| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Department of Cardiac Surgery, University of Catania, Catania, Italy
b Department of Cardiology, University of Chieti, Italy
Accepted for publication May 18, 2007.
* Address correspondence to Dr Calafiore, Division of Cardiac Surgery, University of Catania, Ferarrotto Hospital, Via Citelli, Catania, 95124, Italy (Email: calafiore{at}unich.it).
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and MethodsResultsCommentReferences |
|---|
Methods: From November 1994 to December 2001, 2,943 patients having multivessel surgical disease underwent myocardial revascularization. Ninety patients were excluded because of incompleteness of data, intraoperative death, or preoperative chronic dialysis. The analysis was split: one analysis included 1,724 (862 each group) of 2,618 patients with normal preoperative creatinine (<1.5 mg/dL), and the second analysis included 160 (80 each group) of 215 patients with preoperative abnormal renal function; in both analyses matched groups were selected applying propensity score.
Results: In the group with normal preoperative creatinine, the incidence of 30-day ARF was 5.4% (2.9% off-pump versus 7.9% on-pump; p < 0.001). Stepwise logistic regression confirmed that cardiopulmonary bypass was an independent variable for increased postoperative ARF incidence (odds ratio, 3.3), as well as age and reduced left ventricular ejection fraction. Receiver operating characteristic curves showed that cardiopulmonary bypass duration was a predictor of higher ARF incidence (area under the curve, 0.79) with a cutoff value of 66 minutes. In the patients with abnormal renal function preoperatively, the incidence of ARF was similar between the groups (16.3% on-pump versus 12.5% off-pump; p = 0.499). Acute renal failure had an important impact on early (odds ratio, 3.6) and late mortality (hazard ratio, 4.1).
Conclusions: Off-pump surgery plays an important renoprotective role and provides better early and late outcome in patients with normal preoperative creatinine. When the preoperative creatinine is abnormal, the surgical strategy does not seem to have any influence. The occurrence of ARF significantly impairs early and long-term mortality, and the surgical strategy does not improve outcomes.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentReferences |
|---|
The above-mentioned data justify the efforts in preventing perioperative and postoperative renal failure. As some therapeutic strategies, like dopamine infusion, mannitol, and furosemide, showed controversial results [11–13], several studies focused on the pathogenetic basis of postoperative renal dysfunction. Cardiopulmonary bypass (CPB) was demonstrated to be the main cause of glomerular and tubular dysfunction or damage [14]. In fact, the markers of glomerular function (creatinine clearance) and damage (microalbuminuria) as well as of tubular function (fractional excretion of sodium) and damage (urinary N-acetyl-ß-glucosaminidase) were found to be significantly higher in on-pump patients, during and after CPB [14, 15].
Causes of the detrimental effects of CPB on kidney are inflammatory response, nonpulsatile flow, hemodilution, renal hypoperfusion, low output syndrome, atheroembolism, increased levels of circulating catecholamines, and free hemoglobin [2, 12, 16–19]. However, the impact of CPB on postoperative renal complications is still controversial, whether preoperative renal function is normal or not [5–10, 20, 21]. These conflicting results and the limited number of large comparative studies on this particular topic caused us to analyze retrospectively our experience in patients with normal and abnormal renal function [3]. The impact of postoperative renal function impairment on early and late outcome was also investigated.
|
Material and Methods |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentReferences |
|---|
1.5 mg/dL). This cutoff value was selected as discriminatory of altered preoperative renal function according to the literature [3] and to our laboratory. These patients were included in other previous publications of ours. Use of our database was authorized by our institutional review board on October 2004. The authorization waived patient consent.
Analysis A
From the entire group of 2,618 patients 1,724 were selected by means of propensity score and sample matching to obtain two groups of 862 patients each (on-pump and off-pump coronary artery bypass graft surgery), with similar preoperative and perioperative characteristics.
Analysis B
From the entire group of 215 patients 160 were selected by means of propensity score and sample matching to obtain two groups of 80 patients each (on-pump and off-pump coronary artery bypass graft surgery), with similar preoperative and perioperative characteristics.
