- Email: [email protected]
Strategies that improve renal medullary oxygenation during experimental cardiopulmonary bypass may mitigate postoperative acute kidney injury
basic research
www.kidney-international.org
Strategies that improve renal medullary oxygenation during experimental cardiopulmonary bypass may mitigate postoperative acute kidney injury
see commentary on page 1292
Yugeesh R. Lankadeva1, Andrew D. Cochrane2, Bruno Marino3, Naoya Iguchi1, Sally G. Hood1, Rinaldo Bellomo4, Clive N. May1,6 and Roger G. Evans5,6 1
Pre-Clinical Critical Care Unit, Florey Institute of Neuroscience and Mental Health, University of Melbourne, Melbourne, Victoria, Australia; 2Department of Cardiothoracic Surgery, Monash Health and Department of Surgery (School of Clinical Sciences at Monash Health), Monash University, Melbourne, Victoria, Australia; 3Cellsaving and Perfusion Resources, Melbourne, Victoria, Australia; 4 Department of Intensive Care, Austin Hospital, Heidelberg, Victoria, Australia; and 5Cardiovascular Disease Program, Biomedicine Discovery Institute and Department of Physiology, Monash University, Melbourne, Victoria, Australia
Renal medullary hypoxia may contribute to cardiac surgery–associated acute kidney injury (AKI). However, the effects of cardiopulmonary bypass (CPB) on medullary oxygenation are poorly understood. Here we tested whether CPB causes medullary hypoxia and whether medullary oxygenation during CPB can be improved by increasing pump flow or mean arterial pressure (MAP). Twelve sheep were instrumented to measure whole kidney, medullary, and cortical blood flow and oxygenation. Five days later, under isoflurane anesthesia, CPB was initiated at a pump flow of 80 mL kg–1min–1 and target MAP of 70 mm Hg. Pump flow was then set at 60 and 100 mL kg–1min–1, while MAP was maintained at approximately 70 mm Hg. MAP was then increased by vasopressor (metaraminol, 0.2–0.6 mg/min) infusion at a pump flow of 80 mL kg–1min–1. CPB at 80 mL kg–1min–1 reduced renal blood flow (RBF), -61% less than the conscious state, perfusion in the cortex (–44%) and medulla (–40%), and medullary PO2 from 43 to 27 mm Hg. Decreasing pump flow from 80 to 60 mL kg–1min–1 further decreased RBF (–16%) and medullary PO2 from 25 to 14 mm Hg. Increasing pump flow from 80 to 100 mL kg–1min–1 increased RBF (17%) and medullary PO2 from 20 to 29 mm Hg. Metaraminol (0.2 mg/min) increased MAP from 63 to 90 mm Hg, RBF (47%), and medullary PO2 from 19 to 39 mm Hg. Thus, the renal medulla is susceptible to hypoxia during CPB, but medullary oxygenation can be improved by increasing pump flow or increasing target MAP by infusion of metaraminol. Kidney International (2019) 95, 1338–1346; https://doi.org/10.1016/ j.kint.2019.01.032 Correspondence: R. Evans, Department of Physiology, 26 Innovation Walk, Monash University, Victoria 3800, Australia. E-mail: [email protected] 6
These authors contribute equally to this work.
Received 21 November 2018; revised 23 January 2019; accepted 24 January 2019; published online 15 March 2019
1338
KEYWORDS: acute kidney injury; cardiac surgery; hypoxia; renal circulation; renal medulla Copyright ª 2019, International Society of Nephrology. Published by Elsevier Inc. All rights reserved.
Translational Statement Renal medullary hypoxia is a plausible mechanistic pathway driving the development of AKI after cardiac surgery requiring CPB. We believe that currently accepted bypass flow and pressure may be inadequate in some cases. Our current findings indicate that it is feasible to improve medullary oxygenation on bypass and to potentially mitigate the risk of postoperative AKI. To test this proposition in human patients, a randomized clinical trial could be conducted to test the effects of increasing target pump flow and arterial pressure on the incidence of AKI.
KI occurs in more than 20% of adult patients after cardiac surgery requiring CPB.1 The ability to decrease the risk of AKI associated with cardiac surgery is hampered by our lack of knowledge of the effects of CPB on renal perfusion and oxygenation.2 Evolving evidence shows that renal medullary hypoxia contributes to the development of AKI.3,4 Mathematical models predict renal medullary hypoxia during CPB,5,6 and medullary hypoxia has been observed in a rat model of CPB7 and in a pilot study of 2 pigs subjected to CPB.8 However, there have been no systematic investigations of renal global and regional (i.e., cortical and medullary) perfusion and oxygenation in a clinically relevant large animal model of CPB. Furthermore, although evidence shows that goaldirected perfusion can mitigate risk of postoperative AKI9–11 and that the increase in renal fractional oxygen extraction during CPB can be alleviated by increasing pump flow,12,13 the potential for perfusion conditions to be
A
Kidney International (2019) 95, 1338–1346
YR Lankadeva et al.: Renal oxygenation during cardiopulmonary bypass
altered to improve renal medullary oxygenation has not been investigated. Renal cortical and medullary perfusion and tissue oxygen tension (PO2) can be continuously measured in unanesthetized sheep via chronically implanted fiber-optic probes.14 Global renal oxygen delivery (DO2) and oxygen consumption (VO2) can also be quantified through the combination of transit-time ultrasound flowmetry and arterial and renal venous blood samples.14 Using this model we aimed to test the hypotheses that CPB causes medullary hypoxia and medullary oxygenation during CPB can be improved by increasing pump flow or mean arterial pressure (MAP). Thus, we assessed the effects of clinically relevant perfusion conditions on global and regional renal perfusion and oxygenation. Subsequently, we established the effects of altering pump flow and MAP. RESULTS Part 1: anesthesia and transition to CPB Anesthesia. Anesthesia reduced MAP (–14 5 mm Hg;
Figure 1) and renal vascular conductance (RVC; –21% 4%; Supplementary Table S1), so renal blood flow (RBF) decreased (–32% 3%; Figure 1). Neither cortical nor medullary perfusion changed significantly (Table 1). Arterial PO2 increased due to use of a fractional inspired O2 (FiO2) of 60% (Supplementary Table S1). However, arterial blood hemoglobin concentration fell due to the fluid load given after induction of anesthesia. Thus, due to reduced arterial oxygen content and reduced RBF, renal DO2 fell by 41% 4% (Table 1). Renal VO2 tended to fall (22% 11%)
basic research
despite no decrease in creatinine clearance or sodium reabsorption. Consequently, neither renal venous PO2 nor saturation (SO2), nor renal fractional oxygen extraction, was significantly altered. In contrast, systemic fractional oxygen extraction fell under anesthesia, so that mixed venous PO2 and SO2 increased. There was no significant change in cortical or medullary PO2 (Figure 1). Cardiopulmonary bypass. As in human CPB, body temperature was intentionally reduced by 3.1 0.3 C, but there was no further significant change in MAP from its level in anesthetized sheep (Figure 1, Supplementary Table S1). Arterial blood hemoglobin content increased but was still 12% 3% less than in the conscious sheep. Arterial oxygen content did not differ significantly from that in conscious sheep. Systemic fractional oxygen extraction increased so that mixed venous SO2 (76.0% 1.2%) was compatible with a target of $75% for CPB in human patients.15 RBF (–61% 6%), RVC (–49% 8%), cortical perfusion (–44% 10%), and medullary perfusion (–40% 15%) were all less during CPB than in conscious sheep (Figure 1, Supplementary Table S1). RBF (–41% 8%), RVC (–34% 10%), and medullary perfusion (–39% 17%) were also less than under anesthesia alone. Consequently, despite no further hemodilution and maintenance of mixed venous SO2 $ 75%, renal DO2 was 61% 6% less than in conscious sheep and 32% 10% less than under anesthesia. Renal VO2 was also 53% 9% less on CPB than in conscious sheep, despite no significant changes in creatinine clearance or sodium reabsorption. Nevertheless, medullary PO2 decreased by 22 8 mm Hg compared with anesthesia alone and tended to be
Figure 1 | Effects of anesthesia and cardiopulmonary bypass on renal hemodynamics and renal tissue oxygenation: Black circles and error bars represent the mean ± SEM of each variable across each of the first 3 experimental periods (n [ 12 for mean arterial pressure and renal blood flow, but due to equipment failure n [ 10 for cortical PO2 and medullary PO2). Data for individual sheep are shown by colored lines. P values were derived from Tukey’s multiple comparison procedure. **P # 0.01, ***P # 0.001 for comparison with Conscious; †P # 0.05, ††† P # 0.001 for comparison of On Bypass with Anesthetized. Kidney International (2019) 95, 1338–1346
1339
basic research
YR Lankadeva et al.: Renal oxygenation during cardiopulmonary bypass
Table 1 | Effects of anesthesia and cardiopulmonary bypass on systemic and whole kidney oxygenation, regional renal perfusion, creatinine clearance, and sodium reabsorption Variable
n
Conscious
Anesthetized
Systemic DO2, ml O2 kg–1min–1 Systemic VO2, ml O2 kg–1min–1 Systemic FEO2, % Renal DO2, ml O2 kg–1min–1 Renal VO2, ml O2 kg–1min–1 Renal FEO2, % Cortical perfusion, units Medullary perfusion, units CrCl, ml kg–1min–1 TNaþ, mmol kg–1min–1
12 12 12 12 5 5 12 11 12 12
NA NA 28.3 1.9 0.93 0.06 0.084 0.016 9.4 1.8 1955 379 720 128 2.12 0.26 298 37
NA NA 17.9 1.3a 0.54 0.04a 0.061 0.011 13.5 3.0 1385 364 670 144 2.18 0.39 306 55
On bypass
PTime
<0.001 <0.001 0.01 0.29 0.001 0.007 0.80 0.75
10.3 2.78 27.0 0.36 0.035 13.7 872 326 1.88 258
0.4 0.16 1.1b 0.07a,c 0.006d 3.4 157a 110c,d 0.43 59
NA, not available. a P # 0.001, dP # 0.01 for comparison with the conscious state (Tukey’s test). b P # 0.001, cP # 0.05 for comparison of bypass with the anesthetized state (Tukey’s test). Values are mean SEM. Oxygen delivery (DO2) was calculated as the product of arterial blood oxygen content (see Supplementary Table S1) and pump flow (systemic DO2) or renal blood flow (renal DO2). Systemic oxygen consumption (VO2) was calculated as the product of pump flow and the oxygen concentration difference between arterial blood and mixed venous blood. Renal VO2 was calculated as the product of renal blood flow and the oxygen concentration difference between arterial blood and renal venous blood. Fractional oxygen extraction (FEO2) was calculated as 100 (VO2/DO2). Creatinine clearance (CrCl) was calculated as the product of urine flow and urinary creatinine concentration divided by plasma creatinine concentration. Sodium reabsorption (TNaþ) was calculated as the product of CrCl and plasma sodium concentration minus sodium excretion. P values are the outcomes of repeated-measures analysis of variance for the main effect of “time.”
less (by 19 7 mm Hg, P ¼ 0.09 by Tukey’s test) than in conscious sheep. In contrast, cortical PO2 did not change significantly with transition to CPB. Part 2: altered pump flow Stability of the preparation.
