Chapter 4. The Role of Genomics in Anesthesia Practice

1. Hundreds of genome-wide association studies have been performed in recent years to identify common variants that are associated with complex disease. Although there have been some notable success stories, overall common variation has explained little of the high heritability of these traits. It is likely that the use of whole-genome sequencing to extend the study of rare variation in complex diseases will greatly advance our understanding of perioperative biology.

2. Preliminary candidate gene studies suggest that susceptibility to adverse perioperative events, including cardiac (myocardial infarction [MI], ventricular dysfunction, atrial fibrillation), neurologic, and renal, among others, is genetically determined.

3. Potential applications of biomarkers in perioperative medicine and critical care include prognosis, diagnosis, and monitoring of adverse outcomes, as well as informing and refining therapeutic decisions. Very few so far have been rigorously evaluated to demonstrate additive performance to existing risk stratification models (clinical validity) or change therapy (clinical utility). Most promising among those are natriuretic peptides and C-reactive protein for cardiovascular risk prediction, and procalcitonin to assess infection in the critically ill.

DNA technology has already changed our health care, the food we eat, and our criminal justice system. Underlying these profound societal implications is the process of elucidating the relationship between specific differences encoded in our DNA (ie, genotypes) and our interindividual variability in morphology, development, behavior, athletic prowess, physiology, and disease susceptibility (ie, phenotypes). Many common diseases like atherosclerosis, coronary artery disease, hypertension, diabetes, cancer, asthma, and our responses to stress, injury, drugs, and nonpharmacologic therapies are genetically complex, characteristically involving interplay of many genetic variations in molecular and biochemical pathways (ie, polygenic) and interactions between genes and environment (ie, multifactorial). In other words, complex phenotypes can be viewed as the integrated effect of many susceptibility genes and many environmental exposures. The proportion of phenotypic variance explained by genetic factors is referred to as heritability and can be estimated by examining the increased similarity of a phenotype in related as compared with unrelated individuals.

In 2003 the 50th anniversary of Watson and Crick's description of the DNA double-helix structure also marked the completion of the Human Genome Project.1 This major accomplishment has provided the discipline of genomics with basic resources to study the functions of all genes in a systematic fashion, including their interaction with environmental factors, and to translate the findings into clinical and societal benefits. Functional genomics uses large-scale experimental methodologies and statistical analyses to investigate the regulation of gene expression in response to physiologic, pharmacologic, and pathologic changes. It also uses genetic information from clinical studies to examine the impact of genetic variability on disease characterization and outcomes. One of the major challenges and ongoing research efforts facing the postgenomic period is to connect the nearly 26,000 protein-coding genes of mammalian organisms to the genetic basis of complex polygenic diseases and the integrated function of complex biologic systems. The rapidly evolving field of genomic medicine proposes to use genomic information to assist medical decision making and tailor health care to the individual patient.

A breathtaking acceleration on the path from the Human Genome Project to genomic medicine has been paved by landmark "big biology" efforts concentrated on the main themes of uncovering the function of human genes, identifying all human genomic variation, and elucidating the genetic basis for human diseases (Table 4-1). Furthermore, dramatic improvements in sequencing technologies and cost reductions have fueled even more ambitious initiatives such as the ongoing Personal Genome Project,2 which aims to publish the complete genomes and medical records of a number of volunteers to enable personalized medicine research. Nevertheless, it is fair to say a decade after its completion, that although the Human Genome Project has revolutionized the pace and scope of biomedical science, it has failed so far to fulfill the promise and grand vision of "personalized medicine," leaving the scientific community somewhat sobered and divided.

