Asian Cardiovasc Thorac Ann 2003;11:42-47
© 2003 Asia Publishing EXchange Ltd
Impact of Preoperative Renal Dysfunction on Cardiac Surgery Results
Dan Abramov, MD,
Miguel Tamariz, MD,
Stephen Fremes, MD,
Sheldon Tobe, MD,
George Christakis, MD,
Veena Guru, BSc,
Bernard Goldman, MD
Department of Cardiovascular Surgery, Sunnybrook Health Science Center, University of Toronto, Toronto, Canada
For reprint information contact: Dan Abramov, MD Tel: 972 7 640 0961 Fax: 972 7 640 0961 email: abramov2{at}zahav.net.il Department of Cardiothoracic Surgery, Soroka Medical Center, Beer Sheva 84101, Israel.
 |
ABSTRACT
|
|---|
Results of cardiac surgery were analyzed using a database that included plasma creatinine levels in 2,214 patients, of whom 507 had preoperative renal dysfunction (creatinine clearance < 0.9 mL•s-1•m-2). Logistic regression and propensity score analyses found preoperative renal dysfunction to be an independent predictor of morbidity and mortality. Plotting preoperative creatinine clearance against morbidity and mortality revealed an exponential increase in morbidity and mortality when preoperative creatinine clearance was < 0.84 mL•s-1•m-2. Patients were stratified for age, operative procedure, and comorbidity. In all stratified groups, preoperative creatinine clearance < 0.84 mL•s-1•m-2 was associated with similar exponential increases in morbidity and mortality. In patients with preoperative renal dysfunction, elevated plasma creatinine levels persevered for 6 months postoperatively. Dialysis beyond postoperative day 10 was required in < 2% of patients with preoperative plasma creatinine of 160–200 µmol•L-1 and in 5% in those with creatinine > 200 µmol•L-1 (p < 0.05). Actuarial survival was significantly reduced (< 90% at 18 months postoperatively) in patients with preoperative renal dysfunction.
|
INTRODUCTION
|
|---|
Patients with preexisting renal dysfunction have been observed to have a greater requirement for postoperative dialysis, ventilation, and intensive care unit stay after cardiac operations.1–3 These studies were performed on small numbers of patients, and the operative results according to preoperative creatinine clearance rates were not given. Thus, they failed to define the threshold of preoperative renal dysfunction that affects postoperative results. Additional information is required on the long-term follow-up of renal function in such patients, particularly, the effect of operative renal injury on preoperatively damaged kidneys.
|
PATIENTS AND METHODS
|
|---|
Between August 1996 and July 1998, 2,214 consecutive patients underwent coronary artery bypass grafting (CABG) or valve surgery (isolated CABG, 82.1%; mitral valve procedure, 3.6%; aortic valve procedure, 5.2%; double-valve procedure, 1.9%; combined valve and CABG, 7.2%) at a tertiary-care university-affiliated hospital. Patients on preoperative dialysis or scheduled for off-pump surgery, resection of a ventricular aneurysm, or extracardiac surgical procedures were excluded from this study. Low-dose fentanyl citrate (10 to 15 µg•kg-1), midazolam (2 to 3 mg), isoflurane (0.5% to 2%), and propofol (100 to 150 µg•kg-1•min-1) were used for induction and maintenance of anesthesia. A standard median sternotomy and aorta-to-right atrial or bicaval cannulation were performed for cardiopulmonary bypass (CPB). Most patients (88%) remained normothermic (32°C to 37°C), the others were cooled to 28°C; pulsatile CPB was used in 67%. Perfusion pressures were maintained at 60 to 80 mm Hg. Blood cardioplegia was delivered in an 8:1 or 16:1 blood/crystalloid ratio.4,5 Warm or tepid (33°C) cardioplegia was used in 80% of operations, cold cardioplegia (10°C) was used in 20%.6 Cardioplegia was delivered either antegradely via the aortic root and completed vein grafts, or retrogradely via the coronary sinus. After cardioplegic induction, additional doses of 300 to 500 mL were administered after completion of each distal and proximal anastomoses, or every 10 to 20 min during valve surgery. Neither dopamine nor dopexamine were administered during the operative procedure.
