Intraoperative Renal Regional Oxygen Desaturation Can Be a Predictor for Acute Kidney Injury after Cardiac Surgery Dae-Kee Choi, MD, Wook-Jong Kim, MD, Ji-Hyun Chin, MD, Eun-Ho Lee, MD, Kyung Don Hahm, MD, Ji Yeon Sim, MD, and In Cheol Choi, MD Objective: To evaluate the usefulness of renal regional oxygen saturation (renal rSO2) in predicting the risk of acute kidney injury (AKI) after cardiac surgery. Design: A prospective observational study. Setting: Tertiary care university hospital. Participants: One hundred patients undergoing cardiac surgery. Interventions: Renal rSO2 was monitored continuously by near-infrared spectroscopy (NIRS) throughout the anesthetic period. Measurements and Main Results: Postoperative AKI was defined using the Risk, Injury, Failure, Loss, and End-stage (RIFLE) criteria. Of 95 patients who were included in the final analysis, 34 patients developed AKI after surgery. Recorded renal rSO2 data were used to calculate the total duration of the time when renal rSO2 was below the threshold values of 70%, 65%, 60%, 55%, and 50%. The total periods when the renal rSO2 level was below each of the threshold values were
significantly longer in patients with AKI than in those without AKI (p ¼ 0.001 or p o 0.001). Receiver operating characteristic (ROC) curve analysis was used to evaluate the predictive power of renal rSO2 for AKI. The ROC curve analysis showed that renal rSO2 could predict the risk of AKI with statistical significance and that a renal rSO2 o 55% had the best performance (area under the curve–ROC, 0.777; 95% CI, 0.6690.885; p o 0.001). Multivariate logistic regression analysis revealed that AKI significantly correlated with the duration of renal rSO2 o 55% (p ¼ 0.002) and logistic EuroSCORE (p ¼ 0.007). Conclusions: Intraoperative renal regional oxygen desaturation can be a good predictor of AKI in adult patients undergoing cardiac surgery. & 2014 Elsevier Inc. All rights reserved.
A
elective cardiac surgery with cardiopulmonary bypass (CPB) were enrolled between October 2010 and August 2011. Exclusion criteria were off-pump coronary artery bypass surgery (OPCAB), aortic surgery (total arch replacement and/or descending aorta replacement surgery), emergency surgery, left ventricular ejection fraction (LVEF) o 30%, body mass index (BMI) Z 30 kg/m2, sCr Z 1.5 mg/dL, and end-stage renal disease or renal transplant. Anesthesia was induced with etomidate (0.2 mg/kg), followed by rocuronium (0.8 mg/kg) to facilitate tracheal intubation, and was maintained with a continuous infusion of propofol (1-2 μg/mL) and remifentanil (5-20 ng/mL) using a target-controlled infusion pump (Orchestra Base Primea; Fresenius Kabi, Paris, France). Normal saline and 6% hydroxyethyl starch 130/0.4 (Voluven, Fresenius Kabi, Bad Homburg, Germany) were used for fluid management, and red blood cell (RBC) transfusion was initiated when hemoglobin level was o8 g/dL. A 20-G radial arterial catheter was inserted to monitor arterial blood pressure and to allow blood sampling. A 7.5-Fr pulmonary arterial catheter (Swan-Ganz Combo V CCL/SvO2/CEDV; Edwards Lifesciences, Irvine, CA) was advanced and connected to a Vigilance II device (Edwards Lifesciences) to monitor cardiac output and mixed venous oxygen saturation (SvO2). After heparinization (400 U/kg of heparin), CPB was started using a roller pump (Stöckert SIII; Stöckert Instrumente, Munich, Germany), with the tubing system without surface coating (Medtronic, Minneapolis, MN, USA), a membrane oxygenator (Affinity; Medtronic), and a 38-μm blood filter (Medtronic). The CPB flow rate was maintained between 2.2 and 2.5 L/min/m2, and the mean arterial pressure during CPB was maintained at 50 to 80
CUTE KIDNEY INJURY (AKI) is a common and serious complication of cardiac surgery, with a reported incidence as high as 45%.1,2 It is associated with poor outcomes, prolonged stays in hospital and in an intensive care unit (ICU), and increased mortality.2,3 Small increases in the levels of serum creatinine (sCr) reflect renal injury and predict poor outcomes.4 Nevertheless, given that a detectable increase in the levels of sCr does not immediately follow the onset of AKI, sCr levels alone are of limited value for the early identification of the risk of AKI. Biomarkers of early renal injury, such as neutrophil gelatinase-associated lipocalin (NGAL), cystatin C, kidney injury molecule 1, and interleukin (IL)-18, recently have been shown to facilitate earlier diagnosis of AKI.5,6 Although these biomarkers show some benefits for early detection of AKI, they cannot be measured continuously to monitor renal function in real time, which is needed to optimize the renal condition. Near-infrared spectroscopy (NIRS) is a new noninvasive technique that continuously monitors regional oxygen saturation (rSO2) by measuring the relative concentrations of oxygenated and deoxygenated hemoglobin within a local tissue area.7 Previous studies have suggested that low rSO2 values can predict the development of organ dysfunction in the brain or other organs after surgery.8 Persistent low renal rSO2 has been associated with renal dysfunction after pediatric cardiac surgery.9 However, to date, the relationship between intraoperative renal rSO2 and AKI in adult patients undergoing cardiac surgery has not been determined. The authors hypothesized that renal rSO2 measured by NIRS can reflect renal desaturation or hypoxia, which may contribute to AKI. The aim of this study was to examine whether intraoperative renal rSO2 could predict the risk of AKI after cardiac surgery. METHODS After approval by the Institutional Ethics Committee and obtaining written informed consent, a total of 100 adult patients who underwent
KEY WORDS: acute kidney injury, near-infrared spectroscopy, cardiac surgery
From the Department of Anesthesiology and Pain Medicine, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea. Address reprint requests to In-Cheol Choi, MD, Department of Anesthesiology and Pain Medicine, Asan Medical Center, University of Ulsan College of Medicine, 388-1, Pungnap 2-Dong, Songpa-gu, Seoul 138-736, Korea E-mail: [email protected] © 2014 Elsevier Inc. All rights reserved. 1053-0770/2601-0001$36.00/0 http://dx.doi.org/10.1053/j.jvca.2013.12.005
Journal of Cardiothoracic and Vascular Anesthesia, Vol ], No ] (Month), 2014: pp ]]]–]]]
1
2
mmHg. Alpha-stat pH management was used during CPB. After weaning from CPB, heparin was antagonized by protamine sulfate (4 mg/kg). All patients were transferred to the ICU after surgery, where care was provided by ICU nurses and staff who were not involved in the study in accordance with existing unit protocol. Patients were discharged from the ICU to a general ward when their vital signs were stable and further ICU care and monitoring were not required. Renal rSO2 was monitored with NIRS (INVOS 5100C; Somanetics Co., Troy, MI) throughout the entire anesthetic period. Before the induction of anesthesia, a disposable NIRS sensor was applied to one or both sides of the flank area that overlies the kidney in order to monitor renal rSO2 under sonographic guidance (Z.one Ultra; ZONARE Medical Systems Inc., Mountain View, CA) (Fig 1). To visualize the kidney, an ultrasound probe was placed in the lower intercostal space (below either the 10th or 11th rib) on the posterior axillary line. After obtaining the long-axis image of the kidney, the renal depth (the distance from the skin surface to the capsule of the kidney) was measured. A sensor then was applied to the probe placement site in instances when the renal depth was o40 mm. Patients were excluded from the study if both renal depths were Z40 mm. Renal rSO2 values were recorded continuously at 5- or 6-second intervals. Whenever renal rSO2 was monitored on both sides, the lower value was selected for analysis. The authors believe that the lower value is clinically more important because selecting the higher value or mean value may mask existing hypoxia, which may contribute to AKI. The sensitivity tests showed that the predictive ability of data including the lower value are better than that of data including the higher or mean value (data not shown). The patient whose monitoring of renal rSO2 was not checked continuously for more than 5 minutes during surgery was excluded from the study. No interventions were attempted based on renal rSO2 values. Cerebral rSO2 also was recorded simultaneously. Recorded rSO2 data were analyzed using Excel software (Microsoft, Seattle, WA). Renal rSO2 data were used to calculate the total duration of renal rSO2 below the threshold value. Multiple thresholds were selected to reflect different degrees of renal hypoxia. The thresholds used were renal rSO2 values of 70%, 65%, 60%, 55%, and 50%. These represent the approximate mean value of the baseline (80.0 ⫾ 9.1%) minus 1, 1.5, 2, 2.5, and 3 standard deviations (SDs), respectively. In addition, the other thresholds used were the proportionate decline form baseline value (10%, 15%, 20%, 25%, and 30%). Demographic data included age, sex, BMI, diabetes mellitus, hypertension, peripheral vascular disease, LVEF, logistic EuroSCORE (European System for Cardiac Operative Risk Evaluation), renal depth, baseline renal rSO2, preoperative hemoglobin, preoperative sCr, and preoperative estimated glomerular filtration rate (eGFR). The eGFR was calculated using the modification of diet in the renal disease (MDRD) formula [eGFR ¼ 186 creatinine –1.154 age –0.203 0.742 (if female)].10 Intraoperative variables included the type of surgery, surgical
