Biomarkers: Understanding, Progress, and Implications in the Perioperative Period
Patients with coronary artery disease undergoing major noncardiac surgery have a considerable risk of perioperative cardiac morbidity and mortality and compromised long-term outcome [1,2]. Several preoperative risk stratification scores, as well as strategies based on physical examination and history, have been developed to predict the risk of cardiac complications following various surgical procedures performed in different patient populations [3,4].
Some of these scores have limitations, and generating them may include cardiac stress testing, with its associated risks and costs [4]. Moreover, although there are many risk stratification tools concerning cardiac outcomes, few exist for important noncardiac complications, such as infectious, neurologic, and renal complications, or for noncardiac causes of mortality.
Biomarkers have attracted the attention of clinicians and investigators as a means of stratifying risk in different patient populations, including surgical patients. They generally entail simple, minimally invasive tests (blood draws) that have potentially high yield in assessing risk stratification [5,6]. Therefore, it is important to understand the biology, pathophysiology, and usefulness of biomarkers during the perioperative period, because these indicators may have predictive value for postoperative outcomes, especially when these biomarkers are added to existing risk stratification techniques. For instance, C-reactive protein (CRP) and N-terminal pro-brain natriuretic peptide (NT-proBNP) improved the average sensitivity of predicting perioperative major cardiovascular outcomes from 59% to 77% when added to a clinical risk prediction system [7]. The ultimate goal is to use these biomarkers to develop and evaluate therapies intended to improve surgical outcomes.
This article reviews and evaluates current knowledge and available evidence regarding the usefulness of some of the commonly studied infectious, inflammatory, neurologic, cardiac, and renal biomarkers in the perioperative period (Table 1).
CRP
Surgery induces an intense inflammatory response [8]. The release of inflammatory markers has been documented following cardiac surgery [9], and major noncardiac surgeries such as joint replacement, major vascular surgery, and colorectal surgery [10]. The perioperative inflammatory response is believed to contribute to poor perioperative outcomes. For example, 17% of postoperative deaths are attributed to cardiovascular causes [11]. Inflammation has been implicated in the pathophysiology of myocardial infarction; it is associated with the development of arterial plaque, starting with lipid deposition to plaque rupture and the resulting complications. If the underlying mechanism of perioperative myocardial infarction (PMI) is plaque rupture, then a negative stress test (which indicates that there is no flow-limiting stenosis) may fail to detect a nonobstructing plaque that may potentially rupture and cause cardiac ischemia or infarction [12]. In such circumstances, inflammatory markers such as CRP may more accurately predict perioperative cardiac morbidity and mortality [13]. There are many available markers of inflammation such as tumor necrosis factor (TNF) and the interleukins (IL) IL-2, IL-4. IL-6, IL-8, IL-10, and IL-14; however, this article focuses on CRP.
CRP, an acute-phase reactant produced in the liver, is a well-known marker of systemic inflammation; it greatly increases in response to acute injury. Its stable concentration over a long period of time is governed chiefly by the rate of hepatic production rather than by factors connected with its clearance [14]. CRP assays, being cost-effective, reliable, and fairly sensitive [15,17], seem to meet all criteria for the ideal biomarker. Indeed, the sensitivity of the high sensitivity CRP (hsCRP) assay seems to exceed even that of the original CRP assay.
In nonsurgical settings, CRP functions as a marker of atherosclerosis; when increased, it is considered a biochemical cardiovascular risk factor [18,20].
Cardiac Surgery
In cardiac surgery, postoperative serum concentrations of CRP are associated with the incidence of postoperative arrhythmias [9], and are predictive of septic complications, the need for catecholamine therapy, prolonged respiratory support, and prolonged intensive care unit (ICU) stay [21]. Moreover, Milazzo and colleagues [22] identified preoperative elevated concentrations of CRP as a predictor of recurrent ischemia up to six years postoperatively.
Noncardiac Surgery
The predictive power of CRP seems to be stronger than that of the clinical risk index [7]. Preoperative high levels are associated with short- and long-term morbidity and mortality after noncardiac [7]. Postoperative high levels of this marker are associated with complications after colorectal and bariatric surgeries [23,25].