Data of all 2,214 patients were collected prospectively in a computerized database: preoperative creatinine levels, daily creatinine levels during hospitalization, and follow-up creatinine levels up to 6 months postoperatively. The same data (without detailed creatinine levels) was available for 4,832 patients operated on between January 1990 and August 1998 in the same institution; these data were used for some of the multivariate analyses. The creatinine clearance rate was calculated from the plasma creatinine level according to the Cockcroft-Gault equation: creatinine clearance = (140–age) x weight x 0.85 (x 1.23 for males)/48.9 x plasma creatinine level (µmol•L-1).7 Preoperative renal dysfunction was defined as creatinine clearance < 0.9 mL•s-1•m-2.
The perioperative outcomes of interest were early mortality (< 30 days postoperatively), postoperative myocardial infarction (new Q waves or ischemic ST-T changes on an electrocardiogram, and elevation of cardiac enzymes), perioperative low output syndrome (use of inotropics or intraaortic balloon pumping for more than 30 min to maintain blood pressure above 90 mm Hg with a cardiac index < 2 L•min-1•m-2, and systemic vascular resistance > 1,200 dyne•s•cm-5), perioperative cerebrovascular accident, deep sternal wound infection, postoperative renal dysfunction (15-mL•min-1 change in creatinine clearance and a rate < 60 mL•min-1), and postoperative dialysis. Follow-up information was obtained on 2,048 patients (93% of late survivors) from outpatient visits, contact with the patient’s physician, or response to a patient questionnaire. The extent of follow-up was similar for those with (91%) and without (93%) preoperative renal dysfunction. Creatinine levels at follow-up (9 to 180 days postoperatively) were obtained in 437 of 495 surviving patients with preoperative renal dysfunction (88%). The mean follow-up was 6.3 ± 5.1 months (equal in patients with and without preoperative renal dysfunction). These data were subjected to actuarial analysis to determine late survival.
Data were collected and managed in dBASE IV datasets, and SAS software (SAS Institute Inc., Cary, NC, USA) was used for statistical analyses. Clinical and angiographic features were analyzed by descriptive statistical methods. Continuous variables were compared by analysis of variance, and categorical variables were compared by Fisher’s exact test. Stepwise multiple logistic regression analysis using the maximum likelihood estimates was employed to determine independent predictors of operative mortality and early nonfatal complications. Model discrimination was evaluated by the area under the receiver operating characteristic curve, and the calibration was assessed with the Hosmer-Lemeshow goodness-of-fit statistic.8–10 To adjust for the different preoperative risk profiles of patients with and without preoperative renal dysfunction, propensity scores were generated by logistic regression, ranked, and divided into quintiles. Logistic regression using the quintiles was used to examine the predictability of preoperative renal dysfunction for mortality and morbidity plus mortality. Cox multivariate regression analysis was employed to assess time-dependent late events such as mortality. Survival curves were analyzed using a life-test procedure with product limit (Kaplan-Meier) estimates.
|
RESULTS
|
|---|
Preoperative renal dysfunction was noted in 507 patients. Tables 1
, 2, and 3 show the profiles and outcomes of cardiac surgical patients with and without preoperative renal dysfunction. Patients with preexisting renal dysfunction were significantly older and had non-elective surgery, with a higher proportion of females, valve procedures, and comorbidity. The CPB and aortic crossclamp times were significantly longer in this group. Among the CABG patients, left internal mammary artery usage was lower in those with preoperative renal dysfunction.
Because of the significant difference in perioperative parameters between the two groups, covariate adjustment was performed using logistic regression and propensity score analysis to determine the independent contribution of preoperative renal dysfunction to operative mortality and nonfatal complications. Step-wise logistic regression analysis of the 4,823 cases of isolated CABG in our institution during the last 9 years revealed preoperative renal failure to be an independent risk factor for postoperative mortality (Table 4) and the pre-specified composite endpoint of perioperative mortality, myocardial infarction, low output syndrome, and cerebrovascular accident (Table 5). Propensity scores were generated per patient by logistic regression and ranked. The incidence of cardiac morbidity and mortality was strongly associated with the quintiles (1st quintile, 6.7%; 5th quintile, 27%; p < 0.001). Multivariate analysis using the quintiles found preoperative renal dysfunction to be a predictor of morbidity and mortality (odds ratio, 1.96; 95% confidence interval, 1.13–3.77; p = 0.01).
