Effect of perioperative sodium bicarbonate administration on renal function following cardiac surgery for infective endocarditis: a randomized, placebo-controlled trial

For the treatment of infective endocarditis (IE), recent evidence favors early surgical intervention to reduce the risk of stroke and death [1, 2]. Surgery with cardiopulmonary bypass (CPB), however, is inevitably accompanied by numerous factors that contribute to acute kidney injury (AKI) [3]. An early, exaggerated inflammatory response with oxidative stress and the preoperative use of nephrotoxic antibiotics further predispose patients with IE to renal impairment [4‐6]. Indeed, AKI affects up to 60% of patients with IE undergoing cardiac surgery [6], and the occurrence and severity of AKI have been indisputably associated with mortality and poor outcomes [7].

Sodium bicarbonate has been demonstrated to protect the kidneys through urinary alkalization and increasing the tubular pH [8, 9]. It reduces the pH-dependent conversion of hemoglobin to methemoglobin and production of free radicals via the Haber-Weiss reaction [10, 11]. After a pilot study reported that with sodium bicarbonate infusion there was a reduced incidence of AKI after cardiac surgery [8], numerous clinical trials evaluated its renoprotective efficacy in different types of surgery and patients, yielding conflicting results [12‐15]. Considering the massive oxidative stress associated with systemic inflammation and prolonged use of nephrotoxic antibiotics in patients with IE [16], the abilities of sodium bicarbonate to attenuate oxidative stress may prove especially helpful in preventing AKI in these patients. However, most previous studies investigating the renal effects of sodium bicarbonate have excluded patients with systemic infection, and no study has heretofore been conducted solely in patients with IE. The aim of this prospective, randomized, placebo-controlled trial was to investigate the effect of perioperative sodium bicarbonate administration on postoperative renal function in patients with IE undergoing cardiac surgery with CPB.

The study protocol was approved by the Institutional Review Board and Hospital Research Ethics Committee of Severance Hospital at Yonsei University College of Medicine (#4-2013-0376) and registered at ClinicalTrials.gov on July 2013 (NCT01920126). The trial was conducted at Yonsei University College of Medicine in Seoul, Korea, between August 2013 and July 2016. After providing written informed consent, 70 patients between 20 and 80 years old, who were scheduled for valvular heart surgery for IE, were enrolled. IE was diagnosed according to the modified Duke criteria [17]. Patients were excluded if they met at least one of these criteria preoperatively: (1) estimated glomerular filtration rate (eGFR) <15 ml/min/1.73 m²; (2) receiving renal replacement therapy; (3) receiving intra-aortic balloon pump support; (4) preexisting hypernatremia, alkalosis, or severe pulmonary edema; or (5) chronic moderate dose to high dose corticosteroid therapy (>10 mg/day prednisolone or equivalent).

Enrolled patients were randomly assigned to either the control or bicarbonate group at a 1:1 ratio using a computer-generated random-number table. Assignments were concealed in sealed envelopes. The study drugs were prepared in identical 50-ml syringes by an anesthesia nurse not involved with the study. The bicarbonate group received 154 mEq/l sodium bicarbonate with a 0.5 mmol/kg intravenous (IV) loading dose over 1 h immediately after anesthetic induction, followed by an IV continuous infusion at 0.15 mmol/kg/h over 23 h. The control group received equivalent volumes of IV 0.9% sodium chloride for the same duration. Attending surgeons and anesthesiologists involved in patient management in the operating room and intensive care unit (ICU) were blinded to the group assignment. Group assignment was disclosed after hospital discharge.

All patients received institutionally standardized anesthetic care and CPB management as previously described [18]. Intraoperatively, all patients received balanced crystalloid (Plasma solution A 1000 inj, CJ, Seoul, Korea) at 6 ml/kg/h and balanced synthetic colloid (Volulyte; Fresenius Kabi, Bad Homburg, Germany) to replace estimated blood loss (to a maximum of 500 ml). Mean arterial pressure (MAP) <70 mm Hg was treated with norepinephrine, with the infusion starting at 0.01 μg/kg/min and titrated to a maximum of 0.3 μg/kg/min. If the blood pressure was not maintained with norepinephrine, vasopressin was infused at 2.4–4.0 units/h. Milrinone (0.5 μg/kg/min) was infused when the mixed venous oxygen saturation decreased below 60% for >10 minutes. Postoperatively, when urine output was <1 ml/kg/h for 30 minutes, loop diuretics were administered after a 200-ml IV bolus of crystalloid. If this was inadequate, fluid boluses were repeated up to three times. If urine output remained inadequate 15 minutes after the third bolus, 5 mg of IV furosemide was administered and then repeated every 30 minutes, with escalating doses until the desired effect (urine output >1 ml/kg/h) was achieved. Renal replacement therapy was initiated if any of the following occurred: refractory oliguria (urine output <0.5 ml/kg/h for 12 h), fluid overload, hyperkalemia (>6.5 mEq/l), rapidly rising serum potassium level, signs of uremia, or metabolic acidosis (pH <7.10).