Variables shown in Figures 2 and 3, Table 2, and Supplementary Tables S2 and S3 remained stable across the 3 periods at which pump flow was set at 80 ml kg–1min–1. Reduced pump flow. Systemic DO2 was reduced by 23% 1%, but systemic VO2 did not change significantly (Table 2). Systemic fractional oxygen extraction increased
and mixed venous SO2 decreased from 76.7% 1.3% to 67.5% 2.0% During perfusion at 60 ml kg–1min–1, RBF (–16% 2%) and RVC (–14% 2%) fell from their levels at 80 ml kg– 1 min–1, but neither cortical nor medullary perfusion changed significantly (Figure 2, Table 2, and Supplementary Table S2). Renal DO2 fell by 14% 3%, but neither renal VO2, renal venous SO2, nor the renal fractional extraction of oxygen changed significantly, despite reductions in creatinine clearance and sodium reabsorption (Table 2 and Supplementary Table S2). Nevertheless, both cortical (–21 5 mm Hg) and medullary (–11 3 mm Hg) PO2 fell.
Figure 2 | Effects of decreasing pump flow on cardiopulmonary bypass on renal hemodynamics and renal tissue oxygenation: Target arterial pressure was maintained at 70 mm Hg by administration of metaraminol. Black circles and error bars represent the mean SEM of each variable across 3 experimental periods (n ¼ 12 for mean arterial pressure and renal blood flow, but due to equipment failure n ¼ 10 for cortical PO2 and n ¼ 9 for medullary PO2). Data for individual sheep are shown by colored lines. **P # 0.01, ***P # 0.001 for comparison of the levels of variables during perfusion at 60 ml kg–1min–1 with their mean levels over the 2 control periods of perfusion at 80 ml kg–1min–1 that preceded and followed the intervention (Student’s paired t-test). 1340
Kidney International (2019) 95, 1338–1346
basic research
YR Lankadeva et al.: Renal oxygenation during cardiopulmonary bypass
Figure 3 | Effects of increasing pump flow on cardiopulmonary bypass on renal hemodynamics and renal tissue oxygenation: Target arterial pressure was maintained at 70 mm Hg by administration of metaraminol. Black circles and error bars represent the mean SEM of each variable across 3 experimental periods (n ¼ 12 for mean arterial pressure and renal blood flow, but due to equipment failure n ¼ 10 for cortical PO2 and n ¼ 10 for medullary PO2). Data for individual sheep are shown by colored lines. *P # 0.05, ***P # 0.001 for comparison of the levels of variables during perfusion at 100 ml kg–1min–1 with their mean levels over the 2 control periods of perfusion at 80 ml kg–1min–1 that preceded and followed the intervention (Student’s paired t-test).
Increased pump flow. This did not significantly alter MAP, body temperature, or arterial PO2, SO2, hemoglobin concentration, or oxygen content (Figure 3, Supplementary Table S3). Systemic DO2 increased by 21% 3%, but systemic VO2 did not change significantly (Table 2). Consequently, systemic fractional oxygen extraction decreased. Mixed venous SO2 increased from 77.0% 1.3% to 82.0% 1.5%. During perfusion at 100 ml kg–1min–1, RBF (þ17% 4%), RVC (þ19% 5%), renal DO2 (þ14% 6%), and medullary PO2 (þ9 3 mm Hg) were greater than at 80 ml kg–1min–1 (Figure 3, Table 2, and Supplementary Table S3). There were no significant changes in creatinine clearance, sodium reabsorption, cortical or medullary perfusion, or cortical PO2. Summary. Both systemic vascular conductance and RVC varied positively with pump flow (Figure 4). Thus, compared with levels at 60 ml kg–1min–1, when pump flow was 100 ml kg–1min–1 systemic DO2 was 59% 5% greater (P < 0.001, paired t-test), RBF was 47% 9% greater (P < 0.001), and renal DO2 was 40% 9% greater (P ¼ 0.006). The increased renal DO2 associated with increased pump flow was in turn associated with increased cortical and medullary PO2 (Figure 5a). Thus, compared with levels at 60 ml kg–1min–1, when pump flow was 100 ml kg–1min–1 cortical PO2 was 36.9 12.5 mm Hg greater (P ¼ 0.02) and medullary PO2 was 15.7 6.2 mm Hg greater (P ¼ 0.05). Part 3: progressively increased MAP
Metaraminol dose-dependently decreased systemic vascular conductance (Supplementary Table S4, Figure 4) so MAP was Kidney International (2019) 95, 1338–1346
dose-dependently increased, from its baseline level of 63 3 mm Hg to 90 3 mm Hg (0.2 mg/min), 107 3 mm Hg (0.4 mg/min), and 113 5 mm Hg (0.6 mg/min) (Figure 6). Systemic DO2 increased slightly (<10%) at 0.2 and 0.4 mg/ min because of slight increases in arterial hemoglobin concentration (Tables 3, Supplementary S4). There were no significant changes in systemic VO2 or systemic oxygen extraction. RVC was not significantly altered by metaraminol (Figure 4), so with the increased MAP, RBF and renal DO2 were dose-dependently increased (Figure 6, Table 3, and Supplementary Table S4). Renal VO2 was also increased by all doses of metaraminol, as were creatinine clearance and sodium reabsorption. No changes in cortical or medullary perfusion were detected. At 0.2 mg/min, metaraminol increased medullary PO2 (by 20.2 6.9 mm Hg) and cortical PO2 (by 39.0 10.1 mm Hg) (Figure 6). At 0.4 mg/min cortical PO2 (by 30.3 10.2 mm Hg), but not medullary PO2, was increased. At 0.6 mg/ min, neither cortical nor medullary PO2 differed significantly from control levels. Thus, cortical and medullary tissue PO2 were positively associated with renal DO2 at lower (0.2 mg/ min) but not higher (0.4 and 0.6 mg/min) doses of metaraminol (Figure 5b). DISCUSSION
In a clinically relevant large animal model of CPB we found hypoxia in the renal medulla, even under perfusion conditions that would be considered optimal in the clinical setting. Medullary hypoxia during CPB appeared to be driven by reduced medullary tissue oxygen delivery. The renal cortex 1341
basic research
YR Lankadeva et al.: Renal oxygenation during cardiopulmonary bypass
Table 2 | Effects of decreasing and increasing pump flow during cardiopulmonary bypass on systemic and whole kidney oxygenation, regional renal perfusion, creatinine clearance, and sodium reabsorption Decreasing pump flow
n
Systemic DO2, ml O2 kg–1min–1 Systemic VO2, ml O2 kg–1min–1 Systemic FEO2, % Renal DO2, ml O2 kg–1min–1 Renal VO2, ml O2 kg–1min–1 Renal FEO2, % Cortical perfusion, units Medullary perfusion, units CrCl, ml kg–1min–1 TNaþ , mmol kg–1min–1
12 12 12 12 5 5 12 11 12 12
Increasing pump plow
n
Systemic DO2, ml O2 kg–1min–1 Systemic VO2, ml O2 kg–1min–1 Systemic FEO2, % Renal DO2, ml O2 kg–1min–1 Renal VO2, ml O2 kg–1min–1 Renal FEO2, % Cortical perfusion, units Medullary perfusion, units CrCl, ml kg–1min–1 TNaþ , mmol kg–1min–1
12 12 12 12 5 5 12 11 12 12
80 ml kg–1min–1 10.1 2.68 26.5 0.35 0.041 16.4 824 376 1.44 198
0.3 0.15 1.2 0.07 0.007 4.1 147 121 0.31 44
80 ml kg–1min–1 10.8 2.77 25.8 0.42 0.052 16.7 883 306 1.72 236
0.4 0.16 1.3 0.08 0.007 2.6 179 129 0.39 53
60 ml kg–1min–1 8.0 2.78 34.7 0.32 0.051 21.3 631 321 0.86 120
0.2 0.19 1.9 0.06 0.001 3.9 96 160 0.20 29
100 ml kg–1min–1 12.7 2.60 20.5 0.45 0.054 16.9 991 364 1.91 265
0.4 0.20 1.3 0.08 0.002 1.9 216 116 0.57 79
80 ml kg–1min–1
P
<0.001 0.06 <0.001 <0.001 0.89 0.16 0.07 0.50 <0.001 <0.001
10.7 2.66 24.9 0.40 0.063 20.6 829 329 1.21 168
0.4 0.18 1.3 0.07 0.007 4.3 160 130 0.22 31
80 ml kg–1min–1
P
<0.001 0.55 <0.001 0.004 0.80 0.44 0.09 0.19 0.27 0.27
10.1 2.55 25.1 0.39 0.053 20.1 870 336 1.25 175
0.3 0.17 1.4 0.08 0.003 2.9 184 112 0.22 31
CrCl, creatinine clearance; Do2, oxygen delivery; FEo2, fractional oxygen extraction; TNaþ, sodium reabsorption; Vo2, oxygen consumption. Values are mean SEM. P values are the outcomes of Student’s paired t-test for comparison of variables during the period of perfusion at 60 (top) or 100 ml kg–1min–1 (bottom) against the mean level during the control periods that preceded and followed these interventions, during which sheep were perfused at 80 ml kg–1min–1.
