A Comparison of Renal vs Limb NIRS for Predicting Cardiac Surgery Associated Acute Kidney Injury

Acute kidney injury (AKI) is a common complication of cardiac surgery occurring in up to 40% of patients.1, 2 This is more than 10 times the rate of other serious post-surgical complications in cardiac surgery such as stroke or deep sternal infection and is even greater than the incidence of prolonged ventilation.3, 4 Severe AKI requiring new onset dialysis after cardiac surgery occurs in 1-6% of cardiac surgery patients.3, 5 AKI increases length of hospital stay from 5 to 11 days and increases hospital costs from $18,463 to $37,674 per patient.6 The severity of AKI is strongly associated with in-hospital mortality with even the mildest forms of AKI having more than double the mortality and the more severe forms having 10-30 times higher mortality rates (depending on the definition of severe AKI and the patient population).6, 7 AKI contributes to 300,000 annual deaths in the U.S. and 2 million deaths globally each year.8, 9 Considering only the most relevant patient populations (major aortic and spinal surgery, relevant cardiovascular procedures, intensive care patients on mechanical ventilation, severe trauma patients), in the US alone more than 3.1 million patients per year are at risk of AKI.10-14 Of those, 30.6%12, 13, 15 will suffer AKI - adding annual hospital costs of $19 billion.6 It is for these reasons that the Society of Thoracic Surgeons lists renal failure as one of the 5 major complications of cardiac surgery.

The pathophysiology of AKI after cardiac surgery is multifactorial but the profound hemodynamic changes that occur in cardiac surgery likely play an important role. The renal medulla has the highest metabolic activity and least oxygen reserve in the kidney and medullary hypoxia is thought to be a major determinant of AKI.16, 17 One major limitation in the effort to reduce the incidence of AKI is the lack of a real-time monitor of renal hypoxia. As there is no current therapy for AKI once the disease has occurred, prevention is the mainstay of treatment and an important step in preventing AKI is its timely recognition. Current diagnostic methods based on serum creatinine, oliguria, or even novel biomarkers, however, take hours to days to become diagnostic. The investigators have no accepted standard for diagnosing kidney hypoxia in real-time before the injury becomes irreversible. Most clinicians base intraoperative hemodynamic goals for the kidney on "educated guesses" of adequate renal perfusion, since the exact cardiac output or mean arterial pressure targets for preventing AKI are unknown. Even the practice of monitoring urine output is uncertain because of its non-specific relationship to kidney function.18-20 Near infrared spectroscopy (NIRS) is a promising technology for monitoring tissue hypoxia because it is noninvasive and has a long history of clinical use as a monitor of cerebral oximetry in both cardiac and vascular surgery. Because NIRS measures a mixture of arterial and venous capillary oxygenation, it is particularly well-suited to monitor the balance between oxygen consumption and delivery. A large body of literature describes the use of NIRS both as a systemic perfusion monitor and a regional monitor of kidney perfusion in neonates and small children.21-26 In adult cardiac surgery patients, two studies using ultrasound guided placement of NIRS sensors over the kidney found that intraoperative regional tissue oxygen desaturations were associated with the subsequent development of AKI.27, 28 A third study found a correlation between NIRS oxygen saturation measurements over the kidney and renal venous oxygen saturation in adult cardiac surgery patients.29 The problem with these studies, however, is that they excluded patients for body mass index (BMI) greater than 30kg/m2 or if the surface of their kidney was greater than 4cm from the skin. According to the Center for Disease Control, the incidence of obesity (BMI greater than 30kg/m2) is over 40% in the United States and obese patients having cardiac surgery are at increased risk for AKI.30, 31 Thus, if NIRS placed over the kidney in adult patients is to be an effective tool in monitoring kidney hypoxia in adults in the United States, then the technology must be effective in all adult patients.

The major limitation to the use of NIRS for renal hypoxia monitoring in adults is the depth of penetration of the signal. Manufacturers guidelines suggest that the maximum depth of penetration of NIRS is 2.5 cm. The actual depth of penetration of NIRS measurements depends on the source-detector separation, the specific algorithm used, and the optical properties of the tissue being measured. When NIRS is modeled experimentally, there is a nonlinear relationship between source-detector separation and depth of penetration such that a source-detector separation of 20 mm resulted in a depth of penetration of 10 mm but a source-detector separation of 40mm resulted in a depth of penetration of 15mm.32 While in very thin adults, the surface of the kidney may be within 2 cm from the skin, the renal medulla (where most AKI is thought to occur) is several centimeters deeper and thus unlikely to be monitored accurately.

The question then arises as to how these adult cardiac surgery studies showed a relationship between NIRS placed over the kidney, AKI outcome, and even renal venous oxygen saturation.27-29 The lab developed a porcine model in which the investigators directly compared kidney medullary oxygen concentration to NIRS measurements placed over the kidney and the thigh during hemorrhagic shock and aortic occlusion.33 The investigators found during extreme hemodynamic changes such as aortic occlusion, there was a moderate correlation between NIRS and invasive kidney tissue oxygen concentration, but this was true irrespective of whether the NIRS sensors were placed over the kidney or over the thigh. The relationship between kidney NIRS and thigh NIRS was stronger than that with kidney oxygen concentration. This was in animals with kidney surface to skin distances of less than 2 cm. Taken together, these data suggest that renal NIRS in adults may be measuring subcutaneous tissue oxygenation but that it is still a global perfusion monitor that may predict poor global perfusion and AKI risk. Indeed, a study in adult cardiac surgery patients showed that NIRS sensors placed over the thigh predicted subsequent AKI with receiver operator characteristic analysis suggesting an optimal cut-off of 67% and an area under the curve of 0.84.34 The purpose of the current study proposal is to directly compare NIRS placed over the kidney to NIRS placed over the arm or the thigh in adult patients having cardiac surgery at risk for AKI. The hypothesis is that limb NIRS and kidney NIRS will both predict AKI and that the ability of NIRS to predict AKI in adult cardiac surgery patients will be independent of BMI and kidney surface to skin distance.