So far, evidence provides no consensus on what value constitutes an abnormally increased hsCRP or even the range of detectable hsCRP in the general population. Because different studies have not used identical assays, a wide range exists in the values that are believed to constitute an increased level. Assays for hsCRP are new, only appearing commercially in the last decade. Most of the available clinical studies have accepted as normal or increased the values that have been designated as normal or increased by various laboratories, depending on the assay each laboratory used. However, more recent studies with the newer hsCRP assays have redefined a normal CRP level in a healthy individual as approximately ≤3 mg/L. Other clinicians might recognize an even lower cutoff (≤1 mg/L) as normal. Although the present assays are highly sensitive, these current values may change with future advances in technology and in the criteria governing specific assays [13,26–29].
Interventions that moderate the inflammatory response may reduce adverse outcomes [30]. Clinical studies show that, in some circumstances, a simple antiinflammatory intervention such as corticosteroid administration is associated not only with lower perioperative CRP levels but also with improved clinical outcomes [31,32]. In a randomized trial of 88 patients undergoing laparoscopic cholecystectomy, 8 mg of dexamethasone given 90 minutes before incision correlated with significantly lower CRP levels, significantly reduced postoperative fatigue, less nausea and vomiting, and a faster return to recreational activities [33]. Therefore, interventions targeted to decrease CRP levels, or to prevent their perioperative increase, may be warranted [34].
Brain or B-type natriuretic peptide and NT-proBNP
Congestive heart failure (CHF) has been shown to be a predictor of PMI following vascular surgery [3,35–37]. It increases the odds of dying from a PMI 12-fold if diagnosed within 12 months of a vascular surgery [1]. Diagnosis of CHF is a clinical challenge. Traditionally, the diagnostic approach begins with a thorough history and physical examination, complemented by chest radiographic examination and echocardiography. However, the history and physical examination can be misleading [38], and chest radiographic examination has its own limitations [39]. Echocardiography can also be misleading, in that 30% to 70% of patients with CHF present with a normal (>50%) or only mildly depressed left ventricular ejection fraction. Diastolic dysfunction may underlie CHF in many of these patients [40]. Most patients with asymptomatic left ventricular dysfunction remain undiagnosed [38]. Brain natriuretic peptide (BNP) and NT-proBNP have emerged as promising biomarkers, as simple and reliable tests that may be useful in establishing the diagnosis of CHF and correlating with its severity [38,41].
BNP is a 32-amino acid polypeptide secreted by cardiomyocytes in response to excessive ventricular or atrial stretching, or during ischemia. NT-proBNP is a 76-amino acid N-terminal fragment that is biologically inactive and is coreleased in the same conditions as BNP [42,43]. Biologically, they are distinguished by a significantly different half-life: NT-ProBNP has a longer half-life, 1 to 2 hours, in contrast with 20 minutes for BNP.
In nonsurgical settings, clinical studies have shown that plasma BNP is a powerful predictor of adverse cardiovascular events in patients with heart failure, acute coronary syndromes, primary pulmonary hypertension, and valvular disease [44,46].
Cardiac Surgery
In cardiothoracic surgery patients, 2 recent clinical studies showed the importance of BNP and NT-proBNP. High levels BNP and NT-proBNP have been associated with the development of atrial fibrillation in patients undergoing cardiac or thoracic oncology surgery [47,48]. Moreover, NT-proBNP may be useful in the assessment of potential donor hearts and in the follow-up of patients undergoing lung transplantation [49,50]. A recent study indicates that plasma BNP levels do not correlate with left ventricular function after cardiac surgery but do correlate with the E/E′ ratio, which is an echocardiographic indicator of filling pressures and diastolic function [51].
Noncardiac Surgery
In noncardiac surgery, many investigators have confirmed that either preoperative BNP [52–56] or NT-proBNP [57–60] are helpful in predicting outcomes (Fig. 1). A clear example comes from Breidhardt and colleagues [61], who showed that preoperative increase of BNP predicts major perioperative cardiovascular complications in patients undergoing orthopedic surgery, and that the combination of BNP and the American Society of Anesthesiologists score is superior to BNP alone in predicting in-hospital cardiac events.
