrit (HCT) during CPB and postoperative renal dysfunction or ARF. Since 1994 we could demonstrate an association between an HCT < 25% during CPB and severe postoperative renal dysfunction in coronary patients [9]. Subsequently, Hardy and coworkers confirmed this finding in 1998 [10]. In recent years, the lowest HCT on CPB has been recognized as a risk factor for postoperative low cardiac output state and in-hospital mortality [11], for a series of different adverse outcomes including ARF [12], and in two studies dedicated to acute renal injury in coronary patients it has been recognized as an independent predictor of postoperative serum creatinine levels increase [13] and ARF [14].
Despite the general consensus in identifying the lowest HCT on CPB as a predictor of renal dysfunction after cardiac operations, the mechanism at the basis of this relationship has not yet been established. The most common interpretation is that poor oxygen availability to the renal medulla during CPB may deteriorate renal function causing ischemic or inflammatory organ injury, or both [9, 12–14]. However, the HCT value during CPB is only one of the two determinants of oxygen delivery (DO2), with the pump flow being the second.
The present study addresses the specific role of pump flow, lowest HCT, and lowest DO2 during CPB in determining postoperative ARF (primary endpoint) and serum creatinine increase (secondary endpoint) in adults undergoing coronary surgery with CPB.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentReferences |
|---|
Patient Population
One thousand forty-eight consecutive patients scheduled for surgical coronary revascularization were admitted to the study. Exclusion criteria were chronic dialytic treatment and patients receiving off-pump cardiac surgery.
Data Collection and Definitions
The following preoperative variables were collected for each patient: age (years), gender, weight (kgs), body surface area (m2), ejection fraction, HCT (%), serum creatinine value (mg/dL), unstable angina (need for intravenous nitrates), recent myocardial infarction (< 30 days), previous vascular surgery, chronic obstructive pulmonary disease, previous cerebrovascular accident, diabetes (on medication), previous cardiac surgery, and emergent operation.
The following operative variables were collected: CPB duration (min), use of centrifugal or roller pump, use of biocompatible CPB surfaces, total heparin dose (international units), lowest temperature on CPB (°C), mean perfusion pressure during CPB (mm Hg), lowest HCT on CPB (%), arterial oxygen tension (mm Hg, recorded simultaneously to lowest HCT), pump flow indexed (mL·min–1 ·m–2, mean value during 30 minutes of CPB around the time when the lowest HCT was recorded), and lowest DO2 indexed on CPB (mL·min–1 ·m–2, calculated at the time when the lowest HCT was reached).
Lowest DO2 was calculated on the basis of the pump flow indexed, arterial oxygen tension, hemoglobin value (mg/dL, derived from lowest HCT/3), and hemoglobin saturation (%), according to the equation:
|
| (1) |
Transfusional needs were assessed as the need for packed red cells during or after the operation, until the discharge from the intensive care unit (ICU).
Postoperative renal outcome of the patients was assessed during the ICU stay as: (1) primary endpoint — incidence of ARF, defined as the need for renal replacement therapy (RR-ARF), and (2) secondary endpoint — peak postoperative serum creatinine level (PPSC) (mg/dL)
Anesthesia, Surgery, and CPB Management
All patients were treated with the same anesthesia technique (totally intravenous anesthesia with remifentanil and midazolam plus cisatracurium for muscle relaxation). Cardiopulmonary bypass was established through a standard median sternotomy, aortic root cannulation, and single atrial cannulation for venous return. Body temperature during CPB was maintained between 32°C and 34°C. Antegrade intermittent cold crystalloid or cold blood cardioplegia was used according to the surgeon's preference. The circuit was primed with 700 mL of a gelatin solution (Eufusin; Bieffe Medical, Modena, Italy) and 200 mL of trihydroxymethylaminomethane solution. Roller or centrifugal pumps were used according to availability; a biocompatible treatment (phosphorylcholine coating) and a closed circuit with separation of the blood suctions was used in 43% of the patients. The pump flow was targeted between 2.0 and 2.4 L·min–1
·m–2, and the target mean arterial pressure was settled at 60 mm Hg.
