Renal dysfunction is a major adverse event after cardiovascular surgery. Therefore, the preoperative prediction of which patients will require renal replacement therapy (RRT) after cardiac surgery is an important issue. In the present study, 1,822 consecutive patients who underwent cardiovascular surgery from 2008 and 2013 at a single institution were reviewed. Patients who were already receiving long-term hemodialysis before surgery (n = 134) were excluded. The remaining 1,688 patients were separated into 2 groups: those requiring postoperative RRT and those without RRT requirement. A total of 128 patients (7.6%) required RRT. Patients requiring RRT had greater perioperative blood loss, longer intubation time, and longer hospital stays (p <0.0001 for all). Multivariate analysis revealed that cardiopulmonary bypass use, preoperative body surface area, the left ventricular ejection fraction, serum albumin, and creatinine were independent risk factors for postoperative RRT (odds ratios 2.435, 0.204, 0.976, 0.556, and 5.394, 95% confidence intervals 1.471 to 4.140, 0.054 to 0.841, 0.962 to 1.025, 0.363 to 0.860, and 3.671 to 8.223, respectively, p <0.05 for all). A subgroup of patients with relatively preserved renal function before surgery (creatinine <1.12 mg/dl, a cut-off value for RRT requirement obtained from receiver-operating characteristic curve analysis [area under the curve 0.74748, sensitivity 60.2%, specificity 85.0%]) showed that preoperative serum albumin concentration was most significantly associated with postoperative RRT requirement (odds ratio 0.048, 95% confidence interval 0.023 to 0.095, p <0.0001). In conclusion, cardiopulmonary bypass use, preoperative renal impairment as reflected by elevated creatinine level, small body size, a low left ventricular ejection fraction, and hypoalbuminemia were associated with a requirement for postoperative RRT. In patients with preserved renal function, hypoalbuminemia was most significantly related to requirement for RRT.
Highlights
- •
Postoperative renal failure was investigated in 1,688 patients who underwent cardiac surgery.
- •
Preoperative factors associated with postoperative renal replacement therapy were analyzed.
- •
Hypoalbuminemia was the strong predictor of a need for renal replacement therapy.
- •
A need for renal replacement therapy was related to an increase in postoperative adverse events.
Renal replacement therapy (RRT) after cardiac surgery is a strong predictor of poor postoperative outcome. Previously reported preoperative risk factors for postoperative acute renal dysfunction requiring transient or permanent RRT included age, female gender, preoperative heart failure, diabetes, hypertension, anemia, cardiopulmonary bypass (CPB) use, and preexisting renal dysfunction, especially in patients who undergo emergency surgery under unstable conditions. However, the factors associated with RRT in patients who undergo elective cardiac surgery in stable condition have not been fully evaluated. Therefore, we investigated the factors associated with postoperative RRT requirement in patients who underwent elective cardiac surgery under stable conditions. None of these patients were receiving any preoperative long-term RRT such as hemodialysis or peritoneal dialysis.
Methods
The data collection protocol was approved by the Institutional Review Board of Juntendo University (Tokyo, Japan). A total of 1,822 consecutive patients who underwent elective cardiac surgery at the Juntendo University School of Medicine from December 2008 to October 2013 were eligible for retrospective entry into the study. Patients who underwent vascular surgery alone, such as ascending aorta replacement, were excluded from the study. A total of 134 patients (7.4% of the cohort) who were already receiving long-term dialysis before the surgery were also excluded. Data obtained from the remaining 1,688 patients were used for this retrospective study.
Patient data were obtained from hospital medical records. Laboratory values and echocardiograms obtained within 2 weeks before surgery were collected. If a patient had undergone multiple examinations, the last set of results immediately before surgery was used for the analysis. Postoperative outcome analysis included length of intensive care unit stay and total postoperative hospital stay, duration of ventilator support, amount of perioperative bleeding, and postoperative adverse events such as stroke and new-onset atrial fibrillation. Patients were divided into 2 groups: those who required transient or/and permanent RRT (hemodialysis or continuous hemodiafiltration) postoperatively and those who had no postoperative RRT requirement. Preoperative information and postoperative courses were compared between the groups.
