Acute Kidney Injury following Cardiac Surgery: Role of Perioperative Blood Pressure Control

Postoperative acute kidney injury (AKI) is a serious complication of cardiac surgery with a high mortality rate [1,2]. There has been little progress in recent years in reducing the incidence of this condition, improving the prognosis of affected patients, or identifying it earlier in its pathological course [2,3,4]. Although the persistent morbidity and mortality of perioperative AKI may reflect increased baseline mortality risk among cardiac surgery patients [2,5], it is vital that we work to identify predisposing factors and develop safer perioperative protocols.

A number of factors, including preoperative blood pressure (BP) elevation and perioperative BP changes below presenting BP, are known to predispose to renal dysfunction [4,6]. Preoperative pulse pressure (PP) also has a significant positive relationship with postoperative renal risk [7]. What is not known is whether higher diastolic BP (DBP) is protective or, conversely, whether lower DBP increases risk for AKI independent of systolic BP (SBP). These relationships could potentially confound the relationship between greater PP and adverse outcomes. It is only recently, however, that we have become aware that BP lability (i.e., BP excursions outside an acceptable physiologic range) during cardiac surgery may be an important predictor of subsequent renal dysfunction [8,9]. The studies that have revealed this association include a descriptive analysis of data [10] from 7,247 patients who underwent coronary artery bypass graft (CABG) surgery and an analysis of pooled data from three open-label prospective clinical trials [8]. In this review, we discuss these data in the context of the epidemiology and management of perioperative renal dysfunction.

Although AKI is a well-recognized complication of cardiac surgery [11], estimates of its frequency vary widely. For example, a recent review found that the reported incidence ranged from 1 to 30% [12], a variation that largely depends on the definition of renal dysfunction used in each trial. The metrics most commonly used to define AKI are the absolute or relative serum creatinine level and/or the need for dialysis in previously undialyzed patients. Table 1 [1,2,3,4,8,11,12,13,14,15,16,17,18] shows the range of metrics used in recent studies involving cardiac surgery patients and the corresponding incidence of AKI. Attempts to promote the use of consistent criteria for the definition of AKI resulted in the publication of the Acute Dialysis Quality Initiative Group’s Risk, Injury, Failure, Loss of Kidney Function, End-Stage Kidney Disease criteria, which were subsequently modified to create the Acute Kidney Injury Network criteria [19,20]. These two sets of criteria have similar abilities to predict in-hospital mortality in critically ill patients [19,20]. There is evidence that the incidence of AKI in cardiac surgery patients is changing over time. However, the true nature of the trends reported is obscured by changes in the definition of AKI and in the demographics of the at-risk population. For example, using a consistent definition of postoperative renal dysfunction, Janssen et al. [3] documented a gradual increase in the incidence of this condition in Dutch CABG patients between 1987 (0.6%) and 1995 (3.4%). This increase was associated with increases in patient age and in the prevalence of diabetes and preexisting renal pathology [3], suggesting that changes in patient population may explain the increased incidence. In contrast, a British study [4] found that the percentage of cardiopulmonary bypass patients who required continuous venovenous hemofiltration decreased between 1989/1990 (2.7%) and 1997/1998 (2.0%). This reduction occurred in a patient population that was older and more severely ill in the second study period [4]. In a third, much larger study [2], conducted in the United States, the raw incidence of AKI after CABG increased from 1.5 to 7.2% between 1988 and 2003. This trend remained after adjustment for demographic and comorbidity characteristics (1988, 1.1%; 2003, 4.1%) [2]. The authors suggested that the observed increase in AKI incidence may be attributable, in part, to adoption of less restrictive diagnostic criteria, a theory that is supported by decreases in the percentage of AKI patients who required dialysis (1988, 15.7%; 2003, 8.6%) and who died (adjusted mortality rate: 1988, 39.5%; 2003, 17.9%) [2]. The conflicting data presented in these three studies [2,3,4] make it difficult to determine the presence of and reasons for any change in AKI incidence in recent years.