All patients received tranexamic acid intraoperatively (15 mg/kg intravenously before establishing CPB, and the same dose was received after protamine), and none received aprotinin.
Anticoagulation was established with an initial dose of 300 international units per kilogram of body weight of porcine intestinal heparin injected into a central venous line 10 minutes before the initiation of CPB, and a target activated clotting time of 480 seconds; patients receiving closed and biocompatible circuits received a reduced dose of heparin with a target activated clotting time settled at 300 seconds. At the end of CPB, heparin was reversed by protamine chloride at a 1:1 ratio of the loading dose, regardless of the total heparin dosage.
Hematocrit and hemoglobin levels were measured every 20 minutes during CPB, together with arterial oxygen tension as a part of a routine arterial blood gas analysis protocol.
After the operation, all the patients were sent to the ICU. A specific perioperative transfusion algorithm was applied (ie, the patients received two packed red cells units before CPB whenever the preoperative hematocrit value was below 30%, and they received two or more packed red cells units during CPB in case of excessive hemodilution [hematocrit value below 22%]). After CPB, the patients received packed red cells in order to maintain a hematocrit value higher than 25%. This target value was raised to higher values according to the clinical condition, and namely to the hemodynamic status, the need for inotropic support, and the age of the patient. Fresh frozen plasma was not used before reaching the ICU. Platelets were usually not transfused, unless in patients reaching the operating room under full dose of ticlopidine or clopidogrel and demonstrating severe postoperative bleeding.
Statistical Analysis
All data are expressed as mean ± standard error of the mean, or absolute numbers and percentage as appropriate. A p value < 0.05 was considered significant for all the following statistical tests. The statistical analysis was performed using computer software SPSS 11.0 (SPSS; Chicago, IL).
Step 1: Univariate analyses
The association between preoperative or intraoperative variables and RR-ARF and PPSC level was first explored in a univariate model. The RR-ARF was defined as a dependent binary (0 to 1) variable, and its association with the various independent variables was explored with the Student's t test for unpaired data (continuous independent variables) and with Pearson's 
2 test (categorical independent variables). Peak postoperative serum creatinine value is a continuous dependent variable, and its association with the various independent variables was explored with the Student's t test for unpaired data (categorical binary independent variables) and a linear regression analysis (continuous independent variables).
The univariate relationship between the two independent variables (lowest HCT and lowest DO2) and the two outcome variables (RR-ARF incidence and PPSC level) has been explored and was graphically analyzed with nonlinear regression analyses based on the technique of rolling decile subgroups [12, 15]. The patient population was arranged in increasing lowest HCT and lowest DO2 order, and a total of 37 subgroups (75% overlapping ranges) were analyzed with respect to RR-ARF incidence and mean PPSC value.
Step 2. Multivariate analyses
All factors being associated with RR-ARF incidence and PPSC level were entered in a multivariable analysis. The risk for RR-ARF was assessed using a multivariable stepwise forward logistic regression model, and the PPSC value was assessed using a multivariable stepwise forward linear regression model. To avoid overfitting and collinearity in assessing the multivariate model, independent variables have been tested for intercorrelation. According to the suggested statistical approach for multivariable risk models [16], in presence of intercorrelation, the most clinically relevant (among similar) dependent variables were chosen for inclusion in the multivariable analysis.
Subsequently, the multivariable model created for the primary endpoint variable (RR-ARF incidence) was corrected, introducing the incidence of transfusions as a potential risk factor, and a multicollinearity test with tolerance and variance inflation diagnostics was performed (a tolerance value > 0.4 was considered acceptable to exclude multicollinearity).