CHOI ET AL
time, CPB time, fluid infusion volume (crystalloid and colloid), urine output, number of patients requiring blood transfusion, and intraoperative minimum hemoglobin. The intraoperative vital signs (heart rate, mean arterial pressure, central venous pressure, mean pulmonary arterial pressure, cardiac index, and SvO2) and arterial blood gas analysis data (arterial oxygen pressure, hemoglobin, and lactate) were recorded after induction, before CPB, after CPB, and at the end of surgery. The primary outcome was the development of AKI, which was defined as an absolute increase in sCr by Z50% or a decrease in eGFR by Z25% from the preoperative baseline within 5 days after surgery, according to the Risk, Injury, Failure, Loss, and End-stage (RIFLE) criteria.11 Postoperative sCr was recorded daily for 5 days after surgery when available, and the highest value was used for analysis. Renal replacement therapy (RRT) was started using continuous venovenous hemodialysis when life-threatening conditions such as severe fluid, electrolyte, and acid-base imbalance could not be managed by conservative treatment. The secondary outcomes included time to discharge from the ICU and hospital, RRT, and in-hospital mortality. Data are presented as the means ⫾ SDs, medians (interquartile range), or numbers of patients (%). Statistical analyses were performed using SPSS for Windows version 18.0 (SPSS Inc., Chicago, IL). Continuous variables were compared using the independent Student’s t test or Mann-Whitney U test. Categoric variables were analyzed using the chi-square test or Fisher’s exact test. To compare the changes in cerebral and renal rSO2 tracing using a linear mixed effect model, the lowest values of rSO2 during every minute were selected (i) from 5 minutes before induction until 55 minutes after induction, (ii) from 20 minutes before CPB until 60 minutes after CPB on, and (iii) from 60 minutes before CPB off to 100 minutes after CPB off. The same method was used to compare changes in renal rSO2 tracing between patients with AKI and those without AKI. To determine the predictive power and the best predictive threshold value of renal rSO2 for AKI, receiver operating characteristic (ROC) curve analysis was performed at multiple threshold values of renal rSO2. To calculate the statistical power (1–β) of the ROC curve analysis, power analysis was tested using PASS 11 (NCSS Statistical Software Co., Kaysville, Utah); null hypothesis is that renal rSO2 cannot predict the risk of AKI [area under the curve (AUC)–ROC is 0.5], and the number of positive groups (patients with AKI) and negative groups (patients without AKI), α (p value), and AUC-ROC that were evaluated in the ROC curve analysis were used. The relationship among renal rSO2 and the intraoperative vital sign and arterial blood gas analysis variables at each time point was determined using a linear regression analysis. Initially, all covariates (heart rate, mean arterial pressure, central venous pressure, mean pulmonary arterial pressure, cardiac index, SvO2, arterial oxygen pressure, hemoglobin, and lactate) were evaluated independently and were candidates for the multivariate linear regression model. A stepwise backward elimination process was used to develop the
Fig 1. The application of an NIRS sensor to the flank that lies above the kidney. (A) The location of the kidney was checked by sonography. (B) A NIRS sensor was applied to the probe placement site. (C) Sonographic findings for the long-axis image of the kidney. White arrows indicate the outline of the kidney. Black arrow indicates the measurement line of renal depth. NIRS, near-infrared spectroscopy.
3
RENAL RSO2 AND AKI
multivariate model. The relationship between AKI and the demographic and intraoperative variables including the total duration of renal rSO2 values below the best predictive threshold value were evaluated using univariate and stepwise multivariate logistic regression analysis. A p value less than 0.05 was considered statistically significant.
RESULTS
Between October 2010 and August 2011, 153 patients admitted to this hospital met the inclusion criteria (Fig 2). Of these, 38 were excluded because both renal depths were Z40 mm. An additional 15 patients refused to enroll in the study. Of 100 patients who were enrolled in this study, 5 were excluded because of incomplete renal rSO2 data. Thus, 95 patients were included in the final analysis; 66 patients received bilateral renal rSO2 monitoring and 29 patients received unilateral monitoring. Demographic and intraoperative characteristics are listed in Table 1. Postoperative AKI developed in 34 (35.8%) patients. In patients with AKI, the logistic EuroSCORE value was significantly higher (p ¼ 0.035), and surgery and CPB times were significantly longer than in those without AKI (p ¼ 0.012 and p ¼ 0.035, respectively). Patients with AKI tended to require blood transfusion during surgery more frequently than those without AKI (p ¼ 0.056). There were no other differences in demographic and intraoperative data between the groups. Patients with AKI had worse secondary outcomes, including longer lengths of ICU stay (p ¼ 0.047) and hospital stay (p o 0.001), and greater incidences of RRT (p ¼ 0.005) and in-hospital mortality (p ¼ 0.043) (Table 1). All patients who required RRT displayed a 2- to 4-fold increase in sCr from baseline within 2 days of surgery. Of 5 patients who required RRT, 3 patients died as a result of sepsis after bowel ischemia. The changes in renal rSO2 tracing observed during the anesthetic period were different from changes in cerebral rSO2 tracing (p o 0.001; Fig 3A). Renal rSO2 values, including the baseline levels, were higher than cerebral rSO2 values before
CPB. Immediately before and after the beginning of CPB, renal and cerebral rSO2 decreased because of unstable vital signs during CPB preparation procedure (cardiac manipulation and arterial and venous cannulation) and start of CPB. Renal rSO2 returned to baseline and then slowly decreased thereafter. In contrast, cerebral rSO2 remained lower than the baseline level throughout the CPB period. Furthermore, after weaning from CPB, renal rSO2 gradually decreased over time, whereas cerebral rSO2 returned to baseline levels. The changes in renal rSO2 tracing in patients with and without AKI are shown in Figure 3B. Although the changes in renal rSO2 tracing were similar in the 2 groups, renal rSO2 values in patients with AKI decreased more rapidly after returning to baseline levels during and after CPB when compared with those in patients without AKI (p o 0.001). The total periods when the renal rSO2 levels were lower than the threshold values in patients were significantly longer in patients with AKI than in those without AKI (p o 0.05 at every threshold values) (Table 2). In addition, the numbers of patients who had renal rSO2 levels less than 60%, 55%, 50%, and 20%, 25%, and 30% decline from baseline value were significantly greater in patients with AKI than in those without AKI (70.6% v 37.7%, p ¼ 0.002; 64.7% v 3.1%, p o 0.001; 47.1% v 3.3%, p o 0.001; 73.5% v 49.2%, p ¼ 0.030; 67.6% v 36.1%, p ¼ 0.005; 61.8% v 24.6%, p ¼ 0.001, respectively). ROC curve analysis revealed that the renal rSO2 concentration could predict the risk of AKI with statistical significance (Table 3) and that a renal rSO2 o55% (for more than 1.5 minutes) was the best predictive threshold value for AKI (AUC-ROC, 0.777; 95% CI, 0.669-0.885; p o 0.001) and had sufficient statistical power (1–β ¼ 0.8438). Multivariate linear regression analysis showed that renal rSO2 values significantly correlated with lactate and with SvO2 at every time point (Table 4). However, other intraoperative vital sign and arterial blood gas analysis variables had a lack of relationship with renal rSO2 at each time point. Univariate logistic regression analysis showed that AKI was associated significantly with logistic EuroSCORE, surgical time, blood transfusion, and duration of renal rSO2 o55%. Finally, multivariate analysis revealed that AKI significantly correlated with the duration of renal rSO2 o55% (p ¼ 0.002) and logistic EuroSCORE (p ¼ 0.007) (Table 5). DISCUSSION
Fig 2. Study inclusion/exclusion flow diagram. BMI, body mass index; ESRD, end-stage renal disease; KT, kidney transplantation; LVEF, left ventricular ejection fraction; OPCAB, off-pump coronary artery bypass surgery; sCr, serum creatinine.