appeared to be protected against hypoxia during CPB, despite similar ischemia to that of the medulla. This may be related to the large increase in tissue PO2 induced by hyperoxemia in the
Figure 4 | Relationships between systemic vascular conductance and renal vascular conductance during changes in pump flow and infusion of metaraminol: Symbols and error bars represent mean ± SEM of n [ 12 for altered pump flow and n [ 11 for metaraminol infusion. For altered pump flow the value for 80 ml kg–1min–1 is the average over the 3 experimental periods at which pump flow was maintained at this level during Part 2 of the experimental protocol (see Supplementary Figure S1). Note that pump flow was held constant at 80 ml kg–1min–1 during infusion of metaraminol. However, when pump flow was altered, a target arterial pressure of 70 mm Hg was achieved by administration of metaraminol (see Methods for details). 1342
cortex, which is largely absent in the medulla.16,17 Renal medullary oxygenation was improved, at least in a subset of animals, by increasing pump flow within the limits of clinical feasibility or by the use of the vasopressor metaraminol to increase MAP within a clinically feasible range. Increased medullary oxygenation, both in response to increased pump flow and increased MAP, was associated with increased renal DO2. Our findings provide a potential approach for improving renal medullary oxygenation during CPB and thus potentially decreasing postoperative AKI. They also provide a potential mechanistic basis for the recent clinical finding that avoidance of whole body DO2 < 280 ml kg–1m–2 on bypass reduced the risk of postoperative AKI.11 To our knowledge this is the first detailed analysis of the effects of CPB on renal oxygenation in a clinically relevant large animal model. Hypoxia in the renal medulla has previously been reported in a rat model of CPB7 and in a pilot study of 2 pigs.8 The ovine model recapitulates many aspects of CPB in human cardiothoracic surgery and was specifically designed to achieve what would be considered optimal perfusion conditions in clinical practice. Arterial oxygen content was optimized by adding w400 ml of fresh donor blood in the prime, so that arterial hemoglobin content was only 11.6% 3.1% less during CPB than in the conscious sheep. Furthermore, the FiO2 of 60% achieved an arterial PO2 of w300 mm Hg and an SO2 > 98% in all 12 animals during CPB. We also used a relatively high standard pump flow of 80 ml kg–1min–1. This equates in humans to a flow rate of 5.6 L/ min, or 3.16 L min–1m–2 for a 70-kg human of height 165 cm (body surface area calculated by the method of Du Bois and Kidney International (2019) 95, 1338–1346
YR Lankadeva et al.: Renal oxygenation during cardiopulmonary bypass
Figure 5 | Relationships between renal oxygen delivery and renal tissue PO2 during changes in (a) pump flow and (b) infusion of metaraminol: Symbols and error bars represent mean ± SEM of n [ 11 for cortical PO2 and n [ 9 for medullary PO2. For altered pump flow the value for 80 ml kg–1min–1 is the average over the 3 experimental periods at which pump flow was maintained at this level during Part 2 of the experimental protocol (see Supplementary Figure S1).
Du Bois18). This is greater than the commonly used flow rate in adult humans of w2.4 L min–1m–2.19 Sheep were also cooled by w3 C on bypass, in line with standard practice for coronary artery bypass graft surgery at our associated center.19 Consequently, mixed venous SO2 on CPB was w76%, within the target range for clinical perfusion.15 We also rendered sheep relatively hypercapnic on CPB to promote dilation of renal resistance arteries.20 Nevertheless, and despite the fact that the sheep we studied were young and otherwise healthy, marked renal medullary hypoxia was observed during CPB. Several mechanisms may contribute to medullary hypoxia during CPB. Our experimental design allowed us to separately evaluate the effects of anesthesia (which reduced MAP and RVC) and subsequent transition to CPB (which further reduced RVC). Thus, during CPB, RVC was reduced by w49% and MAP by w23 mm Hg compared with the conscious state. Consequently, RBF and renal DO2 were both reduced, by w61%, along with reductions in local perfusion in both the cortex and medulla. However, renal VO2 was also reduced during CPB (w53%) by a similar magnitude to the reduction Kidney International (2019) 95, 1338–1346
basic research
in renal DO2, despite no significant change in creatinine clearance or sodium reabsorption. Furthermore, in contrast to renal medullary PO2, cortical PO2 was well maintained during CPB, despite decreased cortical perfusion. Thus, the renal medulla is particularly susceptible to the development of hypoxia during CPB. Simulations from computational models of CPB in rats support the notion that this susceptibility arises, at least in part, from the anatomic position of the thick ascending limbs of the loop of Henle at the periphery of vascular bundles in the outer medulla. They are thus susceptible to even relatively small deficits in local oxygen supply in relation to their relatively high oxygen demand.5,6 Our findings indicate that renal medullary oxygenation partly depends on pump flow. When pump flow was reduced to 60 ml kg–1min–1 (equivalent to 2.37 L min–1 m–2 in adult humans, close to standard pump flow in clinical practice), the renal medulla was rendered markedly hypoxic (w14 mm Hg). However, when pump flow was increased to 100 ml kg–1min–1, renal medullary PO2 increased markedly to w29 mm Hg, albeit to levels still w17 mm Hg less than in the conscious sheep. Increased pump flow led to increased RVC, RBF, and renal DO2, providing at least a partial explanation for the improved medullary PO2. These data provide the first evidence of which we are aware that renal medullary oxygenation during CPB can be increased by increased pump flow. Our observations are consistent with those of Mackay and colleagues21 who found that increasing pump flow in a porcine model of CPB, under conditions in which MAP was also allowed to increase, increased RBF as measured by microspheres. Our current findings indicate that even in the absence of increased MAP, increased pump flow can increase renal oxygenation, including in the medulla. They are also consistent with clinical observations that low pump flow or low whole body DO2 on CPB is associated with greater incidence of postoperative AKI9,22,23 and that avoidance of low whole body DO2 during CPB can reduce the incidence of postoperative AKI.10,11 Medullary PO2 was also increased by low-dose infusion of metaraminol to increase MAP from w63 to w90 mm Hg. The renal circulation appears to be relatively insensitive to the vasoconstrictor effects of metaraminol, because RVC did not change significantly in the face of large decreases in systemic vascular conductance. Renal VO2 was also increased by all 3 doses of metaraminol, presumably because of the increase in glomerular filtration rate and thus tubular sodium reabsorption.24 Nevertheless, metaraminol infusion increased renal medullary PO2 during CPB. These data are also consistent with the observation that larger decreases in average MAP during CPB, relative to preoperative MAP, are associated with a greater incidence of AKI.23 They are also consistent with the inability to detect detrimental effects of high-dose metaraminol on renal function in patients with sepsis.25 However, it should be noted that a (albeit small) clinical trial of higher target MAP during CPB was unable to demonstrate reduced incidence of AKI.26 Thus, the potential for increased target MAP during CPB to reduce the risk of AKI remains to be determined. 1343
basic research
YR Lankadeva et al.: Renal oxygenation during cardiopulmonary bypass
Figure 6 | Effects of progressively increasing doses of metaraminol on renal hemodynamics and renal tissue oxygenation: Pump flow was maintained at 80 ml kg–1min–1 throughout this part of the experiment. Black circles and error bars represent the mean SEM of each variable across the 4 experimental periods (n ¼ 11 for mean arterial pressure and renal blood flow, but due to equipment failure n ¼ 10 for cortical PO2 and n ¼ 9 for medullary PO2). Data for individual sheep are shown by colored lines. *P # 0.05, **P # 0.01, ***P # 0.001 for comparison with the control period without metaraminol (Dunnett’s test).
Limitations of our study include the anesthetic used, isoflurane, which is now uncommonly used in human cardiac surgery. However, there is no evidence that more commonly used agents, such as sevoflurane or propofol/fentanyl, are more renoprotective than isoflurane.27 We acknowledge the limitations of laser Doppler flowmetry, which is an index of erythrocyte velocity in highly perfused tissues such as the kidney.28 We studied young adult sheep, so our findings are more relevant to young healthy patients than older patients with comorbidities. Nevertheless, we might expect the deleterious effects of CPB on medullary oxygenation to be worse in elderly patients with comorbidities than in young sheep. Finally, we were only able to collect renal venous blood samples in 5 of the 12 sheep because of loss of patency of the renal venous catheter. This reduced our ability to detect changes in renal VO2 and related parameters.
In conclusion, in a clinically relevant ovine model of CPB we were able to monitor whole kidney and intrarenal tissue perfusion and oxygenation. The renal medulla is susceptible to hypoxia during CPB, but medullary oxygenation can be improved by increasing pump flow or, more effectively, by increasing MAP with a vasopressor agent that does not decrease RVC. Both interventions are associated with increased renal DO2, thus ameliorating the deficit of a variable that appears to be a hallmark of CPB. METHODS Animal preparation All protocols were approved by the Animal Ethics Committee of the Florey Institute of Neuroscience and Mental Health under guidelines of the National Health and Medical Research Council of Australia. Merino ewes (35–48 kg body weight, n ¼ 12) were housed in
Table 3 | Effects of metaraminol on systemic and whole kidney oxygenation, regional renal perfusion, creatinine clearance, and sodium reabsorption Metaraminol Variable –1
–1
Systemic DO2, ml O2 kg min Systemic VO2, ml O2 kg–1min–1 Systemic FEO2, % Renal DO2, ml O2 kg–1min–1 Renal VO2, ml O2 kg–1min–1 Renal FEO2, % Cortical perfusion, units Medullary perfusion, units CrCl, ml kg–1min–1 TNaþ, mmol kg–1min–1
n
0 (control)
200 mg/min
400 mg/min
11 11 11 11 5 5 11 10 11 11
10.1 2.42 23.8 0.24 0.050 21.5 606 365 0.63 89
0.4 0.20 1.3 0.03 0.004 2.9 109 133 0.11 15
10.6 2.45 23.0 0.38 0.058 18.7 1141 439 1.51 212
a
0.3 0.20 1.4 0.06a 0.003b 3.9 310 123 0.32b 45a
10.7 2.63 24.4 0.43 0.076 21.5 1281 437 1.57 219
b
0.3 0.25 1.9 0.06a 0.008b 6.1 344 150 0.22c 31c
600 mg/min
PTime
0.19 0.32 0.49 0.005 0.02 0.31 0.10 0.64 0.001 0.002
10.5 2.47 23.9 0.43 0.064 16.9 1064 372 1.43 199
0.3 0.12 1.6 0.06a 0.004b 4.4 307 136 0.24a 34a
CrCl, creatinine clearance; Do2, oxygen delivery; FEo2, fractional oxygen extraction; TNaþ, sodium reabsorption; Vo2, oxygen consumption. Values are mean SEM. PTime is the outcome of repeated-measures analysis of variance. a P # 0.01, bP # 0.05, cP # 0.001 for comparison against control values before infusion of metaraminol (Dunnett’s test).
1344
Kidney International (2019) 95, 1338–1346
basic research
YR Lankadeva et al.: Renal oxygenation during cardiopulmonary bypass
individual metabolic cages with free access to water and 800 g of oaten chaff daily. Sheep underwent preliminary aseptic surgery under general anesthesia, induced with 15 mg/kg sodium thiopental (Jurox, Rutherford, Australia) and, after intubation, maintained on 2.0% to 2.5% isoflurane (Isoflo; Zoetis, Rhodes, Australia). A transit-time flow probe (4 mm) was placed around the left renal artery, the renal vein was cannulated, and fiber-optic probes (CP-026-001; Oxford Optronix Ltd, Abingdon, UK) were inserted in the renal cortex and medulla to measure tissue PO2 by fluorescence lifetime oximetry, tissue perfusion by laser-Doppler flux, and tissue temperature.14 The carotid artery and jugular vein were cannulated. A bladder catheter (Foley size 12, 30 ml; Euromedical, Malaysia) was inserted for collection of urine. Sheep were given 50 mg flunixin meglumine (Flunixon; Norbrook, Tullarmarine, Australia) for analgesia and 900 mg procaine penicillin (Ilium; Troy Laboratories, Glendinning, Australia) during surgery and then 24 and 48 hours postoperatively and allowed 5 days of recovery before experimentation. Analog signals were digitized as previously described.14 A maintenance infusion of compound sodium lactate (1 ml kg–1h–1; Baxter, Old Toongabbie, NSW, Australia) was given i.v. from 6 pm the night before the experiment.