STEP 3
A receiver operating characteristic curve was constructed, and the area under the receiver operating characteristic curve (area under the curve) was determined to assess the ability of different oxygen flow-related variables to predict RR-ARF after cardiac operations. On the basis of the cutoff values identified, the patient population was divided into four groups, and the incidence of RR-ARF was compared between different groups with a Pearson's 2 test.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentReferences |
|---|
Univariate Analyses
Demographics, preoperative, and intraoperative variables of the patient population are reported in Table 1. For each variable the statistical correlation with RR-ARF and PPSC value is reported. Patients with RR-ARF were significantly older, had a higher preoperative serum creatinine level, a higher rate of chronic obstructive pulmonary disease and diabetes, and received a longer CPB procedure. All three study variables (lowest HCT on CPB; pump flow indexed; DO2 indexed on CPB) were significantly associated with RR-ARF, with a negative regression coefficient (the lower the values, the higher the risk for RR-ARF).
|
The nonlinear relationship (p < 0.001) between lowest HCT and RR-ARF/PPSC and between lowest DO2 and RR-ARF/PPSC is reported in Figure 1, following the rolling decile approach and a nonlinear regression analysis.
|
|
|
|
2 for intergroup differences (Fig 3). There was a significant intergroup difference: group A had a significant lower incidence of RR-ARF than group C (p < 0.01) and D (p < 0.001). No difference was observed between groups A and B, and between groups C and D.
|
|
|
Comment |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentReferences |
|---|
Kidney, namely in its medullar portion, is an organ at particular risk for a low oxygen delivery status. The tubular vascularization is basically a portal type, being the peritubular capillary represented by the residual blood previously nourishing the glomerular capillaries. Moreover, the oxygen consumption of the kidney parenchimal cells is proportional to the blood flow, and the artero-venous oxygen difference is a constant, independent from the blood flow [17].
All the previous articles addressing the association between lowest HCT on CPB and postoperative renal dysfunction failed to consider the role of pump flow and therefore to measure the oxygen delivery. Of course, if the pump flow is a constant, the HCT may be considered a direct indicator of oxygen delivery; however, this is not the case. In their article, Swaminathan and coworkers [13] report a pump flow indexed between 2 and 2.4 L·min–1 ·m–2, Habib and coworkers [12] between 2.5 and 3.0 L·min–1 ·m–2, and Karkouti and coworkers [14] between 2 and 2.4 L·min–1 ·m–2; these values may justify a 20% to 30% difference in oxygen delivery for the same HCT value. Moreover it should be consider that coupling a low pump flow with a low hematocrit is well possible in clinical practice; the pump outflow is usually modulated on the basis of the perfusion pressure [12, 13], and on the basis of the patient's temperature or according to the contingent situation (ie, poor venous return). Normally HCT on pump is not a criterion for changing the pump flow, and in any case this has not been considered in the previously mentioned articles [9–13].
An interesting point is the nonlinear relationship between DO2 indexed and RR-ARF incidence; the predictive analysis for RR-ARF (receiver operating characteristic) identified a cutoff value at 272 mL·min–1 ·m–2.
In critical care patients, it has been demonstrated by many authors [18–20] that below a critical DO2 the oxygen consumption starts to decrease and becomes pathologically dependent on DO2. Below this "dysoxia threshold" the oxygen supply is insufficient to meet the oxygen demands, and therefore the metabolic needs are covered in part by the anaerobic lactacid metabolism. As a result of this condition, lactic acidosis becomes evident, and a multiple organ dysfunction may be initiated. The value of this critical DO2 value varies in different circulatory failure patterns, but with the exception of septic shock it has been identified in the range of 8 to 9 mL·min–1 ·kg–1 (600 mL/min for a patient weighing 75 kgs). This value, if indexed, corresponds to 325 mL·min–1 ·m–2, a value that is 20% higher than the one detected in our population as a predictor for RR-ARF. However, our patients had reduced metabolic needs due to the action of narcotics and to the moderately hypothermic state. Therefore, we are authorized to consider that the dysoxia threshold and the critical value for RR-ARF are probably located at similar values of DO2. A possible mechanism linking DO2 and adverse renal outcomes is that patients who have been submitted to a low DO2 during CPB (either due to a low HCT, a low pump flow, or both) experienced a period of "masked" circulatory failure, with a poor organ oxygen supply, whose consequences become visible in the early phases after recovery in the ICU. This mechanism could explain even the increased nonrenal morbidity and mortality observed in patients who have undergone CPB with a low HCT value [11, 12]. A severe limitation of this hypothesis is that we did not measure oxygen consumption during CPB in our patient, nor blood lactates during and after the operation. Therefore we are unable to demonstrate that our patients suffered from a pathologic oxygen consumption dependency and that they developed lactic acidosis during or after the operation. However, even if there is very little information on the possible existence of a critical DO2 during CPB, a previous article [21] explored the behavior of oxygen consumption, DO2, and blood lactates immediately after CPB, identifying a critical DO2 at a level of 300 mL·min–1 ·m–2, again in the range of our data.