Patient Selection
Allocation to off-pump surgery was on the basis of the vessel size (>1.2 mm) and the absence of diffuse coronary calcifications. In the presence of mechanical or electric instability, the patient was selected preferably for on-pump surgery. The final decision was dependent on the basis and the expertise of the surgeon responsible for the operation. Patients converted to on-pump were considered in the off-pump group (intention to treat).
Surgical Technique
On-pump
Cardiopulmonary bypass was instituted by cannulation of the ascending aorta and right atrium. A standard circuit with a hollow-fiber membrane oxygenator and a roller pump was used. The body temperature was kept at 37°C. Myocardial protection was achieved by means of intermittent antegrade warm blood cardioplegia.
Off-Pump
The technique of exposure and stabilization of the target coronary vessel has already been reported [22]. In the most recent years, an apical suction was used to expose in particular the lateral and the inferior wall (Xpose, Guidant Corporation, Cupertino, CA). When the coronary artery was exposed, stabilization was achieved with a pressure (Acces Ultima System, Guidant Corporation) or suction (Axius Vacuum 2 System, Guidant Corporation) stabilizer.
Clinical Data Collection, Monitoring, and Definition
A set of perioperative data was collected for all patients undergoing myocardial revascularization at our institution. The following variables were recorded and defined. Acute renal failure was defined as postoperative blood creatinine equal or higher than 2.0 mg/dL if the preoperative value was normal (<1.5 mg/dL), or increment of at least 1 mg/dL if preoperative renal function was altered. Mortality included death from any cause during (early mortality) or after (late mortality) the first 30 days after surgery. Postoperative creatinine value was measured every postoperative day for each patient until discharge; the peak value was considered for the analysis.
Follow-Up
All the patients were followed up in our outpatient clinic 3, 6, and 12 months after surgery and thereafter at yearly intervals. The more recent information was obtained by calling the patient or the referring cardiologist. Follow-up was 100% complete and ended on June 30, 2006. Mean follow-up of survivors was 7.5 ± 1.9 years. Two hundred forty-three patients were still alive 10 years after surgery.
Statistical Analysis
Data are reported as mean ± standard deviation for continuous variables, or otherwise as a percentage. Statistical analysis comparing two groups was performed with the Mann-Whitney U test. In case of paired data, the Wilcoxon signed rank test was applied. Pearson 
2 or Fisher’s exact test was used to compare categorical variables. Stepwise logistic regression was used to realize a nonparsimonious model to calculate the propensity score [22] (the probability of being selected for on-pump given a set of preoperative risk factors already reported [23]). The model fit was evaluated using the Hosmer and Lemeshow test (analysis A: 2 = 3.22; p = 0.92; analysis B: 2 = 3.96; p = 0.86). Then, a sample matching was performed: each off-pump patient was matched with the on-pump patient having the closest propensity score.
Stepwise logistic regression was used to verify the impact of CPB on higher incidence of postoperative ARF and early mortality as well as the impact of ARF on early mortality; The results of stepwise logistic regression were reported as odds ratio (OR), 95% confidence limits (95% CL), and probability value. Receiver operating characteristic curve was used to identify the cutoff predictive value of CPB duration; area under the curve, 95% confidence limits, probability value, sensitivity and specificity were reported [24]. Actuarial survival was obtained by means of Kaplan-Meier method; Ten-year survival was reported. Statistical differences were investigated with log-rank test. Time-to-event (mortality) analysis was performed by a multivariate Cox proportional-hazard regression. The results of Cox analysis were reported as instantaneous relative risk ratio (hazard ratio), 95% CL, and probability value. The SPSS software (SPSS, Inc, Chicago, IL) was used. Probability values less than 0.05 were considered significant.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentReferences |
|---|
|
Patients with ARF showed a postoperative peak creatinine value of 2.72 ± 1.81 mg/dL, significantly higher than patients with normal postoperative renal function (1.17±0.26 mg/dL; p < 0.001). Dopamine infusion was used in 91 of 93 patients (97.8%) with postoperative ARF; in the remaining 2 patients, diuretic therapy and fluid supply were able to treat the ARF. Hemofiltration was used in 3 patients (3.2%). Complete recovery of renal function (creatinine <1.5 mg/dL) was achieved in 29 of 88 survivors (33.0%); 1 patient (1.1%) did not recover and started the dialysis on the nephrology ward. Twenty-nine patients (33.0%) were discharged to the cardiology ward with creatinine equal to or greater than 2.0 mg/dL.