In the current analysis the authors have found that intraoperative renal regional oxygen desaturation is associated significantly with postoperative AKI and that use of NIRS for continuous monitoring of the intraoperative renal rSO2 can predict the risk of AKI in adult patients undergoing cardiac surgery with CPB. Postoperative AKI is associated with increased morbidity and mortality in patients undergoing cardiac surgery.2,3 Therefore, early detection of deterioration in renal function and perfusion is important for effective management of patients with AKI after cardiac surgery. In the present study, slow decreases in renal rSO2 were observed during and after CPB. Moreover, renal rSO2 decreased more rapidly in patients with
4
CHOI ET AL
Table 1. Patient Characteristics and Postoperative Outcomes AKI (n ¼ 34)
Demographic data Age (y) Female gender Body mass index (kg/m2) Diabetes mellitus Hypertension Peripheral vascular disease Ejection fraction (%) Logistic EuroSCORE (%) Renal depth from skin (mm) Right Left Baseline renal rSO2 (%) Preoperative hemoglobin (mg/dL) Preoperative creatinine (mg/dL) Preoperative eGFR (mL/min/1.73m2) Intraoperative data Type of surgery Single valve* Multiple valve Valve þ CABG Other† Surgery time (min) Cardiopulmonary bypass time (min) Intraoperative fluid infusion Crystalloid (mL) Colloid (mL) Blood transfusion Urine output (mL) Minimum hemoglobin (mg/dL) Postoperative outcomes Maximum creatinine (mg/dL) Minimum eGFR (mL/min/1.73m2) Intensive care unit stay (h) Hospital stay (d) Renal replacement therapy In-hospital death
No AKI (n ¼ 61)
p Value
62.4 ⫾ 11.3 17 (50.0%) 23.3 ⫾ 3.2 1 (2.9%) 15 (44.1%) 2 (3.3%) 57.5 ⫾ 10.1 3.1 (2.1–4.3)
59.2 31 23.9 8 24 5 59.0 2.2
⫾ 10.1 (50.8%) ⫾ 2.9 (13.1%) (39.3%) (14.7%) ⫾ 8.4 (1.5–3.6)
0.156 1.000 0.337 0.151 0.669 0.093 0.489 0.035
29.6 ⫾ 8.7 31.4 ⫾ 8.6 77.7 ⫾ 11.6 12.6 ⫾ 1.7 0.84 ⫾ 0.20 84.8 ⫾ 20.3
29.8 30.2 81.3 13.3 0.82 87.6
⫾ 7.9 ⫾ 8.6 ⫾ 7.2 ⫾ 1.7 ⫾ 0.19 ⫾ 19.7
0.882 0.281 0.106 0.072 0.608 0.518 0.082
20 (58.8%) 7 (20.6%) 5 (14.7%) 2 (5.9%) 297.7 ⫾ 77.7 145.2 ⫾ 66.9
44 (72.1%) 11 (18.0%) 1 (1.6%) 5 (8.2%) 258.4 ⫾ 67.9 118.5 ⫾ 34.8
0.012 0.035
1307.4 ⫾ 443.3 775.0 ⫾ 354.7 13 (38.2%) 784.3 ⫾ 485.9 8.2 ⫾ 0.8
1252.5 ⫾ 505.8 674.6 ⫾ 311.5 12 (19.7%) 849.1 ⫾ 545.5 8.6 ⫾ 1.0
0.598 0.155 0.056 0.565 0.099
⫾ 0.27 ⫾ 18.6 (27–48) (6-10) 0 0
o0.001 o0.001 0.047 o0.001 0.005 0.043
1.91 ⫾ 0.93 38.2 ⫾ 16.2 48 (30-97) 14 (9-23) 5 (14.7%) 3 (8.8%)
0.96 74.0 45 8
NOTE. Values are expressed as the means ⫾ SD, medians (interquartile range), or numbers of patients (%).Grade of aortic atherosclerosis is evaluated using Davila-Roman’s grading system. Abbreviations: AKI, acute kidney injury; CABG, coronary artery bypass grafting; eGFR, estimated glomerular filtration rate; EuroSCORE, European System for Cardiac Operative Risk Evaluation. *“Valve” surgery included aortic valve replacement or repair, mitral valve replacement or repair, and tricuspid valve repair. †“Other” surgery included atrial septal defect repair, myxomectomy, and myomectomy.
AKI than in those without AKI. In addition, sustained low renal rSO2 values significantly predicted the risk of AKI after surgery. Recently, urinary biomarkers (NGAL, IL-18, and cystatin C) and renal NIRS have been demonstrated to differentiate patients with good versus poor outcomes in the early postoperative period after pediatric cardiac surgery.12 However, unlike biomarkers, which cannot be measured continuously and generally require a substantial amount of time for their assay and analysis, NIRS enables real-time monitoring. Thus, renal rSO2 monitoring may enable earlier prediction of AKI than biomarkers such as NGAL. Of the most recently identified early biomarkers of AKI, NGAL has shown the most promising results, with elevated levels of urinary NGAL detected within 1 or 2 hours after CPB having been shown to predict the risk of postoperative AKI.13,14
To date, NIRS usually is used to monitor cerebral rSO2 during cardiac surgery.8 Intraoperative cerebral oxygen desaturation is associated with poor neurologic outcomes such as postoperative cognitive dysfunction and stroke.8,15 However, NIRS monitoring is not used to monitor kidney function during adult cardiac surgery, and few studies have examined the use of NIRS to monitor renal oxygenation. The use of NIRS to measure low renal oximetry has been reported to correlate with renal dysfunction in infants undergoing cardiac surgery.9 In that previous study, low renal oximetry values were observed mainly during the first 24 hours after surgery, which preceded the peak sCr level by 48 hours. In contrast, in the present study the renal rSO2 concentration was measured during the anesthetic period but not during the postoperative period in adult patients. Hence, although the 2 studies differ in terms of the
5
RENAL RSO2 AND AKI
Fig 3. Changes in regional oxygen saturation (rSO2) tracing of the brain and kidney. (A) Changes in renal and cerebral rSO2 tracing (renal rSO2 v cerebral rSO2, p o 0.001). (B) Changes in renal rSO2 tracing of patients with and without AKI (AKI v No AKI, p o 0.001). AKI, acute kidney injury; CPB, cardiopulmonary bypass.
study populations and timing of rSO2 measurement, both suggest that NIRS monitoring can detect renal oxygen desaturation, which contributes to AKI.
Several intraoperative elements of cardiac surgery contribute to the risk of hypoperfusion and ischemia/reperfusion– mediated AKI.16 Embolism, CPB, low-output syndrome,
Table 2. Comparison of the Duration and Number of Patients at Each Threshold Renal rSO2 Value Duration of Renal rSO2 (min)
Threshold
Number of Patients
AKI
No AKI
(n ¼ 34)
(n ¼ 61)
p Value
43.1 ⫾ 53.4 21.4 ⫾ 38.4 8.9 ⫾ 23.3 3.3 ⫾ 15.3 0.7 ⫾ 4.1
0.001 0.001 o0.001 0.001 0.001
28 25 24 22 16
(82.4) (73.5) (70.6) (64.7) (47.1)
41 34 23 8 2
63.2 ⫾ 75.9 42.7 ⫾ 60.4 25.8 ⫾ 45.0 14.3 ⫾ 30.3 7.0 ⫾ 19.5
0.033 0.023 0.005 0.001 o0.001
29 26 25 23 21
(85.3) (76.5) (73.5) (67.6) (61.8)
46 37 30 22 15
Renal rSO2 o70% 112.9 ⫾ 103.8 o65% 80.0 ⫾ 86.1 o60% 56.9 ⫾ 65.6 o55% 38.7 ⫾ 54.6 o50% 26.2 ⫾ 41.9 Decline from baseline o10% decline 103.0 ⫾ 102.5 o15% decline 78.6 ⫾90.9 o20% decline 60.7 ⫾ 73.9 o25% decline 45.3 ⫾ 61.7 o30% decline 32.3 ⫾ 54.4
AKI
No AKI
(n ¼ 34)
(n ¼ 61)
Odds Ratio
95% CI
p Value
(67.2) (55.7) (37.7) (13.1) (3.3)
2.276 2.206 3.965 12.146 26.222
0.812-6.384 0.884-5.503 1.610-9.765 4.365-33.798 5.500-125.017
0.118 0.090 0.003 o0.001 o0.001
(75.4) (60.7) (49.2) (36.1) (24.6)
1.891 2.108 2.870 3.707 4.954
0.621-5.760 0.820-5.420 1.153-7.148 1.524-9.012 2.005-12.239
0.304 0.174 0.030 0.005 0.001
NOTE. Values are expressed as the mean ⫾ SD or number of patients (%).Abbreviations: AKI, acute kidney injury; renal rSO2, renal regional oxygen saturation.
6
CHOI ET AL
Table 3. Receiver Operating Characteristic (ROC) Curve Analysis for the Predictive Power of Renal rSO2 on AKI Prediction at Each Threshold Value Duration of Renal rSO2 (min) Threshold
AUC-ROC (95% CI)
Renal rSO2 o70% o65% o60% o55% o50% Decline from baseline o10% decline o15% decline o20% decline o25% decline o30% decline
Cutoff
Sensitivity (%)
Specificity (%)
p Value
Power (1β)
0.713 0.716 0.756 0.777 0.723
(0.598-0.829) (0.598-0.834) (0.643–0.868) (0.669-0.885) (0.606-0.840)
116 65 21 1.5 6
50.0 50.0 61.8 64.7 47.1
90.2 90.2 88.5 88.5 96.7
0.001 0.001 o0.001 o0.001 o0.001
0.5970 0.6169 0.6624 0.8438 0.9204
0.631 0.639 0.676 0.694 0.699
(0.513-0.748) (0.522-0.757) (0.561-0.791) (0.579-0.808) (0.584-0.813)
32 16.5 5.5 1 0.5
73.5 70.6 70.6 67.6 61.8
47.5 49.2 41.0 36.1 24.6
0.035 0.025 0.005 0.002 0.001
0.5048 0.5034 0.5174 0.5211 0.4704
NOTE. Values are expressed as means ⫾ SD. Abbreviations: AKI, acute kidney injury; AUC-ROC, area under the curve–receiver operating characteristic curves; renal rSO2, renal regional oxygen saturation.
