Noninvasive Measurements of Renal Perfusion During Cardiac Surgery

Acute kidney injury (AKI) is an unfortunately common complication of cardiac surgery that occurs in up to 40% of patients and results in increased mortality, prolonged ICU and hospital length of stay. AKI after cardiac surgery is not a benign complication, increasing mortality from 1% to 19% in those with AKI and to 63% mortality in those who required renal replacement therapy (RRT). In another study, patients with AKI with RRT after cardiac surgery had 39 times the mortality as patients without AKI (95% CI 32-48).[6] AKI has also been associated with increased morbidity and a larger number of patients requiring discharge to an extended care facility.

The pathophysiology of AKI after cardiac surgery is multifactorial. Patients undergoing cardiac surgery are likely more susceptible to AKI given their tendency to be older patients with multiple co-morbidities including pre-existing chronic kidney disease, cardiac dysfunction, diabetes mellitus, and advanced age. Cardiac surgery patients are also exposed to multiple nephrotoxic agents in the peri-operative period including radiocontrast dyes used for angiography, angiotensin converting enzyme inhibitors, and diuretics. Hemodynamically unstable patients may be subject to low cardiac output and low systemic blood pressure because of the pathological condition that brings them to the cardiac operating room in the first place. Low cardiac output and systemic blood pressure can result in reduced renal perfusion. Induction and the maintenance of anesthesia can result in myocardial depression and hemodynamic instability, reducing renal perfusion even further. In most cases, cardiac surgery is facilitated by cardiopulmonary bypass (CPB) which is the process of placing the patient on a heart and lung machine that pumps, oxygenates, and removes CO2 from the patient's blood for them while the heart is arrested. The CPB circuit itself is known to cause significant inflammation and hemodynamic changes that may cause renal injury particularly with prolonged CPB times. Vasopressor and inotropic agents such as vasopressin, norepinephrine, milrinone, and epinephrine are often used to maintain blood pressure and cardiac output. While some of these agents may increase systemic blood pressure by increasing the systemic vascular resistance, this may in fact result in a decrease in renal perfusion. The affect of these agents on the incidence of AKI is uncertain.

Traditionally, the diagnosis of AKI is based on either a sustained fall in urine output or a rise in serum creatinine. The 2012 Kidney Disease Improving Global Outcomes (KDIGO) classification defines AKI as an increase in serum creatinine by 0.3 mg/dl or more in 48 hours or a rise to at least 1.5 times baseline. Three stages of AKI are then defined based on increasing values of serum creatinine or duration of decreased urine output.

The major limitation of creatinine and urine output as a marker of kidney function is the time lag between injury and diagnosis. It often takes 24-36 hours after renal injury for serum creatinine levels to rise. Peri-operative urine output is affected by volume status, anesthetic drugs, and the use of diuretics and AKI is not diagnosed until oliguria has occurred for at least 6-12 hrs. This renders serum creatinine and urinary output measurements insensitive to acute changes in renal function and relatively useless in the acute diagnosis of AKI during and after cardiac surgery.

More recently, several early biomarkers have been developed to identify patients whom are at risk for developing AKI. Two of these biomarkers, TIMP-2 and IGFBP7, have been used for the early prediction of AKI in ICU and cardiac surgery patients. Nonetheless, "early detection" with these biomarkers is till limited to 3-4 hours (an in some studies 24 hours) after renal injury.

One of major limitation in the efforts to reduce the incidence of AKI in cardiac surgery is the lack of a real-time monitor of renal perfusion. As mentioned above, urine output is well known to be a poor indicator of renal perfusion. While urinary flow rate may be linearly related to blood pressure while on CPB, this is likely related to a phenomenon called pressure diuresis and is unlikely to be a reflection of improved renal perfusion. Renal blood flow can be measured by cannulating the renal vein through a central venous catheter placed in the femoral vein. This, however, is a highly invasive technique and is not used routinely.

As a result of the lack of real-time monitoring of the kidney during cardiac surgery, anesthesiologists are often left to make educated guesses as to what blood pressure and cardiac output are adequate for renal perfusion based on the patients baseline blood pressure and kidney function. In a patient with a long history of hypertension and/or chronic kidney disease the anesthesiologist's goal is often to try to maintain a higher mean arterial pressure (MAP) both on and off CPB than normal in order to "improve renal perfusion". There is very little data to support this practice, particularly if we need to use vasopressor agents to achieve these higher MAP goals, and the exact target MAP in these patients is unknown.

This lack of real-time monitoring of renal perfusion is in stark contrast to the vigilant monitoring of the brain during cardiac surgery. Cerebral oxymeters that are routinely used to measure brain oxygen saturations, transcranial Doppler systems to measure cerebral blood flow, and EEG to measure brain activity. Often the argument is made to use the brain as an index organ for adequate perfusion to the rest of the body, but during periods of hemodynamic instability, brain perfusion is preserved at the cost of other organ. The lower limit of autoregulation of the brain (the BP below which blood flow becomes dependent on blood pressure) is thought to be a BP of 50-55 mmHg. [20, 21] Although the exact number in humans has not been determined and adequate BP for kidney perfusion is widely debated, the lower limit of autoregulation for the kidney is probably significantly higher than the brain. Brain desaturations, therefore, may be specific for poor perfusion to other organs such as the kidney and gut but they are likely not sensitive for these changes.