MAP and RBF were stable a dose–response relationship for metaraminol infusion was established (0, 0.2, 0.4, and 0.6 mg/min). At each dose a 20-minute experimental period began once MAP and RBF were stable. Upon completion of Part 3, metaraminol was withdrawn and sheep were killed with sodium pentobarbital (20 mg/kg i.v., Lethobarb; Virbac, Wetherill Park, Australia). Five minutes later the perfusion pump was turned off to confirm that renal perfusion and PO2 were close to zero. The left kidney was removed, and the positions of the fiber-optic probes confirmed. Statistical analysis Data are expressed as mean SEM. After 1-way repeated-measures analysis of variance, post hoc comparisons29 were made by Student’s paired t-test, Tukey’s test, or Dunnett’s test. Analyses were performed using SYSTAT (version 13; Systat, San Jose, CA) or GraphPad Prism (GraphPad Software, La Jolla, CA). Two-sided P # 0.05 was considered statistically significant. DISCLOSURE All the authors declared no competing interests.
ACKNOWLEDGMENTS
Experimental protocol The experimental protocol had 3 parts (Supplementary Figure S1), each comprising a series of 20- to 30-minute experimental periods. Urine was collected throughout each experimental period. At the midpoint of each experimental period arterial, mixed venous, and renal venous blood samples were collected for oximetry and blood chemistry (ABL Systems 625, Radiometer, Copenhagen, Denmark). Arterial plasma and urine samples were stored at –80 C for later measurement of creatinine and sodium. Part 1: anesthesia and transition to CPB. For the first experimental period (30 minutes) the animal was conscious and unrestrained in its home cage. Anesthesia was then induced (as above) with a FiO2 of 60%. Once anesthesia was established, an infusion of 10 ml/kg of compound sodium lactate solution was administered over a 20-minute period. After this, sheep received a maintenance infusion of compound sodium lactate solution at 2 ml kg–1h–1 throughout the experiment. Once MAP and RBF were stable, the second experimental period began (stable anesthesia). CPB was then established (details in Supplementary Methods) at a flow rate of 80 ml kg–1min–1. If required, metaraminol (Metaraminol Montrose 10 mg/ml; Montrose Life Sciences, NSW, Australia) was given i.v. to maintain MAP at 65 to 75 mm Hg, with a target of 70 mm Hg. Once MAP and RBF were stable, the third experimental period began (on bypass). Part 2: altered pump flow. A target MAP of 70 mm Hg was maintained with titration of metaraminol. Pump flow rate was randomized to 60 or 100 ml kg–1min–1, returning to 80 ml kg–1min–1 after either intervention. The alternate pump flow rate was then used. Thus, each period of altered pump flow (60 or 100 ml kg–1min–1) was preceded and followed by a period at 80 ml kg–1min–1 (Supplementary Figure S1). At each pump flow, once MAP and RBF were stable, a 20-minute experimental period began. Part 3: progressively increased MAP. Metaraminol support from the previous experimental period was withdrawn, and once Kidney International (2019) 95, 1338–1346
We thank Tom Vale and Tony Dornom for technical assistance and Libin Jose and Ilaria Catteneo for perfusion services. Support from the National Health and Medical Research Council of Australia (GNT1122455), the Victorian Government Operational Infrastructure Support Grant, and the National Heart Foundation of Australia (101853) is gratefully acknowledged. SUPPLEMENTARY MATERIAL Supplementary Methods. Table S1. Effects of anesthesia and cardiopulmonary bypass on systemic and renal hemodynamics, blood oximetry and chemistry, and renal function. Table S2. Effects of reduced pump flow during cardiopulmonary bypass on systemic and renal hemodynamics, blood oximetry and chemistry, and renal function. Table S3. Effects of increasing pump flow during cardiopulmonary bypass on systemic and renal hemodynamics, blood oximetry and chemistry, and renal function. Table S4. Effects of metaraminol on systemic and renal hemodynamics, blood oximetry and chemistry, and renal function during cardiopulmonary bypass. Figure S1. Experimental timeline. The experiment comprised 3 parts. In Part 1, sheep were transitioned from the conscious state to stable anesthesia to cardiopulmonary bypass. In Part 2, pump flow was varied while mean arterial pressure was titrated to a target of 70 mm Hg by i.v. administration of metaraminol. Note that each period of low (60 ml kg–1min–1) or high (100 ml kg–1min–1) pump flow was preceded and followed by a period of perfusion at 80 ml kg–1min–1. In Part 3, pump flow was maintained at 80 ml kg–1min–1 during infusion of progressively greater doses of metaraminol (0, 0.2, 0.4, and 0.6 mg/min). Each experimental period commenced once arterial pressure and renal blood flow were stable (10–20 minutes after the change to conditions) and was of 20 to 30 minutes’ duration. Urine was collected during the experimental period, and a blood sample was taken at the mid-point of each experimental period, as shown by the arrows. See the text for further details. Supplementary material is linked to the online version of the paper at www.kidney-international.org. 1345
basic research
REFERENCES 1. Hu J, Chen R, Liu S, et al. Global incidence and outcomes of adult patients with acute kidney injury after cardiac surgery: a systematic review and meta-analysis. J Cardiothorac Vasc Anesth. 2016;30:82–89. 2. Evans RG, Lankadeva YR, Cochrane AD, et al. Renal haemodynamics and oxygenation during and after cardiac surgery and cardiopulmonary bypass. Acta Physiol. 2018;222:e12995. 3. Ow CPC, Ngo JP, Ullah MM, et al. Renal hypoxia in kidney disease: cause or consequence? Acta Physiol. 2018;222:e12999. 4. Post EH, Kellum JA, Bellomo R, et al. Renal perfusion in sepsis: from macro- to microcirculation. Kidney Int. 2017;91:45–60. 5. 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. 6. 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. 7. 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. 8. Stafford-Smith M, Grocott HP. Renal medullary hypoxia during experimental cardiopulmonary bypass: a pilot study. Perfusion. 2005;20: 53–58. 9. 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. 10. Magruder JT, Crawford TC, Harness HL, et al. A pilot goal-directed perfusion initiative is associated with less acute kidney injury after cardiac surgery. J Thorac Cardiovasc Surg. 2017;153:118–125. 11. Ranucci M, Johnson I, Willcox T, et al. Goal-directed perfusion to reduce acute kidney injury: a randomized trial. J Thorac Cardiovasc Surg. 2018;156:1918–1927. 12. Lannemyr L, Bragadottir G, Hjarpe A, et al. Impact of cardiopulmonary bypass flow on renal oxygenation in patients undergoing cardiac surgery. Ann Thorac Surg. 2019;107:505–511. 13. 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. 14. Calzavacca P, Evans RG, Bailey M, et al. Long-term measurement of renal cortical and medullary tissue oxygenation and perfusion in unanesthetized sheep. Am J Physiol Regul Integr Comp Physiol. 2015;308: R832–R839.
1346
YR Lankadeva et al.: Renal oxygenation during cardiopulmonary bypass
15. Svenmarker S, Haggmark S, Ostman M, et al. Central venous oxygen saturation during cardiopulmonary bypass predicts 3-year survival. Interactive Cardiovasc Thorac Surg. 2013;16:21–26. 16. Aperia AC. The influence of arterial PO2 on renal tissue PO2. Acta Physiol Scand. 1969;75:353–359. 17. Leong CL, Anderson WP, O’Connor PM, et al. Evidence that renal arterialvenous oxygen shunting contributes to dynamic regulation of renal oxygenation. Am J Physiol Renal Physiol. 2007;292:F1726–F1733. 18. Du Bois D, Du Bois EF. A formula to estimate the approximate surface area if height and weight be known. Arch Intern Med. 1916;17:863–871. 19. Zhu MZL, Martin A, Cochrane AD, et al. Urinary hypoxia: an intraoperative marker of risk of cardiac surgery-associated acute kidney injury. Nephol Dial Transplant. 2018;33:2191–2201. 20. Daugherty RM Jr, Scott JB, Dabney JM, et al. Local effects of O2 and CO2 on limb, renal, and coronary vascular resistances. Am J Physiol. 1967;213: 1102–1110. 21. Mackay JH, Feerick AE, Woodson LC, et al. Increasing organ blood flow during cardiopulmonary bypass in pigs: comparison of dopamine and perfusion pressure. Crit Care Med. 1995;23:1090–1098. 22. Fischer UM, Weissenberger WK, Warters RD, et al. Impact of cardiopulmonary bypass management on postcardiac surgery renal function. Perfusion. 2002;17:401–406. 23. Kanji HD, Schulze CJ, Hervas-Malo M, et al. Difference between preoperative and cardiopulmonary bypass mean arterial pressure is independently associated with early cardiac surgery-associated acute kidney injury. J Cardiothorac Surg. 2010;5:71. 24. Evans RG, Harrop GK, Ngo JP, et al. Basal renal oxygen consumption and the efficiency of oxygen utilization for sodium reabsorption. Am J Physiol Renal Physiol. 2014;306:F551–F560. 25. Hou LC, Li SZ, Xiong LZ, et al. Effect of dopamine and metaraminol on the renal function of patients with septic shock. Chin Med J. 2007;120:680–683. 26. Urzua J, Troncoso S, Bugedo G, et al. Renal function and cardiopulmonary bypass: effect of perfusion pressure. J Cardiothorac Vasc Anesth. 1992;6:299–303. 27. Tan SI, Brewster DJ, Horrigan D, et al. Pharmacological and non-surgical renal protective strategies for cardiac surgery patients undergoing cardiopulmonary bypass: a systematic review [e-pub ahead of print]. ANZ J Surg. https://doi.org/10.1111/ans.14800. 28. Eppel GA, Bergstrom G, Anderson WP, et al. Autoregulation of renal medullary blood flow in rabbits. Am J Physiol Regul Integr Comp Physiol. 2003;284:R233–R244. 29. Ludbrook J. Multiple comparison procedures updated. Clin Exp Pharmacol Physiol. 1998;25:1032–1037.