anemia, hemodilution, and exogenous catecholamine can reduce arterial oxygen content or impair renal oxygen delivery, all of which may contribute to AKI.3,16–18 Therefore, it is logical to continuously monitor the renal rSO2 level using NIRS, which can measure regional tissue oxygen saturation. The multivariate linear regression analysis between renal rSO2 and intraoperative clinical factors indicates that the renal rSO2 concentration significantly correlates with lactate and SvO2, which are sensitive global indicators of tissue perfusion, despite the lack of an association with other clinical factors. The results are consistent with previous results that have demonstrated a strong relationship between rSO2 at various sites and the levels of blood lactate and SvO2.7,19 These analyses suggest that rSO2 measured by NIRS may be a useful indicator of impaired oxygen delivery to tissues. Thus, renal rSO2 may enable continuous monitoring of renal perfusion in real time, allowing the optimization of the renal condition. It should be noted that renal injury or failure after cardiac surgery is associated with numerous sources of kidney insult beyond renal ischemia, including nephrotoxicity and systemic inflammatory responses.16,18 Nephrotoxins include
contrast media, some antibiotics, aprotinin, free hemoglobin, and α-adrenergic agonist agents.20–24 Abrupt increases in the levels of circulatory inflammatory mediators, such as IL-1b, IL-6, and tumor necrosis factor α, can result from surgical trauma, CPB, endotoxemia, and ischemia-reperfusion injury.16,18,25 Unfortunately, NIRS devices cannot monitor these changes in local tissues. In addition, several postoperative clinical factors contribute to the development of AKI, which were not considered in this study. For these reasons, the ROC curve analysis in the results revealed the relatively low sensitivities (47%-65%) at each of the threshold values examined. Of the 26 patients who never displayed renal rSO2 values o70%, 6 developed AKI after surgery. To overcome these shortcomings, it may be helpful to use other AKI biomarkers simultaneously during and after surgery. This study had several limitations. First, the inclusion criteria were highly selective in order to permit the correct measurement of the renal rSO2 level. In general, NIRS can measure the regional level of tissue oxygen saturation in a local tissue area located 3 to 4 cm beneath the commercially available sensor.7 For this reason, 38 patients who had both
Table 4. Relationship Between Renal rSO2 and Intraoperative Vital Sign and Arterial Blood Gas Analysis Variables at Each Time Point Renal rSO2 After Induction
Before CPB
After CPB
End of surgery
Univariate p
Multivariate p
Univariate p
Multivariate p
Univariate p
Multivariate p
Univariate p
Multivariate p
Variable
Value
Value
Value
value
Value
Value
value
value
Heart rate MBP CVP MPAP CI SvO2 PaO2 Hematocrit Lactate
0.096 0.471 0.596 0.735 0.258 0.061 0.230 0.728 0.038
— — — — — 0.027 — — 0.017
0.669 0.932 0.263 0.889 0.610 0.111 0.368 0.493 0.049
— — — — — 0.043 — — 0.020
0.083 0.391 0.828 0.563 0.010 o0.001 0.002 0.755 0.001
— — — — — o0.001 — — 0.001
0.328 0.749 0.869 0.656 0.125 0.002 0.061 0.379 0.036
— — — — — 0.003 — — 0.019
Abbreviations: CI, cardiac index; CPB, cardiopulmonary bypass; CVP, central venous pressure; MBP, mean blood pressure; MPAP, mean pulmonary arterial pressure; PaO2, arterial oxygen pressure; rSO2, regional oxygen saturation; SvO2, mixed venous oxygen saturation.
7
RENAL RSO2 AND AKI
Table 5. Logistic Regression Analysis for Predictors of Acute Kidney Injury β Variable
Univariate Age Gender Body mass index Diabetes mellitus Hypertension Peripheral vascular disease Ejection fraction Logistic EuroSCORE Baseline renal rSO2 Preoperative hemoglobin Preoperative creatinine Type of surgery Surgery time Cardiopulmonary bypass time Intraoperative fluid infusion Crystalloid Colloid Urine output Blood transfusion Intraoperative minimum hemoglobin Duration of renal rSO2 o 55% Multivariate Logistic EuroSCORE Duration of renal rSO2 o 55%
p
Coefficient
95% CI
Value
— — — — — 5.086 — 1.223 — — — — 1.008 —
— — — — — 0.930-27.814 — 1.036-1.444 — — — — 1.001-1.014 —
0.153 0.939 0.332 0.105 0.650 0.061 0.866 0.017 0.070 0.076 0.674 0.583 0.018 0.212
— — — 2.528 — 1.046
— — — 0.991-6.448 — 1.017-1.076
0.593 0.152 0.561 0.052 0.101 0.002
1.297 1.047
1.073-1.567 1.017-1.077
0.007 0.002
Abbreviations: EuroSCORE, European System for Cardiac Operative Risk Evaluation; rSO2, regional oxygen saturation.
general population (46 mm).26 Therefore, a sensor that can measure rSO2 in the deeper tissue area will be needed to monitor renal rSO2 in all patients without restriction. Second, there was a possibility that the NIRS sensor used could not measure the regional tissue oxygen saturation of the true renal area (renal cortex and medulla). A NIRS sensor might measure rSO2 of the cortical renal area in patients who had renal depth o30 mm. On the other hand, the measurement might be contaminated by rSO2 of the near renal area (perinephric and paranephric fat tissue surrounding the kidney) in those who had renal depth 430 mm. Despite this limitation, the results indicated that the decrease in renal rSO2 level was associated significantly with postoperative AKI and predicted the risk of AKI. However, further studies with a sensor measuring the deeper tissue area are needed to confirm these results. CONCLUSION
In conclusion, the current prospective observational study indicated a predictive ability of intraoperative renal rSO2 monitoring for postoperative AKI in adult patients undergoing cardiac surgery. However, the use of an NIRS sensor that can monitor tissue areas deeper below the skin surface is warranted to generalize the use of this method to all patients undergoing cardiac surgery. In addition, further studies are needed to evaluate effective interventions to correct renal regional oxygen desaturation. ACKNOWLEDGMENTS
renal depths Z40 mm were excluded from the study despite meeting other inclusion criteria. The average renal depths of the study population were 29.7 mm for the right kidney and 30.6 mm for the left kidney. These depths are less than those of the
This study was partly supported by an unrestricted grant by E-Wha Biomedics Co., Seoul, Korea. The authors thank Seungbong Han, PhD, Statistician, Department of Clinical Epidemiology and Biostatistics, Asan Medical Center, Seoul, Korea, for valuable statistical advice and assistance.
REFERENCES 1. Chew ST, Mar WM, Ti LK: Association of ethnicity and acute kidney injury after cardiac surgery in a South East Asian population. Br J Anaesth 110(3):397-401, 2012 2. Hobson CE, Yavas S, Segal MS, et al: Acute kidney injury is associated with increased long-term mortality after cardiothoracic surgery. Circulation 119:2444-2453, 2009 3. Karkouti K, Wijeysundera DN, Yau TM, et al: Acute kidney injury after cardiac surgery: focus on modifiable risk factors. Circulation 119:495-502, 2009 4. Lassnigg A, Schmidlin D, Mouhieddine M, et al: Minimal changes of serum creatinine predict prognosis in patients after cardiothoracic surgery: a prospective cohort study. J Am Soc Nephrol 15:1597-1605, 2004 5. Martensson J, Martling CR, Bell M: Novel biomarkers of acute kidney injury and failure: clinical applicability. Br J Anaesth 109: 843-850, 2012 6. Wyckoff T, Augoustides JG: Advances in acute kidney injury associated with cardiac surgery: the unfolding revolution in early detection. J Cardiothorac Vasc Anesth 26:340-345, 2012 7. Moerman A, Vandenplas G, Bove T, et al: Relation between mixed venous oxygen saturation and cerebral oxygen saturation measured by absolute and relative near-infrared spectroscopy during
off-pump coronary artery bypass grafting. Br J Anaesth 110(2): 258-265, 2012 8. Murkin JM, Arango M: Near-infrared spectroscopy as an index of brain and tissue oxygenation. Br J Anaesth 103(Suppl 1):i3-13, 2009 9. Owens GE, King K, Gurney JG, et al: Low renal oximetry correlates with acute kidney injury after infant cardiac surgery. Pediatr Cardiol 32:183-188, 2011 10. Levey AS, Bosch JP, Lewis JB, et al: A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med 130:461-470, 1999 11. Bellomo R, Ronco C, Kellum JA, et al: Acute renal failure— definition, outcome measures, animal models, fluid therapy and information technology needs: the Second International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group. Crit Care 8:R204-R212, 2004 12. Hazle MA, Gajarski RJ, Aiyagari R, et al: Urinary biomarkers and renal near-infrared spectroscopy predict intensive care unit outcomes after cardiac surgery in infants younger than 6 months of age. J Thorac Cardiovasc Surg 146(4):861-867, 2013 Oct 13. Wagener G, Jan M, Kim M, et al: Association between increases in urinary neutrophil gelatinase-associated lipocalin and acute renal
8
dysfunction after adult cardiac surgery. Anesthesiology 105:485-491, 2006 14. Mishra J, Dent C, Tarabishi R, et al: Neutrophil gelatinaseassociated lipocalin (NGAL) as a biomarker for acute renal injury after cardiac surgery. Lancet 365:1231-1238, 2005 15. Selnes OA, Gottesman RF, Grega MA, et al: Cognitive and neurologic outcomes after coronary-artery bypass surgery. N Engl J Med 366:250-257, 2012 16. Stafford-Smith M, Patel UD, Phillips-Bute BG, et al: Acute kidney injury and chronic kidney disease after cardiac surgery. Adv Chronic Kidney Dis 15:257-277, 2008 17. Shander A, Javidroozi M, Ozawa S, et al: What is really dangerous: anaemia or transfusion? Br J Anaesth 107(Suppl 1):i41-i59, 2011 18. Sear JW: Kidney dysfunction in the postoperative period. Br J Anaesth 95:20-32, 2005 19. Chakravarti SB, Mittnacht AJ, Katz JC, et al: Multisite nearinfrared spectroscopy predicts elevated blood lactate level in children after cardiac surgery. J Cardiothorac Vasc Anesth 23:663-667, 2009 20. Wong GT, Irwin MG: Contrast-induced nephropathy. Br J Anaesth 99:474-483, 2007
CHOI ET AL
21. Nielsen DV, Hjortdal V, Larsson H, et al: Perioperative aminoglycoside treatment is associated with a higher incidence of postoperative dialysis in adult cardiac surgery patients. J Thorac Cardiovasc Surg 142:656-661, 2011 22. Shaw AD, Stafford-Smith M, White WD, et al: The effect of aprotinin on outcome after coronary-artery bypass grafting. N Engl J Med 358:784-793, 2008 23. Haase M, Haase-Fielitz A, Bagshaw SM, et al: Cardiopulmonary bypass-associated acute kidney injury: a pigment nephropathy? Contrib Nephrol 156:340-353, 2007 24. Santos FO, Silveira MA, Maia RB, et al: Acute renal failure after coronary artery bypass surgery with extracorporeal circulation— incidence, risk factors, and mortality. Arq Bras Cardiol 83:150-154, 2004 25. Granfeldt A, Ebdrup L, Tønnesen E, et al: Renal cytokine profile in an endotoxemic porcine model. Acta Anaesthesiol Scand 52: 614-620, 2008 26. Bolton WK, Tully RJ, Lewis EJ, et al: Localization of the kidney for percutaneous biopsy. A comparative study of methods. Ann Intern Med 81:159-164, 1974
significantly longer in patients with AKI than in those without AKI (p ¼ 0.001 or p o 0.001). Receiver operating characteristic (ROC) curve analysis was used to evaluate the predictive power of renal rSO2 for AKI. The ROC curve analysis showed that renal rSO2 could predict the risk of AKI with statistical significance and that a renal rSO2 o 55% had the best performance (area under the curve–ROC, 0.777; 95% CI, 0.6690.885; p o 0.001). Multivariate logistic regression analysis revealed that AKI significantly correlated with the duration of renal rSO2 o 55% (p ¼ 0.002) and logistic EuroSCORE (p ¼ 0.007). Conclusions: Intraoperative renal regional oxygen desaturation can be a good predictor of AKI in adult patients undergoing cardiac surgery. & 2014 Elsevier Inc. All rights reserved.