Stepwise logistic regression confirmed that CPB was an independent variable for increased postoperative ARF incidence (OR, 3.3; 95% CL, 1.9 to 5.1; p < 0.001) as well as age (OR, 1.02; 95% CL, 1.004 to 1.05; p = 0.0299) and ejection fraction equal or lower than 0.35 (OR, 2.6; 95% CL, 1.5 to 6.1; p = 0.0081). Receiver operating characteristic curves showed that CPB duration played an important predictive role (area under the curve, 0.79; 95% CL, 0.72 to 0.85; p < 0.001) and identified a cutoff value of 66 minutes, above which the risk of developing postoperative ARF was higher (sensitivity and specificity, 70%; OR, 5.3; 95% CL, 3.6 to 8.0; p < 0.001).
Analysis B
Preoperative and perioperative characteristics were similar in both groups (Table 2). The incidence of ARF was 13.8% (23 patients), without significant differences between groups (16.3% on-pump versus 12.5% off-pump; p = 0.499). Mean preoperative creatinine value was similar for each group (Table 2). In the postoperative period the peak value of creatinine was 2.87 ± 1.96 mg/dL (p < 0.001 versus preoperative value); in the off-pump group creatinine rose to 2.84 ± 1.65 mg/dL (p < 0.001), whereas in the on-pump group it increased to 2.86 ± 2.22 mg/dL (p < 0.001). The percentage of increment of creatinine was similar in both groups (22.4% versus 21.5%; p = 0.873). The incidence of ARF was similar in patients having or not having lateral wall revascularization either in case of off-pump (11.7% versus 15.0%; p = 0.705) or on-pump surgery (19.7% versus 0%; p = 0.112). Patients with postoperative ARF showed a postoperative peak creatinine value of 4.30 ± 2.04 mg/dL, significantly higher than patients who did not experience ARF (2.60 ± 1.83 mg/dL; p < 0.001). Dopamine infusion was used in 14 of 23 patients (60.9%) with postoperative ARF; in the remaining 9 patients (39.1%), diuretic therapy and fluid supply were enough to treat the ARF. Hemofiltration was used in 1 patient (4.3%). Among the 22 patients who experienced ARF and survived the first postoperative month, 1 patient (4.3%) did not recover and started the dialysis on the nephrology ward. Twenty-one patients (95.5%) were discharged to the cardiology ward with a mean creatinine value of 3.38 ± 1.09 mg/dL.
|
In analysis A, early mortality was 1.6% (28 of 1,724 patients), again significantly higher in patients experiencing ARF (5.5% ARF group versus 1.4% no-ARF group; p = 0.003). Off-pump surgery provided lower early mortality (0.9% off-pump versus 2.3% on-pump; p = 0.022). Multivariate analysis confirmed that postoperative ARF (OR, 3.4; 95% CL, 1.2 to 9.2; p = 0.017) and CPB (OR, 2.3; 95% CL, 1.05 to 5.3; p = 0.041) were risk factors for early mortality.
In analysis B, 4 patients (2.5%) died within the first postoperative month; in this analysis, there were no statistical differences between patients with or without ARF (n = 1; 4.3% ARF group versus n = 2; 2.2% no-ARF group; p = 0.466) and in those undergoing on-pump or off-pump surgery (n = 3; 3.7% off-pump group versus n = 1; 1.3% on-pump group; p = 0.620).