Development of AKI after cardiac surgery carries an extremely poor prognosis. Mortality rates range from 18 to almost 80% in those who develop this condition, and from 0.9 to 6.4% in those who do not (table 1) [1,2,3,4,8,11,12,13,14,15,16,17,18]. As a result, the OR for death associated with postoperative AKI is in the range of 34–39 [13,18]. In cardiac surgery patients, postoperative AKI is also associated with longer intensive care unit and hospital stays [11,15,18,20] and, among survivors, with the need for permanent dialysis [1,16]. Discharge to a care facility, rather than discharge home, is also significantly more common in patients who have experienced postoperative renal failure [15]. The grave prognosis associated with postoperative AKI is widely acknowledged. Less well recognized is the potential association between mortality and smaller declines in renal function. This was explored by Lassnigg et al. [21], who used data from 4,118 patients undergoing cardiac or thoracic aortic surgery. In this population, a decrease of ≤0.3 mg/dl in serum creatinine level within 48 h of surgery was associated with the best prognosis (30-day mortality rate, 2.1%; fig. 1) [21]. Compared with this group, patients with even a small increase in serum creatinine level (≥0 but <0.5 mg/dl) had a hazard ratio (HR) for death of 2.8, and those with a larger increase (≥0.5 mg/dl) had an HR of 18.6 (p =0.001 for both). These relationships were maintained after multivariate analysis (<0.5 mg/dl: HR 1.9; p = 0.0004; ≥0.5 mg/dl: HR 5.8; p =0.0001). Patients with the most substantial decreases in serum creatinine levels also had a higher mortality rate than those with decreases of ≤0.3 mg/dl (fig. 1), a finding that was attributed to the diluting effect of greater volume replacement and blood transfusions in this high-risk group of patients who had higher preoperative serum creatinine levels, more preoperative comorbid conditions, and a more complicated intraoperative course than the remainder of the cohort [21].

Delays in recognizing the condition may contribute to the poor prognosis associated with postoperative AKI [22]. Delays may occur because of the limitations of serum creatinine as a marker of glomerular filtration rate [22] and, in particular, because of the lag time between deterioration in renal function and attainment of the serum creatinine level that triggers intervention. In addition, the data published by Lassnigg et al. [21] demonstrate that even subtle changes in serum creatinine level, many of which may be interpreted as clinically insignificant – and which certainly do not define true AKI – can affect patient outcome. These limitations may be overcome by the introduction of novel serum and urinary biomarkers of AKI. These include serum levels of cystatin C, neutrophil gelatinase-associated lipocalin, and uric acid [23]. Urinary biomarkers include enzymes that are released by damaged tubular cells (e.g., alkaline phosphatase, γ-glutamyl transpeptidase), low-molecular-weight proteins (e.g., cystatin C), and proteins that are produced in the kidney in association with AKI (e.g., Gro-α, IL-18) [23].

There is evidence that the mortality rate in AKI patients has decreased in recent years. Relevant data were published by Thakar et al. [24] who analyzed data from more than 33,000 cardiac surgical procedures performed between 1993 and 2002. This analysis showed that the mortality rate associated with postoperative AKI decreased from 32 to 23% (p < 0.0001) between the first and second halves of the study period [24]. Over the same period, the mortality rate associated with AKI requiring dialysis decreased from 61 to 49% (p < 0.01). These changes in mortality rate in AKI patients may be due to changes in preoperative treatment. For example, it has been suggested that the decrease in mortality risk among patients in this study who had preoperative congestive heart failure (CHF) and who developed AKI requiring dialysis (odds ratio [OR] for death, 2001 vs. 1993: 0.01; p < 0.0001) may have been due to changes over time in the preoperative management of this patient population [24]. Potentially relevant factors include the increased use of angiotensin-converting enzyme (ACE) inhibitors and angiotensin II receptor blockers (ARBs), which, although not administered to CHF patients primarily for their antihypertensive effect, will affect BP. The ability of ACE inhibitors to reduce the lower limit of the renal autoregulatory range may also be relevant in this context [25].

Numerous pre-, intra-, and postoperative factors have been associated with the development of AKI after cardiac surgery (table 2) [1,3,5,6,12,14,15,16,26,27,28,29,30,31,32,33]. Of the numerous preoperative factors that have been shown to affect risk, renal function is particularly important. For example, in one study [4], a preoperative serum creatinine level ≥1.7 mg/dl was associated with a 48-fold increase in the risk of postoperative renal dysfunction. Older age is frequently identified as a risk factor, as are comorbid conditions such as chronic hypertension, diabetes, and CHF.