A
elective cardiac surgery with cardiopulmonary bypass (CPB) were enrolled between October 2010 and August 2011. Exclusion criteria were off-pump coronary artery bypass surgery (OPCAB), aortic surgery (total arch replacement and/or descending aorta replacement surgery), emergency surgery, left ventricular ejection fraction (LVEF) o 30%, body mass index (BMI) Z 30 kg/m2, sCr Z 1.5 mg/dL, and end-stage renal disease or renal transplant. Anesthesia was induced with etomidate (0.2 mg/kg), followed by rocuronium (0.8 mg/kg) to facilitate tracheal intubation, and was maintained with a continuous infusion of propofol (1-2 μg/mL) and remifentanil (5-20 ng/mL) using a target-controlled infusion pump (Orchestra Base Primea; Fresenius Kabi, Paris, France). Normal saline and 6% hydroxyethyl starch 130/0.4 (Voluven, Fresenius Kabi, Bad Homburg, Germany) were used for fluid management, and red blood cell (RBC) transfusion was initiated when hemoglobin level was o8 g/dL. A 20-G radial arterial catheter was inserted to monitor arterial blood pressure and to allow blood sampling. A 7.5-Fr pulmonary arterial catheter (Swan-Ganz Combo V CCL/SvO2/CEDV; Edwards Lifesciences, Irvine, CA) was advanced and connected to a Vigilance II device (Edwards Lifesciences) to monitor cardiac output and mixed venous oxygen saturation (SvO2). After heparinization (400 U/kg of heparin), CPB was started using a roller pump (Stöckert SIII; Stöckert Instrumente, Munich, Germany), with the tubing system without surface coating (Medtronic, Minneapolis, MN, USA), a membrane oxygenator (Affinity; Medtronic), and a 38-μm blood filter (Medtronic). The CPB flow rate was maintained between 2.2 and 2.5 L/min/m2, and the mean arterial pressure during CPB was maintained at 50 to 80
CUTE KIDNEY INJURY (AKI) is a common and serious complication of cardiac surgery, with a reported incidence as high as 45%.1,2 It is associated with poor outcomes, prolonged stays in hospital and in an intensive care unit (ICU), and increased mortality.2,3 Small increases in the levels of serum creatinine (sCr) reflect renal injury and predict poor outcomes.4 Nevertheless, given that a detectable increase in the levels of sCr does not immediately follow the onset of AKI, sCr levels alone are of limited value for the early identification of the risk of AKI. Biomarkers of early renal injury, such as neutrophil gelatinase-associated lipocalin (NGAL), cystatin C, kidney injury molecule 1, and interleukin (IL)-18, recently have been shown to facilitate earlier diagnosis of AKI.5,6 Although these biomarkers show some benefits for early detection of AKI, they cannot be measured continuously to monitor renal function in real time, which is needed to optimize the renal condition. Near-infrared spectroscopy (NIRS) is a new noninvasive technique that continuously monitors regional oxygen saturation (rSO2) by measuring the relative concentrations of oxygenated and deoxygenated hemoglobin within a local tissue area.7 Previous studies have suggested that low rSO2 values can predict the development of organ dysfunction in the brain or other organs after surgery.8 Persistent low renal rSO2 has been associated with renal dysfunction after pediatric cardiac surgery.9 However, to date, the relationship between intraoperative renal rSO2 and AKI in adult patients undergoing cardiac surgery has not been determined. The authors hypothesized that renal rSO2 measured by NIRS can reflect renal desaturation or hypoxia, which may contribute to AKI. The aim of this study was to examine whether intraoperative renal rSO2 could predict the risk of AKI after cardiac surgery. METHODS After approval by the Institutional Ethics Committee and obtaining written informed consent, a total of 100 adult patients who underwent
KEY WORDS: acute kidney injury, near-infrared spectroscopy, cardiac surgery
From the Department of Anesthesiology and Pain Medicine, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea. Address reprint requests to In-Cheol Choi, MD, Department of Anesthesiology and Pain Medicine, Asan Medical Center, University of Ulsan College of Medicine, 388-1, Pungnap 2-Dong, Songpa-gu, Seoul 138-736, Korea E-mail: [email protected] © 2014 Elsevier Inc. All rights reserved. 1053-0770/2601-0001$36.00/0 http://dx.doi.org/10.1053/j.jvca.2013.12.005
Journal of Cardiothoracic and Vascular Anesthesia, Vol ], No ] (Month), 2014: pp ]]]–]]]
1
2
mmHg. Alpha-stat pH management was used during CPB. After weaning from CPB, heparin was antagonized by protamine sulfate (4 mg/kg). All patients were transferred to the ICU after surgery, where care was provided by ICU nurses and staff who were not involved in the study in accordance with existing unit protocol. Patients were discharged from the ICU to a general ward when their vital signs were stable and further ICU care and monitoring were not required. Renal rSO2 was monitored with NIRS (INVOS 5100C; Somanetics Co., Troy, MI) throughout the entire anesthetic period. Before the induction of anesthesia, a disposable NIRS sensor was applied to one or both sides of the flank area that overlies the kidney in order to monitor renal rSO2 under sonographic guidance (Z.one Ultra; ZONARE Medical Systems Inc., Mountain View, CA) (Fig 1). To visualize the kidney, an ultrasound probe was placed in the lower intercostal space (below either the 10th or 11th rib) on the posterior axillary line. After obtaining the long-axis image of the kidney, the renal depth (the distance from the skin surface to the capsule of the kidney) was measured. A sensor then was applied to the probe placement site in instances when the renal depth was o40 mm. Patients were excluded from the study if both renal depths were Z40 mm. Renal rSO2 values were recorded continuously at 5- or 6-second intervals. Whenever renal rSO2 was monitored on both sides, the lower value was selected for analysis. The authors believe that the lower value is clinically more important because selecting the higher value or mean value may mask existing hypoxia, which may contribute to AKI. The sensitivity tests showed that the predictive ability of data including the lower value are better than that of data including the higher or mean value (data not shown). The patient whose monitoring of renal rSO2 was not checked continuously for more than 5 minutes during surgery was excluded from the study. No interventions were attempted based on renal rSO2 values. Cerebral rSO2 also was recorded simultaneously. Recorded rSO2 data were analyzed using Excel software (Microsoft, Seattle, WA). Renal rSO2 data were used to calculate the total duration of renal rSO2 below the threshold value. Multiple thresholds were selected to reflect different degrees of renal hypoxia. The thresholds used were renal rSO2 values of 70%, 65%, 60%, 55%, and 50%. These represent the approximate mean value of the baseline (80.0 ⫾ 9.1%) minus 1, 1.5, 2, 2.5, and 3 standard deviations (SDs), respectively. In addition, the other thresholds used were the proportionate decline form baseline value (10%, 15%, 20%, 25%, and 30%). Demographic data included age, sex, BMI, diabetes mellitus, hypertension, peripheral vascular disease, LVEF, logistic EuroSCORE (European System for Cardiac Operative Risk Evaluation), renal depth, baseline renal rSO2, preoperative hemoglobin, preoperative sCr, and preoperative estimated glomerular filtration rate (eGFR). The eGFR was calculated using the modification of diet in the renal disease (MDRD) formula [eGFR ¼ 186 creatinine –1.154 age –0.203 0.742 (if female)].10 Intraoperative variables included the type of surgery, surgical
CHOI ET AL
time, CPB time, fluid infusion volume (crystalloid and colloid), urine output, number of patients requiring blood transfusion, and intraoperative minimum hemoglobin. The intraoperative vital signs (heart rate, mean arterial pressure, central venous pressure, mean pulmonary arterial pressure, cardiac index, and SvO2) and arterial blood gas analysis data (arterial oxygen pressure, hemoglobin, and lactate) were recorded after induction, before CPB, after CPB, and at the end of surgery. The primary outcome was the development of AKI, which was defined as an absolute increase in sCr by Z50% or a decrease in eGFR by Z25% from the preoperative baseline within 5 days after surgery, according to the Risk, Injury, Failure, Loss, and End-stage (RIFLE) criteria.11 Postoperative sCr was recorded daily for 5 days after surgery when available, and the highest value was used for analysis. Renal replacement therapy (RRT) was started using continuous venovenous hemodialysis when life-threatening conditions such as severe fluid, electrolyte, and acid-base imbalance could not be managed by conservative treatment. The secondary outcomes included time to discharge from the ICU and hospital, RRT, and in-hospital mortality. Data are presented as the means ⫾ SDs, medians (interquartile range), or numbers of patients (%). Statistical analyses were performed using SPSS for Windows version 18.0 (SPSS Inc., Chicago, IL). Continuous variables were compared using the independent Student’s t test or Mann-Whitney U test. Categoric variables were analyzed using the chi-square test or Fisher’s exact test. To compare the changes in cerebral and renal rSO2 tracing using a linear mixed effect model, the lowest values of rSO2 during every minute were selected (i) from 5 minutes before induction until 55 minutes after induction, (ii) from 20 minutes before CPB until 60 minutes after CPB on, and (iii) from 60 minutes before CPB off to 100 minutes after CPB off. The same method was used to compare changes in renal rSO2 tracing between patients with AKI and those without AKI. To determine the predictive power and the best predictive threshold value of renal rSO2 for AKI, receiver operating characteristic (ROC) curve analysis was performed at multiple threshold values of renal rSO2. To calculate the statistical power (1–β) of the ROC curve analysis, power analysis was tested using PASS 11 (NCSS Statistical Software Co., Kaysville, Utah); null hypothesis is that renal rSO2 cannot predict the risk of AKI [area under the curve (AUC)–ROC is 0.5], and the number of positive groups (patients with AKI) and negative groups (patients without AKI), α (p value), and AUC-ROC that were evaluated in the ROC curve analysis were used. The relationship among renal rSO2 and the intraoperative vital sign and arterial blood gas analysis variables at each time point was determined using a linear regression analysis. Initially, all covariates (heart rate, mean arterial pressure, central venous pressure, mean pulmonary arterial pressure, cardiac index, SvO2, arterial oxygen pressure, hemoglobin, and lactate) were evaluated independently and were candidates for the multivariate linear regression model. A stepwise backward elimination process was used to develop the
Fig 1. The application of an NIRS sensor to the flank that lies above the kidney. (A) The location of the kidney was checked by sonography. (B) A NIRS sensor was applied to the probe placement site. (C) Sonographic findings for the long-axis image of the kidney. White arrows indicate the outline of the kidney. Black arrow indicates the measurement line of renal depth. NIRS, near-infrared spectroscopy.