Preoperative factors associated with RRT requirement were investigated by uni- and multivariate logistic analysis. The cut-off value for the preoperative creatinine level associated with RRT requirement was calculated, and a subanalysis was also performed to investigate factors that would predict RRT requirement in patients with preoperative creatinine levels less than the cut-off value.
Data are presented as mean ± SD. Normality was evaluated for each variable using normal distribution plots and histograms. Variables were compared between the groups using Student’s unpaired 2-tailed t test. The Mann-Whitney test was performed with median and interquartile values when the variables were not normally distributed. Categorical variables were compared using the chi-square test. Univariate logistic regression analysis was used to select potential variables associated with RRT requirement. Variables showing significant impacts on RRT in univariate analysis were then entered into a multivariate logistic regression model. Data that showed a bimodal distribution were subjected to logarithmic transformation as needed, to improve normality before performing logistic regression analysis. The cut-off value of preoperative creatinine associated with postoperative RRT requirement was determined using receiver-operating characteristic curve analysis. Sensitivity and specificity were determined and are expressed as percentages. The factors associated with RRT requirement in a subset of patients who had relatively preserved preoperative creatinine concentrations were identified by performing a multivariate analysis in selected patients whose creatinine levels were less than the cut-off value. Statistical analyses were performed using JMP version 11.0 (SAS Institute Inc., Cary, North Carolina).
Results
Of the 1,822 patients who underwent cardiac surgery during the study period, 134 patients who were already receiving long-term hemodialysis preoperatively were excluded from the analysis. Therefore, data obtained from the remaining 1,688 patients were used for this investigation. Of these, 128 patients required transient or permanent RRT after cardiac surgery (7.6%), while 1,560 patients required no RRT. A flowchart of the patient selection strategy is shown in Figure 1 .
Clinical characteristics and preoperative laboratory values in the total cohort of the patients are listed in Table 1 Our cohort consisted of 720 patients (42.7%) who had undergone coronary artery bypass grafting (CABG), 759 (45.0%) who had undergone valvular surgery, and 207 (12.3%) who had undergone a combination of CABG and valvular surgery. Most patients who underwent CABG underwent only off-pump procedures (718 patients [99.7%] of those who underwent CABG); therefore, a total of 968 patients (57.3%) underwent CPB, and the remaining patients did not.
| Patients requiring RRT (n=128) | Patients not requiring RRT (n=1560) | p value | |
|---|---|---|---|
| Clinical characteristics | |||
| Age (years) | 69.6±11.8 | 69.4±12.3 | 0.0057 |
| Men | 81(63.3%) | 1077(69.0%) | 0.1773 |
| Body surface area (m 2 ) | 1.59±0.20 | 1.65±0.20 | 0.0024 |
| Type of surgery | 0.9392 | ||
| Coronary artery bypass surgery alone | 55(43.0%) | 665(42.6%) | |
| Valvular surgery alone | 59(46.1%) | 700(44.9%) | |
| Coronary artery bypass surgery + Valvular surgery | 14 (10.9%) | 193(12.4%) | |
| Others | 0(0%) | 2(0.13%) | |
| Cardiopulmonary bypass usage | 87 (68.0%) | 881 (56.5%) | 0.0120 |
| Left ventricular ejection fraction (%) | 52.4±17.4 | 59.6±13.9 | < 0.0001 |
| Preoperative laboratory value | |||
| Hemoglobin (g/dL) | 11.5±1.93 | 12.9±1.73 | < 0.0001 |
| Total bilirubin (mg/dL) | 0.8±0.7 | 0.8±0.5 | 0.7560 |
| Total protein (g/dL) | 6.6±0.8 | 6.9±0.6 | < 0.0001 |
| Albumin (g/dL) | 3.6±0.7 | 4.1±0.5 | < 0.0001 |
| Potassium (mEq/L) | 4.4±0.5 | 4.3±0.4 | 0.0025 |
| Sodium (mEq/L) | 139.1±3.6 | 140.0±2.3 | 0.0003 |
| Blood urea nitrogen (mg/dL) | 30.7±18.2 | 17.8±7.4 | < 0.0001 |
| Creatinine (mg/dL) | 1.8±1.20 | 0.9±0.3 | <0.0001 |
| Brain natriuretic peptide (pg/mL) | 326.6 (134.1-888.0) | 79.4 (33.0-192.4) | <0.0001 |
| C-reactive protein (mg/dL) | 2.00±3.77(n=105) | 0.47±1.37 (n=1293) | <0.0001 |
Univariate and multivariate logistic regression analysis for the factors associated with RRT on the basis of the total cohort are listed in Table 2 . Multivariate analysis showed that smaller body size, CPB requirement during surgery, a lower left ventricular ejection fraction, lower serum albumin concentration, and higher serum creatinine level remained independent factors associated with the need for postoperative RRT ( Table 2 ). C-reactive protein (CRP) level was obtained from only a limited number of patients; therefore, this parameter was not included in the multivariate analysis.
