Urinary hypoxia: an intraoperative marker of risk of cardiac surgery-associated acute kidney injury

Urinary hypoxia: an intraoperative marker of risk of cardiac surgery-associated acute kidney injury Abstract Background Acute kidney injury (AKI) is common after cardiac surgery and profoundly affects postoperative mortality and morbidity. There are no validated methods to assess risk of AKI intraoperatively. Methods We determined the association between postoperative AKI and intraoperative urinary oxygen tension (PO2), measured via a fiber optic probe in the tip of the urinary catheter, in 65 patients undergoing high-risk cardiac surgery requiring cardiopulmonary bypass (CPB). AKI was diagnosed by modified Kidney Disease: Improving Global Outcomes criteria. Results Urinary PO2 fell during the operation, often reaching its nadir during rewarming or after weaning from CPB. Nadir urinary PO2 was lower in the 26 patients who developed AKI (mean ± SD, 8.9 ± 5.6 mmHg) than in the 39 patients who did not (14.9 ± 10.2 mmHg, P = 0.008). Patients who developed AKI had longer periods of urinary PO2 ≤15 and 10 mmHg than patients who did not. Odds of AKI increased when urinary PO2 fell to ≤10 mmHg {3.60 [95% confidence interval (CI) 1.27–10.21]} or ≤5 mmHg [3.60 (95% CI 1.04–12.42), P = 0.04] during the operation. When urinary PO2 fell to ≤15 mmHg, for more than or equal to  the median duration for all patients (4.8 min/h surgery), the odds of AKI were 4.85 (95% CI 1.64–14.40), P = 0.004. The area under the receiver-operator curve for this parameter alone was 0.69, and was 0.89 when other variables with P ≤ 0.10 in univariable analysis were included in the model. Conclusion Low urinary PO2 during adult cardiac surgery requiring CPB predicts AKI, so may identify patients in which intervention to improve renal oxygenation might reduce the risk of AKI. biomarkers, cardiopulmonary bypass, hypoxia, prognosis, renal failure INTRODUCTION Acute kidney injury (AKI) occurs in ∼30% of patients after cardiac surgery [1]. Even mild AKI after cardiopulmonary bypass (CPB) is associated with a >4-fold increase in risk of in-hospital death [1]. In severe AKI, requiring renal replacement therapy (1–2% of patients after surgery on CPB), mortality exceeds 35% [2]. Available biomarkers [3] offer limited prediction of AKI at arrival in the intensive care unit (ICU) [4, 5], but there are currently no biomarkers with predictive value during surgery, when intervention is still feasible. The etiology of AKI after cardiac surgery is complex [6, 7]. Nevertheless, hypoxia in the renal medulla appears to be a hallmark of AKI [8], being present not just during and after experimental CPB [9–11], but also in AKI induced by sepsis [12], ischemia-reperfusion injury [13], administration of radiocontrast agents [14] and rhabdomyolysis [15]. Additionally, renal oxygen delivery is impaired in humans on CPB [16]. Medullary hypoxia may thus represent a major driver of AKI. It is not feasible to measure renal medullary oxygen tension (PO2) in patients during cardiac surgery. However, it is feasible to measure continuously the PO2 of bladder urine via a fiber optic probe placed in the bladder catheter [17, 18]. Bladder urine PO2 correlates strongly with renal medullary PO2 in sheep during hyperdynamic septic AKI [19, 20] and in rabbits during pharmacologically induced changes in medullary PO2 [21]. Thus, we conducted a pilot study to test the hypothesis that urinary hypoxia during cardiac surgery requiring CPB predicts later development of AKI. MATERIALS AND METHODS Observational protocol The protocol was approved by the Human Research Ethics Committee of Monash Health (Reference: 12375B) and conformed to the Declaration of Helsinki. All patients provided informed consent. Between January 2015 and August 2017, 68 patients undergoing cardiac surgery requiring CPB were prospectively enrolled. Three patients were subsequently excluded, leaving a sample of 65 patients (Supplementary data, Figure S1 and Supplementary Methods). We included patients undergoing coronary artery bypass graft (CABG) surgery, valve repair/replacement surgery or combined ‘CABG–valve’ surgery. We excluded patients: with a baseline serum creatinine >200 μmol/L, or an estimated glomerular filtration rate (eGFR) <30 mL/min/1.73 m2; on hemodialysis (acutely or long term); who were kidney transplant recipients; with preoperative (pre-op) AKI as defined by the Kidney Disease: Improving Global Outcomes (KDIGO) AKI criteria [22]; or who were unable to provide informed consent. The sample of convenience was enriched with patients deemed at high risk of in-hospital death, based on their logistic EuroSCORE [23]. Many of the factors in the EuroSCORE are also risk factors for postoperative AKI [24]. Anesthesia, surgery and perfusion Standard institutional protocols were followed (Supplementary Methods). Continuous measurement of urinary PO2 The urinary bladder was catheterized with a standard Foley urinary catheter (Medtronic, Minneapolis, MN, USA). A sterilized fiber optic luminescence optode (NX-LAS-1/O/E-5 m, Oxford Optronix, Abingdon, UK) was advanced through the lumen of the catheter, to the catheter tip, so that the sensing tip of the probe was in direct contact with bladder urine. The bladder catheter was then connected to a standard urine collection system (see Supplementary Methods). Urinary PO2 was measured at 6-s intervals during surgery. Other data available from patient monitoring systems were recorded every 5 min (see Supplementary Methods). Definition of AKI Serum creatinine concentration was assessed preoperatively, at the time of entry to the ICU, and then daily (between midnight and 5:00 am) thereafter until discharge. The primary outcome was postoperative AKI, defined by modified KDIGO criteria [22], that is, a maximum increase in serum creatinine of ≥0.30 mg/dL (26.5 µmol/L) from pre-op baseline within the first 48 h after surgery, or ≥50% within the first 5 days after surgery. We did not utilize the urine output criterion for diagnosis of AKI, on the basis that it may result in over-diagnosis [25, 26]. Statistical analysis Continuous variables that satisfied the Shapiro–Wilk test are presented as mean ± SD [27]. For variables that failed this test of normality, the data were plotted and if the distribution was skewed, data are presented as median (interquartile range) and nonparametric statistical tests were used. Statistical analyses were performed using SPSS (Version 23.0, IBM Corp., Armonk, NY, USA). Two-sided P ≤ 0.05 was considered statistically significant. Dichotomous comparisons were made using Student’s unpaired t-test or the Mann–Whitney U test. Differences in proportions were assessed using chi-square, or with Fisher’s exact test when subgroups had fewer than five observations. P-values from within-subject factors in repeated measures analysis of variance were adjusted using the Greenhouse–Geisser method [28]. Univariable logistic regression was used to identify associations between various thresholds of intraoperative urinary PO2 and development of postoperative AKI. The sensitivity and specificity of each threshold was assessed using receiver-operator characteristic (ROC) curves. Factors with P ≤ 0.10 were then introduced into multivariable regression analyses. Model 1 included all variables with P ≤ 0.10 from univariable analysis. In multivariable Model 2, variables with P > 0.10 were removed step-wise until only those with P ≤ 0.10 remained. A similar approach was taken to assess factors associated with urinary hypoxia. RESULTS Patient and surgical characteristics Of the 65 patients included in the analyses, 26 (40%) developed AKI (Table 1). Table 1. Baseline characteristics of the sample of patients Characteristic  All Patients  No AKI  AKI  P  (n = 65)  (n = 39)  (n = 26)  Age, years  69.8 ± 8.91  70.8 ± 8.6  68.4 ± 9.4  0.29  Female  11 (16.9)  5 (12.8)  6 (23.1)  0.28  Body mass index, kg/m2  28.8 ± 4.9  28.7 ± 5.2  28.9 ± 4.5  0.90  Type of surgery           Isolated CABG  41 (63.1)  27 (69.2)  14 (53.8)  0.21   Single valve  8 (12.3)  5 (12.8)  3 (11.5)  >0.99b   >1 valve  4 (6.2)  0 (0.0)  4 (15.4)  0.02b   Single valve + CABG  12 (18.5)  7 (17.9)  5 (19.2)  >0.99b  Status of surgery           Elective  44 (67.7)  27 (69.2)  17 (65.4)  0.75   Urgent  20 (30.8)  12 (30.8)  8 (30.8)  >0.99   Emergency  1 (2.1)  0 (0.0)  1 (3.8)  0.40b  Risk factors and comorbidity           Diabetes mellitus  27 (41.5)  15 (38.5)  12 (46.2)  0.54   Hypertension  49 (75.4)  26 (66.7)  23 (88.5)  0.05   Hypercholesterolemia  48 (73.8)  27 (69.2)  21 (80.3)  0.30   Smoking history  41 (63.1)  25 (64.1)  16 (61.5)  0.83   Chronic lung disease  6 (9.2)  3 (7.7)  3 (11.5)  0.68b   Peripheral vascular disease  9 (13.8)  5 (12.8)  4 (15.4)  >0.99b   Cerebrovascular disease  8 (12.3)  5 (12.8)  3 (11.5)  >0.99b   NYHA ≥III  22 (33.8)  9 (23.1)  13 (50.0)  0.03   Congestive cardiac failure  12 (18.8)  6 (15.4)  6 (24.0)  0.43   Pre-op inotropes  2 (3.1)  1 (2.6)  1 (3.8)  >0.99b   Estimated LVEF <30%  6 (9.2)  2 (5.1)  4 (15.4)  0.21b   Arrhythmia  19 (29.2)  10 (25.6)  9 (34.6)  0.44   Redo sternotomy  5 (7.7)  2 (5.1)  3 (11.5)  0.38b   Baseline serum creatinine, μmol/L  87.7 ± 27.3  82.7 ± 26.1  95.2 ± 27.8  0.07   Baseline eGFRa, mL/min/1.73 m2  76.1 ± 20.3  80.9 ± 20.2  69.0 ± 18.8  0.02  Clinical risk scores       Logistic EuroSCORE, median (Q1, Q3)  4.7 (2.6, 9.2)  4.6 (2.6, 9.9)  5.4 (2.5, 8.7)  0.70c   Mehta Score, % risk, median (Q1, Q3)  0.7 (0.4, 1.3)  0.6 (0.4, 0.9)  1.1 (0.4, 1.9)  0.05c   Cleveland Score, % risk, median (Q1, Q3)  0.3 (0.1, 1.1)  0.3 (0.1, 0.6)  0.6 (0.1, 2.2)  0.11c  Characteristic  All Patients  No AKI  AKI  P  (n = 65)  (n = 39)  (n = 26)  Age, years  69.8 ± 8.91  70.8 ± 8.6  68.4 ± 9.4  0.29  Female  11 (16.9)  5 (12.8)  6 (23.1)  0.28  Body mass index, kg/m2  28.8 ± 4.9  28.7 ± 5.2  28.9 ± 4.5  0.90  Type of surgery           Isolated CABG  41 (63.1)  27 (69.2)  14 (53.8)  0.21   Single valve  8 (12.3)  5 (12.8)  3 (11.5)  >0.99b   >1 valve  4 (6.2)  0 (0.0)  4 (15.4)  0.02b   Single valve + CABG  12 (18.5)  7 (17.9)  5 (19.2)  >0.99b  Status of surgery           Elective  44 (67.7)  27 (69.2)  17 (65.4)  0.75   Urgent  20 (30.8)  12 (30.8)  8 (30.8)  >0.99   Emergency  1 (2.1)  0 (0.0)  1 (3.8)  0.40b  Risk factors and comorbidity           Diabetes mellitus  27 (41.5)  15 (38.5)  12 (46.2)  0.54   Hypertension  49 (75.4)  26 (66.7)  23 (88.5)  0.05   Hypercholesterolemia  48 (73.8)  27 (69.2)  21 (80.3)  0.30   Smoking history  41 (63.1)  25 (64.1)  16 (61.5)  0.83   Chronic lung disease  6 (9.2)  3 (7.7)  3 (11.5)  0.68b   Peripheral vascular disease  9 (13.8)  5 (12.8)  4 (15.4)  >0.99b   Cerebrovascular disease  8 (12.3)  5 (12.8)  3 (11.5)  >0.99b   NYHA ≥III  22 (33.8)  9 (23.1)  13 (50.0)  0.03   Congestive cardiac failure  12 (18.8)  6 (15.4)  6 (24.0)  0.43   Pre-op inotropes  2 (3.1)  1 (2.6)  1 (3.8)  >0.99b   Estimated LVEF <30%  6 (9.2)  2 (5.1)  4 (15.4)  0.21b   Arrhythmia  19 (29.2)  10 (25.6)  9 (34.6)  0.44   Redo sternotomy  5 (7.7)  2 (5.1)  3 (11.5)  0.38b   Baseline serum creatinine, μmol/L  87.7 ± 27.3  82.7 ± 26.1  95.2 ± 27.8  0.07   Baseline eGFRa, mL/min/1.73 m2  76.1 ± 20.3  80.9 ± 20.2  69.0 ± 18.8  0.02  Clinical risk scores       Logistic EuroSCORE, median (Q1, Q3)  4.7 (2.6, 9.2)  4.6 (2.6, 9.9)  5.4 (2.5, 8.7)  0.70c   Mehta Score, % risk, median (Q1, Q3)  0.7 (0.4, 1.3)  0.6 (0.4, 0.9)  1.1 (0.4, 1.9)  0.05c   Cleveland Score, % risk, median (Q1, Q3)  0.3 (0.1, 1.1)  0.3 (0.1, 0.6)  0.6 (0.1, 2.2)  0.11c  Values are mean ± SD or n (%) unless otherwise stated. P-values are derived from Student’s unpaired t-test for continuous variables and chi-square test for proportions unless otherwise shown. a eGFR was calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation [29]. Risk scores are shown for (i) The European System for Cardiac Operative Risk Evaluation (Logistic EuroSCORE) [23]; (ii) Mehta Score [24]; and (iii) the Cleveland Clinic Score [30], for prediction of risk of dialysis after cardiac surgery. Of the 26 patients who developed AKI, 15 had Stage 1 injury (increase in serum creatinine >26.5 μmol/L from baseline within 48 h, or >50% within 5 days); 7 had Stage 2 injury (increase in serum creatinine >100% within 5 days); and 4 patients developed Stage 3 injury (increase in serum creatinine >200% within 5 days or need for hemodialysis) [22]. Three patients required postoperative dialysis and were the only patients in the cohort of 65 who died in hospital. For the 26 patients who developed AKI, the diagnostic criterion was reached, on average, 20.6 ± 14.8 h after admission to the ICU. bFisher’s exact test. cMann Whitney U test. LVEF, left ventricular ejection fraction; NYHA, New York Heart Association Heart Failure Classification. Table 1. Baseline characteristics of the sample of patients Characteristic  All Patients  No AKI  AKI  P  (n = 65)  (n = 39)  (n = 26)  Age, years  69.8 ± 8.91  70.8 ± 8.6  68.4 ± 9.4  0.29  Female  11 (16.9)  5 (12.8)  6 (23.1)  0.28  Body mass index, kg/m2  28.8 ± 4.9  28.7 ± 5.2  28.9 ± 4.5  0.90  Type of surgery           Isolated CABG  41 (63.1)  27 (69.2)  14 (53.8)  0.21   Single valve  8 (12.3)  5 (12.8)  3 (11.5)  >0.99b   >1 valve  4 (6.2)  0 (0.0)  4 (15.4)  0.02b   Single valve + CABG  12 (18.5)  7 (17.9)  5 (19.2)  >0.99b  Status of surgery           Elective  44 (67.7)  27 (69.2)  17 (65.4)  0.75   Urgent  20 (30.8)  12 (30.8)  8 (30.8)  >0.99   Emergency  1 (2.1)  0 (0.0)  1 (3.8)  0.40b  Risk factors and comorbidity           Diabetes mellitus  27 (41.5)  15 (38.5)  12 (46.2)  0.54   Hypertension  49 (75.4)  26 (66.7)  23 (88.5)  0.05   Hypercholesterolemia  48 (73.8)  27 (69.2)  21 (80.3)  0.30   Smoking history  41 (63.1)  25 (64.1)  16 (61.5)  0.83   Chronic lung disease  6 (9.2)  3 (7.7)  3 (11.5)  0.68b   Peripheral vascular disease  9 (13.8)  5 (12.8)  4 (15.4)  >0.99b   Cerebrovascular disease  8 (12.3)  5 (12.8)  3 (11.5)  >0.99b   NYHA ≥III  22 (33.8)  9 (23.1)  13 (50.0)  0.03   Congestive cardiac failure  12 (18.8)  6 (15.4)  6 (24.0)  0.43   Pre-op inotropes  2 (3.1)  1 (2.6)  1 (3.8)  >0.99b   Estimated LVEF <30%  6 (9.2)  2 (5.1)  4 (15.4)  0.21b   Arrhythmia  19 (29.2)  10 (25.6)  9 (34.6)  0.44   Redo sternotomy  5 (7.7)  2 (5.1)  3 (11.5)  0.38b   Baseline serum creatinine, μmol/L  87.7 ± 27.3  82.7 ± 26.1  95.2 ± 27.8  0.07   Baseline eGFRa, mL/min/1.73 m2  76.1 ± 20.3  80.9 ± 20.2  69.0 ± 18.8  0.02  Clinical risk scores       Logistic EuroSCORE, median (Q1, Q3)  4.7 (2.6, 9.2)  4.6 (2.6, 9.9)  5.4 (2.5, 8.7)  0.70c   Mehta Score, % risk, median (Q1, Q3)  0.7 (0.4, 1.3)  0.6 (0.4, 0.9)  1.1 (0.4, 1.9)  0.05c   Cleveland Score, % risk, median (Q1, Q3)  0.3 (0.1, 1.1)  0.3 (0.1, 0.6)  0.6 (0.1, 2.2)  0.11c  Characteristic  All Patients  No AKI  AKI  P  (n = 65)  (n = 39)  (n = 26)  Age, years  69.8 ± 8.91  70.8 ± 8.6  68.4 ± 9.4  0.29  Female  11 (16.9)  5 (12.8)  6 (23.1)  0.28  Body mass index, kg/m2  28.8 ± 4.9  28.7 ± 5.2  28.9 ± 4.5  0.90  Type of surgery           Isolated CABG  41 (63.1)  27 (69.2)  14 (53.8)  0.21   Single valve  8 (12.3)  5 (12.8)  3 (11.5)  >0.99b   >1 valve  4 (6.2)  0 (0.0)  4 (15.4)  0.02b   Single valve + CABG  12 (18.5)  7 (17.9)  5 (19.2)  >0.99b  Status of surgery           Elective  44 (67.7)  27 (69.2)  17 (65.4)  0.75   Urgent  20 (30.8)  12 (30.8)  8 (30.8)  >0.99   Emergency  1 (2.1)  0 (0.0)  1 (3.8)  0.40b  Risk factors and comorbidity           Diabetes mellitus  27 (41.5)  15 (38.5)  12 (46.2)  0.54   Hypertension  49 (75.4)  26 (66.7)  23 (88.5)  0.05   Hypercholesterolemia  48 (73.8)  27 (69.2)  21 (80.3)  0.30   Smoking history  41 (63.1)  25 (64.1)  16 (61.5)  0.83   Chronic lung disease  6 (9.2)  3 (7.7)  3 (11.5)  0.68b   Peripheral vascular disease  9 (13.8)  5 (12.8)  4 (15.4)  >0.99b   Cerebrovascular disease  8 (12.3)  5 (12.8)  3 (11.5)  >0.99b   NYHA ≥III  22 (33.8)  9 (23.1)  13 (50.0)  0.03   Congestive cardiac failure  12 (18.8)  6 (15.4)  6 (24.0)  0.43   Pre-op inotropes  2 (3.1)  1 (2.6)  1 (3.8)  >0.99b   Estimated LVEF <30%  6 (9.2)  2 (5.1)  4 (15.4)  0.21b   Arrhythmia  19 (29.2)  10 (25.6)  9 (34.6)  0.44   Redo sternotomy  5 (7.7)  2 (5.1)  3 (11.5)  0.38b   Baseline serum creatinine, μmol/L  87.7 ± 27.3  82.7 ± 26.1  95.2 ± 27.8  0.07   Baseline eGFRa, mL/min/1.73 m2  76.1 ± 20.3  80.9 ± 20.2  69.0 ± 18.8  0.02  Clinical risk scores       Logistic EuroSCORE, median (Q1, Q3)  4.7 (2.6, 9.2)  4.6 (2.6, 9.9)  5.4 (2.5, 8.7)  0.70c   Mehta Score, % risk, median (Q1, Q3)  0.7 (0.4, 1.3)  0.6 (0.4, 0.9)  1.1 (0.4, 1.9)  0.05c   Cleveland Score, % risk, median (Q1, Q3)  0.3 (0.1, 1.1)  0.3 (0.1, 0.6)  0.6 (0.1, 2.2)  0.11c  Values are mean ± SD or n (%) unless otherwise stated. P-values are derived from Student’s unpaired t-test for continuous variables and chi-square test for proportions unless otherwise shown. a eGFR was calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation [29]. Risk scores are shown for (i) The European System for Cardiac Operative Risk Evaluation (Logistic EuroSCORE) [23]; (ii) Mehta Score [24]; and (iii) the Cleveland Clinic Score [30], for prediction of risk of dialysis after cardiac surgery. Of the 26 patients who developed AKI, 15 had Stage 1 injury (increase in serum creatinine >26.5 μmol/L from baseline within 48 h, or >50% within 5 days); 7 had Stage 2 injury (increase in serum creatinine >100% within 5 days); and 4 patients developed Stage 3 injury (increase in serum creatinine >200% within 5 days or need for hemodialysis) [22]. Three patients required postoperative dialysis and were the only patients in the cohort of 65 who died in hospital. For the 26 patients who developed AKI, the diagnostic criterion was reached, on average, 20.6 ± 14.8 h after admission to the ICU. bFisher’s exact test. cMann Whitney U test. LVEF, left ventricular ejection fraction; NYHA, New York Heart Association Heart Failure Classification. Preoperatively, a greater proportion of patients who later developed AKI had hypertension, New York Heart Association (NYHA) Class  ≥III heart failure or were undergoing a double or triple valve procedure (Table 1). Patients who later developed AKI also had a 15% lower baseline eGFR and a greater pre-op risk of dialysis (Mehta Score). The median logistic EuroSCORE did not differ significantly between the two groups of patients. Patients who developed AKI had operations that were 25% longer in duration than patients who did not develop AKI (Supplementary data, Table S1). The durations of CPB (149 ± 88 versus 109 ± 52 min) and aortic cross-clamp (111 ± 63 versus 83 ± 48 min) were also longer for patients who developed AKI. Patients who developed AKI were more often administered furosemide, received greater amounts of norepinephrine after CPB, had 0.6 L more positive net volume status and 4.8 mmHg lower nadir mean arterial pressure (MAP) than patients who did not develop AKI (Supplementary data, Table S2). Postoperatively, patients with AKI required longer periods of ventilation and ICU stay and more often required blood products than patients who did not develop AKI (Supplementary data, Table S1). Intraoperative urinary PO2 Urinary PO2 progressively fell across the course of the surgical procedure (Figure 1A and B). After induction of anesthesia and before the first incision (epoch 1; baseline), the mean urinary PO2 was 65.5 ± 34.5 mmHg. Mean urinary PO2 then fell by 31% from pre-op levels, to 45.2 ± 25.9 mmHg, during the period after the first incision and before going on CPB (epoch 2). It fell by a further 24% from baseline, to a mean of 29.3 ± 19.5 mmHg during the first (mild hypothermic) phase of CPB. Mean urinary PO2 continued to fall by a further 9% during the second (rewarming) phase of CPB, to 23.4 ± 16.0 mmHg. Critically, even once patients were weaned from CPB, urinary PO2 did not recover to pre- or early-bypass levels. Post-bypass, urinary PO2 were on average less than half that of pre-bypass levels (23.5 ± 14.9 mmHg versus 49.1 ± 26.9 mmHg). FIGURE 1: View largeDownload slide Urinary oxygen tension during on-pump cardiac surgery. Urinary PO2 was recorded continuously for 98.8% of the period of CPB and 98.5% of the total duration