Of the 70 patients enrolled, IE was definitively confirmed through surgical findings and pathologic examination in 32 patients in each group. According to the intention-to-treat principle, statistical analyses were performed using data from all enrolled patients. Four patients in each group (P > 0.999) underwent emergency surgery within 24 h after being diagnosed with IE. All patients were treated with antibiotics before the day of surgery. At the time of surgery, 29 (83%) and 27 (77%) patients in the bicarbonate and control groups, respectively, had active infection with a positive blood culture, leukocytosis (>10,800/μl), fever (temperature >38 °C), or elevated C-reactive protein (>8 mg/l) (P = 0.550). The time between diagnosis of IE and surgery was similar in both groups (bicarbonate vs. control: 11 (5, 17) days vs. 8 (5, 17) days, P = 0.857). All baseline patient characteristics were comparable in the two groups, except that more patients received anti-platelet drugs prior to surgery (5 vs. 0, P = 0.020) and had platelet counts below 150,000/μl (11 vs. 3, P = 0.017) in the control group (Table 1).

In this prospective, randomized, placebo-controlled trial we did not observe any beneficial effects of perioperative bicarbonate administration on postoperative renal function in patients with IE undergoing cardiac surgery with CPB. Sodium bicarbonate did not attenuate the peak postoperative SCr level nor reduce the incidence of AKI defined by the AKIN criteria. In fact, the postoperative increase in SCr above baseline was even greater in the bicarbonate group than in the control group.

Acute IE is a challenging disease that is associated with high morbidity and mortality. Although most patients with IE undergo medical treatment, more than 20% require surgery, and the number of surgical treatments has gradually increased as accumulating evidence supports the use of cardiac surgery to improve outcomes [1, 2]. However, morbidity and mortality following cardiac surgery in patients with IE remain significant. AKI is one of the most common morbidities, with a rate of 20–60% [6, 22]. Moreover, AKI clearly increases long-term postoperative mortality and cardiac event rates in patients with IE [6, 23], but definitive strategies to prevent AKI in this patient population have not been established.

The systemic inflammation and oxidative stress associated with septic conditions predispose patients with IE undergoing cardiac surgery to AKI [5]. Furthermore, prolonged use of nephrotoxic antibiotics preoperatively, and administration of contrast media shortly before cardiac surgery, adds to the risk of AKI by enhancing proinflammatory oxidative stress [16, 24]. Sodium bicarbonate administration produces blood and urine alkalization and may slow pH-dependent Haber-Weiss free-radical production, thereby attenuating reactive oxygen species-mediated tubular injury [11, 24, 25]. Sodium bicarbonate may also directly scavenge other reactive species [26]. Tubular alkalosis achieved by sodium bicarbonate has been demonstrated to mitigate the conversion of hemoglobin to methemoglobin, preventing tubular obstruction and tubular cell necrosis [10]. Accordingly, sodium bicarbonate administration has shown benefits in experimental models of ischemic [27], doxorubicin-induced [28], and contrast-medium-induced AKI [29]. However, several trials and meta-analyses investigating the renal effects of sodium bicarbonate in the context of cardiac surgery yielded inconsistent results [8, 14, 15, 21]. Based on the theoretical ability of sodium bicarbonate to reduce oxidative stress in renal tubular cells, we investigated whether perioperative sodium bicarbonate infusion would lessen renal injury in this patient population, an issue that has not been previously addressed.

Contrary to our expectation, sodium bicarbonate infusion did not demonstrate renoprotective effects in this study. The postoperative peak SCr level and incidence of AKI defined by the AKIN criteria were similar in both groups. Indeed, when AKI was diagnosed using the criteria that were applied in the first trial reporting renoprotective effects of sodium bicarbonate in cardiac surgery (increase in SCr >0.5 mg/dl or 25% above baseline) [8], the incidence of AKI was significantly greater in the bicarbonate group than in the control group. Moreover, the increase in postoperative SCr levels above baseline was significantly greater in the bicarbonate group than in the control group.