Kidney International (2019) 95, 1338–1346
www.kidney-international.org
Strategies that improve renal medullary oxygenation during experimental cardiopulmonary bypass may mitigate postoperative acute kidney injury
see commentary on page 1292
Yugeesh R. Lankadeva1, Andrew D. Cochrane2, Bruno Marino3, Naoya Iguchi1, Sally G. Hood1, Rinaldo Bellomo4, Clive N. May1,6 and Roger G. Evans5,6 1
Pre-Clinical Critical Care Unit, Florey Institute of Neuroscience and Mental Health, University of Melbourne, Melbourne, Victoria, Australia; 2Department of Cardiothoracic Surgery, Monash Health and Department of Surgery (School of Clinical Sciences at Monash Health), Monash University, Melbourne, Victoria, Australia; 3Cellsaving and Perfusion Resources, Melbourne, Victoria, Australia; 4 Department of Intensive Care, Austin Hospital, Heidelberg, Victoria, Australia; and 5Cardiovascular Disease Program, Biomedicine Discovery Institute and Department of Physiology, Monash University, Melbourne, Victoria, Australia
Renal medullary hypoxia may contribute to cardiac surgery–associated acute kidney injury (AKI). However, the effects of cardiopulmonary bypass (CPB) on medullary oxygenation are poorly understood. Here we tested whether CPB causes medullary hypoxia and whether medullary oxygenation during CPB can be improved by increasing pump flow or mean arterial pressure (MAP). Twelve sheep were instrumented to measure whole kidney, medullary, and cortical blood flow and oxygenation. Five days later, under isoflurane anesthesia, CPB was initiated at a pump flow of 80 mL kg–1min–1 and target MAP of 70 mm Hg. Pump flow was then set at 60 and 100 mL kg–1min–1, while MAP was maintained at approximately 70 mm Hg. MAP was then increased by vasopressor (metaraminol, 0.2–0.6 mg/min) infusion at a pump flow of 80 mL kg–1min–1. CPB at 80 mL kg–1min–1 reduced renal blood flow (RBF), -61% less than the conscious state, perfusion in the cortex (–44%) and medulla (–40%), and medullary PO2 from 43 to 27 mm Hg. Decreasing pump flow from 80 to 60 mL kg–1min–1 further decreased RBF (–16%) and medullary PO2 from 25 to 14 mm Hg. Increasing pump flow from 80 to 100 mL kg–1min–1 increased RBF (17%) and medullary PO2 from 20 to 29 mm Hg. Metaraminol (0.2 mg/min) increased MAP from 63 to 90 mm Hg, RBF (47%), and medullary PO2 from 19 to 39 mm Hg. Thus, the renal medulla is susceptible to hypoxia during CPB, but medullary oxygenation can be improved by increasing pump flow or increasing target MAP by infusion of metaraminol. Kidney International (2019) 95, 1338–1346; https://doi.org/10.1016/ j.kint.2019.01.032 Correspondence: R. Evans, Department of Physiology, 26 Innovation Walk, Monash University, Victoria 3800, Australia. E-mail: [email protected] 6
These authors contribute equally to this work.
Received 21 November 2018; revised 23 January 2019; accepted 24 January 2019; published online 15 March 2019
1338
KEYWORDS: acute kidney injury; cardiac surgery; hypoxia; renal circulation; renal medulla Copyright ª 2019, International Society of Nephrology. Published by Elsevier Inc. All rights reserved.
Translational Statement Renal medullary hypoxia is a plausible mechanistic pathway driving the development of AKI after cardiac surgery requiring CPB. We believe that currently accepted bypass flow and pressure may be inadequate in some cases. Our current findings indicate that it is feasible to improve medullary oxygenation on bypass and to potentially mitigate the risk of postoperative AKI. To test this proposition in human patients, a randomized clinical trial could be conducted to test the effects of increasing target pump flow and arterial pressure on the incidence of AKI.
KI occurs in more than 20% of adult patients after cardiac surgery requiring CPB.1 The ability to decrease the risk of AKI associated with cardiac surgery is hampered by our lack of knowledge of the effects of CPB on renal perfusion and oxygenation.2 Evolving evidence shows that renal medullary hypoxia contributes to the development of AKI.3,4 Mathematical models predict renal medullary hypoxia during CPB,5,6 and medullary hypoxia has been observed in a rat model of CPB7 and in a pilot study of 2 pigs subjected to CPB.8 However, there have been no systematic investigations of renal global and regional (i.e., cortical and medullary) perfusion and oxygenation in a clinically relevant large animal model of CPB. Furthermore, although evidence shows that goaldirected perfusion can mitigate risk of postoperative AKI9–11 and that the increase in renal fractional oxygen extraction during CPB can be alleviated by increasing pump flow,12,13 the potential for perfusion conditions to be
A
Kidney International (2019) 95, 1338–1346
YR Lankadeva et al.: Renal oxygenation during cardiopulmonary bypass
altered to improve renal medullary oxygenation has not been investigated. Renal cortical and medullary perfusion and tissue oxygen tension (PO2) can be continuously measured in unanesthetized sheep via chronically implanted fiber-optic probes.14 Global renal oxygen delivery (DO2) and oxygen consumption (VO2) can also be quantified through the combination of transit-time ultrasound flowmetry and arterial and renal venous blood samples.14 Using this model we aimed to test the hypotheses that CPB causes medullary hypoxia and medullary oxygenation during CPB can be improved by increasing pump flow or mean arterial pressure (MAP). Thus, we assessed the effects of clinically relevant perfusion conditions on global and regional renal perfusion and oxygenation. Subsequently, we established the effects of altering pump flow and MAP. RESULTS Part 1: anesthesia and transition to CPB Anesthesia. Anesthesia reduced MAP (–14 5 mm Hg;
Figure 1) and renal vascular conductance (RVC; –21% 4%; Supplementary Table S1), so renal blood flow (RBF) decreased (–32% 3%; Figure 1). Neither cortical nor medullary perfusion changed significantly (Table 1). Arterial PO2 increased due to use of a fractional inspired O2 (FiO2) of 60% (Supplementary Table S1). However, arterial blood hemoglobin concentration fell due to the fluid load given after induction of anesthesia. Thus, due to reduced arterial oxygen content and reduced RBF, renal DO2 fell by 41% 4% (Table 1). Renal VO2 tended to fall (22% 11%)
basic research
despite no decrease in creatinine clearance or sodium reabsorption. Consequently, neither renal venous PO2 nor saturation (SO2), nor renal fractional oxygen extraction, was significantly altered. In contrast, systemic fractional oxygen extraction fell under anesthesia, so that mixed venous PO2 and SO2 increased. There was no significant change in cortical or medullary PO2 (Figure 1). Cardiopulmonary bypass. As in human CPB, body temperature was intentionally reduced by 3.1 0.3 C, but there was no further significant change in MAP from its level in anesthetized sheep (Figure 1, Supplementary Table S1). Arterial blood hemoglobin content increased but was still 12% 3% less than in the conscious sheep. Arterial oxygen content did not differ significantly from that in conscious sheep. Systemic fractional oxygen extraction increased so that mixed venous SO2 (76.0% 1.2%) was compatible with a target of $75% for CPB in human patients.15 RBF (–61% 6%), RVC (–49% 8%), cortical perfusion (–44% 10%), and medullary perfusion (–40% 15%) were all less during CPB than in conscious sheep (Figure 1, Supplementary Table S1). RBF (–41% 8%), RVC (–34% 10%), and medullary perfusion (–39% 17%) were also less than under anesthesia alone. Consequently, despite no further hemodilution and maintenance of mixed venous SO2 $ 75%, renal DO2 was 61% 6% less than in conscious sheep and 32% 10% less than under anesthesia. Renal VO2 was also 53% 9% less on CPB than in conscious sheep, despite no significant changes in creatinine clearance or sodium reabsorption. Nevertheless, medullary PO2 decreased by 22 8 mm Hg compared with anesthesia alone and tended to be
Figure 1 | Effects of anesthesia and cardiopulmonary bypass on renal hemodynamics and renal tissue oxygenation: Black circles and error bars represent the mean ± SEM of each variable across each of the first 3 experimental periods (n [ 12 for mean arterial pressure and renal blood flow, but due to equipment failure n [ 10 for cortical PO2 and medullary PO2). Data for individual sheep are shown by colored lines. P values were derived from Tukey’s multiple comparison procedure. **P # 0.01, ***P # 0.001 for comparison with Conscious; †P # 0.05, ††† P # 0.001 for comparison of On Bypass with Anesthetized. Kidney International (2019) 95, 1338–1346
1339
basic research
YR Lankadeva et al.: Renal oxygenation during cardiopulmonary bypass
Table 1 | Effects of anesthesia and cardiopulmonary bypass on systemic and whole kidney oxygenation, regional renal perfusion, creatinine clearance, and sodium reabsorption Variable
n
Conscious
Anesthetized
Systemic DO2, ml O2 kg–1min–1 Systemic VO2, ml O2 kg–1min–1 Systemic FEO2, % Renal DO2, ml O2 kg–1min–1 Renal VO2, ml O2 kg–1min–1 Renal FEO2, % Cortical perfusion, units Medullary perfusion, units CrCl, ml kg–1min–1 TNaþ, mmol kg–1min–1
12 12 12 12 5 5 12 11 12 12
NA NA 28.3 1.9 0.93 0.06 0.084 0.016 9.4 1.8 1955 379 720 128 2.12 0.26 298 37
NA NA 17.9 1.3a 0.54 0.04a 0.061 0.011 13.5 3.0 1385 364 670 144 2.18 0.39 306 55
On bypass
PTime
<0.001 <0.001 0.01 0.29 0.001 0.007 0.80 0.75
10.3 2.78 27.0 0.36 0.035 13.7 872 326 1.88 258
0.4 0.16 1.1b 0.07a,c 0.006d 3.4 157a 110c,d 0.43 59
NA, not available. a P # 0.001, dP # 0.01 for comparison with the conscious state (Tukey’s test). b P # 0.001, cP # 0.05 for comparison of bypass with the anesthetized state (Tukey’s test). Values are mean SEM. Oxygen delivery (DO2) was calculated as the product of arterial blood oxygen content (see Supplementary Table S1) and pump flow (systemic DO2) or renal blood flow (renal DO2). Systemic oxygen consumption (VO2) was calculated as the product of pump flow and the oxygen concentration difference between arterial blood and mixed venous blood. Renal VO2 was calculated as the product of renal blood flow and the oxygen concentration difference between arterial blood and renal venous blood. Fractional oxygen extraction (FEO2) was calculated as 100 (VO2/DO2). Creatinine clearance (CrCl) was calculated as the product of urine flow and urinary creatinine concentration divided by plasma creatinine concentration. Sodium reabsorption (TNaþ) was calculated as the product of CrCl and plasma sodium concentration minus sodium excretion. P values are the outcomes of repeated-measures analysis of variance for the main effect of “time.”
less (by 19 7 mm Hg, P ¼ 0.09 by Tukey’s test) than in conscious sheep. In contrast, cortical PO2 did not change significantly with transition to CPB. Part 2: altered pump flow Stability of the preparation.