of surgery. For each individual patient, urinary PO2 was averaged over time using all available measurements during each epoch. (A) Columns and error bars are mean ± SD of urinary PO2 during the five epochs of surgery. Pre-Op, the period after induction of anesthesia but before the first incision; Pre-CPB, the period of surgery prior to commencement of CPB; CPB-Hypothermic, hypothermic CPB; CPB-Rewarm, the period of rewarming on bypass; and Post-CPB, the period of surgery after weaning from CPB. PGroup, PEpoch and PGroup*Epoch are the outcomes of repeated measures analysis of variance. P-values above each pair of measurements are from Student’s unpaired t-test. (B) Scatterplot showing the mean urinary PO2 for each individual subject, across each epoch of surgery. Solid horizontal lines represent the between-subject mean urinary PO2. (C) The lowest level of urinary PO2 (nadir) recorded during surgery for each subject. Error bars are mean ± SD. The P-values are the outcomes of Student’s unpaired t-test. FIGURE 1: View largeDownload slide Urinary oxygen tension during on-pump cardiac surgery. Urinary PO2 was recorded continuously for 98.8% of the period of CPB and 98.5% of the total duration of surgery. For each individual patient, urinary PO2 was averaged over time using all available measurements during each epoch. (A) Columns and error bars are mean ± SD of urinary PO2 during the five epochs of surgery. Pre-Op, the period after induction of anesthesia but before the first incision; Pre-CPB, the period of surgery prior to commencement of CPB; CPB-Hypothermic, hypothermic CPB; CPB-Rewarm, the period of rewarming on bypass; and Post-CPB, the period of surgery after weaning from CPB. PGroup, PEpoch and PGroup*Epoch are the outcomes of repeated measures analysis of variance. P-values above each pair of measurements are from Student’s unpaired t-test. (B) Scatterplot showing the mean urinary PO2 for each individual subject, across each epoch of surgery. Solid horizontal lines represent the between-subject mean urinary PO2. (C) The lowest level of urinary PO2 (nadir) recorded during surgery for each subject. Error bars are mean ± SD. The P-values are the outcomes of Student’s unpaired t-test. The overall profile of mean urinary PO2 did not differ significantly between patients who developed AKI and those who did not (Figure 1). However, mean urinary PO2 during the post-bypass epoch was lower in patients who later developed AKI (18.7 ± 9.0 versus 26.7 ± 17.2 mmHg, P = 0.03). Furthermore, in patients who later developed AKI, the lowest level of urinary PO2 during surgery (nadir) was 40% less than that of patients in the non-AKI group (8.9 ± 5.6 versus 14.9 ± 10.2 mmHg, P = 0.008). In 72% of the 65 patients, intraoperative nadir urinary PO2 occurred during the rewarming or post-bypass epochs. This pattern was similar in patients who later developed AKI (19 of 26, 73%) and those who did not (28 of 39, 72%). Urinary hypoxia We calculated the time spent at or below the threshold levels of urinary PO2 of 15, 10 and 5 mmHg, as well as the area under the curve for these thresholds (Supplementary data, Figure S2). These variables are expressed as a function of the total duration of each surgical procedure to allow comparison between patients. Patients who later developed AKI experienced 11-fold longer periods of urinary PO2 ≤15 mmHg, and the area under the curve was 51-fold greater, than patients who did not later develop AKI (Table 2). Periods of urinary PO2 ≤10 or ≤5 mmHg also tended to occur more commonly in patients who later developed AKI, but these periods were rare. Thus, we considered a threshold of ≤15 mmHg as useful working definition of urinary hypoxia. Table 2. Intraoperative urinary hypoxia Variable  All Patients  No AKI  AKI  P  (n = 65)  (n = 39)  (n = 26)  Median nadir urinary PO2, mmHg  11.4 (6.2, 15.5)  13.5 (8.3, 15.8)  7.8 (4.2, 13.3)  0.008  Mean nadir urinary PO2, mmHg  12.5 ± 9.1  14.9 ± 10.2  8.9 ± 5.6  0.008  Mean urine flow, mL/min  4.9 ± 2.5  5.4 ± 2.8  4.1 ± 1.7  0.04  Urinary PO2 at or below threshold values (adjusted for the duration of surgery)   Area under curve, mmHg × min/h            ≤15 mmHg  7.2 (0.0, 83.9)  1.2 (0.0, 52.7)  61.5 (4.0, 151.4)  0.01    ≤10 mmHg  0.0 (0.0, 13.4)  0.0 (0.0, 2.3)  4.4 (0.0, 56.0)  0.01    ≤5 mmHg  0.0 (0.0, 0.0)  0.0 (0.0, 0.0)  0.0 (0.0, 0.9)  0.06   Time at or below threshold, min/h            ≤15 mmHg  4.8 (0.0, 21.2)  1.5 (0.0, 12.0)  16.7 (2.5, 23.2)  0.02    ≤10 mmHg  0.0 (0.0, 8.4)  0.0 (0.0, 2.2)  4.7 (0.0, 15.2)  0.02    ≤5 mmHg  0.0 (0.0, 0.0)  0.0 (0.0, 0.0)  0.0 (0.0, 2.0)  0.08  Variable  All Patients  No AKI  AKI  P  (n = 65)  (n = 39)  (n = 26)  Median nadir urinary PO2, mmHg  11.4 (6.2, 15.5)  13.5 (8.3, 15.8)  7.8 (4.2, 13.3)  0.008  Mean nadir urinary PO2, mmHg  12.5 ± 9.1  14.9 ± 10.2  8.9 ± 5.6  0.008  Mean urine flow, mL/min  4.9 ± 2.5  5.4 ± 2.8  4.1 ± 1.7  0.04  Urinary PO2 at or below threshold values (adjusted for the duration of surgery)   Area under curve, mmHg × min/h            ≤15 mmHg  7.2 (0.0, 83.9)  1.2 (0.0, 52.7)  61.5 (4.0, 151.4)  0.01    ≤10 mmHg  0.0 (0.0, 13.4)  0.0 (0.0, 2.3)  4.4 (0.0, 56.0)  0.01    ≤5 mmHg  0.0 (0.0, 0.0)  0.0 (0.0, 0.0)  0.0 (0.0, 0.9)  0.06   Time at or below threshold, min/h            ≤15 mmHg  4.8 (0.0, 21.2)  1.5 (0.0, 12.0)  16.7 (2.5, 23.2)  0.02    ≤10 mmHg  0.0 (0.0, 8.4)  0.0 (0.0, 2.2)  4.7 (0.0, 15.2)  0.02    ≤5 mmHg  0.0 (0.0, 0.0)  0.0 (0.0, 0.0)  0.0 (0.0, 2.0)  0.08  Values are median (Q1, Q3) or mean ± SD. P-values were derived from the Mann–Whitney U test or Student’s unpaired t-test. Urinary PO2 fell to ≤15 mmHg in 47 patients (25 non-AKI, 22 AKI), to ≤10 mmHg in 29 patients (13 non-AKI and 16 AKI) and to ≤5 mmHg in 13 patients (5 non-AKI and 8 AKI). The area under the curve was determined by integrating the magnitude by which urinary PO2 was less than or equal to each specific threshold by the time below the threshold (Supplementary Figure 2). Both time at or below each threshold and the areas under the curve were normalized to the total duration of the surgery. Table 2. Intraoperative urinary hypoxia Variable  All Patients  No AKI  AKI  P  (n = 65)  (n = 39)  (n = 26)  Median nadir urinary PO2, mmHg  11.4 (6.2, 15.5)  13.5 (8.3, 15.8)  7.8 (4.2, 13.3)  0.008  Mean nadir urinary PO2, mmHg  12.5 ± 9.1  14.9 ± 10.2  8.9 ± 5.6  0.008  Mean urine flow, mL/min  4.9 ± 2.5  5.4 ± 2.8  4.1 ± 1.7  0.04  Urinary PO2 at or below threshold values (adjusted for the duration of surgery)   Area under curve, mmHg × min/h            ≤15 mmHg  7.2 (0.0, 83.9)  1.2 (0.0, 52.7)  61.5 (4.0, 151.4)  0.01    ≤10 mmHg  0.0 (0.0, 13.4)  0.0 (0.0, 2.3)  4.4 (0.0, 56.0)  0.01    ≤5 mmHg  0.0 (0.0, 0.0)  0.0 (0.0, 0.0)  0.0 (0.0, 0.9)  0.06   Time at or below threshold, min/h            ≤15 mmHg  4.8 (0.0, 21.2)  1.5 (0.0, 12.0)  16.7 (2.5, 23.2)  0.02    ≤10 mmHg  0.0 (0.0, 8.4)  0.0 (0.0, 2.2)  4.7 (0.0, 15.2)  0.02    ≤5 mmHg  0.0 (0.0, 0.0)  0.0 (0.0, 0.0)  0.0 (0.0, 2.0)  0.08  Variable  All Patients  No AKI  AKI  P  (n = 65)  (n = 39)  (n = 26)  Median nadir urinary PO2, mmHg  11.4 (6.2, 15.5)  13.5 (8.3, 15.8)  7.8 (4.2, 13.3)  0.008  Mean nadir urinary PO2, mmHg  12.5 ± 9.1  14.9 ± 10.2  8.9 ± 5.6  0.008  Mean urine flow, mL/min  4.9 ± 2.5  5.4 ± 2.8  4.1 ± 1.7  0.04  Urinary PO2 at or below threshold values (adjusted for the duration of surgery)   Area under curve, mmHg × min/h            ≤15 mmHg  7.2 (0.0, 83.9)  1.2 (0.0, 52.7)  61.5 (4.0, 151.4)  0.01    ≤10 mmHg  0.0 (0.0, 13.4)  0.0 (0.0, 2.3)  4.4 (0.0, 56.0)  0.01    ≤5 mmHg  0.0 (0.0, 0.0)  0.0 (0.0, 0.0)  0.0 (0.0, 0.9)  0.06   Time at or below threshold, min/h            ≤15 mmHg  4.8 (0.0, 21.2)  1.5 (0.0, 12.0)  16.7 (2.5, 23.2)  0.02    ≤10 mmHg  0.0 (0.0, 8.4)  0.0 (0.0, 2.2)  4.7 (0.0, 15.2)  0.02    ≤5 mmHg  0.0 (0.0, 0.0)  0.0 (0.0, 0.0)  0.0 (0.0, 2.0)  0.08  Values are median (Q1, Q3) or mean ± SD. P-values were derived from the Mann–Whitney U test or Student’s unpaired t-test. Urinary PO2 fell to ≤15 mmHg in 47 patients (25 non-AKI, 22 AKI), to ≤10 mmHg in 29 patients (13 non-AKI and 16 AKI) and to ≤5 mmHg in 13 patients (5 non-AKI and 8 AKI). The area under the curve was determined by integrating the magnitude by which urinary PO2 was less than or equal to each specific threshold by the time below the threshold (Supplementary Figure 2). Both time at or below each threshold and the areas under the curve were normalized to the total duration of the surgery. Clinical parameters monitored across the entire surgical procedure Body temperature, MAP and saturation of arterial hemoglobin with oxygen (SAO2) did not differ significantly between patients who later developed AKI and those who did not (Figure 2). Intraoperative urine flow increased dramatically during CPB. It was greatest during the rewarming phase, reaching 7.4 ± 4.8 mL/min across all patients, before declining to 6.0 ± 4.7 mL/min in the post-bypass period. Intraoperative urine flow was on an average 3.2 mL/min (37%) less during the rewarming phase and 3.0 mL/min (42%) less during the post-bypass period in patients who later developed AKI, than those who did not. FIGURE 2: View largeDownload slide Clinical parameters measured across the five epochs of surgery. (A) Body temperature. (B) Mean arterial pressure (MAP). (C) Saturation of arterial hemoglobin with oxygen (SAO2). (D) Urine flow. Columns and error bars represent mean ± SD. P-values are outcomes of two-way repeated measures analysis of variance with the factors group and epoch. ** denotes P ≤ 0.01 for comparisons between AKI and non-AKI groups, generated from Student’s unpaired t-test (post hoc analysis). The five epochs of surgery are as for Figure 1. Note: urine flow was not measured during the pre-op period, when the urinary catheter was first deployed. FIGURE 2: View largeDownload slide Clinical parameters measured across the five epochs of surgery. (A) Body temperature. (B) Mean arterial pressure (MAP). (C) Saturation of arterial hemoglobin with oxygen (SAO2). (D) Urine flow. Columns and error bars represent mean ± SD. P-values are outcomes of two-way repeated measures analysis of variance with the factors group and epoch. ** denotes P ≤ 0.01 for comparisons between AKI and non-AKI groups, generated from Student’s unpaired t-test (post hoc analysis). The five epochs of surgery are as for Figure 1. Note: urine flow was not measured during the pre-op period, when the urinary catheter was first deployed. Hemodynamics and whole-body oxygenation during CPB No differences in hemodynamic parameters of CPB were detected between patients who later developed AKI and those who did not (Table 3). Table 3. Hemodynamics and whole-body oxygenation during CPB Variable  CPB-Hypothermic (∼34°C)   CPB-Rewarm   PGroup  PEpoch  PGroup*Epoch  No AKI  AKI  No AKI  AKI  (n = 39)  (n = 26)  (n = 39)  (n = 26)  Pump flow, L/min  4.68 ± 0.57  4.68 ± 0.52  4.78 ± 0.54  4.75 ± 0.50  0.92  0.007  0.70  Pump flow, L/min/m2  2.40 ± 0.22  2.40 ± 0.17  2.45 ± 0.21  2.43 ± 0.18  0.97  0.005  0.71  PAO2, mmHg  244 ± 31  249 ± 25  238 ± 27  253 ± 32  0.12  0.85  0.14  PACO2, mmHg  38.8 ± 3.2  38.9 ± 2.6  40.5 ± 2.5  39.3 ± 2.8  0.36  0.008  0.11  Hemoglobin, g/L  87.3 ± 11.0  83.5 ± 13.4  88.0 ± 10.7  82.6 ± 12.2  0.12  0.92  0.14  Hematocrit, %  26.2 ± 3.3  25.0 ± 4.1  26.4 ± 3.2  24.8 ± 3.6  0.12  0.75  0.20  DO2, mL/min/m2  548 ± 113  524 ± 104  551 ± 105  509 ± 109  0.22  0.33  0.16  VO2, mL/min1/m2  101 ± 24  96 ± 23  120 ± 26  111 ± 25  0.24  <0.001  0.15  FEO2  0.19 ± 0.03  0.19 ± 0.04  0.22 ± 0.04  0.22 ± 0.04  0.80  <0.001  0.74  SVO2, %  81.2 ± 3.0  81.0 ± 3.3  78.8 ± 3.4  77.4 ± 3.2  0.66  <0.001  0.90  Variable  CPB-Hypothermic (∼34°C)   CPB-Rewarm   PGroup  PEpoch  PGroup*Epoch  No AKI  AKI  No AKI  AKI  (n = 39)  (n = 26)  (n = 39)  (n = 26)  Pump flow, L/min  4.68 ± 0.57  4.68 ± 0.52  4.78 ± 0.54  4.75 ± 0.50  0.92  0.007  0.70  Pump flow, L/min/m2  2.40 ± 0.22  2.40 ± 0.17  2.45 ± 0.21  2.43 ± 0.18  0.97  0.005  0.71  PAO2, mmHg  244 ± 31  249 ± 25  238 ± 27  253 ± 32  0.12  0.85  0.14  PACO2, mmHg  38.8 ± 3.2  38.9 ± 2.6  40.5 ± 2.5  39.3 ± 2.8  0.36  0.008  0.11  Hemoglobin, g/L  87.3 ± 11.0  83.5 ± 13.4  88.0 ± 10.7  82.6 ± 12.2  0.12  0.92  0.14  Hematocrit, %  26.2 ± 3.3  25.0 ± 4.1  26.4 ± 3.2  24.8 ± 3.6  0.12  0.75  0.20  DO2, mL/min/m2  548 ± 113  524 ± 104  551 ± 105  509 ± 109  0.22  0.33  0.16  VO2, mL/min1/m2  101 ± 24  96 ± 23  120 ± 26  111 ± 25  0.24  <0.001  0.15  FEO2  0.19 ± 0.03  0.19 ± 0.04  0.22 ± 0.04  0.22 ± 0.04  0.80  <0.001  0.74  SVO2, %  81.2 ± 3.0  81.0 ± 3.3  78.8 ± 3.4  77.4 ± 3.2  0.66  <0.001  0.90  Values are mean ± SD. P-values are from two-way repeated measures analysis of variance. DO2, whole-body oxygen delivery; FEO2, whole-body oxygen extraction fraction; PACO2, partial pressure of carbon dioxide in arterial blood; PAO2, partial pressure of oxygen in arterial blood; SVO2, saturation of venous hemoglobin with oxygen; VO2, whole-body oxygen consumption. Table 3. Hemodynamics and whole-body oxygenation during CPB Variable  CPB-Hypothermic (∼34°C)   CPB-Rewarm   PGroup  PEpoch  PGroup*Epoch  No AKI  AKI  No AKI  AKI  (n = 39)  (n = 26)  (n = 39)  (n = 26)  Pump flow, L/min  4.68 ± 0.57  4.68 ± 0.52  4.78 ± 0.54  4.75 ± 0.50  0.92  0.007  0.70  Pump flow, L/min/m2  2.40 ± 0.22  2.40 ± 0.17  2.45 ± 0.21  2.43 ± 0.18  0.97  0.005  0.71  PAO2, mmHg  244 ± 31  249 ± 25  238 ± 27  253 ± 32  0.12  0.85  0.14  PACO2, mmHg  38.8 ± 3.2  38.9 ± 2.6  40.5 ± 2.5  39.3 ± 2.8  0.36  0.008  0.11  Hemoglobin, g/L  87.3 ± 11.0  83.5 ± 13.4  88.0 ± 10.7  82.6 ± 12.2  0.12  0.92  0.14  Hematocrit, %  26.2 ± 3.3  25.0 ± 4.1  26.4 ± 3.2  24.8 ± 3.6  0.12  0.75  0.20  DO2, mL/min/m2  548 ± 113  524 ± 104  551 ± 105  509 ± 109  0.22  0.33  0.16  VO2, mL/min1/m2  101 ± 24  96 ± 23  120 ± 26  111 ± 25  0.24  <0.001  0.15  FEO2  0.19 ± 0.03  0.19 ± 0.04  0.22 ± 0.04  0.22 ± 0.04  0.80  <0.001  0.74  SVO2, %  81.2 ± 3.0  81.0 ± 3.3  78.8 ± 3.4  77.4 ± 3.2  0.66  <0.001  0.90  Variable  CPB-Hypothermic (∼34°C)   CPB-Rewarm   PGroup  PEpoch  PGroup*Epoch  No AKI  AKI  No AKI  AKI  (n = 39)  (n = 26)  (n = 39)  (n = 26)  Pump flow, L/min  4.68 ± 0.57  4.68 ± 0.52  4.78 ± 0.54  4.75 ± 0.50  0.92  0.007  0.70  Pump flow, L/min/m2  2.40 ± 0.22  2.40 ± 0.17  2.45 ± 0.21  2.43 ± 0.18  0.97  0.005  0.71  PAO2, mmHg  244 ± 31  249 ± 25  238 ± 27  253 ± 32  0.12  0.85  0.14  PACO2, mmHg  38.8 ± 3.2  38.9 ± 2.6  40.5 ± 2.5  39.3 ± 2.8  0.36  0.008  0.11  Hemoglobin, g/L  87.3 ± 11.0  83.5 ± 13.4  88.0 ± 10.7  82.6 ± 12.2  0.12  0.92  0.14  Hematocrit, %  26.2 ± 3.3  25.0 ± 4.1  26.4 ± 3.2  24.8 ± 3.6  0.12  0.75  0.20  DO2, mL/min/m2  548 ± 113  524 ± 104  551 ± 105  509 ± 109  0.22  0.33  0.16  VO2, mL/min1/m2  101 ± 24  96 ± 23  120 ± 26  111 ± 25  0.24  <0.001  0.15  FEO2  0.19 ± 0.03  0.19 ± 0.04  0.22 ± 0.04  0.22 ± 0.04  0.80  <0.001  0.74  SVO2, %  81.2 ± 3.0  81.0 ± 3.3  78.8 ± 3.4  77.4 ± 3.2  0.66  <0.001  0.90  Values are mean ± SD. P-values are from two-way repeated measures analysis of variance. DO2, whole-body oxygen delivery; FEO2, whole-body oxygen extraction fraction; PACO2, partial pressure of carbon dioxide in arterial blood; PAO2, partial pressure of oxygen in arterial blood; SVO2, saturation of venous hemoglobin with oxygen; VO2, whole-body oxygen consumption. Variables monitored only before and after CPB When averaged over the periods before and after bypass, central venous pressure (CVP) was 2.8 ± 0.9 mmHg greater, mean pulmonary artery pressure (PAP) was 3.5 ± 2.2 mmHg greater and end-tidal PCO2 was 2.1 ± 0.8 mmHg less, in patients who developed AKI than in those who did not (Supplementary data, Table S3). Intraoperative urinary hypoxia predicts AKI after cardiac surgery In univariable analysis, the odds of developing AKI were 3.60-fold greater when intraoperative urinary PO2 fell to ≤10 mmHg at any time during the surgical procedure (Figure 3; Supplementary data, Table S4). The duration of urinary PO2 at or below specific thresholds, when expressed as continuous variables, were also associated with risk of AKI. For every minute per hour of surgery, urinary PO2 was ≤15 mmHg, the risk of AKI was 4% greater, while it was 7% greater for a threshold of 10 mmHg. The median duration of urinary PO2 ≤15 mmHg was 4.8 min/h of surgery. A duration of urinary hypoxia (i.e. urinary PO2 ≤15 mmHg) of this median value or greater (≥4.8 min/h) was associated with a 4.85-fold greater risk of AKI [95% confidence interval (CI) 1.64–14.40] with an area under the ROC curve of 0.69 (95% CI 0.55–0.82). FIGURE 3: View largeDownload slide Association between indices of urinary hypoxia and risk of AKI. Data are the outcomes of univariable logistic regression, with the 95% CI limits of the ORs shown in parentheses. FIGURE 3: View largeDownload slide Association between indices of urinary hypoxia and risk of AKI. Data are the outcomes of univariable logistic regression, with the 95% CI limits of the ORs shown in parentheses. Apart from urinary hypoxia, other factors associated with the development of AKI were: lower baseline eGFR, NYHA Class ≥III, lower intraoperative urine output, longer duration of CPB and surgery, higher average CVP and average mean PAP measured before and after CPB, administration of more norepinephrine in the post-bypass period, lower nadir MAP during surgery and greater net positive fluid balance at the conclusion of surgery (Table 4). When these, and other variables with P ≤ 0.10 in univariable analysis, were included in a multivariable model, the area under the ROC curve was 0.89 (95% CI 0.81–0.97) (Model 1 in Table 4). In multivariable analysis with backward step-wise elimination (Model 2 in Table 4), only intraoperative urinary hypoxia (≤15 mmHg for ≥4.8 min/h), lower intraoperative urine output and a greater amount of norepinephrine administered post-bypass were found to be independently associated with later development of AKI. In this model, the odds ratio (OR) for urinary hypoxia was 7.48 (95% CI 1.54–36.43) and the area under the ROC curve was 0.87 (95% CI 0.78–0.96). Table 4. Univariable and multivariable logistic regression for predictors of AKI Variable  Univariable analysis   Multivariable model 1   Multivariable model 2   OR  95% CI  P  OR  95% CI  P  OR  95% CI  P  Urinary PO2 ≤15 mmHg, for ≥ the study median of 4.8 min/h of surgery  4.85  1.64–14.36  0.004  7.15  1.26–40.57  0.03  7.48  1.54–36.43  0.01  Hypertension  3.83  0.97–15.16  0.06  3.97  0.60–26.28  0.15  4.58  0.75–27.92  0.10  Baseline eGFR, mL/min/1.73 m2  0.97  0.94–1.00  0.03  0.97  0.94–1.01  0.17  0.97  0.94–1.00  0.06  NYHA ≥III  3.33  1.14–9.72  0.03  2.98  0.54–16.62  0.21  3.86  0.81–18.28  0.09  Mehta Score, % risk  1.41  0.93–2.13  0.11              Intraoperative urine flow, mL/min  0.79  0.63–0.99  0.04  0.64  0.43–0.95  0.03  0.61  0.42–0.90  0.01  CVP, mmHga  1.21  1.05–1.40  0.01              Mean PAP, mmHg  1.11  1.01–1.22  0.03  1.01  0.86–1.19  0.86        ETCO2, mmHg  0.88  0.76–1.01  0.07  0.91  0.76–1.09  0.32        Duration of CPB, mina  1.01  1.00–1.02  0.03              Duration of surgery, min  1.01  1.00–1.02  0.03  1.00  0.99–1.01  0.62        Post-CPB norepinephrine, total, µg/kg  1.38  1.02–1.87  0.04  1.01  1.00–1.01  0.28  1.01  1.00–1.01  0.05  Nadir MAP during surgery, mmHg  0.95  0.89–1.00  0.05  1.00  0.91–1.09  0.92        Net volume status, liters positive  1.50  1.03–2.20  0.04  1.29  0.71–2.32  0.41        Variable  Univariable analysis   Multivariable model 1   Multivariable model 2   OR  95% CI  P  OR  95% CI  P  OR  95% CI  P  Urinary PO2 ≤15 mmHg, for ≥ the study median of 4.8 min/h of surgery  4.85  1.64–14.36  0.004  7.15  1.26–40.57  0.03  7.48  1.54–36.43  0.01  Hypertension  3.83  0.97–15.16  0.06  3.97  0.60–26.28  0.15  4.58  0.75–27.92  0.10  Baseline eGFR, mL/min/1.73 m2  0.97  0.94–1.00  0.03  0.97  0.94–1.01  0.17  0.97  0.94–1.00  0.06  NYHA ≥III  3.33  1.14–9.72  0.03  2.98  0.54–16.62  0.21  3.86  0.81–18.28  0.09  Mehta Score, % risk  1.41  0.93–2.13  0.11              Intraoperative urine flow, mL/min  0.79  0.63–0.99  0.04  0.64  0.43–0.95  0.03  0.61  0.42–0.90  0.01  CVP, mmHga  1.21  1.05–1.40  0.01              Mean PAP, mmHg  1.11  1.01–1.22  0.03  1.01  