Risk may also be affected by the patient’s preoperative pharmacological regimen. For example, preoperative use of renin-angiotensin system (RAS) inhibitors (i.e., ACE inhibitors and ARBs) has been associated with a 28% increase in postoperative AKI risk in cardiac surgery patients [31]. This effect was independent of intra- and postoperative hypotension, suggesting that it is not mediated via an effect on BP, more likely, it is related to effects on glomerular capillary pressure.

The most consistently identified intraoperative factor related to the development of AKI is cardiopulmonary bypass duration. This has led to increased interest in the use of off-pump surgery. In a recent meta-analysis [34] examining the risk of AKI in off-pump CABG, off-pump surgery reduced the risk of AKI (OR 0.57; 95% confidence interval (CI) 0.43–0.76) and the risk of AKI requiring renal replacement therapy (OR 0.55; 95% CI 0.43–0.71) when compared with conventional surgery. An earlier meta-analysis [35] that evaluated multiple outcomes following off-pump CABG also found that compared with conventional CABG, off-pump surgery was associated with a reduction in renal failure (OR 0.62; 95% CI 0.50–0.78). Use of an intra-aortic balloon pump also substantially increases the risk of renal dysfunction or failure (OR 4.3–4.4) [1,6]. Significant postoperative factors mainly involve complications of surgery and a prolonged postoperative course. The majority of the studies summarized in table 2 [1,3,5,6,12,14,15,16,26,27,28,29,30,31,32,33] did not select patients based on preoperative renal function.

Although many of the risk factors for postoperative AKI cannot be modified, recent publications have highlighted the influence of the modifiable factor, perioperative BP.The publication of data showing associations between the development of renal dysfunction or failure and preexisting essential, isolated systolic, and wide PP hypertension is clarifying the role of BP in the pathogenesis of postoperative AKI [1,3,5,6,14]. A comprehensive investigation of the role played by preoperative BP in postoperative renal risk was published by Aronson et al. [6] in 2007. This study determined the influence of isolated systolic hypertension, isolated diastolic hypertension, and wide PP hypertension on the development of postoperative renal dysfunction or failure in 4,801 patients undergoing CABG with cardiopulmonary bypass. Renal dysfunction was defined as a postoperative serum creatinine level of ≥2.0 mg/dl (177 mol/l), plus an increase of ≥0.70 mg/dl (62 mol/l) from preoperative baseline. Renal failure was defined as either renal dysfunction requiring dialysis or post-mortem evidence of renal failure [6]. Wide PP hypertension was a significant predictor of AKI, which increased by 49% for each additional 20 mm Hg increment above a PP of 40 mm Hg (OR 1.49; 95% CI 1.17–1.89; p = 0.001) [6]. In addition, the risk of AKI-related death was more than 3 times higher in patients with a PP >80 mm Hg than in those with a lower PP (3.7 vs. 1.1%). However, in this analysis, there was not a specific examination of SBP or DBP. In another report by Aronson et al. [5], the authors reported that isolated systolic hypertension was associated with adverse outcomes (cardiovascular and renal events and all-cause mortality) after CABG surgery. Interestingly, they noted that those patients with diastolic hypertension (≥90 mm Hg) did not have an increased risk of adverse outcomes regardless of whether they had systolic hypertension. It was only the patients without diastolic hypertension who had an increased risk of adverse outcome (1.4; 95% CI 1.1–1.7)

In an independent investigation [36], wide PP was also associated with adverse cerebral and cardiac outcomes, cardiac death, and all-cause mortality in this patient population. These data suggest an association between preoperative BP and postoperative renal risk. The superiority of PP over SBP as a predictor of renal risk may be related to the pathophysiology of widened PP (increased arterial stiffness and reduced vascular compliance) [36]. Organs such as the kidneys, which have low vascular resistance, are particularly susceptible to increases in PP and decreases perfusion pressure when as might occur with low DBP in vascular compliance is reduced. The deleterious changes in the renal vasculature caused by this increase in pulsatile stress may underlie the adverse renal events that occur in these patients.