Variables shown in Figures 2 and 3, Table 2, and Supplementary Tables S2 and S3 remained stable across the 3 periods at which pump flow was set at 80 ml kg–1min–1. Reduced pump flow. Systemic DO2 was reduced by 23% 1%, but systemic VO2 did not change significantly (Table 2). Systemic fractional oxygen extraction increased
and mixed venous SO2 decreased from 76.7% 1.3% to 67.5% 2.0% During perfusion at 60 ml kg–1min–1, RBF (–16% 2%) and RVC (–14% 2%) fell from their levels at 80 ml kg– 1 min–1, but neither cortical nor medullary perfusion changed significantly (Figure 2, Table 2, and Supplementary Table S2). Renal DO2 fell by 14% 3%, but neither renal VO2, renal venous SO2, nor the renal fractional extraction of oxygen changed significantly, despite reductions in creatinine clearance and sodium reabsorption (Table 2 and Supplementary Table S2). Nevertheless, both cortical (–21 5 mm Hg) and medullary (–11 3 mm Hg) PO2 fell.
Figure 2 | Effects of decreasing pump flow on cardiopulmonary bypass on renal hemodynamics and renal tissue oxygenation: Target arterial pressure was maintained at 70 mm Hg by administration of metaraminol. Black circles and error bars represent the mean SEM of each variable across 3 experimental periods (n ¼ 12 for mean arterial pressure and renal blood flow, but due to equipment failure n ¼ 10 for cortical PO2 and n ¼ 9 for medullary PO2). Data for individual sheep are shown by colored lines. **P # 0.01, ***P # 0.001 for comparison of the levels of variables during perfusion at 60 ml kg–1min–1 with their mean levels over the 2 control periods of perfusion at 80 ml kg–1min–1 that preceded and followed the intervention (Student’s paired t-test). 1340
Kidney International (2019) 95, 1338–1346
basic research
YR Lankadeva et al.: Renal oxygenation during cardiopulmonary bypass
Figure 3 | Effects of increasing pump flow on cardiopulmonary bypass on renal hemodynamics and renal tissue oxygenation: Target arterial pressure was maintained at 70 mm Hg by administration of metaraminol. Black circles and error bars represent the mean SEM of each variable across 3 experimental periods (n ¼ 12 for mean arterial pressure and renal blood flow, but due to equipment failure n ¼ 10 for cortical PO2 and n ¼ 10 for medullary PO2). Data for individual sheep are shown by colored lines. *P # 0.05, ***P # 0.001 for comparison of the levels of variables during perfusion at 100 ml kg–1min–1 with their mean levels over the 2 control periods of perfusion at 80 ml kg–1min–1 that preceded and followed the intervention (Student’s paired t-test).
Increased pump flow. This did not significantly alter MAP, body temperature, or arterial PO2, SO2, hemoglobin concentration, or oxygen content (Figure 3, Supplementary Table S3). Systemic DO2 increased by 21% 3%, but systemic VO2 did not change significantly (Table 2). Consequently, systemic fractional oxygen extraction decreased. Mixed venous SO2 increased from 77.0% 1.3% to 82.0% 1.5%. During perfusion at 100 ml kg–1min–1, RBF (þ17% 4%), RVC (þ19% 5%), renal DO2 (þ14% 6%), and medullary PO2 (þ9 3 mm Hg) were greater than at 80 ml kg–1min–1 (Figure 3, Table 2, and Supplementary Table S3). There were no significant changes in creatinine clearance, sodium reabsorption, cortical or medullary perfusion, or cortical PO2. Summary. Both systemic vascular conductance and RVC varied positively with pump flow (Figure 4). Thus, compared with levels at 60 ml kg–1min–1, when pump flow was 100 ml kg–1min–1 systemic DO2 was 59% 5% greater (P < 0.001, paired t-test), RBF was 47% 9% greater (P < 0.001), and renal DO2 was 40% 9% greater (P ¼ 0.006). The increased renal DO2 associated with increased pump flow was in turn associated with increased cortical and medullary PO2 (Figure 5a). Thus, compared with levels at 60 ml kg–1min–1, when pump flow was 100 ml kg–1min–1 cortical PO2 was 36.9 12.5 mm Hg greater (P ¼ 0.02) and medullary PO2 was 15.7 6.2 mm Hg greater (P ¼ 0.05). Part 3: progressively increased MAP
Metaraminol dose-dependently decreased systemic vascular conductance (Supplementary Table S4, Figure 4) so MAP was Kidney International (2019) 95, 1338–1346
dose-dependently increased, from its baseline level of 63 3 mm Hg to 90 3 mm Hg (0.2 mg/min), 107 3 mm Hg (0.4 mg/min), and 113 5 mm Hg (0.6 mg/min) (Figure 6). Systemic DO2 increased slightly (<10%) at 0.2 and 0.4 mg/ min because of slight increases in arterial hemoglobin concentration (Tables 3, Supplementary S4). There were no significant changes in systemic VO2 or systemic oxygen extraction. RVC was not significantly altered by metaraminol (Figure 4), so with the increased MAP, RBF and renal DO2 were dose-dependently increased (Figure 6, Table 3, and Supplementary Table S4). Renal VO2 was also increased by all doses of metaraminol, as were creatinine clearance and sodium reabsorption. No changes in cortical or medullary perfusion were detected. At 0.2 mg/min, metaraminol increased medullary PO2 (by 20.2 6.9 mm Hg) and cortical PO2 (by 39.0 10.1 mm Hg) (Figure 6). At 0.4 mg/min cortical PO2 (by 30.3 10.2 mm Hg), but not medullary PO2, was increased. At 0.6 mg/ min, neither cortical nor medullary PO2 differed significantly from control levels. Thus, cortical and medullary tissue PO2 were positively associated with renal DO2 at lower (0.2 mg/ min) but not higher (0.4 and 0.6 mg/min) doses of metaraminol (Figure 5b). DISCUSSION
In a clinically relevant large animal model of CPB we found hypoxia in the renal medulla, even under perfusion conditions that would be considered optimal in the clinical setting. Medullary hypoxia during CPB appeared to be driven by reduced medullary tissue oxygen delivery. The renal cortex 1341
basic research
YR Lankadeva et al.: Renal oxygenation during cardiopulmonary bypass
Table 2 | Effects of decreasing and increasing pump flow during cardiopulmonary bypass on systemic and whole kidney oxygenation, regional renal perfusion, creatinine clearance, and sodium reabsorption Decreasing pump flow
n
Systemic DO2, ml O2 kg–1min–1 Systemic VO2, ml O2 kg–1min–1 Systemic FEO2, % Renal DO2, ml O2 kg–1min–1 Renal VO2, ml O2 kg–1min–1 Renal FEO2, % Cortical perfusion, units Medullary perfusion, units CrCl, ml kg–1min–1 TNaþ , mmol kg–1min–1
12 12 12 12 5 5 12 11 12 12
Increasing pump plow
n
Systemic DO2, ml O2 kg–1min–1 Systemic VO2, ml O2 kg–1min–1 Systemic FEO2, % Renal DO2, ml O2 kg–1min–1 Renal VO2, ml O2 kg–1min–1 Renal FEO2, % Cortical perfusion, units Medullary perfusion, units CrCl, ml kg–1min–1 TNaþ , mmol kg–1min–1
12 12 12 12 5 5 12 11 12 12
80 ml kg–1min–1 10.1 2.68 26.5 0.35 0.041 16.4 824 376 1.44 198
0.3 0.15 1.2 0.07 0.007 4.1 147 121 0.31 44
80 ml kg–1min–1 10.8 2.77 25.8 0.42 0.052 16.7 883 306 1.72 236
0.4 0.16 1.3 0.08 0.007 2.6 179 129 0.39 53
60 ml kg–1min–1 8.0 2.78 34.7 0.32 0.051 21.3 631 321 0.86 120
0.2 0.19 1.9 0.06 0.001 3.9 96 160 0.20 29
100 ml kg–1min–1 12.7 2.60 20.5 0.45 0.054 16.9 991 364 1.91 265
0.4 0.20 1.3 0.08 0.002 1.9 216 116 0.57 79
80 ml kg–1min–1
P
<0.001 0.06 <0.001 <0.001 0.89 0.16 0.07 0.50 <0.001 <0.001
10.7 2.66 24.9 0.40 0.063 20.6 829 329 1.21 168
0.4 0.18 1.3 0.07 0.007 4.3 160 130 0.22 31
80 ml kg–1min–1
P
<0.001 0.55 <0.001 0.004 0.80 0.44 0.09 0.19 0.27 0.27
10.1 2.55 25.1 0.39 0.053 20.1 870 336 1.25 175
0.3 0.17 1.4 0.08 0.003 2.9 184 112 0.22 31
CrCl, creatinine clearance; Do2, oxygen delivery; FEo2, fractional oxygen extraction; TNaþ, sodium reabsorption; Vo2, oxygen consumption. Values are mean SEM. P values are the outcomes of Student’s paired t-test for comparison of variables during the period of perfusion at 60 (top) or 100 ml kg–1min–1 (bottom) against the mean level during the control periods that preceded and followed these interventions, during which sheep were perfused at 80 ml kg–1min–1.