0.86–1.19  0.86        ETCO2, mmHg  0.88  0.76–1.01  0.07  0.91  0.76–1.09  0.32        Duration of CPB, mina  1.01  1.00–1.02  0.03              Duration of surgery, min  1.01  1.00–1.02  0.03  1.00  0.99–1.01  0.62        Post-CPB norepinephrine, total, µg/kg  1.38  1.02–1.87  0.04  1.01  1.00–1.01  0.28  1.01  1.00–1.01  0.05  Nadir MAP during surgery, mmHg  0.95  0.89–1.00  0.05  1.00  0.91–1.09  0.92        Net volume status, liters positive  1.50  1.03–2.20  0.04  1.29  0.71–2.32  0.41        Univariable analyses were performed for variables found to differ significantly between patients who developed AKI and those who did not (Tables 1 and 2; Supplementary data, Tables S1 and S2). The only exceptions to this were for the variables ‘>1 valve’ and ‘intraoperative furosemide’, where logistic regression could not be performed due to 100% association with development of postoperative AKI. Univariable outcomes with P ≤ 0.10 were eligible for inclusion in the multivariable logistic regression models. a ‘CVP’ and ‘duration of CPB’ were excluded because these variables shared a multi-collinear relationship with ‘mean PAP’ and ‘duration of surgery’, respectively. Multivariable Model 1: all univariable outcomes with P ≤ 0.10 were included. Multivariable Model 2: variables with P > 0.10 were removed from the model in a step-wise fashion until only those with P ≤ 0.10 remained. The area under the ROC curve for Model 1 was 0.89 (95% CI 0.81–0.97), P ≤ 0.001 with inclusion of the urinary PO2 variable, and 0.85 (95% CI 0.76–0.95), P ≤ 0.001 without urinary PO2. For Model 2 it was 0.87 (0.78–0.96), P ≤ 0.001 with inclusion of the urinary PO2 variable, and 0.83 (95% CI 0.73–0.93), P ≤ 0.001 without urinary PO2. An additional model, which included only those variables known before the operation (hypertension, eGFR and NYHA Class) had an area under the ROC curve of 0.82 with the inclusion of the urinary PO2 variable and 0.71 without inclusion of urinary PO2. CVP, central venous pressure; ETCO2, end-tidal partial pressure of carbon dioxide. Table 4. Univariable and multivariable logistic regression for predictors of AKI Variable  Univariable analysis   Multivariable model 1   Multivariable model 2   OR  95% CI  P  OR  95% CI  P  OR  95% CI  P  Urinary PO2 ≤15 mmHg, for ≥ the study median of 4.8 min/h of surgery  4.85  1.64–14.36  0.004  7.15  1.26–40.57  0.03  7.48  1.54–36.43  0.01  Hypertension  3.83  0.97–15.16  0.06  3.97  0.60–26.28  0.15  4.58  0.75–27.92  0.10  Baseline eGFR, mL/min/1.73 m2  0.97  0.94–1.00  0.03  0.97  0.94–1.01  0.17  0.97  0.94–1.00  0.06  NYHA ≥III  3.33  1.14–9.72  0.03  2.98  0.54–16.62  0.21  3.86  0.81–18.28  0.09  Mehta Score, % risk  1.41  0.93–2.13  0.11              Intraoperative urine flow, mL/min  0.79  0.63–0.99  0.04  0.64  0.43–0.95  0.03  0.61  0.42–0.90  0.01  CVP, mmHga  1.21  1.05–1.40  0.01              Mean PAP, mmHg  1.11  1.01–1.22  0.03  1.01  0.86–1.19  0.86        ETCO2, mmHg  0.88  0.76–1.01  0.07  0.91  0.76–1.09  0.32        Duration of CPB, mina  1.01  1.00–1.02  0.03              Duration of surgery, min  1.01  1.00–1.02  0.03  1.00  0.99–1.01  0.62        Post-CPB norepinephrine, total, µg/kg  1.38  1.02–1.87  0.04  1.01  1.00–1.01  0.28  1.01  1.00–1.01  0.05  Nadir MAP during surgery, mmHg  0.95  0.89–1.00  0.05  1.00  0.91–1.09  0.92        Net volume status, liters positive  1.50  1.03–2.20  0.04  1.29  0.71–2.32  0.41        Variable  Univariable analysis   Multivariable model 1   Multivariable model 2   OR  95% CI  P  OR  95% CI  P  OR  95% CI  P  Urinary PO2 ≤15 mmHg, for ≥ the study median of 4.8 min/h of surgery  4.85  1.64–14.36  0.004  7.15  1.26–40.57  0.03  7.48  1.54–36.43  0.01  Hypertension  3.83  0.97–15.16  0.06  3.97  0.60–26.28  0.15  4.58  0.75–27.92  0.10  Baseline eGFR, mL/min/1.73 m2  0.97  0.94–1.00  0.03  0.97  0.94–1.01  0.17  0.97  0.94–1.00  0.06  NYHA ≥III  3.33  1.14–9.72  0.03  2.98  0.54–16.62  0.21  3.86  0.81–18.28  0.09  Mehta Score, % risk  1.41  0.93–2.13  0.11              Intraoperative urine flow, mL/min  0.79  0.63–0.99  0.04  0.64  0.43–0.95  0.03  0.61  0.42–0.90  0.01  CVP, mmHga  1.21  1.05–1.40  0.01              Mean PAP, mmHg  1.11  1.01–1.22  0.03  1.01  0.86–1.19  0.86        ETCO2, mmHg  0.88  0.76–1.01  0.07  0.91  0.76–1.09  0.32        Duration of CPB, mina  1.01  1.00–1.02  0.03              Duration of surgery, min  1.01  1.00–1.02  0.03  1.00  0.99–1.01  0.62        Post-CPB norepinephrine, total, µg/kg  1.38  1.02–1.87  0.04  1.01  1.00–1.01  0.28  1.01  1.00–1.01  0.05  Nadir MAP during surgery, mmHg  0.95  0.89–1.00  0.05  1.00  0.91–1.09  0.92        Net volume status, liters positive  1.50  1.03–2.20  0.04  1.29  0.71–2.32  0.41        Univariable analyses were performed for variables found to differ significantly between patients who developed AKI and those who did not (Tables 1 and 2; Supplementary data, Tables S1 and S2). The only exceptions to this were for the variables ‘>1 valve’ and ‘intraoperative furosemide’, where logistic regression could not be performed due to 100% association with development of postoperative AKI. Univariable outcomes with P ≤ 0.10 were eligible for inclusion in the multivariable logistic regression models. a ‘CVP’ and ‘duration of CPB’ were excluded because these variables shared a multi-collinear relationship with ‘mean PAP’ and ‘duration of surgery’, respectively. Multivariable Model 1: all univariable outcomes with P ≤ 0.10 were included. Multivariable Model 2: variables with P > 0.10 were removed from the model in a step-wise fashion until only those with P ≤ 0.10 remained. The area under the ROC curve for Model 1 was 0.89 (95% CI 0.81–0.97), P ≤ 0.001 with inclusion of the urinary PO2 variable, and 0.85 (95% CI 0.76–0.95), P ≤ 0.001 without urinary PO2. For Model 2 it was 0.87 (0.78–0.96), P ≤ 0.001 with inclusion of the urinary PO2 variable, and 0.83 (95% CI 0.73–0.93), P ≤ 0.001 without urinary PO2. An additional model, which included only those variables known before the operation (hypertension, eGFR and NYHA Class) had an area under the ROC curve of 0.82 with the inclusion of the urinary PO2 variable and 0.71 without inclusion of urinary PO2. CVP, central venous pressure; ETCO2, end-tidal partial pressure of carbon dioxide. Factors associated with urinary hypoxia Univariable analysis revealed significant associations of urinary hypoxia (urinary PO2 ≤15 mmHg for ≥4.8 min/h) with: longer durations of surgery and CPB, lesser blood hemoglobin concentration and oxygen delivery (DO2) on bypass, intraoperative administration of furosemide and metaraminol, as well as the use of more epinephrine or norepinephrine in the post-bypass period (Table 5). Multivariable analysis revealed independent associations of urinary hypoxia with longer duration of surgery, lesser DO2 on bypass and the use of norepinephrine. Table 5. Univariable and multivariable logistic regression for factors associated with urinary hypoxia (urinary PO2 ≤15 mmHg for ≥4.8 min/h of surgery) Variable  Univariable   Multivariable   OR  95% CI  P  OR  95% CI  P  Medical and demographic variables at baseline   Age, years  0.99  0.93–1.04  0.61         Female  0.77  0.21–2.84  0.70         Body mass index, kg/m2  0.97  0.87–1.07  0.51         Non-CABG  1.62  0.59–4.48  0.35         Non-elective surgery  1.95  0.67–5.64  0.22         Diabetes mellitus  0.65  0.24–1.75  0.39         Hypertension  0.75  0.24–2.32  0.61         Hypercholesterolemia  0.64  0.21–1.97  0.44         Smoking history  1.37  0.50–3.76  0.54         Chronic lung disease  0.45  0.08–2.66  0.38         Peripheral vascular disease  0.43  0.10–1.91  0.27         Cerebrovascular disease  2.15  0.49–9.45  0.31         NYHA ≥III  0.73  0.26–2.03  0.54         Congestive cardiac failure  0.96  0.28–3.37  0.95         Pre-op inotropes  0.97  0.06–16.2  0.98         Estimated EF <30%  2.07  0.35–12.18  0.42         Arrhythmia  1.11  0.38–3.24  0.85         Redo sternotomy  1.50  0.23–9.63  0.67         Serum creatinine, μmol/L  1.01  0.99–1.03  0.27         eGFR, mL/min/1.73 m2  1.00  0.97–1.02  0.79         Logistic EuroSCORE  1.01  0.95–1.07  0.73         Mehta Score, % risk  1.26  0.85–1.87  0.26         Cleveland Clinic Score, % risk  1.27  0.79–2.04  0.33        Variables during surgery and CPB   Duration of surgery, min  1.01  1.00–1.02  0.01  1.01  1.00–1.02  0.02   CPB time, mina  1.01  1.00–1.02  0.05         Aortic cross-clamp time, mina  1.01  1.00–1.02  0.07         Mean arterial pressure, mmHg  1.01  0.93–1.11  0.80         SAO2, %  1.01  0.60–1.70  0.97         Body temperature, °C  0.92  0.22–3.78  0.91         Urine flow, case average, L/min  0.99  0.81–1.20  0.89         Pump flow, L/min  0.61  0.23–1.61  0.32         PAO2, mmHg  1.00  0.98–1.03  0.69         PACO2, mmHg  1.10  0.89–1.37  0.38         Hemoglobin, g/La  0.94  0.90–0.99  0.02         DO2, mL/min/m2  0.99  0.99–1.00  0.03  0.99  0.99–1.00  0.01   VO2, mL/min/m2  0.99  0.97–1.01  0.20         SVO2, %  0.93  0.79–1.10  0.40         Heart rate, bpm  1.03  0.96–1.09  0.42         Central venous pressure, mmHg  1.07  0.94–1.20  0.31         Mean PAP, mmHg  1.05  0.97–1.14  0.26         ETCO2, mmHg  0.98  0.87–1.10  0.67         Net volume status, liters positive  1.06  0.81–1.38  0.68         Nadir MAP during case, mmHg  0.96  0.91–1.01  0.14         Nadir MAP during CPB, mmHg  0.95  0.90–1.00  0.07        Pharmacological interventions               At any time during surgery                Furosemide  8.35  0.96–72.3  0.05          Esmolol  1.50  0.23–9.63  0.67          Milrinone  4.28  0.45–40.53  0.21         Before CPB                Norepinephrine  1.17  0.35–3.95  0.80           Norepinephrine, total given, µg/kg  1.22  0.74–2.01  0.43         During CPB                Metaraminol dose, mg  1.13  1.02–1.24  0.02          Norepinephrine  1.17  0.35–3.95  0.80           Norepinephrine, total given, µg/kg  1.62  0.73–3.58  0.23         After CPB                Epinephrine  1.83  0.67–5.05  0.24           Epinephrine, total given, µg/kg  2.10  1.03–4.26  0.04          Norepinephrine  6.00  1.50–23.99  0.01  5.24  1.04–26.39  0.05     Norepinephrine, total given, µg/kg  1.56  1.06–2.28  0.02        Variable  Univariable   Multivariable   OR  95% CI  P  OR  95% CI  P  Medical and demographic variables at baseline   Age, years  0.99  0.93–1.04  0.61         Female  0.77  0.21–2.84  0.70         Body mass index, kg/m2  0.97  0.87–1.07  0.51         Non-CABG  1.62  0.59–4.48  0.35         Non-elective surgery  1.95  0.67–5.64  0.22         Diabetes mellitus  0.65  0.24–1.75  0.39         Hypertension  0.75  0.24–2.32  0.61         Hypercholesterolemia  0.64  0.21–1.97  0.44         Smoking history  1.37  0.50–3.76  0.54         Chronic lung disease  0.45  0.08–2.66  0.38         Peripheral vascular disease  0.43  0.10–1.91  0.27         Cerebrovascular disease  2.15  0.49–9.45  0.31         NYHA ≥III  0.73  0.26–2.03  0.54         Congestive cardiac failure  0.96  0.28–3.37  0.95         Pre-op inotropes  0.97  0.06–16.2  0.98         Estimated EF <30%  2.07  0.35–12.18  0.42         Arrhythmia  1.11  0.38–3.24  0.85         Redo sternotomy  1.50  0.23–9.63  0.67         Serum creatinine, μmol/L  1.01  0.99–1.03  0.27         eGFR, mL/min/1.73 m2  1.00  0.97–1.02  0.79         Logistic EuroSCORE  1.01  0.95–1.07  0.73         Mehta Score, % risk  1.26  0.85–1.87  0.26         Cleveland Clinic Score, % risk  1.27  0.79–2.04  0.33        Variables during surgery and CPB   Duration of surgery, min  1.01  1.00–1.02  0.01  1.01  1.00–1.02  0.02   CPB time, mina  1.01  1.00–1.02  0.05         Aortic cross-clamp time, mina  1.01  1.00–1.02  0.07         Mean arterial pressure, mmHg  1.01  0.93–1.11  0.80         SAO2, %  1.01  0.60–1.70  0.97         Body temperature, °C  0.92  0.22–3.78  0.91         Urine flow, case average, L/min  0.99  0.81–1.20  0.89         Pump flow, L/min  0.61  0.23–1.61  0.32         PAO2, mmHg  1.00  0.98–1.03  0.69         PACO2, mmHg  1.10  0.89–1.37  0.38         Hemoglobin, g/La  0.94  0.90–0.99  0.02         DO2, mL/min/m2  0.99  0.99–1.00  0.03  0.99  0.99–1.00  0.01   VO2, mL/min/m2  0.99  0.97–1.01  0.20         SVO2, %  0.93  0.79–1.10  0.40         Heart rate, bpm  1.03  0.96–1.09  0.42         Central venous pressure, mmHg  1.07  0.94–1.20  0.31         Mean PAP, mmHg  1.05  0.97–1.14  0.26         ETCO2, mmHg  0.98  0.87–1.10  0.67         Net volume status, liters positive  1.06  0.81–1.38  0.68         Nadir MAP during case, mmHg  0.96  0.91–1.01  0.14         Nadir MAP during CPB, mmHg  0.95  0.90–1.00  0.07        Pharmacological interventions               At any time during surgery                Furosemide  8.35  0.96–72.3  0.05          Esmolol  1.50  0.23–9.63  0.67          Milrinone  4.28  0.45–40.53  0.21         Before CPB                Norepinephrine  1.17  0.35–3.95  0.80           Norepinephrine, total given, µg/kg  1.22  0.74–2.01  0.43         During CPB                Metaraminol dose, mg  1.13  1.02–1.24  0.02          Norepinephrine  1.17  0.35–3.95  0.80           Norepinephrine, total given, µg/kg  1.62  0.73–3.58  0.23         After CPB                Epinephrine  1.83  0.67–5.05  0.24           Epinephrine, total given, µg/kg  2.10  1.03–4.26  0.04          Norepinephrine  6.00  1.50–23.99  0.01  5.24  1.04–26.39  0.05     Norepinephrine, total given, µg/kg  1.56  1.06–2.28  0.02        Univariable outcomes with P ≤ 0.10 were eligible for inclusion in the multivariable logistic regression model. a Hemoglobin during CPB was excluded because it is included in the calculation of DO2. Duration of CPB and aortic cross-clamp time were excluded because these variables are a subset of the total duration of surgery. DO2, whole-body oxygen delivery; EF, ejection fraction; ETCO2, end-tidal partial pressure of carbon dioxide; PACO2, partial pressure of carbon dioxide in arterial blood; PAO2, partial pressure of oxygen in arterial blood; SAO2, saturation of arterial hemoglobin with oxygen; SVO2, saturation of venous hemoglobin with oxygen; VO2, whole-body oxygen consumption. Table 5. Univariable and multivariable logistic regression for factors associated with urinary hypoxia (urinary PO2 ≤15 mmHg for ≥4.8 min/h of surgery) Variable  Univariable   Multivariable   OR  95% CI  P  OR  95% CI  P  Medical and demographic variables at baseline   Age, years  0.99  0.93–1.04  0.61         Female  0.77  0.21–2.84  0.70         Body mass index, kg/m2  0.97  0.87–1.07  0.51         Non-CABG  1.62  0.59–4.48  0.35         Non-elective surgery  1.95  0.67–5.64  0.22         Diabetes mellitus  0.65  0.24–1.75  0.39         Hypertension  0.75  0.24–2.32  0.61         Hypercholesterolemia  0.64  0.21–1.97  0.44         Smoking history  1.37  0.50–3.76  0.54         Chronic lung disease  0.45  0.08–2.66  0.38         Peripheral vascular disease  0.43  0.10–1.91  0.27         Cerebrovascular disease  2.15  0.49–9.45  0.31         NYHA ≥III  0.73  0.26–2.03  0.54         Congestive cardiac failure  0.96  0.28–3.37  0.95         Pre-op inotropes  0.97  0.06–16.2  0.98         Estimated EF <30%  2.07  0.35–12.18  0.42         Arrhythmia  1.11  0.38–3.24  0.85         Redo sternotomy  1.50  0.23–9.63  0.67         Serum creatinine, μmol/L  1.01  0.99–1.03  0.27         eGFR, mL/min/1.73 m2  1.00  0.97–1.02  0.79         Logistic EuroSCORE  1.01  0.95–1.07  0.73         Mehta Score, % risk  1.26  0.85–1.87  0.26         Cleveland Clinic Score, % risk  1.27  0.79–2.04  0.33        Variables during surgery and CPB   Duration of surgery, min  1.01  1.00–1.02  0.01  1.01  1.00–1.02  0.02   CPB time, mina  1.01  1.00–1.02  0.05         Aortic cross-clamp time, mina  1.01  1.00–1.02  0.07         Mean arterial pressure, mmHg  1.01  0.93–1.11  0.80         SAO2, %  1.01  0.60–1.70  0.97         Body temperature, °C  0.92  0.22–3.78  0.91         Urine flow, case average, L/min  0.99  0.81–1.20  0.89         Pump flow, L/min  0.61  0.23–1.61  0.32         PAO2, mmHg  1.00  0.98–1.03  0.69         PACO2, mmHg  1.10  0.89–1.37  0.38         Hemoglobin, g/La  0.94  0.90–0.99  0.02         DO2, mL/min/m2  0.99  0.99–1.00  0.03  0.99  0.99–1.00  0.01   VO2, mL/min/m2  0.99  0.97–1.01  0.20         SVO2, %  0.93  0.79–1.10  0.40         Heart rate, bpm  1.03  0.96–1.09  0.42         Central venous pressure, mmHg  1.07  0.94–1.20  0.31         Mean PAP, mmHg  1.05  0.97–1.14  0.26         ETCO2, mmHg  0.98  0.87–1.10  0.67         Net volume status, liters positive  1.06  0.81–1.38  0.68         Nadir MAP during case, mmHg  0.96  0.91–1.01  0.14         Nadir MAP during CPB, mmHg  0.95  0.90–1.00  0.07        Pharmacological interventions               At any time during surgery                Furosemide  8.35  0.96–72.3  0.05          Esmolol  1.50  0.23–9.63  0.67          Milrinone  4.28  0.45–40.53  0.21         Before CPB                Norepinephrine  1.17  0.35–3.95  0.80           Norepinephrine, total given, µg/kg  1.22  0.74–2.01  0.43         During CPB                Metaraminol dose, mg  1.13  1.02–1.24  0.02          Norepinephrine  1.17  0.35–3.95  0.80           Norepinephrine, total given, µg/kg  1.62  0.73–3.58  0.23         After CPB                Epinephrine  1.83  0.67–5.05  0.24           Epinephrine, total given, µg/kg  2.10  1.03–4.26  0.04          Norepinephrine  6.00  1.50–23.99  0.01  5.24  1.04–26.39  0.05     Norepinephrine, total given, µg/kg  1.56  1.06–2.28  0.02        Variable  Univariable   Multivariable   OR  95% CI  P  OR  95% CI  P  Medical and demographic variables at baseline   Age, years  0.99  0.93–1.04  0.61         Female  0.77  0.21–2.84  0.70         Body mass index, kg/m2  0.97  0.87–1.07  0.51         Non-CABG  1.62  0.59–4.48  0.35         Non-elective surgery  1.95  0.67–5.64  0.22         Diabetes mellitus  0.65  0.24–1.75  0.39         Hypertension  0.75  0.24–2.32  0.61         Hypercholesterolemia  0.64  0.21–1.97  0.44         Smoking history  1.37  0.50–3.76  0.54         Chronic lung disease  0.45  0.08–2.66  0.38         Peripheral vascular disease  0.43  0.10–1.91  0.27         Cerebrovascular disease  2.15  0.49–9.45  0.31         NYHA ≥III  0.73  0.26–2.03  0.54         Congestive cardiac failure  0.96  0.28–3.37  0.95         Pre-op inotropes  0.97  0.06–16.2  0.98         Estimated EF <30%  2.07  0.35–12.18  0.42         Arrhythmia  1.11  0.38–3.24  0.85         Redo sternotomy  1.50  0.23–9.63  0.67         Serum creatinine, μmol/L  1.01  0.99–1.03  0.27         eGFR, mL/min/1.73 m2  1.00  0.97–1.02  0.79         Logistic EuroSCORE  1.01  0.95–1.07  0.73         Mehta Score, % risk  1.26  0.85–1.87  0.26         Cleveland Clinic Score, % risk  1.27  0.79–2.04  0.33        Variables during surgery and CPB   Duration of surgery, min  1.01  1.00–1.02  0.01  1.01  1.00–1.02  0.02   CPB time, mina  1.01  1.00–1.02  0.05         Aortic cross-clamp time, mina  1.01  1.00–1.02  0.07         Mean arterial pressure, mmHg  1.01  0.93–1.11  0.80         SAO2, %  1.01  0.60–1.70  0.97         Body temperature, °C  0.92  0.22–3.78  0.91         Urine flow, case average, L/min  0.99  0.81–1.20  0.89         Pump flow, L/min  0.61  0.23–1.61  0.32         PAO2, mmHg  1.00  0.98–1.03  0.69         PACO2, mmHg  1.10  0.89–1.37  0.38         Hemoglobin, g/La  0.94  0.90–0.99  0.02         DO2, mL/min/m2  0.99  0.99–1.00  0.03  0.99  0.99–1.00  0.01   VO2, mL/min/m2  0.99  0.97–1.01  0.20         SVO2, %  0.93  0.79–1.10  0.40         Heart rate, bpm  1.03  0.96–1.09  0.42         Central venous pressure, mmHg  1.07  0.94–1.20  0.31         Mean PAP, mmHg  1.05  0.97–1.14  0.26         ETCO2, mmHg  0.98  0.87–1.10  0.67         Net volume status, liters positive  1.06  0.81–1.38  0.68         Nadir MAP during case, mmHg  0.96  0.91–1.01  0.14         Nadir MAP during CPB, mmHg  0.95  0.90–1.00  0.07        Pharmacological interventions               At any time during surgery                Furosemide  8.35  0.96–72.3  0.05          Esmolol  1.50  0.23–9.63  0.67          Milrinone  4.28  0.45–40.53  0.21         Before CPB                Norepinephrine  1.17  0.35–3.95  0.80           Norepinephrine, total given, µg/kg  1.22  0.74–2.01  0.43         During CPB                Metaraminol dose, mg  1.13  1.02–1.24  