3
RENAL RSO2 AND AKI
multivariate model. The relationship between AKI and the demographic and intraoperative variables including the total duration of renal rSO2 values below the best predictive threshold value were evaluated using univariate and stepwise multivariate logistic regression analysis. A p value less than 0.05 was considered statistically significant.
RESULTS
Between October 2010 and August 2011, 153 patients admitted to this hospital met the inclusion criteria (Fig 2). Of these, 38 were excluded because both renal depths were Z40 mm. An additional 15 patients refused to enroll in the study. Of 100 patients who were enrolled in this study, 5 were excluded because of incomplete renal rSO2 data. Thus, 95 patients were included in the final analysis; 66 patients received bilateral renal rSO2 monitoring and 29 patients received unilateral monitoring. Demographic and intraoperative characteristics are listed in Table 1. Postoperative AKI developed in 34 (35.8%) patients. In patients with AKI, the logistic EuroSCORE value was significantly higher (p ¼ 0.035), and surgery and CPB times were significantly longer than in those without AKI (p ¼ 0.012 and p ¼ 0.035, respectively). Patients with AKI tended to require blood transfusion during surgery more frequently than those without AKI (p ¼ 0.056). There were no other differences in demographic and intraoperative data between the groups. Patients with AKI had worse secondary outcomes, including longer lengths of ICU stay (p ¼ 0.047) and hospital stay (p o 0.001), and greater incidences of RRT (p ¼ 0.005) and in-hospital mortality (p ¼ 0.043) (Table 1). All patients who required RRT displayed a 2- to 4-fold increase in sCr from baseline within 2 days of surgery. Of 5 patients who required RRT, 3 patients died as a result of sepsis after bowel ischemia. The changes in renal rSO2 tracing observed during the anesthetic period were different from changes in cerebral rSO2 tracing (p o 0.001; Fig 3A). Renal rSO2 values, including the baseline levels, were higher than cerebral rSO2 values before
CPB. Immediately before and after the beginning of CPB, renal and cerebral rSO2 decreased because of unstable vital signs during CPB preparation procedure (cardiac manipulation and arterial and venous cannulation) and start of CPB. Renal rSO2 returned to baseline and then slowly decreased thereafter. In contrast, cerebral rSO2 remained lower than the baseline level throughout the CPB period. Furthermore, after weaning from CPB, renal rSO2 gradually decreased over time, whereas cerebral rSO2 returned to baseline levels. The changes in renal rSO2 tracing in patients with and without AKI are shown in Figure 3B. Although the changes in renal rSO2 tracing were similar in the 2 groups, renal rSO2 values in patients with AKI decreased more rapidly after returning to baseline levels during and after CPB when compared with those in patients without AKI (p o 0.001). The total periods when the renal rSO2 levels were lower than the threshold values in patients were significantly longer in patients with AKI than in those without AKI (p o 0.05 at every threshold values) (Table 2). In addition, the numbers of patients who had renal rSO2 levels less than 60%, 55%, 50%, and 20%, 25%, and 30% decline from baseline value were significantly greater in patients with AKI than in those without AKI (70.6% v 37.7%, p ¼ 0.002; 64.7% v 3.1%, p o 0.001; 47.1% v 3.3%, p o 0.001; 73.5% v 49.2%, p ¼ 0.030; 67.6% v 36.1%, p ¼ 0.005; 61.8% v 24.6%, p ¼ 0.001, respectively). ROC curve analysis revealed that the renal rSO2 concentration could predict the risk of AKI with statistical significance (Table 3) and that a renal rSO2 o55% (for more than 1.5 minutes) was the best predictive threshold value for AKI (AUC-ROC, 0.777; 95% CI, 0.669-0.885; p o 0.001) and had sufficient statistical power (1–β ¼ 0.8438). Multivariate linear regression analysis showed that renal rSO2 values significantly correlated with lactate and with SvO2 at every time point (Table 4). However, other intraoperative vital sign and arterial blood gas analysis variables had a lack of relationship with renal rSO2 at each time point. Univariate logistic regression analysis showed that AKI was associated significantly with logistic EuroSCORE, surgical time, blood transfusion, and duration of renal rSO2 o55%. Finally, multivariate analysis revealed that AKI significantly correlated with the duration of renal rSO2 o55% (p ¼ 0.002) and logistic EuroSCORE (p ¼ 0.007) (Table 5). DISCUSSION
Fig 2. Study inclusion/exclusion flow diagram. BMI, body mass index; ESRD, end-stage renal disease; KT, kidney transplantation; LVEF, left ventricular ejection fraction; OPCAB, off-pump coronary artery bypass surgery; sCr, serum creatinine.
In the current analysis the authors have found that intraoperative renal regional oxygen desaturation is associated significantly with postoperative AKI and that use of NIRS for continuous monitoring of the intraoperative renal rSO2 can predict the risk of AKI in adult patients undergoing cardiac surgery with CPB. Postoperative AKI is associated with increased morbidity and mortality in patients undergoing cardiac surgery.2,3 Therefore, early detection of deterioration in renal function and perfusion is important for effective management of patients with AKI after cardiac surgery. In the present study, slow decreases in renal rSO2 were observed during and after CPB. Moreover, renal rSO2 decreased more rapidly in patients with
4
CHOI ET AL
Table 1. Patient Characteristics and Postoperative Outcomes AKI (n ¼ 34)
Demographic data Age (y) Female gender Body mass index (kg/m2) Diabetes mellitus Hypertension Peripheral vascular disease Ejection fraction (%) Logistic EuroSCORE (%) Renal depth from skin (mm) Right Left Baseline renal rSO2 (%) Preoperative hemoglobin (mg/dL) Preoperative creatinine (mg/dL) Preoperative eGFR (mL/min/1.73m2) Intraoperative data Type of surgery Single valve* Multiple valve Valve þ CABG Other† Surgery time (min) Cardiopulmonary bypass time (min) Intraoperative fluid infusion Crystalloid (mL) Colloid (mL) Blood transfusion Urine output (mL) Minimum hemoglobin (mg/dL) Postoperative outcomes Maximum creatinine (mg/dL) Minimum eGFR (mL/min/1.73m2) Intensive care unit stay (h) Hospital stay (d) Renal replacement therapy In-hospital death
No AKI (n ¼ 61)
p Value
62.4 ⫾ 11.3 17 (50.0%) 23.3 ⫾ 3.2 1 (2.9%) 15 (44.1%) 2 (3.3%) 57.5 ⫾ 10.1 3.1 (2.1–4.3)
59.2 31 23.9 8 24 5 59.0 2.2
⫾ 10.1 (50.8%) ⫾ 2.9 (13.1%) (39.3%) (14.7%) ⫾ 8.4 (1.5–3.6)
0.156 1.000 0.337 0.151 0.669 0.093 0.489 0.035
29.6 ⫾ 8.7 31.4 ⫾ 8.6 77.7 ⫾ 11.6 12.6 ⫾ 1.7 0.84 ⫾ 0.20 84.8 ⫾ 20.3
29.8 30.2 81.3 13.3 0.82 87.6
⫾ 7.9 ⫾ 8.6 ⫾ 7.2 ⫾ 1.7 ⫾ 0.19 ⫾ 19.7
0.882 0.281 0.106 0.072 0.608 0.518 0.082
20 (58.8%) 7 (20.6%) 5 (14.7%) 2 (5.9%) 297.7 ⫾ 77.7 145.2 ⫾ 66.9
44 (72.1%) 11 (18.0%) 1 (1.6%) 5 (8.2%) 258.4 ⫾ 67.9 118.5 ⫾ 34.8
0.012 0.035
1307.4 ⫾ 443.3 775.0 ⫾ 354.7 13 (38.2%) 784.3 ⫾ 485.9 8.2 ⫾ 0.8
1252.5 ⫾ 505.8 674.6 ⫾ 311.5 12 (19.7%) 849.1 ⫾ 545.5 8.6 ⫾ 1.0
0.598 0.155 0.056 0.565 0.099
⫾ 0.27 ⫾ 18.6 (27–48) (6-10) 0 0
o0.001 o0.001 0.047 o0.001 0.005 0.043
1.91 ⫾ 0.93 38.2 ⫾ 16.2 48 (30-97) 14 (9-23) 5 (14.7%) 3 (8.8%)
0.96 74.0 45 8
NOTE. Values are expressed as the means ⫾ SD, medians (interquartile range), or numbers of patients (%).Grade of aortic atherosclerosis is evaluated using Davila-Roman’s grading system. Abbreviations: AKI, acute kidney injury; CABG, coronary artery bypass grafting; eGFR, estimated glomerular filtration rate; EuroSCORE, European System for Cardiac Operative Risk Evaluation. *“Valve” surgery included aortic valve replacement or repair, mitral valve replacement or repair, and tricuspid valve repair. †“Other” surgery included atrial septal defect repair, myxomectomy, and myomectomy.