Concerning 1,907 patients surviving the first postoperative month, ARF prolonged the postoperative hospital stay from 4.6 ± 2.9 to 6.7 ± 5.3 days (p < 0.001).
Late Outcome
Ten-year survival was 93.6% ± 0.6%, significantly higher in patients with preoperative normal creatinine value (94.5% ± 0.6%, analysis A versus 83.2% ± 3.7%, analysis B; p < 0.001) and in patients who experienced postoperative ARF (73.8% ± 5.6% versus 94.7% ± 0.5%; p < 0.001; Fig 1); a worse survival for the ARF group was found both in analysis A (80.5% ± 4.8% versus 95.3% ± 0.5%, p < 0.001) and in analysis B (54.8% ± 13.7% versus 89.1% ± 2.7%; p = 0.005). Cox analysis confirmed the impact of postoperative ARF on late outcome (Table 3).
|
|
show 10-year survival for on-pump and off-pump surgery according to the type of analysis (A or B) and the occurrence or not of postoperative ARF.
|
|
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentReferences |
|---|
Another important result from our analysis is that a CPB longer than roughly 1 hour (66 minutes) raised with good accuracy (70%) the risk of ARF (OR, 5.3). Although others stated that prolonged CPB may produce ARF [1], none of them reported any cutoff value. However, off-pump likely plays the main renoprotective role (ARF: 2.9% for off-pump versus 5.0% for CPB shorter than 66 minutes; p = 0.033; OR, 2.7; p = 0.029). Additionally, advanced age and very low ejection fraction were found to impair renal function, as already demonstrated [1, 5, 7, 9].
Again, in analysis B, the incidence of ARF (13.0%) was consistent with that reported by others (5.9% to 19.5%) [5, 6, 8]. In this case, attempts to find any impact of CPB on ARF failed. Both groups showed the same increment of creatinine value in the postoperative period. Although the small sample size could reduce the statistical power of analysis B, the same results were achieved in another propensity-based analysis comparing larger cohorts [8].
To better understand the results of our study, we should answer the following questions. Why does even off-pump surgery impair renal function? Why does CPB provide a significant higher incidence of ARF in patients with normal preoperative renal function? Why does the renoprotective role of off-pump surgery seem to disappear in patients with preoperative abnormal renal function?
The inflammatory response does not differ greatly between on-pump and off-pump groups [25, 26]. In off-pump patients with good preoperative renal reserve, the inflammatory response may cause a mild renal dysfunction, resulting more in a slight increment of postoperative creatinine than in a clinical picture of ARF. Likewise, preoperative dysfunctioning kidneys are more susceptible, and the inflammatory response may cause an important renal dysfunction or damage, even similar to that produced by CPB.
Transient circulatory failure and global hypoperfusion, occurring in case of exposure of the lateral vessels, could provide another possible explanation for postoperative renal impairment in off-pump surgery [21]. In our series, patients with preoperative normal renal function undergoing off-pump surgery showed a significantly higher rate of ARF in case of lateral wall revascularization (3.6% versus 0.7%). In patients with normal renal function, lateral wall revascularization may be performed on beating heart, being very careful to avoid transient hypotension, but when it occurs, the likelihood of ARF is lower than with on-pump surgery (3.6% versus 7.9%; p = 0.001). In case of preoperative abnormal renal function, transient circulatory failure does not seem to have any particular relevance.
Finally, atheroembolism as a result of aortic manipulation is addressed as another possible cause of renal impairment [27]. This can occur even in off-pump surgery because of side clamping. However, in our series side clamping did not influence the rate of ARF.
These speculations help us not only to depict the role of off-pump surgery in case of either normal or altered preoperative renal function, but also to explain the major impact of CPB on postoperative renal impairment. Actually, although off-pump surgery produces an inflammatory response similar to CPB, the latter is responsible for a more rapid (C5a and C5b-9) and more profound (C5b-9) expression of complement activation, which are responsible for tubular epithelial cell lesion by an ischemia–reperfusion mechanism [18]. Moreover, the nonpulsatile flow of CPB is strongly correlated with increased likelihood of renal dysfunction, especially if associated with hypotension [9]. Recently, the relationship between hemodilution and ARF was confirmed by Karkouti and associates [19].