View this table:
[in this window]
[in a new window]
|
Table 5. Logistic Regression Analysis* of Postoperative Mortality and Complications in 4,839 CABG Patients (1990–8)
|
|
Figure 1 depicts the frequency of mortality and morbidity for various preoperative rates of creatinine clearance. An exponential increase in morbidity and mortality was observed when preoperative creatinine clearance rates were below 0.84 mL•s-1•m-2. Because age is a component of the formula for creatinine clearance calculation, which may bias the results (older patients would be in the lower creatinine clearance groups), two checks were carried out. Firstly, plasma creatinine levels (rather than creatinine clearance) and morbidity and mortality were plotted. Exponential increases in morbidity and mortality were observed when plasma creatinine levels exceeded 160 µmol•L-1 (Figure 2). Secondly, we adjusted for age by stratifying into three age groups: 51–60 years, 61–70 years, 71–80 years; the same exponential rises in morbidity and mortality were seen in each age group when creatinine clearance rates were < 0.84 mL•s-1•m-2. When surgical procedure was considered, the curves for mortality and morbidity in patients with either isolated CABG or valve procedures also showed a sharp rise when creatine clearance was below 0.84 mL•s-1•m-2. To determine the net effect of different levels of preoperative renal dysfunction, the subgroup of patients undergoing their first cardiac procedure who were younger than 65 years of age with good preoperative left ventricular function and no additional comorbidity were analyzed separately. Although this subgroup usually has a very low incidence of mortality and morbidity, the same exponential rises in morbidity and mortality were seen when creatinine clearance was < 0.84 mL•s-1•m-2.
View larger version (23K):
[in this window]
[in a new window]
|
Figure 1. Frequency of complications (low output syndrome, myocardial infarction, cerebrovascular accident) and early mortality in 2,214 patients with different preoperative creatinine clearance rates.
|
|
View larger version (21K):
[in this window]
[in a new window]
|
Figure 2. Frequency of complications (low output syndrome, myocardial infarction, cerebrovascular accident) and early mortality in 2,214 patients with different preoperative plasma creatinine levels.
|
|
Figure 3 shows the variations in mean postoperative plasma creatinine levels over time. Patients with normal preoperative renal function had steady creatinine levels, whereas those with preoperative renal dysfunction tended to have increased creatinine levels up to 6 months postoperatively. Among the patients with preoperative renal dysfunction, 1.8% of those with preoperative plasma creatinine < 200 µmol•L-1, and almost 5% of those with levels > 200 µmol•L-1 required long-term dialysis (more than 10 days postoperatively). As seen in Figure 4, actuarial survival of patients with preoperative renal dysfunction was significantly lower than those with normal renal function (log rank: p = 0.001; Wilcoxon: p = 0.01), falling to < 90% at 18 months postoperatively. Preoperative renal insufficiency was found to be an independent risk factor for mortality and complications during follow-up (Tables 6 and 7).
View larger version (12K):
[in this window]
[in a new window]
|
Figure 3. Mean plasma creatinine levels in the preoperative period, early postoperative days, and late follow-up in 266 patients with preoperative renal dysfunction and those with normal renal function preoperatively.
|
|
View larger version (9K):
[in this window]
[in a new window]
|
Figure 4. Actuarial survival after cardiac surgery in patients with and without preoperative renal dysfunction. Patients undergoing coronary artery bypass grafting (CABG) and valve surgery were considered separately.
|
|
View this table:
[in this window]
[in a new window]
|
Table 7. Cox Proportional Hazard Analysis of Death, Angina Pectoris, or Myocardial Infarction in 4,839 CABG Patients (1990–8)
|
|
|
DISCUSSION
|
|---|
Although patients with preoperative renal dysfunction have a higher risk of postoperative morbidity and mortality, renal dysfunction per se might not explain the big differences in the results of cardiac surgery. A high plasma creatinine level is also an indicator of premorbid conditions such as diffuse atherosclerotic disease with renal involvement, and pre-renal azotemia due to preoperative low cardiac output syndrome; these conditions may better explain the high rate of complications.
The database used in this study contains more than 180 parameters for each of the 2,214 consecutive patients, including detailed preoperative and postoperative creatinine levels. This enabled us to analyze the operative results of patients with different degrees of preoperative renal dysfunction. Preoperative creatinine clearance < 0.84 mL•s-1•m-2 or plasma creatinine > 160 µmol•L-1 was associated with exponential increases in morbidity and mortality. Even when stratified for age, operative procedure, or absence of comorbidity, the same morbidity and mortality thresholds were observed. The group of young healthy patients is of special interest because their postoperative results are usually excellent; however, creatinine clearance < 0.84 mL•s-1•m-2 markedly increased morbidity and mortality even in this subset.