0.02          Norepinephrine  1.17  0.35–3.95  0.80           Norepinephrine, total given, µg/kg  1.62  0.73–3.58  0.23         After CPB                Epinephrine  1.83  0.67–5.05  0.24           Epinephrine, total given, µg/kg  2.10  1.03–4.26  0.04          Norepinephrine  6.00  1.50–23.99  0.01  5.24  1.04–26.39  0.05     Norepinephrine, total given, µg/kg  1.56  1.06–2.28  0.02        Univariable outcomes with P ≤ 0.10 were eligible for inclusion in the multivariable logistic regression model. a Hemoglobin during CPB was excluded because it is included in the calculation of DO2. Duration of CPB and aortic cross-clamp time were excluded because these variables are a subset of the total duration of surgery. DO2, whole-body oxygen delivery; EF, ejection fraction; ETCO2, end-tidal partial pressure of carbon dioxide; PACO2, partial pressure of carbon dioxide in arterial blood; PAO2, partial pressure of oxygen in arterial blood; SAO2, saturation of arterial hemoglobin with oxygen; SVO2, saturation of venous hemoglobin with oxygen; VO2, whole-body oxygen consumption. DISCUSSION Our major findings were: (i) it is feasible to continuously measure bladder urine PO2 in the operating theater using a fiber optic probe; (ii) bladder urine PO2 falls across the course of cardiac surgery, reaching its nadir most commonly during the rewarming phase of CPB or early after weaning from CPB; (iii) patients who develop AKI after cardiac surgery experience longer and more severe periods of low intraoperative urinary PO2 than patients who do not develop AKI; and (iv) low intraoperative urinary PO2 is an independent predictor of postoperative AKI. Thus, urinary PO2 may provide the first available real-time ‘biomarker’ of risk of postoperative AKI in the setting of cardiac surgery that provides prognostic information during the surgical procedure itself, when it may not be too late to intervene to prevent AKI. The detrimental impact of hypoxia could depend on both its duration and severity. Therefore, we calculated the time spent at or below various thresholds of urinary PO2 (15, 10 and 5 mmHg) and the area under the curve for each threshold. At thresholds of 15 and 10 mmHg, the duration spent at or below the threshold and the area under the curve was much greater in patients who later developed AKI than in those who did not. Furthermore, univariable logistic regression showed that the odds of AKI were 3.6 times greater when urinary PO2 fell to ≤10 mmHg at any time during the operation. This analysis provides insight into what we define as ‘urinary hypoxia’; the severity and duration of low intraoperative urinary PO2 that indicates increased risk of AKI. Our analysis suggests that a threshold of 15 mmHg may identify a level of risk that merits intervention. Established risk factors for AKI after cardiac surgery include high baseline serum creatinine and/or impaired eGFR [31, 32], heart failure [24, 30, 32], longer duration of CPB [1, 32], low pre- and intraoperative hemoglobin concentration [1], low DO2 during CPB [33, 34] and low intraoperative urine flow [31]. In multivariable logistic regression controlling for such factors with P ≤ 0.10 in univariable analysis, a strong independent association was found between AKI and intraoperative urinary hypoxia, as defined by urinary PO2 ≤15 mmHg for  more than or equal to the median duration for all 65 patients (4.8 min/h of surgery). In backward step-wise multivariable logistic regression, only urinary hypoxia, low intraoperative urine flow and a greater amount of norepinephrine administered post-bypass were independently associated with AKI. The predictive value of urinary hypoxia compares with those of current and emerging biomarkers of AKI. The area under the curve for urinary hypoxia established in the current study was 0.69. This is comparable to that achieved with available biomarkers of renal injury measured soon after weaning from CPB or within 24 h of surgery [35]. The only exception may be the product of urinary concentrations of tissue inhibitor of metalloproteinase 2 and insulin-like growth factor-binding protein 7, for which an area under the ROC curve of 0.81 (95% CI 0.68–0.93) was recently reported [36]. Nevertheless, unlike urinary PO2, none of these biomarkers has predictive efficacy when measured intraoperatively, or provides real-time information during CPB that might allow improved management of renal health. It is also noteworthy that the predictive efficacy of urinary hypoxia can be significantly improved by inclusion of certain pre-op (hypertension, low baseline eGFR and NYHA Class  ≥III) and intraoperative (urine flow) factors in the model. A number of lines of evidence indicate that urinary hypoxia during cardiac surgery reflects hypoxia in the renal medulla, as detailed in a recent review [37]. Medullary hypoxia has been observed during and after CPB in experimental animals [9–11] and is predicted by simulations using computational models [38, 39]. There is also clinical evidence that renal oxygen delivery falls during CPB [16] and that postoperative AKI is associated with intraoperative desaturation of renal tissue [40]. There is strong evidence that bladder urine PO2 provides a relatively noninvasive estimate of renal medullary PO2. In the human inner medulla, vascular bundles are distributed between clusters of collecting ducts, providing a pathway for oxygen diffusion between vasa recta and urine [41]. Leonhardt et al. showed more than 50 years ago that the PO2 of urine in the human renal pelvis is in equilibrium with renal medullary tissue PO2 [42]. Provided urine flow is relatively high, this ‘signal’ of medullary oxygenation appears to survive the journey from the pelvis to the bladder. This proposition is supported by recent experimental observations in sheep [19, 20] and rabbits [21], showing strong correlations between renal medullary PO2 and bladder urine PO2 under a range of conditions, and simulations using computational models of oxygen diffusion across the ureteric wall [21, 39]. Critically, CPB is usually accompanied by brisk diuresis, so provides an excellent platform to test the utility of bladder urine PO2 as an intraoperative ‘physiological biomarker’ of risk of AKI. Several established risk factors for AKI after CPB were significantly associated with urinary hypoxia by univariable (duration of surgery, hemoglobin concentration and whole-body DO2 during CPB and requirement for epinephrine and/or norepinephrine support after weaning from CPB) and multivariable logistic regression. Thus, urinary hypoxia might capture the combined effects of pre- and intraoperative factors that lead to medullary hypoxia during on-pump cardiac surgery. Our current findings show that urinary hypoxia during cardiac surgery indicates increased risk of AKI. But how might we intervene to ameliorate this risk? There is little evidence that specific pharmacological or non-pharmacological interventions can have a major impact on the risk of AKI associated with cardiac surgery [43, 44]. Continuous measurement of urinary PO2 might provide a translational pathway for investigating novel interventions to prevent AKI, since it can be deployed in both experimental animals, in which regional kidney perfusion and oxygenation can be directly monitored [19–21], and in man. Our current study is limited by a relatively small sample size at a single center. For example, the small sample size precludes a subanalysis in which we might exclude mild renal dysfunction (Stage 1), or employ newer statistical methods such as the net reclassification index [45] or integrated discrimination index [46]. However, strengths include the prospective design and the use of a fiber optic method for continuous measurement of urinary PO2 that could feasibly be applied to routine clinical practice. Our finding that urinary PO2 progressively decreases after the start of CPB is consistent with previous clinical observations using polarographic electrodes [47]. In this previous study, post-CPB recovery of urinary PO2 was associated with better postoperative renal function. Our current observations provide two major advances. First, we show that urinary hypoxia during surgery, rather than just post-CPB recovery of urinary PO2, predicts development of AKI. Secondly, the fiber optic probe we used avoids many of the limitations of polarographic electrodes such as fragility and the need for repeated calibration. Thus, the method we have used is feasible for routine clinical use and provides prognostic information during surgery, when it is still feasible to intervene to prevent AKI. SUPPLEMENTARY DATA Supplementary data are available at ndt online. ACKNOWLEDGEMENTS The authors thank the surgeons, perfusionists, anesthetists and other members of the operation theater and administrative staff at Monash Health who helped with these studies. FUNDING This work was supported by grants from the National Health and Medical Research Council of Australia [GNT1122455] and The National Heart Foundation of Australia [VG101377]. AUTHORS’ CONTRIBUTIONS R.G.E., A.D.C., J.A.S. and A.G.T. were responsible for study concept and design. M.Z.L.Z., A.M., G.K.H., J.P.N. and R.G.E. were responsible for acquisition, analysis and interpretation of data. M.Z.L.Z. and R.G.E. were responsible for drafting of the manuscript. M.Z.L.Z., A.M., A.D.C., J.A.S., A.G.T., G.K.H., J.P.N. and R.G.E. were responsible for critical revision of the manuscript for intellectual content. R.G.E., J.A.S., A.D.C. and A.G.T. were responsible for obtaining funding. M.Z.L.Z. and A.G.T. were responsible for analysis. CONFLICT OF INTEREST STATEMENT None declared. REFERENCES 1 Karkouti K, Wijeysundera DN, Yau TM et al.   Acute kidney injury after cardiac surgery: focus on modifiable risk factors. Circulation  2009; 119: 495– 502 Google Scholar CrossRef Search ADS PubMed  2 Lenihan CR, Montez-Rath ME, Mora Mangano CT et al.   Trends in acute kidney injury, associated use of dialysis, and mortality after cardiac surgery, 1999 to 2008. Ann Thorac Surg  2013; 95: 20– 28 Google Scholar CrossRef Search ADS PubMed  3 de Geus HR, Betjes MG, Bakker J. Biomarkers for the prediction of acute kidney injury: a narrative review on current status and future challenges. Clin Kidney J  2012; 5: 102– 108 Google Scholar CrossRef Search ADS PubMed  4 Koyner JL, Vaidya VS, Bennett MR et al.   Urinary biomarkers in the clinical prognosis and early detection of acute kidney injury. Clin J Am Soc Nephrol  2010; 5: 2154– 2165 Google Scholar CrossRef Search ADS PubMed  5 McIlroy DR, Farkas D, Matto M et al.   Neutrophil gelatinase-associated lipocalin combined with delta serum creatinine provides early risk stratification for adverse outcomes after cardiac surgery: a prospective observational study. Crit Care Med  2015; 43: 1043– 1052 Google Scholar CrossRef Search ADS PubMed  6 Bellomo R, Auriemma S, Fabbri A et al.   The pathophysiology of cardiac surgery-associated acute kidney injury (CSA-AKI). Int J Artif Organs  2008; 31: 166– 178 Google Scholar CrossRef Search ADS PubMed  7 Mao H, Katz N, Ariyanon W et al.   Cardiac surgery-associated acute kidney injury. Blood Purif  2014; 37: 34– 50 Google Scholar CrossRef Search ADS PubMed  8 Evans RG, Ince C, Joles JA et al.   Haemodynamic influences on kidney oxygenation: clinical implications of integrative physiology. Clin Exp Pharmacol Physiol  2013; 40: 106– 122 Google Scholar CrossRef Search ADS PubMed  9 Stafford-Smith M, Grocott HP. Renal medullary hypoxia during experimental cardiopulmonary bypass: a pilot study. Perfusion  2005; 20: 53– 58 Google Scholar CrossRef Search ADS PubMed  10 Patel NN, Toth T, Jones C et al.   Prevention of post-cardiopulmonary bypass acute kidney injury by endothelin A receptor blockade. Crit Care Med  2011; 39: 793– 802 Google Scholar CrossRef Search ADS PubMed  11 Darby PJ, Kim N, Hare GM et al.   Anemia increases the risk of renal cortical and medullary hypoxia during cardiopulmonary bypass. Perfusion  2013; 28: 504– 511 Google Scholar CrossRef Search ADS PubMed  12 Calzavacca P, Evans RG, Bailey M et al.   Cortical and medullary tissue perfusion and oxygenation in experimental septic acute kidney injury. Crit Care Med  2015; 43: e431– e439 Google Scholar CrossRef Search ADS PubMed  13 Abdelkader A, Ho J, Ow CP et al.   Renal oxygenation in acute renal ischemia-reperfusion injury. Am J Physiol Renal Physiol  2014; 306: F1026– F1038 Google Scholar CrossRef Search ADS PubMed  14 Heyman SN, Rosen S, Rosenberger C. Renal parenchymal hypoxia, hypoxia adaptation, and the pathogenesis of radiocontrast nephropathy. Clin J Am Soc Nephrol  2008; 3: 288– 296 Google Scholar CrossRef Search ADS PubMed  15 Rosenberger C, Goldfarb M, Shina A et al.   Evidence for sustained renal hypoxia and transient hypoxia adaptation in experimental rhabdomyolysis-induced acute kidney injury. Nephrol Dial Transplant  2007; 23: 1135– 1143 Google Scholar CrossRef Search ADS PubMed  16 Lannemyr L, Bragadottir G, Krumbholz V et al.   Effects of cardiopulmonary bypass on renal perfusion, filtration, and oxygenation in patients undergoing cardiac surgery. Anesthesiology  2017; 126: 205– 213 Google Scholar CrossRef Search ADS PubMed  17 Evans RG, Smith JA, Gardiner BS et al.   Hypoxia as a biomarker of kidney disease. In: Patel VB, Preedy VR (eds). Diseases of the Kidney. Biomarkers of Disease: Methods, Discoveries and Applications . Berlin: Springer-Verlag, 2016, 83– 105 Google Scholar CrossRef Search ADS   18 Evans RG, Smith JA, Wright C et al.   Urinary oxygen tension: a clinical window on the health of the renal medulla? Am J Physiol Regul Integr Comp Physiol  2014; 306: R45– R50 Google Scholar CrossRef Search ADS PubMed  19 Lankadeva YR, Kosaka J, Evans RG et al.   Intrarenal and urinary oxygenation during norepinephrine resuscitation in ovine septic acute kidney injury. Kidney Int  2016; 90: 100– 108 Google Scholar CrossRef Search ADS PubMed  20 Lankadeva YR, Kosaka J, Evans RG et al.   Urinary oxygenation as a surrogate measure of medullary oxygenation during angiotensin II therapy in septic acute kidney injury. Crit Care Med  2018; 46: e41– e48 Google Scholar CrossRef Search ADS PubMed  21 Sgouralis I, Kett MM, Ow CP et al.   Bladder urine oxygen tension for assessing renal medullary oxygenation in rabbits: experimental and modelling studies. Am J Physiol Regul Integr Comp Physiol  2016; 311: R532– R544 Google Scholar CrossRef Search ADS PubMed  22 Kidney Disease: Improving Global Outcomes (KDIGO) Acute Kidney Injury Working Group. KDIGO clinical practice guideline for acute kidney injury. Kidney Int Suppl  2012; 2: 1– 138 CrossRef Search ADS   23 Roques F, Michel P, Goldstone AR et al.   The logistic EuroSCORE. Eur Heart J  2003; 24: 1– 2 Google Scholar CrossRef Search ADS   24 Mehta RH, Grab JD, O'Brien SM et al.   Bedside tool for predicting the risk of postoperative dialysis in patients undergoing cardiac surgery. Circulation  2006; 114: 2208– 2216 Google Scholar CrossRef Search ADS PubMed  25 McIlroy DR, Argenziano M, Farkas D et al.   Incorporating oliguria into the diagnostic criteria for acute kidney injury after on-pump cardiac surgery: impact on incidence and outcomes. J Cardiothorac Vasc Anesth  2013; 27: 1145– 1152 Google Scholar CrossRef Search ADS PubMed  26 Koeze J, Keus F, Dieperink W et al.   Incidence, timing and outcome of AKI in critically ill patients varies with the definition used and the addition of urine output criteria. BMC Nephrol  2017; 18: 70 Google Scholar CrossRef Search ADS PubMed  27 Pallant J. SPSS Survival Manual: A Step by Step Guide to DATA Analysis Using SPSS for Windows Version 15 . Maidenhead Berkshire, UK: Open University Press, 2007 28 Ludbrook J. Repeated measurements and multiple comparisons in cardiovascular research. Cardiovasc Res  1994; 28: 303– 311 Google Scholar CrossRef Search ADS PubMed  29 Levey AS, Bosch JP, Lewis JB et al.   A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Ann Intern Med  1999; 130: 461– 470 Google Scholar CrossRef Search ADS PubMed  30 Thakar CV, Arrigain S, Worley S et al.   A clinical score to predict acute renal failure after cardiac surgery. J Am Soc Nephrol  2004; 16: 162– 168 Google Scholar CrossRef Search ADS PubMed  31 Parolari A, Pesce LL, Pacini D et al.   Risk factors for perioperative acute kidney injury after adult cardiac surgery: role of perioperative management. Ann Thorac Surg  2012; 93: 584– 591 Google Scholar CrossRef Search ADS PubMed  32 Ng SY, Sanagou M, Wolfe R et al.   Prediction of acute kidney injury within 30 days of cardiac surgery. J Thorac Cardiovasc Surg  2014; 147: 1875– 1883 Google Scholar CrossRef Search ADS PubMed  33 de Somer F, Mulholland JW, Bryan MR et al.   O2 delivery and CO2 production during cardiopulmonary bypass as determinants of acute kidney injury: time for a goal-directed perfusion management? Crit Care  2011; 15: R192 Google Scholar CrossRef Search ADS PubMed  34 Magruder JT, Dungan SP, Grimm JC et al.   Nadir oxygen delivery on bypass and hypotension increase acute kidney injury risk after cardiac operations. Ann Thorac Surg  2015; 100: 1697– 1703 Google Scholar CrossRef Search ADS PubMed  35 Ho J, Tangri N, Komenda P et al.   Urinary, plasma, and serum biomarkers' utility for predicting acute kidney injury associated with cardiac surgery in adults: a meta-analysis. Am J Kidney Dis  2015; 66: 993– 1005 Google Scholar CrossRef Search ADS PubMed  36 Meersch M, Schmidt C, Van Aken H et al.   Urinary TIMP-2 and IGFBP7 as early biomarkers of acute kidney injury and renal recovery following cardiac surgery. PLoS One  2014; 9: e93460 Google Scholar CrossRef Search ADS PubMed  37 Evans RG, Lankadeva YR, Cochrane AD et al.   Renal haemodynamics and oxygenation during and after cardiac surgery and cardiopulmonary bypass. Acta Physiol (Oxf)  2018; 222: e12995 Google Scholar CrossRef Search ADS   38 Sgouralis I, Evans RG, Gardiner BS et al.   Renal hemodynamics, function and oxygenation during cardiac surgery performed on cardiopulmonary bypass: a modeling study. Physiol Rep  2015; 3: e12260 Google Scholar CrossRef Search ADS PubMed  39 Sgouralis I, Evans RG, Layton AT. Renal medullary and urinary oxygen tension during cardiopulmonary bypass in the rat. Math Med Biol  2017; 34: 313– 333 Google Scholar PubMed  40 Choi DK, Kim WJ, Chin JH et al.   Intraoperative renal regional oxygen desaturation can be a predictor for acute kidney injury after cardiac surgery. J Cardiothorac Vasc Anesth  2014; 28: 564– 571 Google Scholar CrossRef Search ADS PubMed  41 Wei G, Rosen S, Dantzler WH et al.   Architecture of the human renal inner medulla and functional implications. Am J Physiol Renal Physiol  2015; 309: F627– F637 Google Scholar CrossRef Search ADS PubMed  42 Leonhardt KO, Landes RR, McCauley RT. Anatomy and physiology of intrarenal oxygen tension: preliminary study of the effects of anesthetics. Anesthesiology  1965; 26: 648– 658 Google Scholar CrossRef Search ADS PubMed  43 Patel NN, Rogers CA, Angelini GD et al.   Pharmacological therapies for the prevention of acute kidney injury following cardiac surgery: a systematic review. Heart Fail Rev  2011; 16: 553– 567 Google Scholar CrossRef Search ADS PubMed  44 O'Neal JB, Shaw AD, Billings FT. Acute kidney injury following cardiac surgery: current understanding and future directions. Crit Care  2016; 20: 187 Google Scholar CrossRef Search ADS PubMed  45 Kerr KF, Wang Z, Janes H et al.   Net reclassification indices for evaluating risk prediction instruments: a critical review. Epidemiology  2014; 25: 114– 121 Google Scholar CrossRef Search ADS PubMed  46 Kerr KF, McClelland RL, Brown ER et al.   Evaluating the incremental value of new biomarkers with integrated discrimination improvement. Am J Epidemiol  2011; 174: 364– 374 Google Scholar CrossRef Search ADS PubMed  47 Kainuma M, Yamada M, Miyake T. Continuous urine oxygen tension monitoring in patients undergoing cardiac surgery. J Cardiothorac Vasc Anesth  1996; 10: 603– 608 Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nephrology Dialysis Transplantation Oxford University Press

Urinary hypoxia: an intraoperative marker of risk of cardiac surgery-associated acute kidney injury

Loading next page...