These results imply that the effects of sodium bicarbonate on renal function may not just be neutral, but they may even be harmful in patients with IE undergoing cardiac surgery, which argues against the routine use of sodium bicarbonate. Notably, intraoperative platelet transfusion requirement was higher in the control group than in the bicarbonate group, which might be attributed to the fact that more patients received anti-platelet drugs (5 vs. 0, P = 0.020) and had platelet counts below 150,000/μl (11 vs. 3, P = 0.017) in the control group. In addition, packed erythrocyte transfusion requirement during the 24 h postoperatively was also higher in the control group than in the bicarbonate group. Erythrocyte and platelet transfusions are associated with an increased risk of AKI and dialysis [30, 31]. Thus, the higher incidence of AKI in the bicarbonate group despite less transfusion further supports the possibly harmful impact of bicarbonate infusions.

In the context of cardiac surgery, previous findings demonstrate that the renoprotective effects of sodium bicarbonate were limited to low-risk patients [12, 32]. Furthermore, a recent trial of prophylactic sodium bicarbonate therapy to prevent AKI in high-risk patients was terminated early because of increased mortality in the bicarbonate group [15]. This difference according to the patient’s risk is consistent with our results, as our patients had a high risk of AKI because of infection, severe inflammation, receipt of nephrotoxic antibiotics, and surgery-related injury. The likelihood of postoperative AKI is higher in patients with more risk factors [6], and a single treatment strategy might be insufficient to attenuate renal injury in patients with multiple risk factors. This might be the most cogent explanation for our results.

The inability of sodium bicarbonate to reduce AKI was discussed from the perspective of alkalemia and resultant physiologic changes in previous studies [13‐15]. Alkalemia can lead to impaired tissue perfusion [33, 34] and oxygen delivery [35], and to increased cell death and apoptosis associated with superoxide formation [36]. Most patients enrolled in our study had active inflammation and received prolonged treatment with multiple nephrotoxic antibiotics, which would aggravate renal injury associated with increased renal vascular resistance and tubular cell impairment [5, 37, 38]. During active systemic inflammation, increased renal vascular resistance (uncoupled from decreased systemic vascular resistance) may diminish renal perfusion, thereby causing ischemia [5, 38]. Indeed, multiple areas of renal infarction were observed in an animal model of IE, suggesting hypoperfusion-induced ischemia [39]. Nephrotoxic antibiotics have also been demonstrated to reduce renal blood flow by increasing renal vascular resistance rather than by lowering perfusion pressure [40]. Alkalosis following sodium bicarbonate infusion could further increase renal vascular resistance [34] and exacerbate renal tubular injury by producing intracellular acidosis. A rapid influx of CO₂ following sodium bicarbonate infusion could worsen intracellular acidosis [41], accelerating the influx of sodium and calcium, and causing cell swelling and renal tubule dysfunction [42].

There has been a paucity of randomized, controlled trials in patients with IE undergoing cardiac surgery. This may be at least partly because these patients are often unstable or require urgent surgery, leading to a preponderance of observational studies. Thus, the primary strength of this study is that it is the first to address the renoprotective efficacy of sodium bicarbonate through randomization, solely in patients with IE undergoing cardiac surgery, although the sample size was not sufficient to compare the incidence of AKI, which is the most clinically relevant index.

The limitations of this study are as follows. First, we did not directly measure urine pH to confirm urinary alkalization by IV sodium bicarbonate administration. However, both plasma and urinary alkalization were achieved in previous studies using similar doses and durations of sodium bicarbonate infusion [8, 15, 21]. Second, we can only speculate on the reasons for the lack of beneficial effect of sodium bicarbonate in this study, because no mechanistic investigations were conducted. Third, we used balanced 130/0.4 hydroxyethyl starch solution (Volulyte) for volume replacement, upon which concerns have been raised regarding its nephrotoxicity in patients with sepsis. However, its influence on renal function is far more controversial when used in the operating theatre. Indeed, there is solid evidence favoring its use in the presence of overt volume loss [43‐45]. Furthermore, the median volume used in this study did not exceed 500 ml in either group throughout the perioperative 24 h, without any intergroup difference. Thus, its influence on the observed results should be negligible.

Perioperative sodium bicarbonate administration did not improve postoperative renal function or other outcomes in patients with IE undergoing cardiac surgery. Rather, it was associated with potentially harmful renal effects, as demonstrated by a greater postoperative increase in SCr above baseline, compared to control. Further research on renoprotective strategies in patients with IE is warranted.