We concluded that patients with preoperative creatinine clearance < 0.84 mL•s-1•m-2 should be evaluated more carefully before being considered for cardiac surgery, even young and otherwise healthy patients. Higher early postoperative morbidity and mortality should be expected as well as poorer long-term survival. Either a nonsurgical approach or a less extensive operation should be contemplated. Off-pump surgery should be assessed as an option for these patients. Further deterioration of renal function can be expected in the long-term follow-up, and patients with preoperative plasma creatinine above 200 µmol•L-1 should be aware of the high incidence of long-term dialysis postoperatively. These patients may benefit from implementation of greater renal protective measures during CPB (discontinuation of nephrotoxic antibiotics and use of mannitol, loop diuretics, and low-dose dopamine) with closer monitoring of renal function afterwards.11–16
|
REFERENCES
|
|---|
- Rao V, Weisel RD, Buth KJ, Cohen GC, Borger MA, Shiono N, et al. Coronary artery bypass grafting in patients with non-dialysis-dependent renal insuficiency. Circulation
1997;96(Suppl II):38–45.
- Sutton RG. Renal considerations, dialysis, and ultrafiltration during cardiopulmonary bypass. Int Anesthesiol Clin
1996;34:165–76.[Medline]
- Lahey SJ, Borlass BC, Lavin PT, Levitsky S. Preoperative risk factors that predict hospital length of stay in coronary artery bypass patients > 60 years old. Circulation
1992; 86(Suppl II):181–5.
- Fremes SE, Levy SL, Christakis GT, Walker SE, Iazzetta J, Mallidi HR, et al. A phase I human trial of adenosine-potassium cardioplegia. Circulation
1996;94(Suppl II):370–5.
- Cohen G, Feder-Elituv R, Iazetta J, Bunting P, Mallidi H, Bozinovski J, et al. Phase 2 studies of adenosine cardioplegia. Circulation
1998;98(Suppl II):225–33.
- Normothermic versus hypothermic blood cardioplegia for coronary bypass surgery: a randomized trial in 1732 patients. The Warm Heart Investigators. Lancet
1994;343:559–63.[Medline]
- Spinler SA, Nawarskas JJ, Boyce EG, Connors JE, Charland SL, Goldfarb S. Predictive performance of ten equations for estimating creatinine clearance in cardiac patients. Iohexol Cooperative Study Group. Ann Pharmacother
1998;32:1275–83.[Abstract]
- Hanley JA, McNeil BJ. The meaning and use of area under a receiver operating characteristic (ROC) curve. Radiology
1982;143:29–36.[Abstract/Free Full Text]
- Hanley JA, McNeil BJ. A method for comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology
1983;148:839–43.[Abstract/Free Full Text]
- Hosmer DW, Lemeshow S. Applied logistic regression. New York: Wiley, 1989.
- Leurs PB, Mulder AW, Fiers HA, Hoorntje SJ. Acute renal failure after cardiovascular surgery. Current concepts in pathophysiology, prevention, and treatment. Eur Heart J
1989;10(Suppl H):38–42.
- Berendes E, Mollhof T, Van Aken H, Schmidt C, Erren M, Deng MC, et al. Efects of dopexamine on creatinine clearance, systemic inflammation, and splanchnic oxygenation in patients undergoing coronary artery bypass grafting. Anesth Analg
1997;84:950–7.[Abstract]
- Gelman S. Ischemic insult, kidney viability, and renal function. Anesth Analg
1998;86:1–2.[Medline]
- Hashimoto K, Nomura K, Nakano M, Sasaki T, Kurosawa H. Pharmacological intervention for renal protection during cardiopulmonary bypass. Heart Vessels
1993;8:203–10.[Medline]
- Kainuma M, Yamada M, Miyaka T. Continuous urine oxygen tension monitoring in patients undergoing cardiac surgery. J Cardiothorac Vasc Anesth
1996;10:803–8.
- Lema G, Urzua J, Jalil R, Canessa R, Moran S, Sacco C, et al. Renal protection in patients undergoing cardiopulmonary bypass with preoperative abnormal renal function. Anesth Analg
1998;86:3–8.[Abstract]
This article has been cited by other articles:
|
|
|
|
 
Determinants of operative mortality in valvular heart surgery.
J. Thorac. Cardiovasc. Surg.,
March 1, 2006;
131(3):
547 - 557.
|
|
|
TOP
ABSTRACT
INTRODUCTION