The Transfusion Role
One of the major problems in all the previous studies was the intercorrelation between HCT, transfusional needs, and renal dysfunction. In other words, transfusions are triggered by low HCT values, and transfusions are a recognized risk factor for renal dysfunction. This intercorrelation makes it difficult to separate the effects of HCT and transfusions on RR-ARF and renal dysfunction. The various authors have tried to overcome this problem with different approaches (ie, using multivariable models [9, 10, 12] or, more correctly, introducing the transfusion variable as a corrective factor of the multivariable models with a careful statistical check of multicollinearity risk [13]). Using a similar approach, we could improve the quality of our multivariable models, and after correction for transfusional risk, only the lowest DO2 indexed on CPB remains an independent predictor of RR-ARF; pump flow indexed becomes statistically not significant, and lowest HCT on CPB maintains a border line p value of 0.06. Therefore this approach confirms the superiority of the lowest DO2 versus the lowest HCT on CPB as predictors of RR-ARF.
Practical CPB Management Considerations
The results of the present study introduce possible changes to the current CPB practice. The simple information that low levels of HCT during CPB may induce a renal damage leaves few possible renal-saving strategies. Basically the most reasonable option is to minimize the CPB circuit to reduce the intraoperative hemodilution as much as possible or to use specific techniques such as retrograde autologous priming of the CPB circuit. Of course blood transfusion before or during CPB are another possible option, but we should consider the potential renal damages induced by blood transfusions themselves. Moreover, the oxygen carrying capacity of bank blood is usually limited.
Conversely, if we consider that a correct DO2 rather than the simple HCT is our target, more options are possible. Basically, the message is that we should adapt the pump flow to the HCT, being ready to increase it in order to perfuse the patient with an oxygen delivery higher than 270 to 280 mL·min–1 ·m–2. The effectiveness of this strategy is well explained in figure 3; patients with a low DO2 have a high risk of RR-ARF regardless of their HCT value; conversely, if the DO2 is maintained above the critical threshold, the RR-ARF risk is low, again regardless of the HCT.
Of course, maintaining a high pump flow is not always possible. In case of very low venous return, it is almost impossible. In this case, we can hypothesize another possible strategy, that is the reduction of the metabolic needs by cooling the patient. However, the effectiveness of this approach is not demonstrated in our study in which all the patients were moderately hypothermic, and we must consider that a previous study dealing with renal injury at different CPB temperatures failed to demonstrate a higher risk in normothermic patients [22].