appeared to be protected against hypoxia during CPB, despite similar ischemia to that of the medulla. This may be related to the large increase in tissue PO2 induced by hyperoxemia in the
Figure 4 | Relationships between systemic vascular conductance and renal vascular conductance during changes in pump flow and infusion of metaraminol: Symbols and error bars represent mean ± SEM of n [ 12 for altered pump flow and n [ 11 for metaraminol infusion. For altered pump flow the value for 80 ml kg–1min–1 is the average over the 3 experimental periods at which pump flow was maintained at this level during Part 2 of the experimental protocol (see Supplementary Figure S1). Note that pump flow was held constant at 80 ml kg–1min–1 during infusion of metaraminol. However, when pump flow was altered, a target arterial pressure of 70 mm Hg was achieved by administration of metaraminol (see Methods for details). 1342
cortex, which is largely absent in the medulla.16,17 Renal medullary oxygenation was improved, at least in a subset of animals, by increasing pump flow within the limits of clinical feasibility or by the use of the vasopressor metaraminol to increase MAP within a clinically feasible range. Increased medullary oxygenation, both in response to increased pump flow and increased MAP, was associated with increased renal DO2. Our findings provide a potential approach for improving renal medullary oxygenation during CPB and thus potentially decreasing postoperative AKI. They also provide a potential mechanistic basis for the recent clinical finding that avoidance of whole body DO2 < 280 ml kg–1m–2 on bypass reduced the risk of postoperative AKI.11 To our knowledge this is the first detailed analysis of the effects of CPB on renal oxygenation in a clinically relevant large animal model. Hypoxia in the renal medulla has previously been reported in a rat model of CPB7 and in a pilot study of 2 pigs.8 The ovine model recapitulates many aspects of CPB in human cardiothoracic surgery and was specifically designed to achieve what would be considered optimal perfusion conditions in clinical practice. Arterial oxygen content was optimized by adding w400 ml of fresh donor blood in the prime, so that arterial hemoglobin content was only 11.6% 3.1% less during CPB than in the conscious sheep. Furthermore, the FiO2 of 60% achieved an arterial PO2 of w300 mm Hg and an SO2 > 98% in all 12 animals during CPB. We also used a relatively high standard pump flow of 80 ml kg–1min–1. This equates in humans to a flow rate of 5.6 L/ min, or 3.16 L min–1m–2 for a 70-kg human of height 165 cm (body surface area calculated by the method of Du Bois and Kidney International (2019) 95, 1338–1346
YR Lankadeva et al.: Renal oxygenation during cardiopulmonary bypass
Figure 5 | Relationships between renal oxygen delivery and renal tissue PO2 during changes in (a) pump flow and (b) infusion of metaraminol: Symbols and error bars represent mean ± SEM of n [ 11 for cortical PO2 and n [ 9 for medullary PO2. For altered pump flow the value for 80 ml kg–1min–1 is the average over the 3 experimental periods at which pump flow was maintained at this level during Part 2 of the experimental protocol (see Supplementary Figure S1).
Du Bois18). This is greater than the commonly used flow rate in adult humans of w2.4 L min–1m–2.19 Sheep were also cooled by w3 C on bypass, in line with standard practice for coronary artery bypass graft surgery at our associated center.19 Consequently, mixed venous SO2 on CPB was w76%, within the target range for clinical perfusion.15 We also rendered sheep relatively hypercapnic on CPB to promote dilation of renal resistance arteries.20 Nevertheless, and despite the fact that the sheep we studied were young and otherwise healthy, marked renal medullary hypoxia was observed during CPB. Several mechanisms may contribute to medullary hypoxia during CPB. Our experimental design allowed us to separately evaluate the effects of anesthesia (which reduced MAP and RVC) and subsequent transition to CPB (which further reduced RVC). Thus, during CPB, RVC was reduced by w49% and MAP by w23 mm Hg compared with the conscious state. Consequently, RBF and renal DO2 were both reduced, by w61%, along with reductions in local perfusion in both the cortex and medulla. However, renal VO2 was also reduced during CPB (w53%) by a similar magnitude to the reduction Kidney International (2019) 95, 1338–1346
basic research
in renal DO2, despite no significant change in creatinine clearance or sodium reabsorption. Furthermore, in contrast to renal medullary PO2, cortical PO2 was well maintained during CPB, despite decreased cortical perfusion. Thus, the renal medulla is particularly susceptible to the development of hypoxia during CPB. Simulations from computational models of CPB in rats support the notion that this susceptibility arises, at least in part, from the anatomic position of the thick ascending limbs of the loop of Henle at the periphery of vascular bundles in the outer medulla. They are thus susceptible to even relatively small deficits in local oxygen supply in relation to their relatively high oxygen demand.5,6 Our findings indicate that renal medullary oxygenation partly depends on pump flow. When pump flow was reduced to 60 ml kg–1min–1 (equivalent to 2.37 L min–1 m–2 in adult humans, close to standard pump flow in clinical practice), the renal medulla was rendered markedly hypoxic (w14 mm Hg). However, when pump flow was increased to 100 ml kg–1min–1, renal medullary PO2 increased markedly to w29 mm Hg, albeit to levels still w17 mm Hg less than in the conscious sheep. Increased pump flow led to increased RVC, RBF, and renal DO2, providing at least a partial explanation for the improved medullary PO2. These data provide the first evidence of which we are aware that renal medullary oxygenation during CPB can be increased by increased pump flow. Our observations are consistent with those of Mackay and colleagues21 who found that increasing pump flow in a porcine model of CPB, under conditions in which MAP was also allowed to increase, increased RBF as measured by microspheres. Our current findings indicate that even in the absence of increased MAP, increased pump flow can increase renal oxygenation, including in the medulla. They are also consistent with clinical observations that low pump flow or low whole body DO2 on CPB is associated with greater incidence of postoperative AKI9,22,23 and that avoidance of low whole body DO2 during CPB can reduce the incidence of postoperative AKI.10,11 Medullary PO2 was also increased by low-dose infusion of metaraminol to increase MAP from w63 to w90 mm Hg. The renal circulation appears to be relatively insensitive to the vasoconstrictor effects of metaraminol, because RVC did not change significantly in the face of large decreases in systemic vascular conductance. Renal VO2 was also increased by all 3 doses of metaraminol, presumably because of the increase in glomerular filtration rate and thus tubular sodium reabsorption.24 Nevertheless, metaraminol infusion increased renal medullary PO2 during CPB. These data are also consistent with the observation that larger decreases in average MAP during CPB, relative to preoperative MAP, are associated with a greater incidence of AKI.23 They are also consistent with the inability to detect detrimental effects of high-dose metaraminol on renal function in patients with sepsis.25 However, it should be noted that a (albeit small) clinical trial of higher target MAP during CPB was unable to demonstrate reduced incidence of AKI.26 Thus, the potential for increased target MAP during CPB to reduce the risk of AKI remains to be determined. 1343
basic research
YR Lankadeva et al.: Renal oxygenation during cardiopulmonary bypass
Figure 6 | Effects of progressively increasing doses of metaraminol on renal hemodynamics and renal tissue oxygenation: Pump flow was maintained at 80 ml kg–1min–1 throughout this part of the experiment. Black circles and error bars represent the mean SEM of each variable across the 4 experimental periods (n ¼ 11 for mean arterial pressure and renal blood flow, but due to equipment failure n ¼ 10 for cortical PO2 and n ¼ 9 for medullary PO2). Data for individual sheep are shown by colored lines. *P # 0.05, **P # 0.01, ***P # 0.001 for comparison with the control period without metaraminol (Dunnett’s test).
Limitations of our study include the anesthetic used, isoflurane, which is now uncommonly used in human cardiac surgery. However, there is no evidence that more commonly used agents, such as sevoflurane or propofol/fentanyl, are more renoprotective than isoflurane.27 We acknowledge the limitations of laser Doppler flowmetry, which is an index of erythrocyte velocity in highly perfused tissues such as the kidney.28 We studied young adult sheep, so our findings are more relevant to young healthy patients than older patients with comorbidities. Nevertheless, we might expect the deleterious effects of CPB on medullary oxygenation to be worse in elderly patients with comorbidities than in young sheep. Finally, we were only able to collect renal venous blood samples in 5 of the 12 sheep because of loss of patency of the renal venous catheter. This reduced our ability to detect changes in renal VO2 and related parameters.
In conclusion, in a clinically relevant ovine model of CPB we were able to monitor whole kidney and intrarenal tissue perfusion and oxygenation. The renal medulla is susceptible to hypoxia during CPB, but medullary oxygenation can be improved by increasing pump flow or, more effectively, by increasing MAP with a vasopressor agent that does not decrease RVC. Both interventions are associated with increased renal DO2, thus ameliorating the deficit of a variable that appears to be a hallmark of CPB. METHODS Animal preparation All protocols were approved by the Animal Ethics Committee of the Florey Institute of Neuroscience and Mental Health under guidelines of the National Health and Medical Research Council of Australia. Merino ewes (35–48 kg body weight, n ¼ 12) were housed in
Table 3 | Effects of metaraminol on systemic and whole kidney oxygenation, regional renal perfusion, creatinine clearance, and sodium reabsorption Metaraminol Variable –1
–1
Systemic DO2, ml O2 kg min Systemic VO2, ml O2 kg–1min–1 Systemic FEO2, % Renal DO2, ml O2 kg–1min–1 Renal VO2, ml O2 kg–1min–1 Renal FEO2, % Cortical perfusion, units Medullary perfusion, units CrCl, ml kg–1min–1 TNaþ, mmol kg–1min–1
n
0 (control)
200 mg/min
400 mg/min
11 11 11 11 5 5 11 10 11 11
10.1 2.42 23.8 0.24 0.050 21.5 606 365 0.63 89
0.4 0.20 1.3 0.03 0.004 2.9 109 133 0.11 15
10.6 2.45 23.0 0.38 0.058 18.7 1141 439 1.51 212
a
0.3 0.20 1.4 0.06a 0.003b 3.9 310 123 0.32b 45a
10.7 2.63 24.4 0.43 0.076 21.5 1281 437 1.57 219
b
0.3 0.25 1.9 0.06a 0.008b 6.1 344 150 0.22c 31c
600 mg/min
PTime
0.19 0.32 0.49 0.005 0.02 0.31 0.10 0.64 0.001 0.002
10.5 2.47 23.9 0.43 0.064 16.9 1064 372 1.43 199
0.3 0.12 1.6 0.06a 0.004b 4.4 307 136 0.24a 34a
CrCl, creatinine clearance; Do2, oxygen delivery; FEo2, fractional oxygen extraction; TNaþ, sodium reabsorption; Vo2, oxygen consumption. Values are mean SEM. PTime is the outcome of repeated-measures analysis of variance. a P # 0.01, bP # 0.05, cP # 0.001 for comparison against control values before infusion of metaraminol (Dunnett’s test).
1344
Kidney International (2019) 95, 1338–1346
basic research
YR Lankadeva et al.: Renal oxygenation during cardiopulmonary bypass
individual metabolic cages with free access to water and 800 g of oaten chaff daily. Sheep underwent preliminary aseptic surgery under general anesthesia, induced with 15 mg/kg sodium thiopental (Jurox, Rutherford, Australia) and, after intubation, maintained on 2.0% to 2.5% isoflurane (Isoflo; Zoetis, Rhodes, Australia). A transit-time flow probe (4 mm) was placed around the left renal artery, the renal vein was cannulated, and fiber-optic probes (CP-026-001; Oxford Optronix Ltd, Abingdon, UK) were inserted in the renal cortex and medulla to measure tissue PO2 by fluorescence lifetime oximetry, tissue perfusion by laser-Doppler flux, and tissue temperature.14 The carotid artery and jugular vein were cannulated. A bladder catheter (Foley size 12, 30 ml; Euromedical, Malaysia) was inserted for collection of urine. Sheep were given 50 mg flunixin meglumine (Flunixon; Norbrook, Tullarmarine, Australia) for analgesia and 900 mg procaine penicillin (Ilium; Troy Laboratories, Glendinning, Australia) during surgery and then 24 and 48 hours postoperatively and allowed 5 days of recovery before experimentation. Analog signals were digitized as previously described.14 A maintenance infusion of compound sodium lactate (1 ml kg–1h–1; Baxter, Old Toongabbie, NSW, Australia) was given i.v. from 6 pm the night before the experiment.