The perioperative period represents a unique and extreme example of gene–environment interaction. As we appreciate in our daily practice in the operating rooms and intensive care units, one hallmark of perioperative physiology is the striking variability in patient responses to the acute, robust, and systemic perturbations induced by surgical injury, hemodynamic challenges, vascular cannulation, extracorporeal circulation, intra-aortic balloon counterpulsation, mechanical ventilation, partial/total organ resection, transient limb/organ ischemia, transfusions, anesthetic agents, and the pharmacopoeia used in the perioperative period. This translates into substantial interindividual variability in the incidence or severity of immediate perioperative adverse events, as well as long-term outcomes (Table 4-2). For decades we have attributed this variability to many complexities such as age, nutritional state, comorbidities—what we colloquially call "protoplasm." Now we are beginning to appreciate that genetic variation is also in part responsible for this observed variability in outcomes. Overall, an individual's genetic susceptibility to adverse perioperative events stems not only from genetic contributions to the development of comorbid risk factors (like coronary artery disease [CAD] and reduced preoperative cardiopulmonary reserve) during his or her lifetime but also from genetic variability in specific biologic pathways participating in the host response surgery (Fig. 4-1). With increasing evidence suggesting that genetic variation can significantly modulate risk of adverse perioperative events,3-6 the emerging field of perioperative genomics aims to apply functional genomic approaches to discover underlying biologic mechanisms that explain why similar patients have such dramatically different outcomes after surgery. It is justified by a unique combination of environmental insults and postoperative phenotypes that characterize surgical and critically ill patient populations. To take full advantage of the unique opportunities offered by the genomic revolution, the cycle of innovation in perioperative medicine must include a comprehensive and standardized definition of the phenotypes of interest (including short- and long-term adverse outcomes such as organ injury/dysfunction, adverse drug responses, transition to chronic pain), followed by identification of the underlying genes, characterization of the mechanism from DNA to phenotype, and rigorous development and validation of diagnostic or therapeutic applications ultimately having an impact on our patients and families.

To integrate this new generation of molecular tests into clinical practice, perioperative physicians need to become familiar with several key concepts, including patterns of human genome variation, gene regulation, basic population genetic methodology, gene and protein expression analysis, and, most importantly, the general principles for evaluating biomarker performance. This chapter serves as a primer in genomic medicine by highlighting the rapidly evolving current and future applications of molecular (including genomic) technologies for perioperative risk stratification, outcome prediction, mechanistic understanding of surgical stress responses, as well as identification and validation of novel targets for perioperative organ protection.

Human Genomic Variation

In elucidating the genetic basis of disease, much of what was investigated before the Human Genome Project era focused on identifying rare genetic variants (mutations) responsible for >1500 monogenic disorders such as hypertrophic cardiomyopathy, long-QT syndrome, sickle cell anemia, cystic fibrosis, or familial hypercholesterolemia, which are highly penetrant (carriers of the mutant gene will likely have the disease) and inherited in Mendelian fashion (hence termed Mendelian diseases). However, most of the genetic diversity in the population is attributable to more widespread DNA sequence variations (polymorphisms), typically single nucleotide base substitutions (single nucleotide polymorphisms [SNPs]), or to a broader category of previously overlooked structural genetic variants that include short sequence repeats (microsatellites), insertion/deletion of one or more nucleotides (indels), inversions, and the recently discovered copy number variants (CNVs; large segments of DNA that vary in number of copies),7 all of which may or may not be associated with a specific phenotype (Fig. 4-2). The first published drafts of the human genome were of haploid genomes, and they identified only a small amount of variation (0.1%) between individuals. However, with the completion of the first diploid genome sequence of an individual human,8 it became apparent that the two parental genomes surprisingly differed from each other by 0.5% when insertion and deletions were included along with SNPs. It was subsequently discovered that the genomes of different individuals differ by 1% to 3%.

To be classified as a polymorphism, the DNA sequence alternatives (ie, alleles) must exist with a frequency of at least 1% in the population. About 15 million SNPs are estimated to exist in the human genome, approximately once every 300 base pairs, located in genes as well as in the surrounding regions of the genome. Polymorphisms may directly alter the amino acid sequence and therefore potentially alter protein function or alter regulatory DNA sequences that modulate protein expression. Sets of nearby SNPs on a chromosome are inherited in blocks, referred to as haplotypes. As shown later, haplotype analysis is a useful way of applying genotype information in disease gene discovery. However, CNVs involve approximately 12% of the human genome, often encompass genes (especially regulating inflammation and brain development), and may influence disease susceptibility through dosage imbalances. In this chapter, we review the common strategies used to incorporate genetic analysis into clinical studies.