In conclusion, we think that more insights in this topic may be obtained by exploring the relationship between DO2, oxygen consumption, and blood lactates during and after coronary operations, to better establish the mechanism determining renal injury in patients perfused at critical DO2 levels, and by investigating the same relationship at different levels of temperature-induced metabolic needs.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  G. J. Murphy, H. Lin, R. J. Coward, T. Toth, R. Holmes, D. Hall, and G. D. Angelini An initial evaluation of post-cardiopulmonary bypass acute kidney injury in swine Eur. J. Cardiothorac. Surg., November 1, 2009; 36(5): 849 - 855. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. De Santo, G. Romano, A. Della Corte, V. de Simone, F. Grimaldi, M. Cotrufo, and M. de Feo Preoperative anemia in patients undergoing coronary artery bypass grafting predicts acute kidney injury J. Thorac. Cardiovasc. Surg., October 1, 2009; 138(4): 965 - 970. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Poullis and R. Poole Mathematical Modeling in Cardiac Surgery: Helping Clinical Trials Answer the Question Seminars in Cardiothoracic and Vascular Anesthesia, June 1, 2009; 13(2): 81 - 86. [Abstract] [PDF]
|
|
|
|
|
|
F. De Somer Optimal Versus Suboptimal Perfusion During Cardiopulmonary Bypass and the Inflammatory Response Seminars in Cardiothoracic and Vascular Anesthesia, June 1, 2009; 13(2): 113 - 117. [Abstract] [PDF]
|
|
|
|
 |
|
 G. S. Murphy, E. A. Hessel II, and R. C. Groom Optimal Perfusion During Cardiopulmonary Bypass: An Evidence-Based Approach Anesth. Analg., May 1, 2009; 108(5): 1394 - 1417. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H Vermeer, S Teerenstra, R. de Sevaux, H. van Swieten, and P. Weerwind The effect of hemodilution during normothermic cardiac surgery on renal physiology and function: a review Perfusion, November 1, 2008; 23(6): 329 - 338. [Abstract] [PDF]
|
|
|
|
 |
|
 D. Morrice, B. G. Loef, R. Henning, G. Navis, A. Rankin, W. van Oeveren, T. Ebels, and A. Epema Renal oxygen delivery during cardiopulmonary bypass Br. J. Anaesth., September 1, 2008; 101(3): 430 - 431. [Full Text] [PDF]
|
|
|
|
 |
|
 M. Ranucci, A. Pazzaglia, C. Bianchini, G. Bozzetti, and G. Isgro Body Size, Gender, and Transfusions as Determinants of Outcome After Coronary Operations Ann. Thorac. Surg., February 1, 2008; 85(2): 481 - 486. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Karkouti, D. N. Wijeysundera, W. S. Beattie, and for the Reducing Bleeding in Cardiac Surgery (RBC) Risk Associated With Preoperative Anemia in Cardiac Surgery: A Multicenter Cohort Study Circulation, January 29, 2008; 117(4): 478 - 484. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Ranucci Perioperative Renal Failure: Hypoperfusion During Cardiopulmonary Bypass? Seminars in Cardiothoracic and Vascular Anesthesia, December 1, 2007; 11(4): 265 - 268. [Abstract] [PDF]
|
|
|
|
|
|
F. Pappalardo, C. Corno, A. Franco, G. Giardina, A.M. Scandroglio, G. Landoni, G. Crescenzi, and A. Zangrillo Reduction of hemodilution in small adults undergoing open heart surgery: a prospective, randomized trial Perfusion, September 1, 2007; 22(5): 317 - 322. [Abstract] [PDF]
|
|
|
|
|
|
M. Ranucci, C. Bellucci, D. Conti, A. Cazzaniga, and B. Maugeri Determinants of Early Discharge From the Intensive Care Unit After Cardiac Operations Ann. Thorac. Surg., March 1, 2007; 83(3): 1089 - 1095. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Z. Ricci and C. Ronco Kidney diseases beyond nephrology--intensive care Nephrol. Dial. Transplant., March 1, 2007; 22(3): 708 - 711. [Full Text] [PDF]
|
|
|
|
 |
|
 T. Johannes, E. G. Mik, B. Nohe, K. E. Unertl, and C. Ince Acute decrease in renal microvascular PO2 during acute normovolemic hemodilution Am J Physiol Renal Physiol, February 1, 2007; 292(2): F796 - F803. [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]
|
|
|
|
|
|
M. Ranucci, G. Isgro, F. Romitti, S. Mele, B. Biagioli, and P. Giomarelli Anaerobic Metabolism During Cardiopulmonary Bypass: Predictive Value of Carbon Dioxide Derived Parameters Ann. Thorac. Surg., June 1, 2006; 81(6): 2189 - 2195. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Weinstein Invited commentary Ann. Thorac. Surg., December 1, 2005; 80(6): 2220 - 2220. [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 |