The foregoing discussion summarized data that demonstrate an effect of preoperative BP on postoperative renal risk. However, there is evidence that the risk of this outcome is also affected by intraoperative BP levels. Fischer et al. [27] reported that patients with normal preoperative renal function who developed AKI after cardiac surgery involving cardiopulmonary bypass had spent a significantly longer time on cardiopulmonary bypass at a mean arterial pressure (MAP) of <60 mm Hg. It is important to note that relatively small changes in perioperative BP can substantially affect postoperative risk. In one study [37], a cut-off level for change from preoperative level in MAP of >20 mm Hg or >20% was sufficient to identify surgical patients at significantly increased cardiac or renal risk (composite endpoint). This level of change can be clinically important, as experimental studies in dogs indicate that autoregulatory mechanisms are only capable of maintaining perfusion of the coronary and renal circulations within physiological limits if systemic pressures remain within the ‘autoregulatory range’ (for the kidney, approximately 75 to >160 mm Hg) [38]. This range may be ‘reset’ by chronic hypertension [39], acute ischemia [40], or arteriosclerosis [39]. Decreases in BP below this range are likely to result in ischemia of critical arterial beds, whereas increases may lead to hyperemia [37]. Moreover, patients with impaired renal function preoperatively (i.e., those with chronic kidney disease) may also have reduced ability to autoregulate afferent renal blood flow and thus be more likely to experience acute declines in renal function when exposed to low systemic BP. The resulting deterioration in renal function may thus reflect the state of renal vascular disease.

Although it is evident that substantial changes in BP in the perioperative period may have clinically significant effects on renal function, few studies have investigated the role of perioperative BP control in the pathogenesis of postoperative AKI. Moreover, to our knowledge, the effect of different BP targets on the incidence of AKI has never been investigated in a controlled prospective study. However, recent analyses from the Evaluation of Clevidipine in the Perioperative Treatment of Hypertension Assessing Safety Events (ECLIPSE) program suggest that greater perioperative BP lability may increase renal risk [29]. The ECLIPSE program comprised three prospective, randomized, open-label, parallel-group studies designed to compare clevidipine, a recently introduced intravenous dihydropyridine calcium channel blocker, with nitroglycerin or sodium nitroprusside perioperatively, or nicardipine postoperatively, in patients undergoing cardiac surgery at 61 medical centers. The primary trial endpoint was the incidence of death, myocardial infarction, stroke, or renal dysfunction at 30 days [29]. The results generated by this program represent the largest randomized clinical trial safety database involving intravenous antihypertensive agents in patients undergoing cardiac surgery in the United States. The degree of BP control achieved in each patient in the ECLIPSE program was assessed by determining the magnitude and duration of SBP excursions both above and below a defined range using area under the curve (AUC) analysis (fig. 2) [29]. The parameter ‘AUC’ captured the degree to which SBP fell outside predefined pre-, intra-, and postoperative ranges. The predefined ranges were chosen with the aim of demonstrating efficacy in a drug development protocol, and were therefore intentionally broad.

There was no difference in the incidence of the combined primary endpoint (death, myocardial infarction, stroke, or renal dysfunction at 30 days) between the clevidipine-treated patients and those in the other three treatment groups combined. However, the mortality rate was significantly higher in the patients who received sodium nitroprusside (13/274, 4.7%) than in those who received clevidipine (5/286, 1.7%; p = 0.04). There were no other significant differences between the clevidipine group and the other individual treatment groups in the incidence of components of the primary endpoint [29].

In a post-hoc analysis [29], multiple logistic regression was used to determine whether a range of potential variables were significant predictors of renal dysfunction (postoperative serum creatinine ≥2.0 mg/dl, with a post-baseline increase of ≥0.7 mg/dl and/or the need for hemodialysis, venovenous filtration, arterial venous filtration, or peritoneal dialysis postoperatively) within 30 days of surgery. The variables tested included AUC, as well as those based on demographic data, baseline characteristics, medical history, treatment group, and procedural characteristics. In this particular analysis, the defined ranges for SBP were 85 to 145 mm Hg pre- and postoperatively and 75 to 135 mm Hg intraoperatively. The analysis demonstrated that, among patients who showed substantial BP lability (i.e., those in the fourth quartile of AUC), increased magnitude and/or duration of excursions outside the SBP range was a significant independent predictor of postoperative AKI (OR 1.8; p = 0.0119) [29].