Basic Genetic Epidemiology

Most ongoing research on complex disorders focuses on identifying genetic variants that alter susceptibility to given conditions. Often the design of such studies is complicated by the presence of multiple risk factors, gene–environment interactions, and a lack of even rough estimates of the number of genes underlying such complex traits. Two broad strategies have been used to identify complex trait loci. The candidate gene approach is motivated by what is known about the trait biologically and can be characterized as a hypothesis-testing approach but is intrinsically biased. The second strategy is the genome-wide scan, in which thousands of markers uniformly distributed throughout the genome are used to locate regions that may harbor genes influencing phenotypic variability. This is a hypothesis-free and unbiased approach, in the sense that no prior assumptions are being made about the biologic processes involved and no weight is given to known genes, thus allowing the detection of previously unknown trait loci. Both the candidate gene and genome scan approaches can be implemented using one of two fundamental methods of identifying polymorphisms affecting common diseases: linkage analysis or association studies in human populations. Importantly, the ability of either one of these mapping strategies to discover genetic susceptibility for complex traits depends on the joint contributions of two independent factors: the frequency of the risk allele in the population and the magnitude of its effect.

Linkage analysis maps the chromosomal location of gene variants related to a given disease by studying the distribution of disease alleles in affected individuals throughout a pedigree. Traditionally, family-based linkage studies have been successful at identifying rare genetic variants of large effects involved in rare monogenic (ie, classical Mendelian) disorders. The approach is less effective for common complex diseases, however, which are characterized by a multitude of genes with rare and/or common alleles creating an apparently chaotic pattern of heterogeneity within and between families. Nevertheless, a few positive findings have emerged, including susceptibility loci for stroke (chromosome 5q12),9 myocardial infarction (chromosome 14),10 and pulse pressure variation, an emerging risk factor for perioperative adverse outcomes (chromosomes 22 and 10).11 Furthermore, the nature of most complex diseases, in general, and perioperative adverse outcomes in elderly patients, in particular, makes the study of extended multigenerational family pedigrees impractical (with few exceptions, eg, malignant hyperthermia), due to the lack of availability of pedigree information and/or DNA samples. As such, most genetic susceptibility variants have been identified so far through population studies of association and are relatively frequent in the population but have individually small effects.

Genetic association studies examine the frequency of specific genetic polymorphisms in a population-based sample of unrelated diseased individuals and appropriately matched unaffected controls. As previously discussed, the increased statistical power to uncover small clinical effects of multiple genes12 and the fact that they do not require family-based sample collections are main advantages of this approach over linkage analysis. Until recently, most significant results were gathered from candidate gene association studies, with genes selected because of a priori hypotheses about their potential etiologic role in disease based on current understanding of the disease pathophysiology.13 For example, genetic variants within the renin-angiotensin system, nitric oxide synthase, and β2-adrenergic receptors, known to modulate vascular tone, were tested and found to be associated with hypertension. Similarly, the possible effects of polymorphisms on genetic predisposition for CAD14 or restenosis after angioplasty15 have been extensively investigated; more recently, two large-scale association studies identified gene variants that might affect susceptibility to myocardial infarction.16 As presented in more detail later, accumulating evidence from candidate gene association studies also suggests that specific genotypes are associated with a variety of organ-specific perioperative adverse outcomes, including myocardial infarction,17,18 neurocognitive dysfunction,19-21 renal compromise,22-24 vein graft restenosis,25,26 postoperative thrombosis,27 vascular reactivity,28 severe sepsis,29,30 transplant rejection,31 and death (for reviews, see 3,6).

One of the main weaknesses of the genetic association approach is that, unless the marker of interest "travels" (ie, is in linkage disequilibrium) with a functional variant or the marker allele is the actual functional variant, the power to detect and map complex trait loci will be reduced. Other known limitations of genetic association studies include potential false-positive findings resulting from population stratification (ie, admixture of different ethnic or genetic backgrounds in the case and control groups), and multiple comparison issues when large numbers of candidate genes are being assessed.32 Replication of findings across different populations or related phenotypes remains the most reliable method of validating a true relationship between genetic polymorphisms and disease,13 but poor reproducibility in subsequent studies has been one of the main criticisms of the candidate gene association approach.33 However, a recent meta-analysis suggested that lack of statistical power may be the main contributor to this inconsistent replication, and it proposed more stringent statistical criteria to exclude false-positive results and the design of large collaborative association studies.34