Adequate renal perfusion is much more complicated than simply measuring renal blood flow or renal venous oxygenation. Grossly the kidney has three anatomical areas: the renal cortex where most of the filtration is done, the renal medulla where urine is actively concentrated, and the renal pelvis where urine is collected. Approximately ¼ of our cardiac output goes to our kidneys and most of that perfuses the cortex for filtration. In contrast, the medulla receives only a limited blood supply. This combined with the high metabolic activity of the renal medulla results in a relatively hypoxic medullary environment with a normal pO2 of 10-20 mmHg and very little oxygen reserve. Medullary hypoxia may be a consequence of decreased oxygen delivery or increased oxygen consumption and is a major determinant of AKI and chronic kidney disease. The relatively hypoxic environment of the renal medulla and it's role in renal injury suggest that global measures of systemic venous oxygenation through a central venous catheter or even renal venous oxygenation through an invasive renal vein catheter may be poor monitors of adequate renal perfusion. The more ideal monitor of renal hypoxia and kidney injury would be a measure of medullary oxygenation.

Due to the physical proximity of the vasa recta in the renal medulla with the urinary collecting ducts, medullary oxygen tension is more closely related to urinary oxygen tension than renal venous oxygenation. Medullary oxygen tension has been measured invasively in animal studies and correlates with both renal pelvic urine and bladder urine oxygenation. In pigs, bladder urinary oxygenation was shown to decrease with increasing degrees of hypoxemia and hemorrhage, then restored with resuscitation. These changes were found to proceed other global hemodynamic changes and increases in base deficit or lactic acidosis. In a sheep model of sepsis, both medullary and urinary oxygen was decreased and restoration of systemic blood pressure with norepinephrine further reduced oxygen tension in both the renal medulla and urine. Both medullary and pelvic urinary oxygen tension were found to decline significantly with the onset of CPB in pigs, gradually increasing after cessation of bypass but remaining lower than pre-CPB levels suggesting that the hemodynamics of CPB may be a significant contributor to the development of AKI in cardiac surgery.

In 1996 Kainuma et al. placed an oxygen electrode inside the urinary catheter of 96 patients undergoing cardiac surgery with CPB. In their set up, there was 20 ml of deadspace between the tip of the catheter and the oxymeter. Samples were drawn from a stop-cock near the oxymeter every two hours for calibration. They found marked decreases in urine oxygen tension in all patients during CPB. Some patients recovered their urine oxygenation after CPB, but in 34% of patients, urinary oxygen tension continued to decline after bypass and these patients had significantly higher incidence of AKI. They concluded that urinary oxygen monitoring may be superior to other more invasive measures of renal perfusion, but to date, no further work has been published on urine oxygen monitoring in humans during cardiac surgery

Fiber-optic technology has been use in endoscopy for cancer detection. Continuous blood gas measurements has also been performed in patients undergoing cardiopulmonary bypass by implanting fiber optic oxygen sensing probes in the internal jugular vein through a central line catheter or in the radial artery though a radial artery catheter. Continuous blood gas measurements has also been performed in patients undergoing cardiopulmonary bypass by implanting fiber optic oxygen sensing probes in the internal jugular vein through a central line catheter or in the radial artery though a radial artery catheter.

More recently Evans et al have placed a fiber optic oxygen sensing probe in the tip of a urinary catheter in 35 patients undergoing cardiac surgery with cardiopulmonary bypass. Their study design was very similar to ours in that they placed a fiber-optic oxygen sensing probe in the tip of a urinary catheter. The difference in our design is that we have created a flow through chamber with oxygen sensor in it as well.

The later, we hope will correlate closely with that in the tip of the urinary catheter but will be a less invasive approach to urine oxygen monitoring. In our study, we plan to place two fiber optic oxygen tension probes in a standard urinary catheter. The first probe will be placed within the urinary catheter and threaded to the tip of the catheter (though still within the catheter and not in the body) in order to measure bladder urine oxygen tension. The second will be a flow through chamber with a fiber optic oxygen sensor and urinary flow rate sensor within it. This flow through chamber will be placed between the urinary catheter and the collecting tubing (see figure 1, attached under "other documents"). A one-way valve in the flow through chamber will allow passage of urine but prevent back flow of either urine or air.

The hypothesis is that a the less invasive flow through chamber oxymeter will provide similar measurements to the more proximately placed fiber at the tip of the urinary catheter, that both will provide reliable measurement of urine oxygenation, and that these measurements will predict post-operative acute kidney injury in cardiac surgery patients.