| Univariate | Multivariate | |||
|---|---|---|---|---|
| OR (95%CI) | p value | OR (95%CI) | p value | |
| Age (years) | 1.024 (1.007-1.042) | 0.0039 | 1.004 (0.981-1.029) | 0.7421 |
| Gender (men=1, women=0) | 0.773 (0.534-1.132) | 0.1828 | ||
| Body surface area (m 2 ) | 0.250 (0.101-0.616) | 0.0027 | 0.204 (0.054-0.841) | 0.0281 |
| Type of surgery | ||||
| Coronary artery bypass surgery alone=1, | 1.151 (0.121-2.543) | 0.729 | ||
| Other surgeries=2 | ||||
| Cardiopulmonary bypass (yes=1, no=0) | 1.635 (1.121-2.423) | 0.0010 | 2.435 (1.471-4.140) | 0.0004 |
| Left ventricular ejection fraction (%) | 0.969 (0.958-0.981) | < 0.0001 | 0.976 (0.962-1.025) | 0.0012 |
| Hemoglobin (g/dL) | 0.626(0.561-0.695) | < 0.0001 | 0.926 (0.803-1.066) | 0.2881 |
| Total bilirubin (mg/dL) | 1.049 (0.735-1.367) | 0.7618 | ||
| Total protein (g/dL) | 0.414 (0.318-0.539) | < 0.0001 | 0.783 (0.531-1.005) | 0.2881 |
| Albumin (g/dL) | 0.233 (0.172-0.314) | < 0.0001 | 0.556 (0.363-0.860) | 0.0082 |
| Potassium (mEq/L) | 1.866 (1.237-2.789) | 0.0032 | 0.984 (0.664-1.007) | 0.4835 |
| Sodium (mEq/L) | 0.905 (0.858-0.957) | 0.0006 | 1.007 (0.936-1.086) | 0.8620 |
| Blood urea nitrogen (mg/dL) | 1.090(1.073-1.108) | < 0.0001 | 1.006(0.995-1.014) | 0.1650 |
| Creatinine (mg/dL) | 7.027 (5.081-9.965) | < 0.0001 | 5.394 (3.671-8.223) | < 0.0001 |
| Brain natriuretic peptide (pg/mL) | 1.001 (1.001-1.002) | < 0.0001 | 1.000(1.000-1.000) | 0.9295 |
| C-reactive protein (mg/mL) (n=1398) | 1.283 (1.189-1.391) | < 0.0001 | ||
The receiver-operating characteristic curve analysis identified 1.12 mg/dl as the optimal cut-off value for preoperative creatinine concentration associated with RRT requirement before surgery; this value was associated with sensitivity and specificity of 60.2% and 85.0%, respectively ( Figure 2 ). Patients whose preoperative creatinine levels were less than this cut-off value were selected as having relatively preserved renal function. A total of 1,376 patients had preoperative creatinine levels <1.12 mg/dl (81.5% of the total cohort). In this subgroup, 51 patients (3.7%) required RRT postoperatively ( Figure 1 ).
A comparison of clinical characteristics and laboratory values between these patients with and without RRT requirements who had relatively preserved renal function (creatinine <1.12 mg/dl) is listed in Table 3 . Patients requiring RRT were more often women and had smaller body sizes compared with those who did not need RRT. Hemoglobin, total protein, and albumin were lower and blood urea nitrogen and CRP were higher in patients with RRT than those who did not require RRT.