At last, after several decades of frustrating limitations in the ability to find genetic variations responsible for common disease risk, with the completion of the International HapMap Project (a high-resolution map of human genetic variation and haplotypes)35 and advances in high-throughput genotyping technologies, the past 3 years have marked an explosion of adequately powered and successfully replicated genome-wide association studies (GWAS) that identified very significant genetic contributors to risk for common polygenic diseases like CAD,36-38 MI,39 diabetes (type 1 and 2),40,41 atrial fibrillation,42 obesity, asthma, common cancers, rheumatoid arthritis, Crohn disease, and others. At the time of this publication, the NHGRI GWAS catalog included 782 publications and 3902 associations, with the list growing every day.43 GWAS make use of the known linkage disequilibrium pattern between SNPs from the human HapMap and the new high-density SNP chip technology to interrogate comprehensively and accurately between 65% and 80% of common variation across the genome, with even higher coverage being possible using statistical imputation techniques. One of the most comprehensive GWAS to date was conducted by the Wellcome Trust Case-Control Consortium, investigating the association between 500,000 SNPs and seven common diseases in 2000 cases and 3000 shared controls. It identified 25 independent association signals at stringent levels of significance (p < 5 × 10−7).36 Interestingly, variants in or near CDKN2A/B (cyclin-dependent kinase inhibitor 2 A/B) conferred increased risk for both type 2 diabetes (odds ratio [OR]: 1.2; p = 7.8 × 10−15) and MI (OR: 1.64, p = 1.2 × 10−20), which may lead to a mechanistic explanation for the link between the two disorders. This finding also highlights the power of GWAS to identify variants outside described genes: While one of the signals occurs in the CDKN2A/B region, the other much stronger association signal occurs more than 200 kb from these genes, in a gene desert, and thus would not have been picked up by a candidate gene approach. Identifying the mechanism by which this variant may affect CDKN2A/B expression will provide new insights into the regulation of these genes. To maximize the power of identifying novel susceptibility loci for CAD and MI, even larger consortia have been assembled, like the Coronary ARtery DIsease Genome-wide Replication And Meta-analysis (CARDIoGRAM),44 which includes 22,000 cases and 60,000 controls. It has already successfully mapped 27 new loci.

An important theme has emerged from GWAS: Most SNPs associated with common diseases collectively explain only a small proportion of the observed contribution of heredity to the risk of disease (eg, 6% for type 2 diabetes) or other complex traits (eg, 2% for body mass index and 5% for height). This missing heritability problem limits substantially the potential application of genomic markers to risk prediction, and it remains an important obstacle to overcome to develop evidence-based guidelines recommending the use of SNPs in clinical care. Two broad explanations have been proposed and considerable resources invested in discovering the unknown sources of heritable risk, which are briefly reviewed here.

First, the underlying rationale for GWAS is the "common disease-common variant" hypothesis, which postulates that common diseases may arise secondary to cumulative effects of common variants. Most genes, however, lack a common functional coding variant with a detectable functional effect, yet they typically contain several rare variants. A counterhypothesis has emerged stating there are additional novel genes harboring such low-frequency variants (possibly with larger effects) that may be the primary drivers of common disease. Currently these variants are poorly detected by genotyping microarrays, but with the advent of sequencing technologies, the potential exists to revolutionize complex traits genetics by identifying and typing rare variants and thus rendering virtually every gene susceptible for genetic analysis.

The second theme emerging from GWAS that may offer clues into the missing heritability problem is that most variants identified so far are located either in intergenic regions or in genes of unknown function. This among other findings has challenged the very concept of the "gene" as the traditional unit of heredity. Discovery of the diverse and ubiquitous roles of new classes of RNA (including microRNAs and short interfering RNAs) led to an emerging picture of gene regulation as interdependent layers of control consisting of interactions of DNA with regulatory proteins and RNA (Fig. 4-3).45 Furthermore, on top of the DNA sequence lies another (epigenetic) code that influences when and what genes should be transcribed or silenced. The epigenome is laid down during prenatal and postnatal development, and it is heritable through cell divisions. Given previous studies implicating changes in DNA methylation patterns or histone modification (the most commonly studied epigenetic marks), for instance as prognostic biomarkers in various cancers, one can assume that epigenetics will soon find a role in the perioperative management of cancer patients.