 
/lp/ou_press/urinary-hypoxia-an-intraoperative-marker-of-risk-of-cardiac-surgery-0EbpwLkGXw
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved.
ISSN
0931-0509
eISSN
1460-2385
D.O.I.
10.1093/ndt/gfy047
Publisher site
See Article on Publisher Site

Abstract

Abstract Background Acute kidney injury (AKI) is common after cardiac surgery and profoundly affects postoperative mortality and morbidity. There are no validated methods to assess risk of AKI intraoperatively. Methods We determined the association between postoperative AKI and intraoperative urinary oxygen tension (PO2), measured via a fiber optic probe in the tip of the urinary catheter, in 65 patients undergoing high-risk cardiac surgery requiring cardiopulmonary bypass (CPB). AKI was diagnosed by modified Kidney Disease: Improving Global Outcomes criteria. Results Urinary PO2 fell during the operation, often reaching its nadir during rewarming or after weaning from CPB. Nadir urinary PO2 was lower in the 26 patients who developed AKI (mean ± SD, 8.9 ± 5.6 mmHg) than in the 39 patients who did not (14.9 ± 10.2 mmHg, P = 0.008). Patients who developed AKI had longer periods of urinary PO2 ≤15 and 10 mmHg than patients who did not. Odds of AKI increased when urinary PO2 fell to ≤10 mmHg {3.60 [95% confidence interval (CI) 1.27–10.21]} or ≤5 mmHg [3.60 (95% CI 1.04–12.42), P = 0.04] during the operation. When urinary PO2 fell to ≤15 mmHg, for more than or equal to  the median duration for all patients (4.8 min/h surgery), the odds of AKI were 4.85 (95% CI 1.64–14.40), P = 0.004. The area under the receiver-operator curve for this parameter alone was 0.69, and was 0.89 when other variables with P ≤ 0.10 in univariable analysis were included in the model. Conclusion Low urinary PO2 during adult cardiac surgery requiring CPB predicts AKI, so may identify patients in which intervention to improve renal oxygenation might reduce the risk of AKI. biomarkers, cardiopulmonary bypass, hypoxia, prognosis, renal failure INTRODUCTION Acute kidney injury (AKI) occurs in ∼30% of patients after cardiac surgery [1]. Even mild AKI after cardiopulmonary bypass (CPB) is associated with a >4-fold increase in risk of in-hospital death [1]. In severe AKI, requiring renal replacement therapy (1–2% of patients after surgery on CPB), mortality exceeds 35% [2]. Available biomarkers [3] offer limited prediction of AKI at arrival in the intensive care unit (ICU) [4, 5], but there are currently no biomarkers with predictive value during surgery, when intervention is still feasible. The etiology of AKI after cardiac surgery is complex [6, 7]. Nevertheless, hypoxia in the renal medulla appears to be a hallmark of AKI [8], being present not just during and after experimental CPB [9–11], but also in AKI induced by sepsis [12], ischemia-reperfusion injury [13], administration of radiocontrast agents [14] and rhabdomyolysis [15]. Additionally, renal oxygen delivery is impaired in humans on CPB [16]. Medullary hypoxia may thus represent a major driver of AKI. It is not feasible to measure renal medullary oxygen tension (PO2) in patients during cardiac surgery. However, it is feasible to measure continuously the PO2 of bladder urine via a fiber optic probe placed in the bladder catheter [17, 18]. Bladder urine PO2 correlates strongly with renal medullary PO2 in sheep during hyperdynamic septic AKI [19, 20] and in rabbits during pharmacologically induced changes in medullary PO2 [21]. Thus, we conducted a pilot study to test the hypothesis that urinary hypoxia during cardiac surgery requiring CPB predicts later development of AKI. MATERIALS AND METHODS Observational protocol The protocol was approved by the Human Research Ethics Committee of Monash Health (Reference: 12375B) and conformed to the Declaration of Helsinki. All patients provided informed consent. Between January 2015 and August 2017, 68 patients undergoing cardiac surgery requiring CPB were prospectively enrolled. Three patients were subsequently excluded, leaving a sample of 65 patients (Supplementary data, Figure S1 and Supplementary Methods). We included patients undergoing coronary artery bypass graft (CABG) surgery, valve repair/replacement surgery or combined ‘CABG–valve’ surgery. We excluded patients: with a baseline serum creatinine >200 μmol/L, or an estimated glomerular filtration rate (eGFR) <30 mL/min/1.73 m2; on hemodialysis (acutely or long term); who were kidney transplant recipients; with preoperative (pre-op) AKI as defined by the Kidney Disease: Improving Global Outcomes (KDIGO) AKI criteria [22]; or who were unable to provide informed consent. The sample of convenience was enriched with patients deemed at high risk of in-hospital death, based on their logistic EuroSCORE [23]. Many of the factors in the EuroSCORE are also risk factors for postoperative AKI [24]. Anesthesia, surgery and perfusion Standard institutional protocols were followed (Supplementary Methods). Continuous measurement of urinary PO2 The urinary bladder was catheterized with a standard Foley urinary catheter (Medtronic, Minneapolis, MN, USA). A sterilized fiber optic luminescence optode (NX-LAS-1/O/E-5 m, Oxford Optronix, Abingdon, UK) was advanced through the lumen of the catheter, to the catheter tip, so that the sensing tip of the probe was in direct contact with bladder urine. The bladder catheter was then connected to a standard urine collection system (see Supplementary Methods). Urinary PO2 was measured at 6-s intervals during surgery. Other data available from patient monitoring systems were recorded every 5 min (see Supplementary Methods). Definition of AKI Serum creatinine concentration was assessed preoperatively, at the time of entry to the ICU, and then daily (between midnight and 5:00 am) thereafter until discharge. The primary outcome was postoperative AKI, defined by modified KDIGO criteria [22], that is, a maximum increase in serum creatinine of ≥0.30 mg/dL (26.5 µmol/L) from pre-op baseline within the first 48 h after surgery, or ≥50% within the first 5 days after surgery. We did not utilize the urine output criterion for diagnosis of AKI, on the basis that it may result in over-diagnosis [25, 26]. Statistical analysis Continuous variables that satisfied the Shapiro–Wilk test are presented as mean ± SD [27]. For variables that failed this test of normality, the data were plotted and if the distribution was skewed, data are presented as median (interquartile range) and nonparametric statistical tests were used. Statistical analyses were performed using SPSS (Version 23.0, IBM Corp., Armonk, NY, USA). Two-sided P ≤ 0.05 was considered statistically significant. Dichotomous comparisons were made using Student’s unpaired t-test or the Mann–Whitney U test. Differences in proportions were assessed using chi-square, or with Fisher’s exact test when subgroups had fewer than five observations. P-values from within-subject factors in repeated measures analysis of variance were adjusted using the Greenhouse–Geisser method [28]. Univariable logistic regression was used to identify associations between various thresholds of intraoperative urinary PO2 and development of postoperative AKI. The sensitivity and specificity of each threshold was assessed using receiver-operator characteristic (ROC) curves. Factors with P ≤ 0.10 were then introduced into multivariable regression analyses. Model 1 included all variables with P ≤ 0.10 from univariable analysis. In multivariable Model 2, variables with P > 0.10 were removed step-wise until only those with P ≤ 0.10 remained. A similar approach was taken to assess factors associated with urinary hypoxia. RESULTS Patient and surgical characteristics Of the 65 patients included in the analyses, 26 (40%) developed AKI (Table 1). Table 1. Baseline characteristics of the sample of patients Characteristic  All Patients  No AKI  AKI  P  (n = 65)  (n = 39)  (n = 26)  Age, years  69.8 ± 8.91  70.8 ± 8.6  68.4 ± 9.4  0.29  Female  11 (16.9)  5 (12.8)  6 (23.1)  0.28  Body mass index, kg/m2  28.8 ± 4.9  28.7 ± 5.2  28.9 ± 4.5  0.90  Type of surgery           Isolated CABG  41 (63.1)  27 (69.2)  14 (53.8)  0.21   Single valve  8 (12.3)  5 (12.8)  3 (11.5)  >0.99b   >1 valve  4 (6.2)  0 (0.0)  4 (15.4)  0.02b   Single valve + CABG  12 (18.5)  7 (17.9)  5 (19.2)  >0.99b  Status of surgery           Elective  44 (67.7)  27 (69.2)  17 (65.4)  0.75   Urgent  20 (30.8)  12 (30.8)  8 (30.8)  >0.99   Emergency  1 (2.1)  0 (0.0)  1 (3.8)  0.40b  Risk factors and comorbidity           Diabetes mellitus  27 (41.5)  15 (38.5)  12 (46.2)  0.54   Hypertension  49 (75.4)  26 (66.7)  23 (88.5)  0.05   Hypercholesterolemia  48 (73.8)  27 (69.2)  21 (80.3)  0.30   Smoking history  41 (63.1)  25 (64.1)  16 (61.5)  0.83   Chronic lung disease  6 (9.2)  3 (7.7)  3 (11.5)  0.68b   Peripheral vascular disease  9 (13.8)  5 (12.8)  4 (15.4)  >0.99b   Cerebrovascular disease  8 (12.3)  5 (12.8)  3 (11.5)  >0.99b   NYHA ≥III  22 (33.8)  9 (23.1)  13 (50.0)  0.03   Congestive cardiac failure  12 (18.8)  6 (15.4)  6 (24.0)  0.43   Pre-op inotropes  2 (3.1)  1 (2.6)  1 (3.8)  >0.99b   Estimated LVEF <30%  6 (9.2)  2 (5.1)  4 (15.4)  0.21b   Arrhythmia  19 (29.2)  10 (25.6)  9 (34.6)  0.44   Redo sternotomy  5 (7.7)  2 (5.1)  3 (11.5)  0.38b   Baseline serum creatinine, μmol/L  87.7 ± 27.3  82.7 ± 26.1  95.2 ± 27.8  0.07   Baseline eGFRa, mL/min/1.73 m2  76.1 ± 20.3  80.9 ± 20.2  69.0 ± 18.8  0.02  Clinical risk scores       Logistic EuroSCORE, median (Q1, Q3)  4.7 (2.6, 9.2)  4.6 (2.6, 9.9)  5.4 (2.5, 8.7)  0.70c   Mehta Score, % risk, median (Q1, Q3)  0.7 (0.4, 1.3)  0.6 (0.4, 0.9)  1.1 (0.4, 1.9)  0.05c   Cleveland Score, % risk, median (Q1, Q3)  0.3 (0.1, 1.1)  0.3 (0.1, 0.6)  0.6 (0.1, 2.2)  0.11c  Characteristic  All Patients  No AKI  AKI  P  (n = 65)  (n = 39)  (n = 26)  Age, years  69.8 ± 8.91  70.8 ± 8.6  68.4 ± 9.4  0.29  Female  11 (16.9)  5 (12.8)  6 (23.1)  0.28  Body mass index, kg/m2  28.8 ± 4.9  28.7 ± 5.2  28.9 ± 4.5  0.90  Type of surgery           Isolated CABG  41 (63.1)  27 (69.2)  14 (53.8)  0.21   Single valve  8 (12.3)  5 (12.8)  3 (11.5)  >0.99b   >1 valve  4 (6.2)  0 (0.0)  4 (15.4)  0.02b   Single valve + CABG  12 (18.5)  7 (17.9)  5 (19.2)  >0.99b  Status of surgery           Elective  44 (67.7)  27 (69.2)  17 (65.4)  0.75   Urgent  20 (30.8)  12 (30.8)  8 (30.8)  >0.99   Emergency  1 (2.1)  0 (0.0)  1 (3.8)  0.40b  Risk factors and comorbidity           Diabetes mellitus  27 (41.5)  15 (38.5)  12 (46.2)  0.54   Hypertension  49 (75.4)  26 (66.7)  23 (88.5)  0.05   Hypercholesterolemia  48 (73.8)  27 (69.2)  21 (80.3)  0.30   Smoking history  41 (63.1)  25 (64.1)  16 (61.5)  0.83   Chronic lung disease  6 (9.2)  3 (7.7)  3 (11.5)  0.68b   Peripheral vascular disease  9 (13.8)  5 (12.8)  4 (15.4)  >0.99b   Cerebrovascular disease  8 (12.3)  5 (12.8)  3 (11.5)  >0.99b   NYHA ≥III  22 (33.8)  9 (23.1)  13 (50.0)  0.03   Congestive cardiac failure  12 (18.8)  6 (15.4)  6 (24.0)  0.43   Pre-op inotropes  2 (3.1)  1 (2.6)  1 (3.8)  >0.99b   Estimated LVEF <30%  6 (9.2)  2 (5.1)  4 (15.4)  0.21b   Arrhythmia  19 (29.2)  10 (25.6)  9 (34.6)  0.44   Redo sternotomy  5 (7.7)  2 (5.1)  3 (11.5)  0.38b   Baseline serum creatinine, μmol/L  87.7 ± 27.3  82.7 ± 26.1  95.2 ± 27.8  0.07   Baseline eGFRa, mL/min/1.73 m2  76.1 ± 20.3  80.9 ± 20.2  69.0 ± 18.8  0.02  Clinical risk scores       Logistic EuroSCORE, median (Q1, Q3)  4.7 (2.6, 9.2)  4.6 (2.6, 9.9)  5.4 (2.5, 8.7)  0.70c   Mehta Score, % risk, median (Q1, Q3)  0.7 (0.4, 1.3)  0.6 (0.4, 0.9)  1.1 (0.4, 1.9)  0.05c   Cleveland Score, % risk, median (Q1, Q3)  0.3 (0.1, 1.1)  0.3 (0.1, 0.6)  0.6 (0.1, 2.2)  0.11c  Values are mean ± SD or n (%) unless otherwise stated. P-values are derived from Student’s unpaired t-test for continuous variables and chi-square test for proportions unless otherwise shown. a eGFR was calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation [29]. Risk scores are shown for (i) The European System for Cardiac Operative Risk Evaluation (Logistic EuroSCORE) [23]; (ii) Mehta Score [24]; and (iii) the Cleveland Clinic Score [30], for prediction of risk of dialysis after cardiac surgery. Of the 26 patients who developed AKI, 15 had Stage 1 injury (increase in serum creatinine >26.5 μmol/L from baseline within 48 h, or >50% within 5 days); 7 had Stage 2 injury (increase in serum creatinine >100% within 5 days); and 4 patients developed Stage 3 injury (increase in serum creatinine >200% within 5 days or need for hemodialysis) [22]. Three patients required postoperative dialysis and were the only patients in the cohort of 65 who died in hospital. For the 26 patients who developed AKI, the diagnostic criterion was reached, on average, 20.6 ± 14.8 h after admission to the ICU. bFisher’s exact test. cMann Whitney U test. LVEF, left ventricular ejection fraction; NYHA, New York Heart Association Heart Failure Classification. Table 1. Baseline characteristics of the sample of patients Characteristic  All Patients  No AKI  AKI  P  (n = 65)  (n = 39)  (n = 26)  Age, years  69.8 ± 8.91  70.8 ± 8.6  68.4 ± 9.4  0.29  Female  11 (16.9)  5 (12.8)  6 (23.1)  0.28  Body mass index, kg/m2  28.8 ± 4.9  28.7 ± 5.2  28.9 ± 4.5  0.90  Type of surgery           Isolated CABG  41 (63.1)  27 (69.2)  14 (53.8)  0.21   Single valve  8 (12.3)  5 (12.8)  3 (11.5)  >0.99b   >1 valve  4 (6.2)  0 (0.0)  4 (15.4)  0.02b   Single valve + CABG  12 (18.5)  7 (17.9)  5 (19.2)  >0.99b  Status of surgery           Elective  44 (67.7)  27 (69.2)  17 (65.4)  0.75   Urgent  20 (30.8)  12 (30.8)  8 (30.8)  >0.99   Emergency  1 (2.1)  0 (0.0)  1 (3.8)  0.40b  Risk factors and comorbidity           Diabetes mellitus  27 (41.5)  15 (38.5)  12 (46.2)  0.54   Hypertension  49 (75.4)  26 (66.7)  23 (88.5)  0.05   Hypercholesterolemia  48 (73.8)  27 (69.2)  21 (80.3)  0.30   Smoking history  41 (63.1)  25 (64.1)  16 (61.5)  0.83   Chronic lung disease  6 (9.2)  3 (7.7)  3 (11.5)  0.68b   Peripheral vascular disease  9 (13.8)  5 (12.8)  4 (15.4)  >0.99b   Cerebrovascular disease  8 (12.3)  5 (12.8)  3 (11.5)  >0.99b   NYHA ≥III  22 (33.8)  9 (23.1)  13 (50.0)  0.03   Congestive cardiac failure  12 (18.8)  6 (15.4)  6 (24.0)  0.43   Pre-op inotropes  2 (3.1)  1 (2.6)  1 (3.8)  >0.99b   Estimated LVEF <30%  6 (9.2)  2 (5.1)  4 (15.4)  0.21b   Arrhythmia  19 (29.2)  10 (25.6)  9 (34.6)  0.44   Redo sternotomy  5 (7.7)  2 (5.1)  3 (11.5)  0.38b   Baseline serum creatinine, μmol/L  87.7 ± 27.3  82.7 ± 26.1  95.2 ± 27.8  0.07   Baseline eGFRa, mL/min/1.73 m2  76.1 ± 20.3  80.9 ± 20.2  69.0 ± 18.8  0.02  Clinical risk scores       Logistic EuroSCORE, median (Q1, Q3)  4.7 (2.6, 9.2)  4.6 (2.6, 9.9)  5.4 (2.5, 8.7)  0.70c   Mehta Score, % risk, median (Q1, Q3)  0.7 (0.4, 1.3)  0.6 (0.4, 0.9)  1.1 (0.4, 1.9)  0.05c   Cleveland Score, % risk, median (Q1, Q3)  0.3 (0.1, 1.1)  0.3 (0.1, 0.6)  0.6 (0.1, 2.2)  0.11c  Characteristic  All Patients  No AKI  AKI  P  (n = 65)  (n = 39)  (n = 26)  Age, years  69.8 ± 8.91  70.8 ± 8.6  68.4 ± 9.4  0.29  Female  11 (16.9)  5 (12.8)  6 (23.1)  0.28  Body mass index, kg/m2  28.8 ± 4.9  28.7 ± 5.2  28.9 ± 4.5  0.90  Type of surgery           Isolated CABG  41 (63.1)  27 (69.2)  14 (53.8)  0.21   Single valve  8 (12.3)  5 (12.8)  3 (11.5)  >0.99b   >1 valve  4 (6.2)  0 (0.0)  4 (15.4)  0.02b   Single valve + CABG  12 (18.5)  7 (17.9)  5 (19.2)  >0.99b  Status of surgery           Elective  44 (67.7)  27 (69.2)  17 (65.4)  0.75   Urgent  20 (30.8)  12 (30.8)  8 (30.8)  >0.99   Emergency  1 (2.1)  0 (0.0)  1 (3.8)  0.40b  Risk factors and comorbidity           Diabetes mellitus  27 (41.5)  15 (38.5)  12 (46.2)  0.54   Hypertension  49 (75.4)  26 (66.7)  23 (88.5)  0.05   Hypercholesterolemia  48 (73.8)  27 (69.2)  21 (80.3)  0.30   Smoking history  41 (63.1)  25 (64.1)  16 (61.5)  0.83   Chronic lung disease  6 (9.2)  3 (7.7)  3 (11.5)  0.68b   Peripheral vascular disease  9 (13.8)  5 (12.8)  4 (15.4)  >0.99b   Cerebrovascular disease  8 (12.3)  5 (12.8)  3 (11.5)  >0.99b   NYHA ≥III  22 (33.8)  9 (23.1)  13 (50.0)  0.03   Congestive cardiac failure  12 (18.8)  6 (15.4)  6 (24.0)  0.43   Pre-op inotropes  2 (3.1)  1 (2.6)  1 (3.8)  >0.99b   Estimated LVEF <30%  6 (9.2)  2 (5.1)  4 (15.4)  0.21b   Arrhythmia  19 (29.2)  10 (25.6)  9 (34.6)  0.44   Redo sternotomy  5 (7.7)  2 (5.1)  3 (11.5)  0.38b   Baseline serum creatinine, μmol/L  87.7 ± 27.3  82.7 ± 26.1  95.2 ± 27.8  0.07   Baseline eGFRa, mL/min/1.73 m2  76.1 ± 20.3  80.9 ± 20.2  69.0 ± 18.8  0.02  Clinical risk scores       Logistic EuroSCORE, median (Q1, Q3)  4.7 (2.6, 9.2)  4.6 (2.6, 9.9)  5.4 (2.5, 8.7)  0.70c   Mehta Score, % risk, median (Q1, Q3)  0.7 (0.4, 1.3)  0.6 (0.4, 0.9)  1.1 (0.4, 1.9)  0.05c   Cleveland Score, % risk, median (Q1, Q3)  0.3 (0.1, 1.1)  0.3 (0.1, 0.6)  0.6 (0.1, 2.2)  0.11c  Values are mean ± SD or n (%) unless otherwise stated. P-values are derived from Student’s unpaired t-test for continuous variables and chi-square test for proportions unless otherwise shown. a eGFR was calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation [29]. Risk scores are shown for (i) The European System for Cardiac Operative Risk Evaluation (Logistic EuroSCORE) [23]; (ii) Mehta Score [24]; and (iii) the Cleveland Clinic Score [30], for prediction of risk of dialysis after cardiac surgery. Of the 26 patients who developed AKI, 15 had Stage 1 injury (increase in serum creatinine >26.5 μmol/L from baseline within 48 h, or >50% within 5 days); 7 had Stage 2 injury (increase in serum creatinine >100% within 5 days); and 4 patients developed Stage 3 injury (increase in serum creatinine >200% within 5 days or need for hemodialysis) [22]. Three patients required postoperative dialysis and were the only patients in the cohort of 65 who died in hospital. For the 26 patients who developed AKI, the diagnostic criterion was reached, on average, 20.6 ± 14.8 h after admission to the ICU. bFisher’s exact test. cMann Whitney U test. LVEF, left ventricular ejection fraction; NYHA, New York Heart Association Heart Failure Classification. Preoperatively, a greater proportion of patients who later developed AKI had hypertension, New York Heart Association (NYHA) Class  ≥III heart failure or were undergoing a double or triple valve procedure (Table 1). Patients who later developed AKI also had a 15% lower baseline eGFR and a greater pre-op risk of dialysis (Mehta Score). The median logistic EuroSCORE did not differ significantly between the two groups of patients. Patients who developed AKI had operations that were 25% longer in duration than patients who did not develop AKI (Supplementary data, Table S1). The durations of CPB (149 ± 88 versus 109 ± 52 min) and aortic cross-clamp (111 ± 63 versus 83 ± 48 min) were also longer for patients who developed AKI. Patients who developed AKI were more often administered furosemide, received greater amounts of norepinephrine after CPB, had 0.6 L more positive net volume status and 4.8 mmHg lower nadir mean arterial pressure (MAP) than patients who did not develop AKI (Supplementary data, Table S2). Postoperatively, patients with AKI required longer periods of ventilation and ICU stay and more often required blood products than patients who did not develop AKI (Supplementary data, Table S1). Intraoperative urinary PO2 Urinary PO2 progressively fell across the course of the surgical procedure (Figure 1A and B). After induction of anesthesia and before the first incision (epoch 1; baseline), the mean urinary PO2 was 65.5 ± 34.5 mmHg. Mean urinary PO2 then fell by 31% from pre-op levels, to 45.2 ± 25.9 mmHg, during the period after the first incision and before going on CPB (epoch 2). It fell by a further 24% from baseline, to a mean of 29.3 ± 19.5 mmHg during the first (mild hypothermic) phase of CPB. Mean urinary PO2 continued to fall by a further 9% during the second (rewarming) phase of CPB, to 23.4 ± 16.0 mmHg. Critically, even once patients were weaned from CPB, urinary PO2 did not recover to pre- or early-bypass levels. Post-bypass, urinary PO2 were on average less than half that of pre-bypass levels (23.5 ± 14.9 mmHg versus 49.1 ± 26.9 mmHg). FIGURE 1: View largeDownload slide Urinary oxygen tension during on-pump cardiac surgery. Urinary PO2 was recorded continuously