In light of these studies, both BNP and NT-proBNP are promising markers for postoperative cardiac outcomes. The current evidence does not support a universal cutoff point for BNP or NT-proBNP levels beyond which poor outcomes are anticipated, because existing studies have used different assays and various statistical methods to obtain their cutoffs; moreover, some did not indicate how they obtained their cutoff [62]. The specified cutoff points have varied between 40 and 189 pg/mL for BNP [52–56] and between 280 and 533 pg/mL for NT-proBNP [57–60]. Additional investigations are needed in order for consensus to be reached on what constitutes increased and predictive levels of both markers. Interventions should be sought that reduce concentrations of these 2 markers, and trials designed to assess the effects of these interventions on surgical outcomes.
Cardiac troponins
Troponins are proteins of the contractile system of the cardiac cells that act together through the tropomyosin complex to regulate muscle contraction [63]. Cardiac troponins have become the biomarker of choice for the diagnosis of acute myocardial ischemia and infarction [64]. Cardiac troponin T (cTnT) levels are detectable 3 to 12 hours after myocardial injury, and the concentration is in direct proportion to the extent of myocardial injury [65,67].
Cardiac troponins have also been used in the perioperative period to evaluate myocardial ischemia and infarction. Early studies considered them the best tool for diagnosing PMI in patients having noncardiac and cardiac surgery. Several investigators have shown that increased preoperative and postoperative levels of cTnI are associated with poor postoperative clinical outcomes [68]. However, at present, no single cutoff value has been recommended for use in the diagnosis of PMI. Moreover, cTnI may increase even before aortic cross-clamping and cardiac manipulation [69]. Martin and colleagues [70] suggested that cTnI levels lower than 0.15 μg/L were not associated with myocardial ischemia in the perioperative period of cardiac surgery. In contrast, cTnI levels higher than 0.15 μg/L were a strong indicator of ischemic damage.
Thielmann and colleagues [71] showed that preoperative cTnI levels higher than 1.15 ng/mL were associated with longer ICU stay and with higher in-hospital mortality. Moreover, levels of 0.15 ng/mL or higher were associated with higher morbidity and mortality 6 months after surgery [72]. In a study conducted by Lehrke and colleagues [73], cTnT measurement 48 hours after surgery was a strong predictor of severe cardiac failure and high postoperative mortality. A cTnT value exceeding 0.46 μg/L 48 hours after surgery carried a 6.7-fold risk of cardiac death.
Cardiac troponins have also been used to identify clinical outcomes in patients undergoing major vascular surgery [74]. Early studies by Lee and colleagues [75] and Meltzer and colleagues [76] suggested an association between increased levels of cTn and postoperative outcomes. More recently, Landesberg and colleagues [77] showed that levels of cTnI greater than 0.6 ng/mL and cTnT greater than 0.03 ng/mL were independently associated with a respectively 2.15-fold and 1.89-fold increase in mortality. The investigators of the Coronary Artery Revascularization Prophylaxis (CARP) trial used cTnI to define PMI. They found that cTnI levels of 0.1 μg/L or higher occurred in 27% of patients undergoing elective vascular surgery and that these levels were a strong predictor of long-term risk of death among patients with diabetes [78]. Preoperative concentrations of cTnI were predictive of postoperative cardiac outcomes in major noncardiac surgery patients and their predictive power was mildly improved by adding preoperative BNP concentration into the model (Fig. 2).
In agreement with these last 2 studies, Filipovic and colleagues [79] found an association between cTnI levels and all-cause mortality 1 year after major noncardiac surgery, but a subsequent study by the same group of investigators could not show this association [80]. Controversies surround the validity of cTnT in patients with impaired renal function. However, Feringa and colleagues [81] showed that, after adjustment for estimate of the glomerular filtration rate, minor increases in troponin T from 0.03 to 0.09 ng/mL were strongly associated with late mortality and major adverse cardiac events after major vascular surgery.