As discussed previously, both family-based linkage studies and genetic association studies depend fundamentally on the concept of linkage for their practical execution (ie, the identified variant is in linkage disequilibrium with, but not necessarily the, causal variant). Newer ("next-generation") approaches based on direct whole-genome sequencing depart from the concept of linkage by attempting to directly identify causal alleles.46 One particular application of genome-scale sequencing is called whole-exome sequencing. The exome, defined as the protein-coding portion of the genome, is composed of approximately 30 megabases (1%) split among about 200,000 exons in humans. Aside from the ability to identify rare variants and the obvious substantial cost reduction (~20-fold), this approach has the advantage of focusing on nonsynonymous variants in coding genes for which there are well-established methods of functional validation and interpretation of biologic effects, thus enabling their implication as causal variants. However, it completely misses noncoding and structural variation in the genome. Early results suggest that whole-exome sequencing is an effective approach to identify causal mutations for monogenic disorders but also to distinguish signal (causal rare variants) from noise (background rate of rare mutations) for complex traits. Successful studies so far (eg, early-onset MI) sequenced individuals were informed by the following key observations: The younger the age when developing MI, the greater the heritability; selecting extremes of the phenotype distribution (eg, young with MI versus old without MI as a "hypernormal" control group) is likely to improve power; that genetic discovery may be enhanced by studying multiple ethnicities. These studies have demanded the development of novel statistical methods to associate rare variants with the phenotype. One such promising solution for overcoming the statistical challenges revolving around sequencing low-frequency variants is to combine all nonsynonymous SNPs (by gene or biologic pathway) into a single statistical test. The first integrated analysis of a complete human genome in a clinical context, in a patient with a family history of vascular disease and early sudden death, was reported in the Lancet in 2010.47 The analysis revealed increased genetic risk for CAD, MI, type 2 diabetes, and some cancers, as well as rare variants in three genes clinically associated with sudden cardiac death. Furthermore, the patient had variants associated with clopidogrel resistance, a positive response to lipid-lowering therapies, and a low initial dosing requirement for warfarin, suggesting that routine whole-genome sequencing can yield clinically relevant information for individual patients.

A further advantage of high-quality genome sequence generated by the next-generation sequencing technologies stems from the ability to assemble the sequence data independently and sort it into the two sets of parental chromosomes in a process termed haplotype phasing. Establishing the complete set of genetic information that we received from each parent is crucial to understanding the links between heritability, gene function, regulatory sequences, and our predisposition to disease.48 Several additional next-generation technologies involve sequencing of multiple genomes per person, such as matched tumor and blood DNA samples from 20 common types of cancer in the Cancer Genome Atlas, which enable development of targeted therapeutics based on a detailed molecular understanding of pathogenesis. Equally important for medical progress is the sequencing of genomes of the billions of microorganisms that dwell within us as part of the Human Microbiome Project.

Biomarker Discovery: Profiling Gene/Protein Expression and Metabolites

Genomic approaches are anchored in the "central dogma" of molecular biology, the concept of transcription of messenger RNA (mRNA) from a DNA template, followed by translation of RNA into protein (Fig. 4-3). Because transcription is a key regulatory step that may eventually signal many other cascades of events, the study of RNA levels in a cell or organ (ie, quantifying gene expression) can improve the understanding of a wide variety of biologic systems. Furthermore, although the human genome contains only about 26,000 genes, functional variability at the protein level is far more diverse, resulting from extensive posttranscriptional, translational, and posttranslational modifications. It is believed that there are approximately 200,000 distinct proteins in humans, which are further modified posttranslationally by phosphorylation, glycosylation, oxidation, and disulfide structures. There is increasing evidence that variability in gene expression levels underlies complex disease and is determined by regulatory DNA polymorphisms affecting transcription, splicing, and translation efficiency in a tissue- and stimulus-specific manner.49 Thus, in addition to the assessment of genetic variability at the DNA sequence level using various genotyping techniques as described in previous sections (static genomics), analysis of large-scale variability in the pattern of RNA and protein expression both at baseline and in response to the multidimensional perioperative stimuli (dynamic genomics) using microarray and proteomic approaches provides a much needed complementary understanding of the overall regulatory networks involved in the pathophysiology of adverse postoperative outcomes. Such dynamic genomic markers can be incorporated in genomic classifiers and used clinically to improve perioperative risk stratification or monitor postoperative recovery.50 This concept of molecular classification involves the description of informational features in a training data set using changes in relative RNA and protein abundance in the context of genetic predisposition and applying to a test data set to recognize a defined "fingerprint" characteristic of a particular perioperative phenotype (Table 4-3). For example, Feezor et al used a combined genomic and proteomic approach to identify expression patterns of 138 genes from peripheral blood leukocytes and the concentrations of 7 circulating plasma proteins that discriminated patients who developed multiple organ dysfunction syndrome (MODS) after thoracoabdominal aortic aneurysm repair from those who did not. Importantly, these patterns of genome-wide gene expression and plasma protein concentration were observed before surgical trauma and visceral ischemia-reperfusion injury, suggesting that patients who developed MODS differed in either their genetic predisposition or their preexisting inflammatory state.51