for 98.8% of the period of CPB and 98.5% of the total duration of surgery. For each individual patient, urinary PO2 was averaged over time using all available measurements during each epoch. (A) Columns and error bars are mean ± SD of urinary PO2 during the five epochs of surgery. Pre-Op, the period after induction of anesthesia but before the first incision; Pre-CPB, the period of surgery prior to commencement of CPB; CPB-Hypothermic, hypothermic CPB; CPB-Rewarm, the period of rewarming on bypass; and Post-CPB, the period of surgery after weaning from CPB. PGroup, PEpoch and PGroup*Epoch are the outcomes of repeated measures analysis of variance. P-values above each pair of measurements are from Student’s unpaired t-test. (B) Scatterplot showing the mean urinary PO2 for each individual subject, across each epoch of surgery. Solid horizontal lines represent the between-subject mean urinary PO2. (C) The lowest level of urinary PO2 (nadir) recorded during surgery for each subject. Error bars are mean ± SD. The P-values are the outcomes of Student’s unpaired t-test. FIGURE 1: View largeDownload slide Urinary oxygen tension during on-pump cardiac surgery. Urinary PO2 was recorded continuously for 98.8% of the period of CPB and 98.5% of the total duration of surgery. For each individual patient, urinary PO2 was averaged over time using all available measurements during each epoch. (A) Columns and error bars are mean ± SD of urinary PO2 during the five epochs of surgery. Pre-Op, the period after induction of anesthesia but before the first incision; Pre-CPB, the period of surgery prior to commencement of CPB; CPB-Hypothermic, hypothermic CPB; CPB-Rewarm, the period of rewarming on bypass; and Post-CPB, the period of surgery after weaning from CPB. PGroup, PEpoch and PGroup*Epoch are the outcomes of repeated measures analysis of variance. P-values above each pair of measurements are from Student’s unpaired t-test. (B) Scatterplot showing the mean urinary PO2 for each individual subject, across each epoch of surgery. Solid horizontal lines represent the between-subject mean urinary PO2. (C) The lowest level of urinary PO2 (nadir) recorded during surgery for each subject. Error bars are mean ± SD. The P-values are the outcomes of Student’s unpaired t-test. The overall profile of mean urinary PO2 did not differ significantly between patients who developed AKI and those who did not (Figure 1). However, mean urinary PO2 during the post-bypass epoch was lower in patients who later developed AKI (18.7 ± 9.0 versus 26.7 ± 17.2 mmHg, P = 0.03). Furthermore, in patients who later developed AKI, the lowest level of urinary PO2 during surgery (nadir) was 40% less than that of patients in the non-AKI group (8.9 ± 5.6 versus 14.9 ± 10.2 mmHg, P = 0.008). In 72% of the 65 patients, intraoperative nadir urinary PO2 occurred during the rewarming or post-bypass epochs. This pattern was similar in patients who later developed AKI (19 of 26, 73%) and those who did not (28 of 39, 72%). Urinary hypoxia We calculated the time spent at or below the threshold levels of urinary PO2 of 15, 10 and 5 mmHg, as well as the area under the curve for these thresholds (Supplementary data, Figure S2). These variables are expressed as a function of the total duration of each surgical procedure to allow comparison between patients. Patients who later developed AKI experienced 11-fold longer periods of urinary PO2 ≤15 mmHg, and the area under the curve was 51-fold greater, than patients who did not later develop AKI (Table 2). Periods of urinary PO2 ≤10 or ≤5 mmHg also tended to occur more commonly in patients who later developed AKI, but these periods were rare. Thus, we considered a threshold of ≤15 mmHg as useful working definition of urinary hypoxia. Table 2. Intraoperative urinary hypoxia Variable  All Patients  No AKI  AKI  P  (n = 65)  (n = 39)  (n = 26)  Median nadir urinary PO2, mmHg  11.4 (6.2, 15.5)  13.5 (8.3, 15.8)  7.8 (4.2, 13.3)  0.008  Mean nadir urinary PO2, mmHg  12.5 ± 9.1  14.9 ± 10.2  8.9 ± 5.6  0.008  Mean urine flow, mL/min  4.9 ± 2.5  5.4 ± 2.8  4.1 ± 1.7  0.04  Urinary PO2 at or below threshold values (adjusted for the duration of surgery)   Area under curve, mmHg × min/h            ≤15 mmHg  7.2 (0.0, 83.9)  1.2 (0.0, 52.7)  61.5 (4.0, 151.4)  0.01    ≤10 mmHg  0.0 (0.0, 13.4)  0.0 (0.0, 2.3)  4.4 (0.0, 56.0)  0.01    ≤5 mmHg  0.0 (0.0, 0.0)  0.0 (0.0, 0.0)  0.0 (0.0, 0.9)  0.06   Time at or below threshold, min/h            ≤15 mmHg  4.8 (0.0, 21.2)  1.5 (0.0, 12.0)  16.7 (2.5, 23.2)  0.02    ≤10 mmHg  0.0 (0.0, 8.4)  0.0 (0.0, 2.2)  4.7 (0.0, 15.2)  0.02    ≤5 mmHg  0.0 (0.0, 0.0)  0.0 (0.0, 0.0)  0.0 (0.0, 2.0)  0.08  Variable  All Patients  No AKI  AKI  P  (n = 65)  (n = 39)  (n = 26)  Median nadir urinary PO2, mmHg  11.4 (6.2, 15.5)  13.5 (8.3, 15.8)  7.8 (4.2, 13.3)  0.008  Mean nadir urinary PO2, mmHg  12.5 ± 9.1  14.9 ± 10.2  8.9 ± 5.6  0.008  Mean urine flow, mL/min  4.9 ± 2.5  5.4 ± 2.8  4.1 ± 1.7  0.04  Urinary PO2 at or below threshold values (adjusted for the duration of surgery)   Area under curve, mmHg × min/h            ≤15 mmHg  7.2 (0.0, 83.9)  1.2 (0.0, 52.7)  61.5 (4.0, 151.4)  0.01    ≤10 mmHg  0.0 (0.0, 13.4)  0.0 (0.0, 2.3)  4.4 (0.0, 56.0)  0.01    ≤5 mmHg  0.0 (0.0, 0.0)  0.0 (0.0, 0.0)  0.0 (0.0, 0.9)  0.06   Time at or below threshold, min/h            ≤15 mmHg  4.8 (0.0, 21.2)  1.5 (0.0, 12.0)  16.7 (2.5, 23.2)  0.02    ≤10 mmHg  0.0 (0.0, 8.4)  0.0 (0.0, 2.2)  4.7 (0.0, 15.2)  0.02    ≤5 mmHg  0.0 (0.0, 0.0)  0.0 (0.0, 0.0)  0.0 (0.0, 2.0)  0.08  Values are median (Q1, Q3) or mean ± SD. P-values were derived from the Mann–Whitney U test or Student’s unpaired t-test. Urinary PO2 fell to ≤15 mmHg in 47 patients (25 non-AKI, 22 AKI), to ≤10 mmHg in 29 patients (13 non-AKI and 16 AKI) and to ≤5 mmHg in 13 patients (5 non-AKI and 8 AKI). The area under the curve was determined by integrating the magnitude by which urinary PO2 was less than or equal to each specific threshold by the time below the threshold (Supplementary Figure 2). Both time at or below each threshold and the areas under the curve were normalized to the total duration of the surgery. Table 2. Intraoperative urinary hypoxia Variable  All Patients  No AKI  AKI  P  (n = 65)  (n = 39)  (n = 26)  Median nadir urinary PO2, mmHg  11.4 (6.2, 15.5)  13.5 (8.3, 15.8)  7.8 (4.2, 13.3)  0.008  Mean nadir urinary PO2, mmHg  12.5 ± 9.1  14.9 ± 10.2  8.9 ± 5.6  0.008  Mean urine flow, mL/min  4.9 ± 2.5  5.4 ± 2.8  4.1 ± 1.7  0.04  Urinary PO2 at or below threshold values (adjusted for the duration of surgery)   Area under curve, mmHg × min/h            ≤15 mmHg  7.2 (0.0, 83.9)  1.2 (0.0, 52.7)  61.5 (4.0, 151.4)  0.01    ≤10 mmHg  0.0 (0.0, 13.4)  0.0 (0.0, 2.3)  4.4 (0.0, 56.0)  0.01    ≤5 mmHg  0.0 (0.0, 0.0)  0.0 (0.0, 0.0)  0.0 (0.0, 0.9)  0.06   Time at or below threshold, min/h            ≤15 mmHg  4.8 (0.0, 21.2)  1.5 (0.0, 12.0)  16.7 (2.5, 23.2)  0.02    ≤10 mmHg  0.0 (0.0, 8.4)  0.0 (0.0, 2.2)  4.7 (0.0, 15.2)  0.02    ≤5 mmHg  0.0 (0.0, 0.0)  0.0 (0.0, 0.0)  0.0 (0.0, 2.0)  0.08  Variable  All Patients  No AKI  AKI  P  (n = 65)  (n = 39)  (n = 26)  Median nadir urinary PO2, mmHg  11.4 (6.2, 15.5)  13.5 (8.3, 15.8)  7.8 (4.2, 13.3)  0.008  Mean nadir urinary PO2, mmHg  12.5 ± 9.1  14.9 ± 10.2  8.9 ± 5.6  0.008  Mean urine flow, mL/min  4.9 ± 2.5  5.4 ± 2.8  4.1 ± 1.7  0.04  Urinary PO2 at or below threshold values (adjusted for the duration of surgery)   Area under curve, mmHg × min/h            ≤15 mmHg  7.2 (0.0, 83.9)  1.2 (0.0, 52.7)  61.5 (4.0, 151.4)  0.01    ≤10 mmHg  0.0 (0.0, 13.4)  0.0 (0.0, 2.3)  4.4 (0.0, 56.0)  0.01    ≤5 mmHg  0.0 (0.0, 0.0)  0.0 (0.0, 0.0)  0.0 (0.0, 0.9)  0.06   Time at or below threshold, min/h            ≤15 mmHg  4.8 (0.0, 21.2)  1.5 (0.0, 12.0)  16.7 (2.5, 23.2)  0.02    ≤10 mmHg  0.0 (0.0, 8.4)  0.0 (0.0, 2.2)  4.7 (0.0, 15.2)  0.02    ≤5 mmHg  0.0 (0.0, 0.0)  0.0 (0.0, 0.0)  0.0 (0.0, 2.0)  0.08  Values are median (Q1, Q3) or mean ± SD. P-values were derived from the Mann–Whitney U test or Student’s unpaired t-test. Urinary PO2 fell to ≤15 mmHg in 47 patients (25 non-AKI, 22 AKI), to ≤10 mmHg in 29 patients (13 non-AKI and 16 AKI) and to ≤5 mmHg in 13 patients (5 non-AKI and 8 AKI). The area under the curve was determined by integrating the magnitude by which urinary PO2 was less than or equal to each specific threshold by the time below the threshold (Supplementary Figure 2). Both time at or below each threshold and the areas under the curve were normalized to the total duration of the surgery. Clinical parameters monitored across the entire surgical procedure Body temperature, MAP and saturation of arterial hemoglobin with oxygen (SAO2) did not differ significantly between patients who later developed AKI and those who did not (Figure 2). Intraoperative urine flow increased dramatically during CPB. It was greatest during the rewarming phase, reaching 7.4 ± 4.8 mL/min across all patients, before declining to 6.0 ± 4.7 mL/min in the post-bypass period. Intraoperative urine flow was on an average 3.2 mL/min (37%) less during the rewarming phase and 3.0 mL/min (42%) less during the post-bypass period in patients who later developed AKI, than those who did not. FIGURE 2: View largeDownload slide Clinical parameters measured across the five epochs of surgery. (A) Body temperature. (B) Mean arterial pressure (MAP). (C) Saturation of arterial hemoglobin with oxygen (SAO2). (D) Urine flow. Columns and error bars represent mean ± SD. P-values are outcomes of two-way repeated measures analysis of variance with the factors group and epoch. ** denotes P ≤ 0.01 for comparisons between AKI and non-AKI groups, generated from Student’s unpaired t-test (post hoc analysis). The five epochs of surgery are as for Figure 1. Note: urine flow was not measured during the pre-op period, when the urinary catheter was first deployed. FIGURE 2: View largeDownload slide Clinical parameters measured across the five epochs of surgery. (A) Body temperature. (B) Mean arterial pressure (MAP). (C) Saturation of arterial hemoglobin with oxygen (SAO2). (D) Urine flow. Columns and error bars represent mean ± SD. P-values are outcomes of two-way repeated measures analysis of variance with the factors group and epoch. ** denotes P ≤ 0.01 for comparisons between AKI and non-AKI groups, generated from Student’s unpaired t-test (post hoc analysis). The five epochs of surgery are as for Figure 1. Note: urine flow was not measured during the pre-op period, when the urinary catheter was first deployed. Hemodynamics and whole-body oxygenation during CPB No differences in hemodynamic parameters of CPB were detected between patients who later developed AKI and those who did not (Table 3). Table 3. Hemodynamics and whole-body oxygenation during CPB Variable  CPB-Hypothermic (∼34°C)   CPB-Rewarm   PGroup  PEpoch  PGroup*Epoch  No AKI  AKI  No AKI  AKI  (n = 39)  (n = 26)  (n = 39)  (n = 26)  Pump flow, L/min  4.68 ± 0.57  4.68 ± 0.52  4.78 ± 0.54  4.75 ± 0.50  0.92  0.007  0.70  Pump flow, L/min/m2  2.40 ± 0.22  2.40 ± 0.17  2.45 ± 0.21  2.43 ± 0.18  0.97  0.005  0.71  PAO2, mmHg  244 ± 31  249 ± 25  238 ± 27  253 ± 32  0.12  0.85  0.14  PACO2, mmHg  38.8 ± 3.2  38.9 ± 2.6  40.5 ± 2.5  39.3 ± 2.8  0.36  0.008  0.11  Hemoglobin, g/L  87.3 ± 11.0  83.5 ± 13.4  88.0 ± 10.7  82.6 ± 12.2  0.12  0.92  0.14  Hematocrit, %  26.2 ± 3.3  25.0 ± 4.1  26.4 ± 3.2  24.8 ± 3.6  0.12  0.75  0.20  DO2, mL/min/m2  548 ± 113  524 ± 104  551 ± 105  509 ± 109  0.22  0.33  0.16  VO2, mL/min1/m2  101 ± 24  96 ± 23  120 ± 26  111 ± 25  0.24  <0.001  0.15  FEO2  0.19 ± 0.03  0.19 ± 0.04  0.22 ± 0.04  0.22 ± 0.04  0.80  <0.001  0.74  SVO2, %  81.2 ± 3.0  81.0 ± 3.3  78.8 ± 3.4  77.4 ± 3.2  0.66  <0.001  0.90  Variable  CPB-Hypothermic (∼34°C)   CPB-Rewarm   PGroup  PEpoch  PGroup*Epoch  No AKI  AKI  No AKI  AKI  (n = 39)  (n = 26)  (n = 39)  (n = 26)  Pump flow, L/min  4.68 ± 0.57  4.68 ± 0.52  4.78 ± 0.54  4.75 ± 0.50  0.92  0.007  0.70  Pump flow, L/min/m2  2.40 ± 0.22  2.40 ± 0.17  2.45 ± 0.21  2.43 ± 0.18  0.97  0.005  0.71  PAO2, mmHg  244 ± 31  249 ± 25  238 ± 27  253 ± 32  0.12  0.85  0.14  PACO2, mmHg  38.8 ± 3.2  38.9 ± 2.6  40.5 ± 2.5  39.3 ± 2.8  0.36  0.008  0.11  Hemoglobin, g/L  87.3 ± 11.0  83.5 ± 13.4  88.0 ± 10.7  82.6 ± 12.2  0.12  0.92  0.14  Hematocrit, %  26.2 ± 3.3  25.0 ± 4.1  26.4 ± 3.2  24.8 ± 3.6  0.12  0.75  0.20  DO2, mL/min/m2  548 ± 113  524 ± 104  551 ± 105  509 ± 109  0.22  0.33  0.16  VO2, mL/min1/m2  101 ± 24  96 ± 23  120 ± 26  111 ± 25  0.24  <0.001  0.15  FEO2  0.19 ± 0.03  0.19 ± 0.04  0.22 ± 0.04  0.22 ± 0.04  0.80  <0.001  0.74  SVO2, %  81.2 ± 3.0  81.0 ± 3.3  78.8 ± 3.4  77.4 ± 3.2  0.66  <0.001  0.90  Values are mean ± SD. P-values are from two-way repeated measures analysis of variance. DO2, whole-body oxygen delivery; FEO2, whole-body oxygen extraction fraction; PACO2, partial pressure of carbon dioxide in arterial blood; PAO2, partial pressure of oxygen in arterial blood; SVO2, saturation of venous hemoglobin with oxygen; VO2, whole-body oxygen consumption. Table 3. Hemodynamics and whole-body oxygenation during CPB Variable  CPB-Hypothermic (∼34°C)   CPB-Rewarm   PGroup  PEpoch  PGroup*Epoch  No AKI  AKI  No AKI  AKI  (n = 39)  (n = 26)  (n = 39)  (n = 26)  Pump flow, L/min  4.68 ± 0.57  4.68 ± 0.52  4.78 ± 0.54  4.75 ± 0.50  0.92  0.007  0.70  Pump flow, L/min/m2  2.40 ± 0.22  2.40 ± 0.17  2.45 ± 0.21  2.43 ± 0.18  0.97  0.005  0.71  PAO2, mmHg  244 ± 31  249 ± 25  238 ± 27  253 ± 32  0.12  0.85  0.14  PACO2, mmHg  38.8 ± 3.2  38.9 ± 2.6  40.5 ± 2.5  39.3 ± 2.8  0.36  0.008  0.11  Hemoglobin, g/L  87.3 ± 11.0  83.5 ± 13.4  88.0 ± 10.7  82.6 ± 12.2  0.12  0.92  0.14  Hematocrit, %  26.2 ± 3.3  25.0 ± 4.1  26.4 ± 3.2  24.8 ± 3.6  0.12  0.75  0.20  DO2, mL/min/m2  548 ± 113  524 ± 104  551 ± 105  509 ± 109  0.22  0.33  0.16  VO2, mL/min1/m2  101 ± 24  96 ± 23  120 ± 26  111 ± 25  0.24  <0.001  0.15  FEO2  0.19 ± 0.03  0.19 ± 0.04  0.22 ± 0.04  0.22 ± 0.04  0.80  <0.001  0.74  SVO2, %  81.2 ± 3.0  81.0 ± 3.3  78.8 ± 3.4  77.4 ± 3.2  0.66  <0.001  0.90  Variable  CPB-Hypothermic (∼34°C)   CPB-Rewarm   PGroup  PEpoch  PGroup*Epoch  No AKI  AKI  No AKI  AKI  (n = 39)  (n = 26)  (n = 39)  (n = 26)  Pump flow, L/min  4.68 ± 0.57  4.68 ± 0.52  4.78 ± 0.54  4.75 ± 0.50  0.92  0.007  0.70  Pump flow, L/min/m2  2.40 ± 0.22  2.40 ± 0.17  2.45 ± 0.21  2.43 ± 0.18  0.97  0.005  0.71  PAO2, mmHg  244 ± 31  249 ± 25  238 ± 27  253 ± 32  0.12  0.85  0.14  PACO2, mmHg  38.8 ± 3.2  38.9 ± 2.6  40.5 ± 2.5  39.3 ± 2.8  0.36  0.008  0.11  Hemoglobin, g/L  87.3 ± 11.0  83.5 ± 13.4  88.0 ± 10.7  82.6 ± 12.2  0.12  0.92  0.14  Hematocrit, %  26.2 ± 3.3  25.0 ± 4.1  26.4 ± 3.2  24.8 ± 3.6  0.12  0.75  0.20  DO2, mL/min/m2  548 ± 113  524 ± 104  551 ± 105  509 ± 109  0.22  0.33  0.16  VO2, mL/min1/m2  101 ± 24  96 ± 23  120 ± 26  111 ± 25  0.24  <0.001  0.15  FEO2  0.19 ± 0.03  0.19 ± 0.04  0.22 ± 0.04  0.22 ± 0.04  0.80  <0.001  0.74  SVO2, %  81.2 ± 3.0  81.0 ± 3.3  78.8 ± 3.4  77.4 ± 3.2  0.66  <0.001  0.90  Values are mean ± SD. P-values are from two-way repeated measures analysis of variance. DO2, whole-body oxygen delivery; FEO2, whole-body oxygen extraction fraction; PACO2, partial pressure of carbon dioxide in arterial blood; PAO2, partial pressure of oxygen in arterial blood; SVO2, saturation of venous hemoglobin with oxygen; VO2, whole-body oxygen consumption. Variables monitored only before and after CPB When averaged over the periods before and after bypass, central venous pressure (CVP) was 2.8 ± 0.9 mmHg greater, mean pulmonary artery pressure (PAP) was 3.5 ± 2.2 mmHg greater and end-tidal PCO2 was 2.1 ± 0.8 mmHg less, in patients who developed AKI than in those who did not (Supplementary data, Table S3). Intraoperative urinary hypoxia predicts AKI after cardiac surgery In univariable analysis, the odds of developing AKI were 3.60-fold greater when intraoperative urinary PO2 fell to ≤10 mmHg at any time during the surgical procedure (Figure 3; Supplementary data, Table S4). The duration of urinary PO2 at or below specific thresholds, when expressed as continuous variables, were also associated with risk of AKI. For every minute per hour of surgery, urinary PO2 was ≤15 mmHg, the risk of AKI was 4% greater, while it was 7% greater for a threshold of 10 mmHg. The median duration of urinary PO2 ≤15 mmHg was 4.8 min/h of surgery. A duration of urinary hypoxia (i.e. urinary PO2 ≤15 mmHg) of this median value or greater (≥4.8 min/h) was associated with a 4.85-fold greater risk of AKI [95% confidence interval (CI) 1.64–14.40] with an area under the ROC curve of 0.69 (95% CI 0.55–0.82). FIGURE 3: View largeDownload slide Association between indices of urinary hypoxia and risk of AKI. Data are the outcomes of univariable logistic regression, with the 95% CI limits of the ORs shown in parentheses. FIGURE 3: View largeDownload slide Association between indices of urinary hypoxia and risk of AKI. Data are the outcomes of univariable logistic regression, with the 95% CI limits of the ORs shown in parentheses. Apart from urinary hypoxia, other factors associated with the development of AKI were: lower baseline eGFR, NYHA Class ≥III, lower intraoperative urine output, longer duration of CPB and surgery, higher average CVP and average mean PAP measured before and after CPB, administration of more norepinephrine in the post-bypass period, lower nadir MAP during surgery and greater net positive fluid balance at the conclusion of surgery (Table 4). When these, and other variables with P ≤ 0.10 in univariable analysis, were included in a multivariable model, the area under the ROC curve was 0.89 (95% CI 0.81–0.97) (Model 1 in Table 4). In multivariable analysis with backward step-wise elimination (Model 2 in Table 4), only intraoperative urinary hypoxia (≤15 mmHg for ≥4.8 min/h), lower intraoperative urine output and a greater amount of norepinephrine administered post-bypass were found to be independently associated with later development of AKI. In this model, the odds ratio (OR) for urinary hypoxia was 7.48 (95% CI 1.54–36.43) and the area under the ROC curve was 0.87 (95% CI 0.78–0.96). Table 4. Univariable and multivariable logistic regression for predictors of AKI Variable  Univariable analysis   Multivariable model 1   Multivariable model 2   OR  95% CI  P  OR  95% CI  P  OR  95% CI  P  Urinary PO2 ≤15 mmHg, for ≥ the study median of 4.8 min/h of surgery  4.85  1.64–14.36  0.004  7.15  1.26–40.57  0.03  7.48  1.54–36.43  0.01  Hypertension  3.83  0.97–15.16  0.06  3.97  0.60–26.28  0.15  4.58  0.75–27.92  0.10  Baseline eGFR, mL/min/1.73 m2  0.97  0.94–1.00  0.03  0.97  0.94–1.01  0.17  0.97  0.94–1.00  0.06  NYHA ≥III  3.33  1.14–9.72  0.03  2.98  0.54–16.62  0.21  3.86  0.81–18.28  0.09  Mehta Score, % risk  1.41  0.93–2.13  0.11              Intraoperative urine flow, mL/min  0.79  0.63–0.99  0.04  0.64  0.43–0.95  0.03  0.61  0.42–0.90  0.01  CVP, mmHga  1.21  1.05–1.40  0.01              Mean PAP, mmHg  1.11  1.01–1.22  0.03  1.01  0.86–1.19  0.86        ETCO2, mmHg  0.88  0.76–1.01  0.07  0.91  0.76–1.09  0.32        Duration of CPB, mina  1.01  1.00–1.02  0.03              Duration of surgery, min  1.01  1.00–1.02  0.03  1.00  0.99–1.01  0.62        Post-CPB norepinephrine, total, µg/kg  1.38  1.02–1.87  0.04  1.01  1.00–1.01  0.28  1.01  1.00–1.01  0.05  Nadir MAP during surgery, mmHg  0.95  0.89–1.00  0.05  1.00  0.91–1.09  0.92        Net volume status, liters positive  1.50  1.03–2.20  0.04  1.29  0.71–2.32  0.41        Variable  Univariable analysis   Multivariable model 1   Multivariable model 2   OR  95% CI  P  OR  95% CI  P  OR  95% CI  P  Urinary PO2 ≤15 mmHg, for ≥ the study median of 4.8 min/h of surgery  4.85  1.64–14.36  0.004  7.15  1.26–40.57  0.03  7.48  1.54–36.43  0.01  Hypertension  3.83  0.97–15.16  0.06  3.97  0.60–26.28  0.15  4.58  0.75–27.92  0.10  Baseline eGFR, mL/min/1.73 m2  0.97  0.94–1.00  0.03  0.97  0.94–1.01  0.17  0.97  0.94–1.00  0.06  NYHA ≥III  3.33  1.14–9.72  0.03  2.98  0.54–16.62  0.21  3.86  0.81–18.28  0.09  Mehta Score, % risk  1.41  0.93–2.13  0.11              Intraoperative urine flow, mL/min  0.79  0.63–0.99  0.04  0.64  0.43–0.95  0.03  0.61  0.42–0.90  0.01  CVP, mmHga  1.21  1.05–1.40  