Sodium bicarbonate to prevent cardiac surgery-associated kidney injury: the end of a dream?

The rationale of urine alkalinization through intravenous sodium bicarbonate to prevent cardiac surgery-associated acute kidney injury relies on several pathophysiological arguments. Urine alkalinization is easily feasible in the ICU setting and is often considered to be associated with few side effects. In a previous issue of Critical Care, a retrospective study evaluates the effect of routine intravenous bicarbonate use to prevent cardiac surgery-associated acute kidney injury with cardiopulmonary bypass. This commentary discusses recent data on the use of bicarbonate to prevent cardiac surgery-associated acute kidney injury.

In a previous issue of Critical Care, Heringlake and colleagues [1] report important results of a cohort analysis of use of bicarbonate to prevent cardiac surgeryassociated acute kidney injury (CSA-AKI). Acute kidney injury (AKI) is common in hospitalized patients and is associated with increased morbidity and mortality. Cardiac surgery with cardiopulmonary bypass (CPB) is the second leading cause of AKI in the ICU, just after sepsis [2]. CSA-AKI incidence varies widely in the literature, usually in between 3% and 50% of patients undergoing CPB [3,4]. Its occurrence depends on several procedure-related (type of cardiac surgery, duration of aortic cross-clamping, CPB time) and patient-dependent factors (age, chronic kidney disease, low left ventricular ejection fraction, peripheral vascular disease, diabetes). Th e incidence of CSA-AKI is only 2.3% in isolated coronary artery bypass graft procedures in patients without comorbidities but its incidence rises up to 15% in patients also with chronic kidney disease [5]. Due to its signifi cant incidence and infl uence on resource utilization and potentially on mortality, the prevention and/or treatment of CSA-AKI is a relevant issue.
Pathophysiology of CPB-AKI is complex, mainly relying on hemodynamic and infl ammatory disturbances [3,6]. CPB is associated with ischemia-reperfusion injury but also with signifi cant release of free hemoglobin and iron. Nevertheless, the respective importance of the diff erent pathophysiological processes in CSA-AKI remains unclear. In addition, nephrotoxic drugs, anemia and blood transfusions may further exacerbate CSA-AKI.
Experimental data have shown that higher tubular pH could be protective in the presence of hemoglobinuria or myoglobinuria [7], especially through inhibition of hydroxyl radical generation and lipid peroxidation, which could be central in AKI [8][9][10]. In analogy with the benefi cial eff ects of urine alkalinization after rhabdomyolysis [11], urine alkalinization after intravenous bicarbonate was thought to prevent CSA-AKI. In a pilot randomized study including 100 patients submitted to cardiac surgery with CPB, perioperative bicarbonate administration decreased the proportion of patients experiencing an increase in creatinine from 52 to 32% (P = 0.043) [12]. In addition, bicarbonate blunted the increase in biomarkers of AKI.
However, bicarbonate may not be the right agent to prevent CSA-AKI. Indeed, bicarbonate failed to prevent experimental ischemia-induced AKI [13]. It also fails to prevent contrast-induced nephropathy [14,15], another condition where ischemia is thought to occur. Hence, confi rmatory trials in CSA-AKI are warranted.
In this issue, Heringlake and colleagues [1] reported the absence of superiority of intravenous bicarbonate for the prevention of CSA-AKI in comparison to saline. Th e authors retrospectively compared a cohort of 280 patients who received 4 mmol sodium bicarbonate per kilogram to a control cohort. Th e proportion of patients experiencing an increase in creatinine by 25% was similar in both groups (41.1% in bicarbonate versus 32.9% in control, P = not signifi cant). Strikingly, this study is in contrast to the pilot study by Haase and colleagues [12], even though intravenous bicarbonate was used in a

Abstract
The rationale of urine alkalinization through intravenous sodium bicarbonate to prevent cardiac surgery-associated acute kidney injury relies on several pathophysiological arguments. Urine alkalinization is easily feasible in the ICU setting and is often considered to be associated with few side eff ects. In a previous issue of Critical Care, a retrospective study evaluates the eff ect of routine intravenous bicarbonate use to prevent cardiac surgery-associated acute kidney injury with cardiopulmonary bypass. This commentary discusses recent data on the use of bicarbonate to prevent cardiac surgery-associated acute kidney injury. comparable manner and similarly alkalinized blood pH in both studies.
Several factors could explain the diff erences between these two studies. First, the Haase study included prospectively selected patients with signifi cant risk for CSA-AKI while the Heringlake study was retrospective and included nearly all patients undergoing cardiacsurgery with CPB in their center. As mentioned above, heterogeneity in patient populations is an important factor determining the risk for CSA-AKI. Th is may be explained by inclusion in the latter trial of 38% of patients undergoing isolated coronary artery bypass graft, which shows the lowest CSA-AKI risk. Th is limits of course the chances to see a benefi cial impact of bicarbonate. Nevertheless the incidence of CSA-AKI was still close to 33%, which is not negligible. Second, the Heringlake trial had the advantage of including almost all patients admitted to their center, which increases external validity of the results. Interestingly, bicarbonate may also be associated with detrimental eff ects. Mean arterial blood pressure was lower after induction of anesthesia, and patients required more fl uids and vasopressors and spent more time in a high dependency unit [1]. Previous work has already drawn attention to the side eff ects of bi carbonate: overcorrection of acidosis signifi cantly increased mor tality [16] and several experiments have reported that intracellular alkalinization hastened cell death after anoxia, stimulated superoxide formation, enhanced proinfl ammatory cytokine release and apoptosis, and increased blood lactate and ketone bodies [17]. Th erefore, use of bicarbonate for CSA-AKI prevention could be a double-edged sword where risk may overwhelm benefi ts if applied in a non-selected patient population.
Th ese contradictory data raise questions about bicarbonate effi ciency for CSA-AKI prevention. From the data published, it seems that bicarbonate is useless in an unselected population [1]. Bicarbonate may still be eff ective in patients at high risk of CSA-AKI and results of confi rmatory trials are needed. Nevertheless, preliminary results of a randomized trial in 427 patients at high risk of CSA-AKI reported a similar incidence of CSA-AKI of 45% in the bicarbonate group and 44% in the saline group [18]. Th e debate is therefore still open but data do not actually support routine use of bicarbonate for CSA-AKI prevention.