Procalcitonin
Procalcitonin (PCT) is a 116-amino acid peptide produced by the c-cells in the thyroid. Its concentration in the serum of healthy individuals is low (<0.1 ng/mL); however, it may be considered an acute-phase reactant because of its production and release during inflammatory stress [63,82]. Traditionally identified as a marker for infectious conditions, PCT at increased levels has been shown more recently to be associated with noninfectious processes such as trauma, surgery, cardiogenic shock, burns, heat stroke, acute respiratory distress syndrome, and rejection after transplantation [83–87]. In these conditions it is produced outside the thyroid by the liver and circulating mononuclear cells.
Cardiac Surgery
Procalcitonin has been investigated in the context of cardiac surgery. It is not clear whether the type of cardiac surgery influences the surge of PCT after surgery. Prat and colleagues [88] could not identify differences in PCT levels between patients undergoing coronary artery bypass graft (CABG) and open heart valvular surgery. In contrast, Sponholz and colleagues [89] and Franke and colleagues [90] reported that valvular surgery induces larger increments of PCT levels than on-pump CABG, and this in turn induces larger increases than off-pump CABG.
Procalcitonin seems to be predictive of poor outcomes after cardiac surgery. An early study by Loebe and colleagues [91] showed that an increased PCT level, higher than 1 ng/mL, after cardiac surgery was predictive of poor postoperative outcomes. Two recent studies showed that procalcitonin significantly increased in patients who underwent cardiac surgery and subsequently developed infectious and noninfectious complications, compared with patients who did not experience complications [88,92]. Procalcitonin proved more accurate than CRP in predicting those complications [92]. In heart transplantation, PCT is also a reliable marker in the diagnosis and monitoring of postoperative infections [93,94]. Moreover, PCT levels were not affected by the use of immunosuppressants [94,95].
Noncardiac Surgery
Schneider and colleagues [96] retrospectively analyzed the association of PCT concentration with postoperative mortality, morbidity, and length of stay in 220 patients who were admitted after surgery to an ICU. The researchers found a significant and logarithmic association between PCT concentration and outcome. Moreover, PCT was an independent predictor of mortality and duration of hospital stay after surgery in the survivors. In patients undergoing oncologic surgery, PCT seems to be a more useful marker than CRP for monitoring the postoperative course and diagnosing severe bacterial infections [97,98]. Procalcitonin is also superior to high sensitivity CRP (hsCRP) as an independent predictor of graft failure in renal transplantation [99]. However, Mommertz and colleagues [100] found no association between perioperative neurologic deficit and PCT in patients who had carotid endarterectomy.
Biomarkers of kidney dysfunction
Acute kidney injury (AKI) remains a common postoperative complication. The incidence of dialysis-dependent acute renal failure after cardiac surgery is approximately 1%, and approximately half of these patients die from this complication [101,103]. Early detection of AKI in the perioperative period may permit early renoprotective interventions, which may in turn translate into more favorable postoperative outcomes; thus, simple and specific biomarkers would be useful to monitor AKI.
Cystatin C
Cystatin C is a 13-kDa endogenous cysteine proteinase inhibitor that is synthesized at a constant rate and released into the plasma by all nucleated cells in the body. It is freely filtered at the glomerulus, not secreted or reabsorbed, and nearly completely catabolized by proximal renal tubular cells. Cystatin C can be measured in plasma and urine [104]. Cystatin C has a lower variability of measurements, a shorter half-life, and a lower distribution volume than creatinine [105]. Because of these properties, it has been suggested that it may be more sensitive than creatinine to early and mild changes in kidney function [106,108]. However, factors such as age, gender, and body mass index may still affect the serum levels of this protein [104,109,110]. A recent study by Villa and colleagues [111] showed that cystatin C correlated better with glomerular filtration rate than creatinine in patients in ICUs with mild kidney dysfunction. Moreover, a meta-analysis conducted by Dharnidharka and colleagues [106] indicated that the serum cystatin C concentration was a more sensitive indicator of renal dysfunction than serum creatinine.