0.01              Mean PAP, mmHg  1.11  1.01–1.22  0.03  1.01  0.86–1.19  0.86        ETCO2, mmHg  0.88  0.76–1.01  0.07  0.91  0.76–1.09  0.32        Duration of CPB, mina  1.01  1.00–1.02  0.03              Duration of surgery, min  1.01  1.00–1.02  0.03  1.00  0.99–1.01  0.62        Post-CPB norepinephrine, total, µg/kg  1.38  1.02–1.87  0.04  1.01  1.00–1.01  0.28  1.01  1.00–1.01  0.05  Nadir MAP during surgery, mmHg  0.95  0.89–1.00  0.05  1.00  0.91–1.09  0.92        Net volume status, liters positive  1.50  1.03–2.20  0.04  1.29  0.71–2.32  0.41        Univariable analyses were performed for variables found to differ significantly between patients who developed AKI and those who did not (Tables 1 and 2; Supplementary data, Tables S1 and S2). The only exceptions to this were for the variables ‘>1 valve’ and ‘intraoperative furosemide’, where logistic regression could not be performed due to 100% association with development of postoperative AKI. Univariable outcomes with P ≤ 0.10 were eligible for inclusion in the multivariable logistic regression models. a ‘CVP’ and ‘duration of CPB’ were excluded because these variables shared a multi-collinear relationship with ‘mean PAP’ and ‘duration of surgery’, respectively. Multivariable Model 1: all univariable outcomes with P ≤ 0.10 were included. Multivariable Model 2: variables with P > 0.10 were removed from the model in a step-wise fashion until only those with P ≤ 0.10 remained. The area under the ROC curve for Model 1 was 0.89 (95% CI 0.81–0.97), P ≤ 0.001 with inclusion of the urinary PO2 variable, and 0.85 (95% CI 0.76–0.95), P ≤ 0.001 without urinary PO2. For Model 2 it was 0.87 (0.78–0.96), P ≤ 0.001 with inclusion of the urinary PO2 variable, and 0.83 (95% CI 0.73–0.93), P ≤ 0.001 without urinary PO2. An additional model, which included only those variables known before the operation (hypertension, eGFR and NYHA Class) had an area under the ROC curve of 0.82 with the inclusion of the urinary PO2 variable and 0.71 without inclusion of urinary PO2. CVP, central venous pressure; ETCO2, end-tidal partial pressure of carbon dioxide. Table 4. Univariable and multivariable logistic regression for predictors of AKI Variable  Univariable analysis   Multivariable model 1   Multivariable model 2   OR  95% CI  P  OR  95% CI  P  OR  95% CI  P  Urinary PO2 ≤15 mmHg, for ≥ the study median of 4.8 min/h of surgery  4.85  1.64–14.36  0.004  7.15  1.26–40.57  0.03  7.48  1.54–36.43  0.01  Hypertension  3.83  0.97–15.16  0.06  3.97  0.60–26.28  0.15  4.58  0.75–27.92  0.10  Baseline eGFR, mL/min/1.73 m2  0.97  0.94–1.00  0.03  0.97  0.94–1.01  0.17  0.97  0.94–1.00  0.06  NYHA ≥III  3.33  1.14–9.72  0.03  2.98  0.54–16.62  0.21  3.86  0.81–18.28  0.09  Mehta Score, % risk  1.41  0.93–2.13  0.11              Intraoperative urine flow, mL/min  0.79  0.63–0.99  0.04  0.64  0.43–0.95  0.03  0.61  0.42–0.90  0.01  CVP, mmHga  1.21  1.05–1.40  0.01              Mean PAP, mmHg  1.11  1.01–1.22  0.03  1.01  0.86–1.19  0.86        ETCO2, mmHg  0.88  0.76–1.01  0.07  0.91  0.76–1.09  0.32        Duration of CPB, mina  1.01  1.00–1.02  0.03              Duration of surgery, min  1.01  1.00–1.02  0.03  1.00  0.99–1.01  0.62        Post-CPB norepinephrine, total, µg/kg  1.38  1.02–1.87  0.04  1.01  1.00–1.01  0.28  1.01  1.00–1.01  0.05  Nadir MAP during surgery, mmHg  0.95  0.89–1.00  0.05  1.00  0.91–1.09  0.92        Net volume status, liters positive  1.50  1.03–2.20  0.04  1.29  0.71–2.32  0.41        Variable  Univariable analysis   Multivariable model 1   Multivariable model 2   OR  95% CI  P  OR  95% CI  P  OR  95% CI  P  Urinary PO2 ≤15 mmHg, for ≥ the study median of 4.8 min/h of surgery  4.85  1.64–14.36  0.004  7.15  1.26–40.57  0.03  7.48  1.54–36.43  0.01  Hypertension  3.83  0.97–15.16  0.06  3.97  0.60–26.28  0.15  4.58  0.75–27.92  0.10  Baseline eGFR, mL/min/1.73 m2  0.97  0.94–1.00  0.03  0.97  0.94–1.01  0.17  0.97  0.94–1.00  0.06  NYHA ≥III  3.33  1.14–9.72  0.03  2.98  0.54–16.62  0.21  3.86  0.81–18.28  0.09  Mehta Score, % risk  1.41  0.93–2.13  0.11              Intraoperative urine flow, mL/min  0.79  0.63–0.99  0.04  0.64  0.43–0.95  0.03  0.61  0.42–0.90  0.01  CVP, mmHga  1.21  1.05–1.40  0.01              Mean PAP, mmHg  1.11  1.01–1.22  0.03  1.01  0.86–1.19  0.86        ETCO2, mmHg  0.88  0.76–1.01  0.07  0.91  0.76–1.09  0.32        Duration of CPB, mina  1.01  1.00–1.02  0.03              Duration of surgery, min  1.01  1.00–1.02  0.03  1.00  0.99–1.01  0.62        Post-CPB norepinephrine, total, µg/kg  1.38  1.02–1.87  0.04  1.01  1.00–1.01  0.28  1.01  1.00–1.01  0.05  Nadir MAP during surgery, mmHg  0.95  0.89–1.00  0.05  1.00  0.91–1.09  0.92        Net volume status, liters positive  1.50  1.03–2.20  0.04  1.29  0.71–2.32  0.41        Univariable analyses were performed for variables found to differ significantly between patients who developed AKI and those who did not (Tables 1 and 2; Supplementary data, Tables S1 and S2). The only exceptions to this were for the variables ‘>1 valve’ and ‘intraoperative furosemide’, where logistic regression could not be performed due to 100% association with development of postoperative AKI. Univariable outcomes with P ≤ 0.10 were eligible for inclusion in the multivariable logistic regression models. a ‘CVP’ and ‘duration of CPB’ were excluded because these variables shared a multi-collinear relationship with ‘mean PAP’ and ‘duration of surgery’, respectively. Multivariable Model 1: all univariable outcomes with P ≤ 0.10 were included. Multivariable Model 2: variables with P > 0.10 were removed from the model in a step-wise fashion until only those with P ≤ 0.10 remained. The area under the ROC curve for Model 1 was 0.89 (95% CI 0.81–0.97), P ≤ 0.001 with inclusion of the urinary PO2 variable, and 0.85 (95% CI 0.76–0.95), P ≤ 0.001 without urinary PO2. For Model 2 it was 0.87 (0.78–0.96), P ≤ 0.001 with inclusion of the urinary PO2 variable, and 0.83 (95% CI 0.73–0.93), P ≤ 0.001 without urinary PO2. An additional model, which included only those variables known before the operation (hypertension, eGFR and NYHA Class) had an area under the ROC curve of 0.82 with the inclusion of the urinary PO2 variable and 0.71 without inclusion of urinary PO2. CVP, central venous pressure; ETCO2, end-tidal partial pressure of carbon dioxide. Factors associated with urinary hypoxia Univariable analysis revealed significant associations of urinary hypoxia (urinary PO2 ≤15 mmHg for ≥4.8 min/h) with: longer durations of surgery and CPB, lesser blood hemoglobin concentration and oxygen delivery (DO2) on bypass, intraoperative administration of furosemide and metaraminol, as well as the use of more epinephrine or norepinephrine in the post-bypass period (Table 5). Multivariable analysis revealed independent associations of urinary hypoxia with longer duration of surgery, lesser DO2 on bypass and the use of norepinephrine. Table 5. Univariable and multivariable logistic regression for factors associated with urinary hypoxia (urinary PO2 ≤15 mmHg for ≥4.8 min/h of surgery) Variable  Univariable   Multivariable   OR  95% CI  P  OR  95% CI  P  Medical and demographic variables at baseline   Age, years  0.99  0.93–1.04  0.61         Female  0.77  0.21–2.84  0.70         Body mass index, kg/m2  0.97  0.87–1.07  0.51         Non-CABG  1.62  0.59–4.48  0.35         Non-elective surgery  1.95  0.67–5.64  0.22         Diabetes mellitus  0.65  0.24–1.75  0.39         Hypertension  0.75  0.24–2.32  0.61         Hypercholesterolemia  0.64  0.21–1.97  0.44         Smoking history  1.37  0.50–3.76  0.54         Chronic lung disease  0.45  0.08–2.66  0.38         Peripheral vascular disease  0.43  0.10–1.91  0.27         Cerebrovascular disease  2.15  0.49–9.45  0.31         NYHA ≥III  0.73  0.26–2.03  0.54         Congestive cardiac failure  0.96  0.28–3.37  0.95         Pre-op inotropes  0.97  0.06–16.2  0.98         Estimated EF <30%  2.07  0.35–12.18  0.42         Arrhythmia  1.11  0.38–3.24  0.85         Redo sternotomy  1.50  0.23–9.63  0.67         Serum creatinine, μmol/L  1.01  0.99–1.03  0.27         eGFR, mL/min/1.73 m2  1.00  0.97–1.02  0.79         Logistic EuroSCORE  1.01  0.95–1.07  0.73         Mehta Score, % risk  1.26  0.85–1.87  0.26         Cleveland Clinic Score, % risk  1.27  0.79–2.04  0.33        Variables during surgery and CPB   Duration of surgery, min  1.01  1.00–1.02  0.01  1.01  1.00–1.02  0.02   CPB time, mina  1.01  1.00–1.02  0.05         Aortic cross-clamp time, mina  1.01  1.00–1.02  0.07         Mean arterial pressure, mmHg  1.01  0.93–1.11  0.80         SAO2, %  1.01  0.60–1.70  0.97         Body temperature, °C  0.92  0.22–3.78  0.91         Urine flow, case average, L/min  0.99  0.81–1.20  0.89         Pump flow, L/min  0.61  0.23–1.61  0.32         PAO2, mmHg  1.00  0.98–1.03  0.69         PACO2, mmHg  1.10  0.89–1.37  0.38         Hemoglobin, g/La  0.94  0.90–0.99  0.02         DO2, mL/min/m2  0.99  0.99–1.00  0.03  0.99  0.99–1.00  0.01   VO2, mL/min/m2  0.99  0.97–1.01  0.20         SVO2, %  0.93  0.79–1.10  0.40         Heart rate, bpm  1.03  0.96–1.09  0.42         Central venous pressure, mmHg  1.07  0.94–1.20  0.31         Mean PAP, mmHg  1.05  0.97–1.14  0.26         ETCO2, mmHg  0.98  0.87–1.10  0.67         Net volume status, liters positive  1.06  0.81–1.38  0.68         Nadir MAP during case, mmHg  0.96  0.91–1.01  0.14         Nadir MAP during CPB, mmHg  0.95  0.90–1.00  0.07        Pharmacological interventions               At any time during surgery                Furosemide  8.35  0.96–72.3  0.05          Esmolol  1.50  0.23–9.63  0.67          Milrinone  4.28  0.45–40.53  0.21         Before CPB                Norepinephrine  1.17  0.35–3.95  0.80           Norepinephrine, total given, µg/kg  1.22  0.74–2.01  0.43         During CPB                Metaraminol dose, mg  1.13  1.02–1.24  0.02          Norepinephrine  1.17  0.35–3.95  0.80           Norepinephrine, total given, µg/kg  1.62  0.73–3.58  0.23         After CPB                Epinephrine  1.83  0.67–5.05  0.24           Epinephrine, total given, µg/kg  2.10  1.03–4.26  0.04          Norepinephrine  6.00  1.50–23.99  0.01  5.24  1.04–26.39  0.05     Norepinephrine, total given, µg/kg  1.56  1.06–2.28  0.02        Variable  Univariable   Multivariable   OR  95% CI  P  OR  95% CI  P  Medical and demographic variables at baseline   Age, years  0.99  0.93–1.04  0.61         Female  0.77  0.21–2.84  0.70         Body mass index, kg/m2  0.97  0.87–1.07  0.51         Non-CABG  1.62  0.59–4.48  0.35         Non-elective surgery  1.95  0.67–5.64  0.22         Diabetes mellitus  0.65  0.24–1.75  0.39         Hypertension  0.75  0.24–2.32  0.61         Hypercholesterolemia  0.64  0.21–1.97  0.44         Smoking history  1.37  0.50–3.76  0.54         Chronic lung disease  0.45  0.08–2.66  0.38         Peripheral vascular disease  0.43  0.10–1.91  0.27         Cerebrovascular disease  2.15  0.49–9.45  0.31         NYHA ≥III  0.73  0.26–2.03  0.54         Congestive cardiac failure  0.96  0.28–3.37  0.95         Pre-op inotropes  0.97  0.06–16.2  0.98         Estimated EF <30%  2.07  0.35–12.18  0.42         Arrhythmia  1.11  0.38–3.24  0.85         Redo sternotomy  1.50  0.23–9.63  0.67         Serum creatinine, μmol/L  1.01  0.99–1.03  0.27         eGFR, mL/min/1.73 m2  1.00  0.97–1.02  0.79         Logistic EuroSCORE  1.01  0.95–1.07  0.73         Mehta Score, % risk  1.26  0.85–1.87  0.26         Cleveland Clinic Score, % risk  1.27  0.79–2.04  0.33        Variables during surgery and CPB   Duration of surgery, min  1.01  1.00–1.02  0.01  1.01  1.00–1.02  0.02   CPB time, mina  1.01  1.00–1.02  0.05         Aortic cross-clamp time, mina  1.01  1.00–1.02  0.07         Mean arterial pressure, mmHg  1.01  0.93–1.11  0.80         SAO2, %  1.01  0.60–1.70  0.97         Body temperature, °C  0.92  0.22–3.78  0.91         Urine flow, case average, L/min  0.99  0.81–1.20  0.89         Pump flow, L/min  0.61  0.23–1.61  0.32         PAO2, mmHg  1.00  0.98–1.03  0.69         PACO2, mmHg  1.10  0.89–1.37  0.38         Hemoglobin, g/La  0.94  0.90–0.99  0.02         DO2, mL/min/m2  0.99  0.99–1.00  0.03  0.99  0.99–1.00  0.01   VO2, mL/min/m2  0.99  0.97–1.01  0.20         SVO2, %  0.93  0.79–1.10  0.40         Heart rate, bpm  1.03  0.96–1.09  0.42         Central venous pressure, mmHg  1.07  0.94–1.20  0.31         Mean PAP, mmHg  1.05  0.97–1.14  0.26         ETCO2, mmHg  0.98  0.87–1.10  0.67         Net volume status, liters positive  1.06  0.81–1.38  0.68         Nadir MAP during case, mmHg  0.96  0.91–1.01  0.14         Nadir MAP during CPB, mmHg  0.95  0.90–1.00  0.07        Pharmacological interventions               At any time during surgery                Furosemide  8.35  0.96–72.3  0.05          Esmolol  1.50  0.23–9.63  0.67          Milrinone  4.28  0.45–40.53  0.21         Before CPB                Norepinephrine  1.17  0.35–3.95  0.80           Norepinephrine, total given, µg/kg  1.22  0.74–2.01  0.43         During CPB                Metaraminol dose, mg  1.13  1.02–1.24  0.02          Norepinephrine  1.17  0.35–3.95  0.80           Norepinephrine, total given, µg/kg  1.62  0.73–3.58  0.23         After CPB                Epinephrine  1.83  0.67–5.05  0.24           Epinephrine, total given, µg/kg  2.10  1.03–4.26  0.04          Norepinephrine  6.00  1.50–23.99  0.01  5.24  1.04–26.39  0.05     Norepinephrine, total given, µg/kg  1.56  1.06–2.28  0.02        Univariable outcomes with P ≤ 0.10 were eligible for inclusion in the multivariable logistic regression model. a Hemoglobin during CPB was excluded because it is included in the calculation of DO2. Duration of CPB and aortic cross-clamp time were excluded because these variables are a subset of the total duration of surgery. DO2, whole-body oxygen delivery; EF, ejection fraction; ETCO2, end-tidal partial pressure of carbon dioxide; PACO2, partial pressure of carbon dioxide in arterial blood; PAO2, partial pressure of oxygen in arterial blood; SAO2, saturation of arterial hemoglobin with oxygen; SVO2, saturation of venous hemoglobin with oxygen; VO2, whole-body oxygen consumption. Table 5. Univariable and multivariable logistic regression for factors associated with urinary hypoxia (urinary PO2 ≤15 mmHg for ≥4.8 min/h of surgery) Variable  Univariable   Multivariable   OR  95% CI  P  OR  95% CI  P  Medical and demographic variables at baseline   Age, years  0.99  0.93–1.04  0.61         Female  0.77  0.21–2.84  0.70         Body mass index, kg/m2  0.97  0.87–1.07  0.51         Non-CABG  1.62  0.59–4.48  0.35         Non-elective surgery  1.95  0.67–5.64  0.22         Diabetes mellitus  0.65  0.24–1.75  0.39         Hypertension  0.75  0.24–2.32  0.61         Hypercholesterolemia  0.64  0.21–1.97  0.44         Smoking history  1.37  0.50–3.76  0.54         Chronic lung disease  0.45  0.08–2.66  0.38         Peripheral vascular disease  0.43  0.10–1.91  0.27         Cerebrovascular disease  2.15  0.49–9.45  0.31         NYHA ≥III  0.73  0.26–2.03  0.54         Congestive cardiac failure  0.96  0.28–3.37  0.95         Pre-op inotropes  0.97  0.06–16.2  0.98         Estimated EF <30%  2.07  0.35–12.18  0.42         Arrhythmia  1.11  0.38–3.24  0.85         Redo sternotomy  1.50  0.23–9.63  0.67         Serum creatinine, μmol/L  1.01  0.99–1.03  0.27         eGFR, mL/min/1.73 m2  1.00  0.97–1.02  0.79         Logistic EuroSCORE  1.01  0.95–1.07  0.73         Mehta Score, % risk  1.26  0.85–1.87  0.26         Cleveland Clinic Score, % risk  1.27  0.79–2.04  0.33        Variables during surgery and CPB   Duration of surgery, min  1.01  1.00–1.02  0.01  1.01  1.00–1.02  0.02   CPB time, mina  1.01  1.00–1.02  0.05         Aortic cross-clamp time, mina  1.01  1.00–1.02  0.07         Mean arterial pressure, mmHg  1.01  0.93–1.11  0.80         SAO2, %  1.01  0.60–1.70  0.97         Body temperature, °C  0.92  0.22–3.78  0.91         Urine flow, case average, L/min  0.99  0.81–1.20  0.89         Pump flow, L/min  0.61  0.23–1.61  0.32         PAO2, mmHg  1.00  0.98–1.03  0.69         PACO2, mmHg  1.10  0.89–1.37  0.38         Hemoglobin, g/La  0.94  0.90–0.99  0.02         DO2, mL/min/m2  0.99  0.99–1.00  0.03  0.99  0.99–1.00  0.01   VO2, mL/min/m2  0.99  0.97–1.01  0.20         SVO2, %  0.93  0.79–1.10  0.40         Heart rate, bpm  1.03  0.96–1.09  0.42         Central venous pressure, mmHg  1.07  0.94–1.20  0.31         Mean PAP, mmHg  1.05  0.97–1.14  0.26         ETCO2, mmHg  0.98  0.87–1.10  0.67         Net volume status, liters positive  1.06  0.81–1.38  0.68         Nadir MAP during case, mmHg  0.96  0.91–1.01  0.14         Nadir MAP during CPB, mmHg  0.95  0.90–1.00  0.07        Pharmacological interventions               At any time during surgery                Furosemide  8.35  0.96–72.3  0.05          Esmolol  1.50  0.23–9.63  0.67          Milrinone  4.28  0.45–40.53  0.21         Before CPB                Norepinephrine  1.17  0.35–3.95  0.80           Norepinephrine, total given, µg/kg  1.22  0.74–2.01  0.43         During CPB                Metaraminol dose, mg  1.13  1.02–1.24  0.02          Norepinephrine  1.17  0.35–3.95  0.80           Norepinephrine, total given, µg/kg  1.62  0.73–3.58  0.23         After CPB                Epinephrine  1.83  0.67–5.05  0.24           Epinephrine, total given, µg/kg  2.10  1.03–4.26  0.04          Norepinephrine  6.00  1.50–23.99  0.01  5.24  1.04–26.39  0.05     Norepinephrine, total given, µg/kg  1.56  1.06–2.28  0.02        Variable  Univariable   Multivariable   OR  95% CI  P  OR  95% CI  P  Medical and demographic variables at baseline   Age, years  0.99  0.93–1.04  0.61         Female  0.77  0.21–2.84  0.70         Body mass index, kg/m2  0.97  0.87–1.07  0.51         Non-CABG  1.62  0.59–4.48  0.35         Non-elective surgery  1.95  0.67–5.64  0.22         Diabetes mellitus  0.65  0.24–1.75  0.39         Hypertension  0.75  0.24–2.32  0.61         Hypercholesterolemia  0.64  0.21–1.97  0.44         Smoking history  1.37  0.50–3.76  0.54         Chronic lung disease  0.45  0.08–2.66  0.38         Peripheral vascular disease  0.43  0.10–1.91  0.27         Cerebrovascular disease  2.15  0.49–9.45  0.31         NYHA ≥III  0.73  0.26–2.03  0.54         Congestive cardiac failure  0.96  0.28–3.37  0.95         Pre-op inotropes  0.97  0.06–16.2  0.98         Estimated EF <30%  2.07  0.35–12.18  0.42         Arrhythmia  1.11  0.38–3.24  0.85         Redo sternotomy  1.50  0.23–9.63  0.67         Serum creatinine, μmol/L  1.01  0.99–1.03  0.27         eGFR, mL/min/1.73 m2  1.00  0.97–1.02  0.79         Logistic EuroSCORE  1.01  0.95–1.07  0.73         Mehta Score, % risk  1.26  0.85–1.87  0.26         Cleveland Clinic Score, % risk  1.27  0.79–2.04  0.33        Variables during surgery and CPB   Duration of surgery, min  1.01  1.00–1.02  0.01  1.01  1.00–1.02  0.02   CPB time, mina  1.01  1.00–1.02  0.05         Aortic cross-clamp time, mina  1.01  1.00–1.02  0.07         Mean arterial pressure, mmHg  1.01  0.93–1.11  0.80         SAO2, %  1.01  0.60–1.70  0.97         Body temperature, °C  0.92  0.22–3.78  0.91         Urine flow, case average, L/min  0.99  0.81–1.20  0.89         Pump flow, L/min  0.61  0.23–1.61  0.32         PAO2, mmHg  1.00  0.98–1.03  0.69         PACO2, mmHg  1.10  0.89–1.37  0.38         Hemoglobin, g/La  0.94  0.90–0.99  0.02         DO2, mL/min/m2  0.99  0.99–1.00  0.03  0.99  0.99–1.00  0.01   VO2, mL/min/m2  0.99  0.97–1.01  0.20         SVO2, %  0.93  0.79–1.10  0.40         Heart rate, bpm  1.03  0.96–1.09  0.42         Central venous pressure, mmHg  1.07  0.94–1.20  0.31         Mean PAP, mmHg  1.05  0.97–1.14  0.26         ETCO2, mmHg  0.98  0.87–1.10  0.67         Net volume status, liters positive  1.06  0.81–1.38  0.68         Nadir MAP during case, mmHg  0.96  0.91–1.01  0.14         Nadir MAP during CPB, mmHg  0.95  0.90–1.00  0.07        Pharmacological interventions               At any time during surgery                Furosemide  8.35  0.96–72.3  0.05          Esmolol  1.50  0.23–9.63  0.67          Milrinone  4.28  0.45–40.53  0.21         Before CPB                Norepinephrine  1.17  0.35–3.95  0.80           Norepinephrine, total given, µg/kg  1.22  0.74–2.01  0.43         During