This exploratory analysis of the ECLIPSE data represents the first indication of a link between perioperative BP control and postoperative AKI risk. Its results are supported by those of a more recent retrospective study [41] involving data from 4,098 patients who underwent CABG surgery at Duke University. This analysis identified a significant association between percentage change from baseline in serum creatinine level and mean SBP incursion nadir below a threshold of 95 mm Hg (p < 0.01). SBP AUC >20% above or below baseline (defined as the mean SBP over the first 5 recordings after arterial line placement) was also associated with AKI. It should be noted that there is also preliminary evidence of an association between increased BP lability and mortality in patients undergoing cardiac surgery [41]. The results of these studies suggest that the effect of intraoperative BP variability on postoperative morbidity is worthy of examination in future studies. A factor that merits further investigation is the potential confounding effect of preoperative antihypertensive treatment on postoperative AKI risk, a factor that was not taken into account in the studies cited above.

The growing body of evidence linking perioperative BP to subsequent AKI suggests that antihypertensive regimens that offer more precise BP control may reduce renal risk. Management of hypertension in surgical patients is particularly challenging because the anesthesiologist must attempt to limit BP elevations without reducing MAP to such an extent that end-organ hypoperfusion and damage result. Current antihypertensive agents do not always allow such precise control. The ideal perioperative antihypertensive agent would have rapid onset and short duration of activity with easy titration to effect and a low risk of overshoot hypotension [42]. The benefits and limitations of currently used agents are summarized in tables 3 and 4[29,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75]. Where available, comparison studies have been included; however, it should be noted that rigorous studies involving commonly used antihypertensives are few and far between and, with the exception of the ECLIPSE program, were completed more than a decade ago. Another limitation of many studies is the small sample size.

In 2000, Vuylsteke et al. [76] published the results of a prospective survey in which 200 American and European anesthesiologists recorded details of their cardiac anesthetic practice. The study found that deepening the plane of anesthesia was the technique used most commonly to reduce intraoperative BP. This strategy was often combined with nitroglycerin- or sodium nitroprusside-based vasodilation [76]. Other antihypertensive agents that may be used for perioperative BP control include dihydropyridine calcium channel antagonists (e.g., clevidipine, nicardipine, isradipine), cardioselective (β₁) β-adrenoreceptor antagonists (e.g., atenolol, esmolol, metoprolol), non-cardioselective β-adrenoreceptor antagonists (e.g., propranolol), mixed α/β-adrenoreceptor antagonists (e.g., labetalol), dopamine receptor agonists (e.g., fenoldopam), and smooth-muscle relaxants (e.g., hydralazine) [42]. The antihypertensive agent(s) used in any given anesthetic protocol will depend on physician familiarity and preference, the specific circumstances of the case, and the advantages and disadvantages of the agents in question.

Postoperative AKI is a serious and not uncommon complication of cardiac surgery. The risk of this adverse event is significantly increased by a number of pre-, intra-, and postoperative factors. It is likely that elevated SBP, coupled with lower DBP (wide PP) as a reflection of vascular disease and aortic stiffness, correlated with increased risk for AKI. It is as likely that diastolic hypertension may be protective, which could suggest that impaired renal autoregulation in patients with vascular disease is a clinical concern. In balance, we still lack the data to demonstrate that control of BP reduces cardiovascular events, AKI, or mortality. Recent preliminary data show that these include greater BP lability (i.e., greater excursions of BP outside the desired range), a factor under clinician control that has also been associated with increased postoperative morbidity and mortality. These observations may reflect the net state of vascular disease predisposing to the adverse consequences of both evaluated and reduced (below the autoregulatory threshold) levels of BP. If these data are confirmed by ongoing and future studies, they would suggest that anesthesiologists should aim individualized levels of BP in the perioperative period, avoiding both hypertension and overshoot hypotension.

Editorial assistance was provided by Excerpta Medica, Bridgewater, N.J. We thank Tia A. Paul, University of Maryland School of Medicine, Baltimore, Md., for expert secretarial support.

M.R.W. sits on the Steering Committee for the Studying the Treatment of Acute Hypertension (STAT Registry); S.A. has received support for work contracted by The Medicines Company as a consultant; E.G.A. has conducted research funded by The Medicines Company, Cubist Pharmaceuticals and Covidien as well as served as a consultant for these companies, and C.V.P. has conducted research funded by The Medicines Company, Sanofi-Aventis, and Glaxo-Smith Kline. He has served as a consultant for Schering-Plough and Sanofi-Aventis.