Postoperative ARF was demonstrated to increase early mortality (7% to 38%) [1–4, 7]. The occurrence of ARF impacted negatively on early mortality (OR, 3.7). This result was mainly related to patients with normal preoperative renal function; in fact, ARF was found to be a risk factor for early mortality in analysis A but not in analysis B. Furthermore, ARF prolonged postoperative hospital stay from 4.6 ± 2.9 days to 6.7 ± 5.3 days (p < 0.001). Postoperative ARF caused a significant impairment of 10-year survival (hazard ratio, 4.1), confirmed in both analysis A (3.7) and analysis B (3.0). Moreover, off-pump surgery can provide an important protective role on both early and late outcome of patients with normal preoperative renal function.
In conclusion, avoiding CPB can be the most important item to prevent as much as possible ARF in patients with normal preoperative creatinine. In patients with abnormal preoperative creatinine, off-pump coronary artery bypass graft surgery does not seem to add any benefit, but further analysis on a larger cohort of patients is necessary to better clarify its role in this subset of patients. When CPB has to be necessarily used, we should remember that the longer the CPB duration, the more important the renal damage is, with a cutoff point of roughly 1 hour. In case of off-pump lateral wall revascularization, it is very important to keep arterial pressure stable to prevent even transient renal hypoperfusion. The occurrence of postoperative renal failure has a great impact on early and late mortality, and the surgical strategy cannot improve them.
The main limitation is the retrospective nature of the analysis. Moreover, the results found in analysis B should be confirmed in a larger cohort of patients. The use of creatininemia as a marker of renal function is not perfect, but it is by far the most widely used in the centers of cardiac surgery, as creatinine clearance is difficult to determine and it is often based on formulas rather than direct observation.
|
References |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentReferences |
|---|
Related Article
This article has been cited by other articles:
|
|
 |
|
  G. Lema Renal Outcome Following On- and Off-pump Coronary Artery Bypass Graft Surgery: Few Patients, Cautious Conclusions Asian Cardiovasc Thorac Ann, August 1, 2009; 17(4): 440 - 441. [Full Text] [PDF]
|
|
|
|
 |
|
 D. L. Ngaage, M. E. Cowen, and A. R. Cale Cardiopulmonary bypass and left ventricular systolic dysfunction impacts operative mortality differently in elderly and young patients Eur. J. Cardiothorac. Surg., February 1, 2009; 35(2): 235 - 240. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Knapik, P. Rozentryt, P. Nadziakiewicz, L. Polonski, and M. Zembala Retrospective cross-validation of simplified predictive index for renal replacement therapy after cardiac surgery Interactive CardioVascular and Thoracic Surgery, December 1, 2008; 7(6): 1101 - 1106. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Lema, J. Urzua, R. Jalil, R. Canessa, B. G. Loef, R. Henning, G. Navis, A. Rankin, W. van Oeveren, T. Ebels, et al. Cardiac surgery, cardiopulmonary bypass, and preoperative renal dysfunction Br. J. Anaesth., September 1, 2008; 101(3): 429 - 430. [Full Text] [PDF]
|
|
|
|
 |
|
 M. Di Mauro and A. M. Calafiore Reply Ann. Thorac. Surg., July 1, 2008; 86(1): 350 - 350. [Full Text] [PDF]
|
|
|
|
|
|
G. Lema, R. Canessa, and J. Urzua Off-Pump Surgery, Cardiopulmonary Bypass, and Renal Dysfunction Ann. Thorac. Surg., July 1, 2008; 86(1): 349 - 350. [Full Text] [PDF]
|
|
|
|
|
|
J. R. Brown Invited commentary Ann. Thorac. Surg., November 1, 2007; 84(5): 1502 - 1503. [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 |