AKI than in those without AKI. In addition, sustained low renal rSO2 values significantly predicted the risk of AKI after surgery. Recently, urinary biomarkers (NGAL, IL-18, and cystatin C) and renal NIRS have been demonstrated to differentiate patients with good versus poor outcomes in the early postoperative period after pediatric cardiac surgery.12 However, unlike biomarkers, which cannot be measured continuously and generally require a substantial amount of time for their assay and analysis, NIRS enables real-time monitoring. Thus, renal rSO2 monitoring may enable earlier prediction of AKI than biomarkers such as NGAL. Of the most recently identified early biomarkers of AKI, NGAL has shown the most promising results, with elevated levels of urinary NGAL detected within 1 or 2 hours after CPB having been shown to predict the risk of postoperative AKI.13,14
To date, NIRS usually is used to monitor cerebral rSO2 during cardiac surgery.8 Intraoperative cerebral oxygen desaturation is associated with poor neurologic outcomes such as postoperative cognitive dysfunction and stroke.8,15 However, NIRS monitoring is not used to monitor kidney function during adult cardiac surgery, and few studies have examined the use of NIRS to monitor renal oxygenation. The use of NIRS to measure low renal oximetry has been reported to correlate with renal dysfunction in infants undergoing cardiac surgery.9 In that previous study, low renal oximetry values were observed mainly during the first 24 hours after surgery, which preceded the peak sCr level by 48 hours. In contrast, in the present study the renal rSO2 concentration was measured during the anesthetic period but not during the postoperative period in adult patients. Hence, although the 2 studies differ in terms of the
5
RENAL RSO2 AND AKI
Fig 3. Changes in regional oxygen saturation (rSO2) tracing of the brain and kidney. (A) Changes in renal and cerebral rSO2 tracing (renal rSO2 v cerebral rSO2, p o 0.001). (B) Changes in renal rSO2 tracing of patients with and without AKI (AKI v No AKI, p o 0.001). AKI, acute kidney injury; CPB, cardiopulmonary bypass.
study populations and timing of rSO2 measurement, both suggest that NIRS monitoring can detect renal oxygen desaturation, which contributes to AKI.
Several intraoperative elements of cardiac surgery contribute to the risk of hypoperfusion and ischemia/reperfusion– mediated AKI.16 Embolism, CPB, low-output syndrome,
Table 2. Comparison of the Duration and Number of Patients at Each Threshold Renal rSO2 Value Duration of Renal rSO2 (min)
Threshold
Number of Patients
AKI
No AKI
(n ¼ 34)
(n ¼ 61)
p Value
43.1 ⫾ 53.4 21.4 ⫾ 38.4 8.9 ⫾ 23.3 3.3 ⫾ 15.3 0.7 ⫾ 4.1
0.001 0.001 o0.001 0.001 0.001
28 25 24 22 16
(82.4) (73.5) (70.6) (64.7) (47.1)
41 34 23 8 2
63.2 ⫾ 75.9 42.7 ⫾ 60.4 25.8 ⫾ 45.0 14.3 ⫾ 30.3 7.0 ⫾ 19.5
0.033 0.023 0.005 0.001 o0.001
29 26 25 23 21
(85.3) (76.5) (73.5) (67.6) (61.8)
46 37 30 22 15
Renal rSO2 o70% 112.9 ⫾ 103.8 o65% 80.0 ⫾ 86.1 o60% 56.9 ⫾ 65.6 o55% 38.7 ⫾ 54.6 o50% 26.2 ⫾ 41.9 Decline from baseline o10% decline 103.0 ⫾ 102.5 o15% decline 78.6 ⫾90.9 o20% decline 60.7 ⫾ 73.9 o25% decline 45.3 ⫾ 61.7 o30% decline 32.3 ⫾ 54.4
AKI
No AKI
(n ¼ 34)
(n ¼ 61)
Odds Ratio
95% CI
p Value
(67.2) (55.7) (37.7) (13.1) (3.3)
2.276 2.206 3.965 12.146 26.222
0.812-6.384 0.884-5.503 1.610-9.765 4.365-33.798 5.500-125.017
0.118 0.090 0.003 o0.001 o0.001
(75.4) (60.7) (49.2) (36.1) (24.6)
1.891 2.108 2.870 3.707 4.954
0.621-5.760 0.820-5.420 1.153-7.148 1.524-9.012 2.005-12.239
0.304 0.174 0.030 0.005 0.001
NOTE. Values are expressed as the mean ⫾ SD or number of patients (%).Abbreviations: AKI, acute kidney injury; renal rSO2, renal regional oxygen saturation.
6
CHOI ET AL
Table 3. Receiver Operating Characteristic (ROC) Curve Analysis for the Predictive Power of Renal rSO2 on AKI Prediction at Each Threshold Value Duration of Renal rSO2 (min) Threshold
AUC-ROC (95% CI)
Renal rSO2 o70% o65% o60% o55% o50% Decline from baseline o10% decline o15% decline o20% decline o25% decline o30% decline
Cutoff
Sensitivity (%)
Specificity (%)
p Value
Power (1β)
0.713 0.716 0.756 0.777 0.723
(0.598-0.829) (0.598-0.834) (0.643–0.868) (0.669-0.885) (0.606-0.840)
116 65 21 1.5 6
50.0 50.0 61.8 64.7 47.1
90.2 90.2 88.5 88.5 96.7
0.001 0.001 o0.001 o0.001 o0.001
0.5970 0.6169 0.6624 0.8438 0.9204
0.631 0.639 0.676 0.694 0.699
(0.513-0.748) (0.522-0.757) (0.561-0.791) (0.579-0.808) (0.584-0.813)
32 16.5 5.5 1 0.5
73.5 70.6 70.6 67.6 61.8
47.5 49.2 41.0 36.1 24.6
0.035 0.025 0.005 0.002 0.001
0.5048 0.5034 0.5174 0.5211 0.4704
NOTE. Values are expressed as means ⫾ SD. Abbreviations: AKI, acute kidney injury; AUC-ROC, area under the curve–receiver operating characteristic curves; renal rSO2, renal regional oxygen saturation.
anemia, hemodilution, and exogenous catecholamine can reduce arterial oxygen content or impair renal oxygen delivery, all of which may contribute to AKI.3,16–18 Therefore, it is logical to continuously monitor the renal rSO2 level using NIRS, which can measure regional tissue oxygen saturation. The multivariate linear regression analysis between renal rSO2 and intraoperative clinical factors indicates that the renal rSO2 concentration significantly correlates with lactate and SvO2, which are sensitive global indicators of tissue perfusion, despite the lack of an association with other clinical factors. The results are consistent with previous results that have demonstrated a strong relationship between rSO2 at various sites and the levels of blood lactate and SvO2.7,19 These analyses suggest that rSO2 measured by NIRS may be a useful indicator of impaired oxygen delivery to tissues. Thus, renal rSO2 may enable continuous monitoring of renal perfusion in real time, allowing the optimization of the renal condition. It should be noted that renal injury or failure after cardiac surgery is associated with numerous sources of kidney insult beyond renal ischemia, including nephrotoxicity and systemic inflammatory responses.16,18 Nephrotoxins include
contrast media, some antibiotics, aprotinin, free hemoglobin, and α-adrenergic agonist agents.20–24 Abrupt increases in the levels of circulatory inflammatory mediators, such as IL-1b, IL-6, and tumor necrosis factor α, can result from surgical trauma, CPB, endotoxemia, and ischemia-reperfusion injury.16,18,25 Unfortunately, NIRS devices cannot monitor these changes in local tissues. In addition, several postoperative clinical factors contribute to the development of AKI, which were not considered in this study. For these reasons, the ROC curve analysis in the results revealed the relatively low sensitivities (47%-65%) at each of the threshold values examined. Of the 26 patients who never displayed renal rSO2 values o70%, 6 developed AKI after surgery. To overcome these shortcomings, it may be helpful to use other AKI biomarkers simultaneously during and after surgery. This study had several limitations. First, the inclusion criteria were highly selective in order to permit the correct measurement of the renal rSO2 level. In general, NIRS can measure the regional level of tissue oxygen saturation in a local tissue area located 3 to 4 cm beneath the commercially available sensor.7 For this reason, 38 patients who had both
Table 4. Relationship Between Renal rSO2 and Intraoperative Vital Sign and Arterial Blood Gas Analysis Variables at Each Time Point Renal rSO2 After Induction
Before CPB
After CPB
End of surgery
Univariate p
Multivariate p
Univariate p
Multivariate p
Univariate p
Multivariate p
Univariate p
Multivariate p
Variable
Value
Value
Value
value
Value
Value
value
value
Heart rate MBP CVP MPAP CI SvO2 PaO2 Hematocrit Lactate
0.096 0.471 0.596 0.735 0.258 0.061 0.230 0.728 0.038
— — — — — 0.027 — — 0.017
0.669 0.932 0.263 0.889 0.610 0.111 0.368 0.493 0.049
— — — — — 0.043 — — 0.020
0.083 0.391 0.828 0.563 0.010 o0.001 0.002 0.755 0.001
— — — — — o0.001 — — 0.001
0.328 0.749 0.869 0.656 0.125 0.002 0.061 0.379 0.036
— — — — — 0.003 — — 0.019
Abbreviations: CI, cardiac index; CPB, cardiopulmonary bypass; CVP, central venous pressure; MBP, mean blood pressure; MPAP, mean pulmonary arterial pressure; PaO2, arterial oxygen pressure; rSO2, regional oxygen saturation; SvO2, mixed venous oxygen saturation.