CPB                Metaraminol dose, mg  1.13  1.02–1.24  0.02          Norepinephrine  1.17  0.35–3.95  0.80           Norepinephrine, total given, µg/kg  1.62  0.73–3.58  0.23         After CPB                Epinephrine  1.83  0.67–5.05  0.24           Epinephrine, total given, µg/kg  2.10  1.03–4.26  0.04          Norepinephrine  6.00  1.50–23.99  0.01  5.24  1.04–26.39  0.05     Norepinephrine, total given, µg/kg  1.56  1.06–2.28  0.02        Univariable outcomes with P ≤ 0.10 were eligible for inclusion in the multivariable logistic regression model. a Hemoglobin during CPB was excluded because it is included in the calculation of DO2. Duration of CPB and aortic cross-clamp time were excluded because these variables are a subset of the total duration of surgery. DO2, whole-body oxygen delivery; EF, ejection fraction; ETCO2, end-tidal partial pressure of carbon dioxide; PACO2, partial pressure of carbon dioxide in arterial blood; PAO2, partial pressure of oxygen in arterial blood; SAO2, saturation of arterial hemoglobin with oxygen; SVO2, saturation of venous hemoglobin with oxygen; VO2, whole-body oxygen consumption. DISCUSSION Our major findings were: (i) it is feasible to continuously measure bladder urine PO2 in the operating theater using a fiber optic probe; (ii) bladder urine PO2 falls across the course of cardiac surgery, reaching its nadir most commonly during the rewarming phase of CPB or early after weaning from CPB; (iii) patients who develop AKI after cardiac surgery experience longer and more severe periods of low intraoperative urinary PO2 than patients who do not develop AKI; and (iv) low intraoperative urinary PO2 is an independent predictor of postoperative AKI. Thus, urinary PO2 may provide the first available real-time ‘biomarker’ of risk of postoperative AKI in the setting of cardiac surgery that provides prognostic information during the surgical procedure itself, when it may not be too late to intervene to prevent AKI. The detrimental impact of hypoxia could depend on both its duration and severity. Therefore, we calculated the time spent at or below various thresholds of urinary PO2 (15, 10 and 5 mmHg) and the area under the curve for each threshold. At thresholds of 15 and 10 mmHg, the duration spent at or below the threshold and the area under the curve was much greater in patients who later developed AKI than in those who did not. Furthermore, univariable logistic regression showed that the odds of AKI were 3.6 times greater when urinary PO2 fell to ≤10 mmHg at any time during the operation. This analysis provides insight into what we define as ‘urinary hypoxia’; the severity and duration of low intraoperative urinary PO2 that indicates increased risk of AKI. Our analysis suggests that a threshold of 15 mmHg may identify a level of risk that merits intervention. Established risk factors for AKI after cardiac surgery include high baseline serum creatinine and/or impaired eGFR [31, 32], heart failure [24, 30, 32], longer duration of CPB [1, 32], low pre- and intraoperative hemoglobin concentration [1], low DO2 during CPB [33, 34] and low intraoperative urine flow [31]. In multivariable logistic regression controlling for such factors with P ≤ 0.10 in univariable analysis, a strong independent association was found between AKI and intraoperative urinary hypoxia, as defined by urinary PO2 ≤15 mmHg for  more than or equal to the median duration for all 65 patients (4.8 min/h of surgery). In backward step-wise multivariable logistic regression, only urinary hypoxia, low intraoperative urine flow and a greater amount of norepinephrine administered post-bypass were independently associated with AKI. The predictive value of urinary hypoxia compares with those of current and emerging biomarkers of AKI. The area under the curve for urinary hypoxia established in the current study was 0.69. This is comparable to that achieved with available biomarkers of renal injury measured soon after weaning from CPB or within 24 h of surgery [35]. The only exception may be the product of urinary concentrations of tissue inhibitor of metalloproteinase 2 and insulin-like growth factor-binding protein 7, for which an area under the ROC curve of 0.81 (95% CI 0.68–0.93) was recently reported [36]. Nevertheless, unlike urinary PO2, none of these biomarkers has predictive efficacy when measured intraoperatively, or provides real-time information during CPB that might allow improved management of renal health. It is also noteworthy that the predictive efficacy of urinary hypoxia can be significantly improved by inclusion of certain pre-op (hypertension, low baseline eGFR and NYHA Class  ≥III) and intraoperative (urine flow) factors in the model. A number of lines of evidence indicate that urinary hypoxia during cardiac surgery reflects hypoxia in the renal medulla, as detailed in a recent review [37]. Medullary hypoxia has been observed during and after CPB in experimental animals [9–11] and is predicted by simulations using computational models [38, 39]. There is also clinical evidence that renal oxygen delivery falls during CPB [16] and that postoperative AKI is associated with intraoperative desaturation of renal tissue [40]. There is strong evidence that bladder urine PO2 provides a relatively noninvasive estimate of renal medullary PO2. In the human inner medulla, vascular bundles are distributed between clusters of collecting ducts, providing a pathway for oxygen diffusion between vasa recta and urine [41]. Leonhardt et al. showed more than 50 years ago that the PO2 of urine in the human renal pelvis is in equilibrium with renal medullary tissue PO2 [42]. Provided urine flow is relatively high, this ‘signal’ of medullary oxygenation appears to survive the journey from the pelvis to the bladder. This proposition is supported by recent experimental observations in sheep [19, 20] and rabbits [21], showing strong correlations between renal medullary PO2 and bladder urine PO2 under a range of conditions, and simulations using computational models of oxygen diffusion across the ureteric wall [21, 39]. Critically, CPB is usually accompanied by brisk diuresis, so provides an excellent platform to test the utility of bladder urine PO2 as an intraoperative ‘physiological biomarker’ of risk of AKI. Several established risk factors for AKI after CPB were significantly associated with urinary hypoxia by univariable (duration of surgery, hemoglobin concentration and whole-body DO2 during CPB and requirement for epinephrine and/or norepinephrine support after weaning from CPB) and multivariable logistic regression. Thus, urinary hypoxia might capture the combined effects of pre- and intraoperative factors that lead to medullary hypoxia during on-pump cardiac surgery. Our current findings show that urinary hypoxia during cardiac surgery indicates increased risk of AKI. But how might we intervene to ameliorate this risk? There is little evidence that specific pharmacological or non-pharmacological interventions can have a major impact on the risk of AKI associated with cardiac surgery [43, 44]. Continuous measurement of urinary PO2 might provide a translational pathway for investigating novel interventions to prevent AKI, since it can be deployed in both experimental animals, in which regional kidney perfusion and oxygenation can be directly monitored [19–21], and in man. Our current study is limited by a relatively small sample size at a single center. For example, the small sample size precludes a subanalysis in which we might exclude mild renal dysfunction (Stage 1), or employ newer statistical methods such as the net reclassification index [45] or integrated discrimination index [46]. However, strengths include the prospective design and the use of a fiber optic method for continuous measurement of urinary PO2 that could feasibly be applied to routine clinical practice. Our finding that urinary PO2 progressively decreases after the start of CPB is consistent with previous clinical observations using polarographic electrodes [47]. In this previous study, post-CPB recovery of urinary PO2 was associated with better postoperative renal function. Our current observations provide two major advances. First, we show that urinary hypoxia during surgery, rather than just post-CPB recovery of urinary PO2, predicts development of AKI. Secondly, the fiber optic probe we used avoids many of the limitations of polarographic electrodes such as fragility and the need for repeated calibration. Thus, the method we have used is feasible for routine clinical use and provides prognostic information during surgery, when it is still feasible to intervene to prevent AKI. SUPPLEMENTARY DATA Supplementary data are available at ndt online. ACKNOWLEDGEMENTS The authors thank the surgeons, perfusionists, anesthetists and other members of the operation theater and administrative staff at Monash Health who helped with these studies. FUNDING This work was supported by grants from the National Health and Medical Research Council of Australia [GNT1122455] and The National Heart Foundation of Australia [VG101377]. AUTHORS’ CONTRIBUTIONS R.G.E., A.D.C., J.A.S. and A.G.T. were responsible for study concept and design. M.Z.L.Z., A.M., G.K.H., J.P.N. and R.G.E. were responsible for acquisition, analysis and interpretation of data. M.Z.L.Z. and R.G.E. were responsible for drafting of the manuscript. M.Z.L.Z., A.M., A.D.C., J.A.S., A.G.T., G.K.H., J.P.N. and R.G.E. were responsible for critical revision of the manuscript for intellectual content. R.G.E., J.A.S., A.D.C. and A.G.T. were responsible for obtaining funding. M.Z.L.Z. and A.G.T. were responsible for analysis. CONFLICT OF INTEREST STATEMENT None declared. REFERENCES 1 Karkouti K, Wijeysundera DN, Yau TM et al.   Acute kidney injury after cardiac surgery: focus on modifiable risk factors. Circulation  2009; 119: 495– 502 Google Scholar CrossRef Search ADS PubMed  2 Lenihan CR, Montez-Rath ME, Mora Mangano CT et al.   Trends in acute kidney injury, associated use of dialysis, and mortality after cardiac surgery, 1999 to 2008. Ann Thorac Surg  2013; 95: 20– 28 Google Scholar CrossRef Search ADS PubMed  3 de Geus HR, Betjes MG, Bakker J. Biomarkers for the prediction of acute kidney injury: a narrative review on current status and future challenges. Clin Kidney J  2012; 5: 102– 108 Google Scholar CrossRef Search ADS PubMed  4 Koyner JL, Vaidya VS, Bennett MR et al.   Urinary biomarkers in the clinical prognosis and early detection of acute kidney injury. Clin J Am Soc Nephrol  2010; 5: 2154– 2165 Google Scholar CrossRef Search ADS PubMed  5 McIlroy DR, Farkas D, Matto M et al.   Neutrophil gelatinase-associated lipocalin combined with delta serum creatinine provides early risk stratification for adverse outcomes after cardiac surgery: a prospective observational study. Crit Care Med  2015; 43: 1043– 1052 Google Scholar CrossRef Search ADS PubMed  6 Bellomo R, Auriemma S, Fabbri A et al.   The pathophysiology of cardiac surgery-associated acute kidney injury (CSA-AKI). Int J Artif Organs  2008; 31: 166– 178 Google Scholar CrossRef Search ADS PubMed  7 Mao H, Katz N, Ariyanon W et al.   Cardiac surgery-associated acute kidney injury. Blood Purif  2014; 37: 34– 50 Google Scholar CrossRef Search ADS PubMed  8 Evans RG, Ince C, Joles JA et al.   Haemodynamic influences on kidney oxygenation: clinical implications of integrative physiology. Clin Exp Pharmacol Physiol  2013; 40: 106– 122 Google Scholar CrossRef Search ADS PubMed  9 Stafford-Smith M, Grocott HP. Renal medullary hypoxia during experimental cardiopulmonary bypass: a pilot study. Perfusion  2005; 20: 53– 58 Google Scholar CrossRef Search ADS PubMed  10 Patel NN, Toth T, Jones C et al.   Prevention of post-cardiopulmonary bypass acute kidney injury by endothelin A receptor blockade. Crit Care Med  2011; 39: 793– 802 Google Scholar CrossRef Search ADS PubMed  11 Darby PJ, Kim N, Hare GM et al.   Anemia increases the risk of renal cortical and medullary hypoxia during cardiopulmonary bypass. Perfusion  2013; 28: 504– 511 Google Scholar CrossRef Search ADS PubMed  12 Calzavacca P, Evans RG, Bailey M et al.   Cortical and medullary tissue perfusion and oxygenation in experimental septic acute kidney injury. Crit Care Med  2015; 43: e431– e439 Google Scholar CrossRef Search ADS PubMed  13 Abdelkader A, Ho J, Ow CP et al.   Renal oxygenation in acute renal ischemia-reperfusion injury. Am J Physiol Renal Physiol  2014; 306: F1026– F1038 Google Scholar CrossRef Search ADS PubMed  14 Heyman SN, Rosen S, Rosenberger C. Renal parenchymal hypoxia, hypoxia adaptation, and the pathogenesis of radiocontrast nephropathy. Clin J Am Soc Nephrol  2008; 3: 288– 296 Google Scholar CrossRef Search ADS PubMed  15 Rosenberger C, Goldfarb M, Shina A et al.   Evidence for sustained renal hypoxia and transient hypoxia adaptation in experimental rhabdomyolysis-induced acute kidney injury. Nephrol Dial Transplant  2007; 23: 1135– 1143 Google Scholar CrossRef Search ADS PubMed  16 Lannemyr L, Bragadottir G, Krumbholz V et al.   Effects of cardiopulmonary bypass on renal perfusion, filtration, and oxygenation in patients undergoing cardiac surgery. Anesthesiology  2017; 126: 205– 213 Google Scholar CrossRef Search ADS PubMed  17 Evans RG, Smith JA, Gardiner BS et al.   Hypoxia as a biomarker of kidney disease. In: Patel VB, Preedy VR (eds). Diseases of the Kidney. Biomarkers of Disease: Methods, Discoveries and Applications . Berlin: Springer-Verlag, 2016, 83– 105 Google Scholar CrossRef Search ADS   18 Evans RG, Smith JA, Wright C et al.   Urinary oxygen tension: a clinical window on the health of the renal medulla? Am J Physiol Regul Integr Comp Physiol  2014; 306: R45– R50 Google Scholar CrossRef Search ADS PubMed  19 Lankadeva YR, Kosaka J, Evans RG et al.   Intrarenal and urinary oxygenation during norepinephrine resuscitation in ovine septic acute kidney injury. Kidney Int  2016; 90: 100– 108 Google Scholar CrossRef Search ADS PubMed  20 Lankadeva YR, Kosaka J, Evans RG et al.   Urinary oxygenation as a surrogate measure of medullary oxygenation during angiotensin II therapy in septic acute kidney injury. Crit Care Med  2018; 46: e41– e48 Google Scholar CrossRef Search ADS PubMed  21 Sgouralis I, Kett MM, Ow CP et al.   Bladder urine oxygen tension for assessing renal medullary oxygenation in rabbits: experimental and modelling studies. Am J Physiol Regul Integr Comp Physiol  2016; 311: R532– R544 Google Scholar CrossRef Search ADS PubMed  22 Kidney Disease: Improving Global Outcomes (KDIGO) Acute Kidney Injury Working Group. KDIGO clinical practice guideline for acute kidney injury. Kidney Int Suppl  2012; 2: 1– 138 CrossRef Search ADS   23 Roques F, Michel P, Goldstone AR et al.   The logistic EuroSCORE. Eur Heart J  2003; 24: 1– 2 Google Scholar CrossRef Search ADS   24 Mehta RH, Grab JD, O'Brien SM et al.   Bedside tool for predicting the risk of postoperative dialysis in patients undergoing cardiac surgery. Circulation  2006; 114: 2208– 2216 Google Scholar CrossRef Search ADS PubMed  25 McIlroy DR, Argenziano M, Farkas D et al.   Incorporating oliguria into the diagnostic criteria for acute kidney injury after on-pump cardiac surgery: impact on incidence and outcomes. J Cardiothorac Vasc Anesth  2013; 27: 1145– 1152 Google Scholar CrossRef Search ADS PubMed  26 Koeze J, Keus F, Dieperink W et al.   Incidence, timing and outcome of AKI in critically ill patients varies with the definition used and the addition of urine output criteria. BMC Nephrol  2017; 18: 70 Google Scholar CrossRef Search ADS PubMed  27 Pallant J. SPSS Survival Manual: A Step by Step Guide to DATA Analysis Using SPSS for Windows Version 15 . Maidenhead Berkshire, UK: Open University Press, 2007 28 Ludbrook J. Repeated measurements and multiple comparisons in cardiovascular research. Cardiovasc Res  1994; 28: 303– 311 Google Scholar CrossRef Search ADS PubMed  29 Levey AS, Bosch JP, Lewis JB et al.   A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Ann Intern Med  1999; 130: 461– 470 Google Scholar CrossRef Search ADS PubMed  30 Thakar CV, Arrigain S, Worley S et al.   A clinical score to predict acute renal failure after cardiac surgery. J Am Soc Nephrol  2004; 16: 162– 168 Google Scholar CrossRef Search ADS PubMed  31 Parolari A, Pesce LL, Pacini D et al.   Risk factors for perioperative acute kidney injury after adult cardiac surgery: role of perioperative management. Ann Thorac Surg  2012; 93: 584– 591 Google Scholar CrossRef Search ADS PubMed  32 Ng SY, Sanagou M, Wolfe R et al.   Prediction of acute kidney injury within 30 days of cardiac surgery. J Thorac Cardiovasc Surg  2014; 147: 1875– 1883 Google Scholar CrossRef Search ADS PubMed  33 de Somer F, Mulholland JW, Bryan MR et al.   O2 delivery and CO2 production during cardiopulmonary bypass as determinants of acute kidney injury: time for a goal-directed perfusion management? Crit Care  2011; 15: R192 Google Scholar CrossRef Search ADS PubMed  34 Magruder JT, Dungan SP, Grimm JC et al.   Nadir oxygen delivery on bypass and hypotension increase acute kidney injury risk after cardiac operations. Ann Thorac Surg  2015; 100: 1697– 1703 Google Scholar CrossRef Search ADS PubMed  35 Ho J, Tangri N, Komenda P et al.   Urinary, plasma, and serum biomarkers' utility for predicting acute kidney injury associated with cardiac surgery in adults: a meta-analysis. Am J Kidney Dis  2015; 66: 993– 1005 Google Scholar CrossRef Search ADS PubMed  36 Meersch M, Schmidt C, Van Aken H et al.   Urinary TIMP-2 and IGFBP7 as early biomarkers of acute kidney injury and renal recovery following cardiac surgery. PLoS One  2014; 9: e93460 Google Scholar CrossRef Search ADS PubMed  37 Evans RG, Lankadeva YR, Cochrane AD et al.   Renal haemodynamics and oxygenation during and after cardiac surgery and cardiopulmonary bypass. Acta Physiol (Oxf)  2018; 222: e12995 Google Scholar CrossRef Search ADS   38 Sgouralis I, Evans RG, Gardiner BS et al.   Renal hemodynamics, function and oxygenation during cardiac surgery performed on cardiopulmonary bypass: a modeling study. Physiol Rep  2015; 3: e12260 Google Scholar CrossRef Search ADS PubMed  39 Sgouralis I, Evans RG, Layton AT. Renal medullary and urinary oxygen tension during cardiopulmonary bypass in the rat. Math Med Biol  2017; 34: 313– 333 Google Scholar PubMed  40 Choi DK, Kim WJ, Chin JH et al.   Intraoperative renal regional oxygen desaturation can be a predictor for acute kidney injury after cardiac surgery. J Cardiothorac Vasc Anesth  2014; 28: 564– 571 Google Scholar CrossRef Search ADS PubMed  41 Wei G, Rosen S, Dantzler WH et al.   Architecture of the human renal inner medulla and functional implications. Am J Physiol Renal Physiol  2015; 309: F627– F637 Google Scholar CrossRef Search ADS PubMed  42 Leonhardt KO, Landes RR, McCauley RT. Anatomy and physiology of intrarenal oxygen tension: preliminary study of the effects of anesthetics. Anesthesiology  1965; 26: 648– 658 Google Scholar CrossRef Search ADS PubMed  43 Patel NN, Rogers CA, Angelini GD et al.   Pharmacological therapies for the prevention of acute kidney injury following cardiac surgery: a systematic review. Heart Fail Rev  2011; 16: 553– 567 Google Scholar CrossRef Search ADS PubMed  44 O'Neal JB, Shaw AD, Billings FT. Acute kidney injury following cardiac surgery: current understanding and future directions. Crit Care  2016; 20: 187 Google Scholar CrossRef Search ADS PubMed  45 Kerr KF, Wang Z, Janes H et al.   Net reclassification indices for evaluating risk prediction instruments: a critical review. Epidemiology  2014; 25: 114– 121 Google Scholar CrossRef Search ADS PubMed  46 Kerr KF, McClelland RL, Brown ER et al.   Evaluating the incremental value of new biomarkers with integrated discrimination improvement. Am J Epidemiol  2011; 174: 364– 374 Google Scholar CrossRef Search ADS PubMed  47 Kainuma M, Yamada M, Miyake T. Continuous urine oxygen tension monitoring in patients undergoing cardiac surgery. J Cardiothorac Vasc Anesth  1996; 10: 603– 608 Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Nephrology Dialysis TransplantationOxford University Press

Published: Mar 14, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off