| Patients requiring RRT (n=51) | Patients not requiring RRT (n=1325) | p value | |
|---|---|---|---|
| Clinical characteristics | |||
| Age (years) | 67.6±14.1 | 65.7±12.6 | 0.2812 |
| Men | 26(51.0%) | 880(66.4%) | 0.0226 |
| Body surface area (m 2 ) | 1.58±0.20 | 1.65±0.20 | 0.0146 |
| Type of surgery | 0.4038 | ||
| Coronary artery bypass surgery alone | 16(31.4%) | 572(43.2%) | |
| Valvular surgery alone | 27(52.9%) | 586(44.2%) | |
| Coronary artery bypass surgery + Valvular surgery | 8 (15.7%) | 165(12.5%) | |
| Others | 0(0%) | 2(0.15%) | |
| Cardiopulmonary bypass requirement (%) | 38 (74.5%) | 772 (58.3%) | 0.0206 |
| Left ventricular ejection fraction (%) | 56.4±16.5 | 60.3±13.5 | 0.0555 |
| Preoperative laboratory value | |||
| Hemoglobin (g/dL) | 11.9±1.78 | 13.1±1.63 | < 0.0001 |
| Total bilirubin (mg/dL) | 0.8±0.6 | 0.8±0.5 | 0.3199 |
| Total protein (g/dL) | 6.6±0.9 | 6.9±0.6 | < 0.0001 |
| Albumin (g/dL) | 3.3±0.7 | 4.1±0.4 | < 0.0001 |
| Potassium (mEq/L) | 4.2±0.5 | 4.6±0.8 | 0.7859 |
| Sodium (mEq/L) | 139.8±2.5 | 140.0±4.7 | 0.6784 |
| Blood urea nitrogen (mg/dL) | 23.5±37.2 | 16.6±13.9 | 0.0001 |
| Creatinine (mg/dL) | 0.76±0.21 | 0.77±0.17 | 0.8550 |
| Brain natriuretic peptide (pg/mL) | 188.2 (71.7-452.7) | 72.0 (29.3-175.2) | <0.0001 |
| C-reactive protein (mg/dL) | 1.85±3.27 (n=45) | 0.45±1.40 (n=1099) | <0.0001 |
Table 4 lists the results of uni- and multivariate logistic regression analyses of the factors associated with RRT in the subgroup of patients with preserved renal function. Multivariate analysis revealed that albumin concentration was most significantly associated with postoperative RRT.
| Univariate | Multivariate | |||
|---|---|---|---|---|
| OR (95%CI) | p value | OR (95%CI) | p value | |
| Age (years) | 1.013 (0.990-1.039) | 0.2684 | ||
| Gender (men=1, women=0) | 0.526 (0.299-0.925) | 0.0268 | 0.788 (0.330-1.849) | 0.5868 |
| Body surface area (m 2 ) | 0.203 (0.056-0.745) | 0.0168 | 0.713 (0.090-6.090) | 0.7578 |
| Type of surgery | ||||
| Coronary artery bypass surgery alone=1, | 0.604 (0.322-1.083) | 0.0916 | ||
| Other surgeries=2 | ||||
| Cardiopulmonary bypass (yes=1, no=0) | 2.093 (1.134-4.118) | 0.0173 | 2.096 (1.032-4.503 | 0.0405 |
| Left ventricular ejection fraction (%) | 0.982 (0.963-1.002) | 0.0693 | ||
| Hemoglobin (g/dL) | 0.657 (0.555-0.776) | ＜0.0001 | 1.054 (0.852-1.305) | 0.6243 |
| Total bilirubin (mg/dL) | 0.993 (0.701-1.010) | 0.7200 | ||
| Total protein (g/dL) | 0.481 (0.335-0.684) | ＜0.0001 | 1.921(1.146-3.260) | 0.0132 |
| Albumin (g/dL) | 0.233 (0.172-0.314) | ＜0.0001 | 0.048 (0.023-0.095) | <0.0001 |
| Potassium (mEq/L) | 0.984 (0.553-1.012) | 0.6897 | ||
| Sodium (mEq/L) | 0.992 (0.961-1.005) | 0.7184 | ||
| Blood urea nitrogen (mg/dL) | 1.008(1.000-1.018) | 0.0584 | ||
| Creatinine (mg/dL) | 0.856 (0.162-4.530) | 0.8550 | ||
| Brain natriuretic peptide (pg/mL) | 1.000 (1.000-1.000) | 0.8815 | ||
| C-reactive protein (mg/mL) (n=1398) | 1.254 (1.115-1.378) | < 0.0001 | ||
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