7
RENAL RSO2 AND AKI
Table 5. Logistic Regression Analysis for Predictors of Acute Kidney Injury β Variable
Univariate Age Gender Body mass index Diabetes mellitus Hypertension Peripheral vascular disease Ejection fraction Logistic EuroSCORE Baseline renal rSO2 Preoperative hemoglobin Preoperative creatinine Type of surgery Surgery time Cardiopulmonary bypass time Intraoperative fluid infusion Crystalloid Colloid Urine output Blood transfusion Intraoperative minimum hemoglobin Duration of renal rSO2 o 55% Multivariate Logistic EuroSCORE Duration of renal rSO2 o 55%
p
Coefficient
95% CI
Value
— — — — — 5.086 — 1.223 — — — — 1.008 —
— — — — — 0.930-27.814 — 1.036-1.444 — — — — 1.001-1.014 —
0.153 0.939 0.332 0.105 0.650 0.061 0.866 0.017 0.070 0.076 0.674 0.583 0.018 0.212
— — — 2.528 — 1.046
— — — 0.991-6.448 — 1.017-1.076
0.593 0.152 0.561 0.052 0.101 0.002
1.297 1.047
1.073-1.567 1.017-1.077
0.007 0.002
Abbreviations: EuroSCORE, European System for Cardiac Operative Risk Evaluation; rSO2, regional oxygen saturation.
general population (46 mm).26 Therefore, a sensor that can measure rSO2 in the deeper tissue area will be needed to monitor renal rSO2 in all patients without restriction. Second, there was a possibility that the NIRS sensor used could not measure the regional tissue oxygen saturation of the true renal area (renal cortex and medulla). A NIRS sensor might measure rSO2 of the cortical renal area in patients who had renal depth o30 mm. On the other hand, the measurement might be contaminated by rSO2 of the near renal area (perinephric and paranephric fat tissue surrounding the kidney) in those who had renal depth 430 mm. Despite this limitation, the results indicated that the decrease in renal rSO2 level was associated significantly with postoperative AKI and predicted the risk of AKI. However, further studies with a sensor measuring the deeper tissue area are needed to confirm these results. CONCLUSION
In conclusion, the current prospective observational study indicated a predictive ability of intraoperative renal rSO2 monitoring for postoperative AKI in adult patients undergoing cardiac surgery. However, the use of an NIRS sensor that can monitor tissue areas deeper below the skin surface is warranted to generalize the use of this method to all patients undergoing cardiac surgery. In addition, further studies are needed to evaluate effective interventions to correct renal regional oxygen desaturation. ACKNOWLEDGMENTS
renal depths Z40 mm were excluded from the study despite meeting other inclusion criteria. The average renal depths of the study population were 29.7 mm for the right kidney and 30.6 mm for the left kidney. These depths are less than those of the
This study was partly supported by an unrestricted grant by E-Wha Biomedics Co., Seoul, Korea. The authors thank Seungbong Han, PhD, Statistician, Department of Clinical Epidemiology and Biostatistics, Asan Medical Center, Seoul, Korea, for valuable statistical advice and assistance.
REFERENCES 1. Chew ST, Mar WM, Ti LK: Association of ethnicity and acute kidney injury after cardiac surgery in a South East Asian population. Br J Anaesth 110(3):397-401, 2012 2. Hobson CE, Yavas S, Segal MS, et al: Acute kidney injury is associated with increased long-term mortality after cardiothoracic surgery. Circulation 119:2444-2453, 2009 3. Karkouti K, Wijeysundera DN, Yau TM, et al: Acute kidney injury after cardiac surgery: focus on modifiable risk factors. Circulation 119:495-502, 2009 4. Lassnigg A, Schmidlin D, Mouhieddine M, et al: Minimal changes of serum creatinine predict prognosis in patients after cardiothoracic surgery: a prospective cohort study. J Am Soc Nephrol 15:1597-1605, 2004 5. Martensson J, Martling CR, Bell M: Novel biomarkers of acute kidney injury and failure: clinical applicability. Br J Anaesth 109: 843-850, 2012 6. Wyckoff T, Augoustides JG: Advances in acute kidney injury associated with cardiac surgery: the unfolding revolution in early detection. J Cardiothorac Vasc Anesth 26:340-345, 2012 7. Moerman A, Vandenplas G, Bove T, et al: Relation between mixed venous oxygen saturation and cerebral oxygen saturation measured by absolute and relative near-infrared spectroscopy during
off-pump coronary artery bypass grafting. Br J Anaesth 110(2): 258-265, 2012 8. Murkin JM, Arango M: Near-infrared spectroscopy as an index of brain and tissue oxygenation. Br J Anaesth 103(Suppl 1):i3-13, 2009 9. Owens GE, King K, Gurney JG, et al: Low renal oximetry correlates with acute kidney injury after infant cardiac surgery. Pediatr Cardiol 32:183-188, 2011 10. Levey AS, Bosch JP, Lewis JB, et al: A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med 130:461-470, 1999 11. Bellomo R, Ronco C, Kellum JA, et al: Acute renal failure— definition, outcome measures, animal models, fluid therapy and information technology needs: the Second International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group. Crit Care 8:R204-R212, 2004 12. Hazle MA, Gajarski RJ, Aiyagari R, et al: Urinary biomarkers and renal near-infrared spectroscopy predict intensive care unit outcomes after cardiac surgery in infants younger than 6 months of age. J Thorac Cardiovasc Surg 146(4):861-867, 2013 Oct 13. Wagener G, Jan M, Kim M, et al: Association between increases in urinary neutrophil gelatinase-associated lipocalin and acute renal
8
dysfunction after adult cardiac surgery. Anesthesiology 105:485-491, 2006 14. Mishra J, Dent C, Tarabishi R, et al: Neutrophil gelatinaseassociated lipocalin (NGAL) as a biomarker for acute renal injury after cardiac surgery. Lancet 365:1231-1238, 2005 15. Selnes OA, Gottesman RF, Grega MA, et al: Cognitive and neurologic outcomes after coronary-artery bypass surgery. N Engl J Med 366:250-257, 2012 16. Stafford-Smith M, Patel UD, Phillips-Bute BG, et al: Acute kidney injury and chronic kidney disease after cardiac surgery. Adv Chronic Kidney Dis 15:257-277, 2008 17. Shander A, Javidroozi M, Ozawa S, et al: What is really dangerous: anaemia or transfusion? Br J Anaesth 107(Suppl 1):i41-i59, 2011 18. Sear JW: Kidney dysfunction in the postoperative period. Br J Anaesth 95:20-32, 2005 19. Chakravarti SB, Mittnacht AJ, Katz JC, et al: Multisite nearinfrared spectroscopy predicts elevated blood lactate level in children after cardiac surgery. J Cardiothorac Vasc Anesth 23:663-667, 2009 20. Wong GT, Irwin MG: Contrast-induced nephropathy. Br J Anaesth 99:474-483, 2007
CHOI ET AL
21. Nielsen DV, Hjortdal V, Larsson H, et al: Perioperative aminoglycoside treatment is associated with a higher incidence of postoperative dialysis in adult cardiac surgery patients. J Thorac Cardiovasc Surg 142:656-661, 2011 22. Shaw AD, Stafford-Smith M, White WD, et al: The effect of aprotinin on outcome after coronary-artery bypass grafting. N Engl J Med 358:784-793, 2008 23. Haase M, Haase-Fielitz A, Bagshaw SM, et al: Cardiopulmonary bypass-associated acute kidney injury: a pigment nephropathy? Contrib Nephrol 156:340-353, 2007 24. Santos FO, Silveira MA, Maia RB, et al: Acute renal failure after coronary artery bypass surgery with extracorporeal circulation— incidence, risk factors, and mortality. Arq Bras Cardiol 83:150-154, 2004 25. Granfeldt A, Ebdrup L, Tønnesen E, et al: Renal cytokine profile in an endotoxemic porcine model. Acta Anaesthesiol Scand 52: 614-620, 2008 26. Bolton WK, Tully RJ, Lewis EJ, et al: Localization of the kidney for percutaneous biopsy. A comparative study of methods. Ann Intern Med 81:159-164, 1974