Cardiac surgery
Several investigators have studied the clinical value of cystatin C in identifying patients with, or at risk for, renal dysfunction after cardiac surgery [112,113]. In a cohort study of 110 elderly patients having cardiac surgery requiring cardiopulmonary bypass, serum cystatin C could not detect mild renal injury earlier than plasma creatinine [114]. Similar results were reported by Liangos and colleagues [115], who found that cystatin C levels 2 hours after cardiopulmonary bypass could not predict AKI. However, these results are disputed by a recent study demonstrating that serum cystatin C predicted AKI about 2 days earlier than clinically significant increases in creatinine [116]. Increased levels of urinary cystatin C have also been associated with early detection of kidney dysfunction after cardiac surgery [117]. In a prospective study of 376 patients, Ledoux and colleagues [113] found that preoperative estimation of renal function from serum cystatin C was strongly associated with hospital morbidity and mortality.
Noncardiac surgery
In contrast with cardiac surgery, cystatin C seems to correlate better with kidney dysfunction in the setting of noncardiac surgery. Lebkowska and colleagues [118] showed that patients with delayed graft function after cadaveric renal transplantation showed significantly higher levels of cystatin C than those with normal graft function. Cystatin C correlated with perioperative levels of neutrophil gelatinase–associated lipocalin (NGAL) and creatinine. Moreover, cystatin C has also been used to detect early kidney dysfunction in living-related donor kidney transplantation. Gourishankar and colleagues [119] found that cystatin C correlated with all other markers of kidney function and detected acute changes in kidney function immediately after donor nephrectomy.
Cystatin C has been used clinically to monitor the perioperative effect of nephrotoxic agents such as cyclooxygenase inhibitors or the effect of a variety of perioperative interventions on kidney function. In a randomized, controlled, double-blinded trial, Puolakka and colleagues [120] found that cystatin C levels were not affected by the administration of parecoxib in patients undergoing laparoscopic hysterectomy, and accordingly the researchers suggested that 80 mg of this analgesic was not associated with perioperative kidney dysfunction. Cystatin C has also been used to evaluate the effects of perioperative fluid optimization in patients undergoing surgery [121]. Boldt and colleagues [122] assessed the effect of different kinds of colloid therapy (albumin vs hydroxyethylstarch) on patients undergoing cardiac surgery and found that differences in the increase of cystatin C after surgery did not reach statistical significance between the 2 groups being studied.
Cystatin C has also been used to investigate the effects of sevoflurane on kidney function. Laisalmi and colleagues [123] showed in a double-blinded controlled study that cystatin C levels did not differ in healthy women undergoing breast surgery who received ketorolac during high fresh gas flow of sevoflurane anesthesia. They concluded that, in healthy women, the use of ketorolac is not associated with kidney injury in patients having general anesthesia with sevoflurane.
Neutrophil Gelatinase-associated Lipocalin
NGAL, a 25-kDa protein that belongs to the superfamily of lipocalins, has been measured in plasma and urine. NGAL binds to 2 receptors, the megalin multiscavenger complex and the 24p3 receptors. It has been suggested that NGAL also acts as an acute-phase reactant that is increased not only during the infective process but also during noninfective systemic diseases. The kidney can release NGAL in response to damage or stress, and thus serum or urinary NGAL can be used to detect AKI [124,125]. It has also been suggested that serum and urinary NGAL may be useful in distinguishing between septic and nonseptic AKI. Bagshaw and colleagues [126] found that septic kidney injury was associated with higher initial and peak values of urinary and serum NGAL than nonseptic kidney dysfunction. Several investigators have also studied the predictive value of NGAL in the perioperative setting, as discussed later [127–130].
Cardiac surgery
Urinary and serum levels of NGAL were increased in patients with postoperative kidney injury after cardiac surgery [131,132]. In addition, a study by Che and colleagues [133] found that urinary NGAL was highly predictive of kidney dysfunction. Its predictive value was higher than that of other markers such as cystatin C, urinary interleukin 18, serum creatinine and N-acetyl-β-d-glucosaminidase. These results were partially replicated by McIlroy and colleagues [134], who observed that urinary NGAL detected AKI only in patients with a preoperative estimated glomerular filtration rates of 90 to 120 mL/min.