MAP and RBF were stable a dose–response relationship for metaraminol infusion was established (0, 0.2, 0.4, and 0.6 mg/min). At each dose a 20-minute experimental period began once MAP and RBF were stable. Upon completion of Part 3, metaraminol was withdrawn and sheep were killed with sodium pentobarbital (20 mg/kg i.v., Lethobarb; Virbac, Wetherill Park, Australia). Five minutes later the perfusion pump was turned off to confirm that renal perfusion and PO2 were close to zero. The left kidney was removed, and the positions of the fiber-optic probes confirmed. Statistical analysis Data are expressed as mean SEM. After 1-way repeated-measures analysis of variance, post hoc comparisons29 were made by Student’s paired t-test, Tukey’s test, or Dunnett’s test. Analyses were performed using SYSTAT (version 13; Systat, San Jose, CA) or GraphPad Prism (GraphPad Software, La Jolla, CA). Two-sided P # 0.05 was considered statistically significant. DISCLOSURE All the authors declared no competing interests.
ACKNOWLEDGMENTS
Experimental protocol The experimental protocol had 3 parts (Supplementary Figure S1), each comprising a series of 20- to 30-minute experimental periods. Urine was collected throughout each experimental period. At the midpoint of each experimental period arterial, mixed venous, and renal venous blood samples were collected for oximetry and blood chemistry (ABL Systems 625, Radiometer, Copenhagen, Denmark). Arterial plasma and urine samples were stored at –80 C for later measurement of creatinine and sodium. Part 1: anesthesia and transition to CPB. For the first experimental period (30 minutes) the animal was conscious and unrestrained in its home cage. Anesthesia was then induced (as above) with a FiO2 of 60%. Once anesthesia was established, an infusion of 10 ml/kg of compound sodium lactate solution was administered over a 20-minute period. After this, sheep received a maintenance infusion of compound sodium lactate solution at 2 ml kg–1h–1 throughout the experiment. Once MAP and RBF were stable, the second experimental period began (stable anesthesia). CPB was then established (details in Supplementary Methods) at a flow rate of 80 ml kg–1min–1. If required, metaraminol (Metaraminol Montrose 10 mg/ml; Montrose Life Sciences, NSW, Australia) was given i.v. to maintain MAP at 65 to 75 mm Hg, with a target of 70 mm Hg. Once MAP and RBF were stable, the third experimental period began (on bypass). Part 2: altered pump flow. A target MAP of 70 mm Hg was maintained with titration of metaraminol. Pump flow rate was randomized to 60 or 100 ml kg–1min–1, returning to 80 ml kg–1min–1 after either intervention. The alternate pump flow rate was then used. Thus, each period of altered pump flow (60 or 100 ml kg–1min–1) was preceded and followed by a period at 80 ml kg–1min–1 (Supplementary Figure S1). At each pump flow, once MAP and RBF were stable, a 20-minute experimental period began. Part 3: progressively increased MAP. Metaraminol support from the previous experimental period was withdrawn, and once Kidney International (2019) 95, 1338–1346
We thank Tom Vale and Tony Dornom for technical assistance and Libin Jose and Ilaria Catteneo for perfusion services. Support from the National Health and Medical Research Council of Australia (GNT1122455), the Victorian Government Operational Infrastructure Support Grant, and the National Heart Foundation of Australia (101853) is gratefully acknowledged. SUPPLEMENTARY MATERIAL Supplementary Methods. Table S1. Effects of anesthesia and cardiopulmonary bypass on systemic and renal hemodynamics, blood oximetry and chemistry, and renal function. Table S2. Effects of reduced pump flow during cardiopulmonary bypass on systemic and renal hemodynamics, blood oximetry and chemistry, and renal function. Table S3. Effects of increasing pump flow during cardiopulmonary bypass on systemic and renal hemodynamics, blood oximetry and chemistry, and renal function. Table S4. Effects of metaraminol on systemic and renal hemodynamics, blood oximetry and chemistry, and renal function during cardiopulmonary bypass. Figure S1. Experimental timeline. The experiment comprised 3 parts. In Part 1, sheep were transitioned from the conscious state to stable anesthesia to cardiopulmonary bypass. In Part 2, pump flow was varied while mean arterial pressure was titrated to a target of 70 mm Hg by i.v. administration of metaraminol. Note that each period of low (60 ml kg–1min–1) or high (100 ml kg–1min–1) pump flow was preceded and followed by a period of perfusion at 80 ml kg–1min–1. In Part 3, pump flow was maintained at 80 ml kg–1min–1 during infusion of progressively greater doses of metaraminol (0, 0.2, 0.4, and 0.6 mg/min). Each experimental period commenced once arterial pressure and renal blood flow were stable (10–20 minutes after the change to conditions) and was of 20 to 30 minutes’ duration. Urine was collected during the experimental period, and a blood sample was taken at the mid-point of each experimental period, as shown by the arrows. See the text for further details. Supplementary material is linked to the online version of the paper at www.kidney-international.org. 1345
basic research
REFERENCES 1. Hu J, Chen R, Liu S, et al. Global incidence and outcomes of adult patients with acute kidney injury after cardiac surgery: a systematic review and meta-analysis. J Cardiothorac Vasc Anesth. 2016;30:82–89. 2. Evans RG, Lankadeva YR, Cochrane AD, et al. Renal haemodynamics and oxygenation during and after cardiac surgery and cardiopulmonary bypass. Acta Physiol. 2018;222:e12995. 3. Ow CPC, Ngo JP, Ullah MM, et al. Renal hypoxia in kidney disease: cause or consequence? Acta Physiol. 2018;222:e12999. 4. Post EH, Kellum JA, Bellomo R, et al. Renal perfusion in sepsis: from macro- to microcirculation. Kidney Int. 2017;91:45–60. 5. 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. 6. 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. 7. 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. 8. Stafford-Smith M, Grocott HP. Renal medullary hypoxia during experimental cardiopulmonary bypass: a pilot study. Perfusion. 2005;20: 53–58. 9. 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. 10. Magruder JT, Crawford TC, Harness HL, et al. A pilot goal-directed perfusion initiative is associated with less acute kidney injury after cardiac surgery. J Thorac Cardiovasc Surg. 2017;153:118–125. 11. Ranucci M, Johnson I, Willcox T, et al. Goal-directed perfusion to reduce acute kidney injury: a randomized trial. J Thorac Cardiovasc Surg. 2018;156:1918–1927. 12. Lannemyr L, Bragadottir G, Hjarpe A, et al. Impact of cardiopulmonary bypass flow on renal oxygenation in patients undergoing cardiac surgery. Ann Thorac Surg. 2019;107:505–511. 13. 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. 14. Calzavacca P, Evans RG, Bailey M, et al. Long-term measurement of renal cortical and medullary tissue oxygenation and perfusion in unanesthetized sheep. Am J Physiol Regul Integr Comp Physiol. 2015;308: R832–R839.
1346
YR Lankadeva et al.: Renal oxygenation during cardiopulmonary bypass
15. Svenmarker S, Haggmark S, Ostman M, et al. Central venous oxygen saturation during cardiopulmonary bypass predicts 3-year survival. Interactive Cardiovasc Thorac Surg. 2013;16:21–26. 16. Aperia AC. The influence of arterial PO2 on renal tissue PO2. Acta Physiol Scand. 1969;75:353–359. 17. Leong CL, Anderson WP, O’Connor PM, et al. Evidence that renal arterialvenous oxygen shunting contributes to dynamic regulation of renal oxygenation. Am J Physiol Renal Physiol. 2007;292:F1726–F1733. 18. Du Bois D, Du Bois EF. A formula to estimate the approximate surface area if height and weight be known. Arch Intern Med. 1916;17:863–871. 19. Zhu MZL, Martin A, Cochrane AD, et al. Urinary hypoxia: an intraoperative marker of risk of cardiac surgery-associated acute kidney injury. Nephol Dial Transplant. 2018;33:2191–2201. 20. Daugherty RM Jr, Scott JB, Dabney JM, et al. Local effects of O2 and CO2 on limb, renal, and coronary vascular resistances. Am J Physiol. 1967;213: 1102–1110. 21. Mackay JH, Feerick AE, Woodson LC, et al. Increasing organ blood flow during cardiopulmonary bypass in pigs: comparison of dopamine and perfusion pressure. Crit Care Med. 1995;23:1090–1098. 22. Fischer UM, Weissenberger WK, Warters RD, et al. Impact of cardiopulmonary bypass management on postcardiac surgery renal function. Perfusion. 2002;17:401–406. 23. Kanji HD, Schulze CJ, Hervas-Malo M, et al. Difference between preoperative and cardiopulmonary bypass mean arterial pressure is independently associated with early cardiac surgery-associated acute kidney injury. J Cardiothorac Surg. 2010;5:71. 24. Evans RG, Harrop GK, Ngo JP, et al. Basal renal oxygen consumption and the efficiency of oxygen utilization for sodium reabsorption. Am J Physiol Renal Physiol. 2014;306:F551–F560. 25. Hou LC, Li SZ, Xiong LZ, et al. Effect of dopamine and metaraminol on the renal function of patients with septic shock. Chin Med J. 2007;120:680–683. 26. Urzua J, Troncoso S, Bugedo G, et al. Renal function and cardiopulmonary bypass: effect of perfusion pressure. J Cardiothorac Vasc Anesth. 1992;6:299–303. 27. Tan SI, Brewster DJ, Horrigan D, et al. Pharmacological and non-surgical renal protective strategies for cardiac surgery patients undergoing cardiopulmonary bypass: a systematic review [e-pub ahead of print]. ANZ J Surg. https://doi.org/10.1111/ans.14800. 28. Eppel GA, Bergstrom G, Anderson WP, et al. Autoregulation of renal medullary blood flow in rabbits. Am J Physiol Regul Integr Comp Physiol. 2003;284:R233–R244. 29. Ludbrook J. Multiple comparison procedures updated. Clin Exp Pharmacol Physiol. 1998;25:1032–1037.
Kidney International (2019) 95, 1338–1346