Alternatively, dynamic genomic markers can be used to improve mechanistic understanding of perioperative stress, and to evaluate and catalog organ-specific responses to surgical stress and severe systemic stimuli such as cardiopulmonary bypass (CPB) and endotoxemia, which can be subsequently used to identify and validate novel targets for organ-protective strategies.52 Using a similar integrated approach of transcriptomic and proteomic analyses, Tomic et al53 characterized the molecular response signatures in peripheral blood to cardiac surgery with and without CPB, a robust trigger of systemic inflammation. The authors demonstrated that, rather than being the primary source of serum cytokines, peripheral blood leukocytes only assume a "primed" phenotype upon contact with the extracorporeal circuit that facilitate their trapping and subsequent tissue-associated inflammatory response. Interestingly, many inflammatory mediators achieved similar systemic levels following off-pump surgery but with delayed kinetics, offering novel insights into the concepts of contact activation and compartmentalization of inflammatory responses to major surgery.

Several studies have profiled myocardial gene expression in the ischemic heart, demonstrating alterations in the expression of immediate-early genes (c-fos, junB), as well as genes coding for calcium-handling proteins (calsequestrin, phospholamban), extracellular matrix, and cytoskeletal proteins.54 Upregulation of transcripts mechanistically involved in cytoprotection (heat shock proteins), resistance to apoptosis, and cell growth has been found in stunned myocardium.55 Moreover, cardiac gene expression profiling after CPB and cardioplegic arrest has identified the upregulation of inflammatory and transcription activators, apoptotic genes, and stress genes,56 which appear to be age related.57 Microarray technology has also been used in the quest for novel cardioprotective genes, with the ultimate goal of designing strategies to activate these genes and prevent myocardial injury. Preconditioning is one of such well-studied models of cardioprotection, which can be induced by various triggers, including intermittent ischemia, osmotic or redox stress, heat shock, toxins, and inhaled anesthetics. The main functional categories of genes identified as potentially involved in cardioprotective pathways include a host of transcription factors, heat shock proteins, antioxidant genes (heme-oxygenase, glutathione peroxidase), and growth factors, but different gene programs appear to be activated in ischemic versus anesthetic preconditioning, resulting in two distinct cardioprotective phenotypes.58 More recently, a transcriptional response pattern consistent with late preconditioning was reported in peripheral blood leukocytes following sevoflurane administration in healthy volunteers, characterized by reduced expression of L-selectin as well as downregulation of genes involved in fatty acid oxidation and the peroxisome activated receptor gamma coactivator (PCG)1α pathway,59 which mirrors changes observed in the myocardium from patients undergoing off-pump coronary artery bypass grafting (CABG) (Table 4-3).60 Deregulation of these novel survival pathways thus appears to generalize across tissues, making them important targets for cardioprotection, but further studies are needed to correlate perioperative gene expression response patterns in end organs such as the myocardium to those in readily available surrogate tissues such as peripheral blood leukocytes.

The transcriptome (the complete collection of transcribed elements of the genome) is not fully representative of the proteome (the complete complement of proteins encoded by the genome) because many transcripts are not targeted for translation, as evidenced recently with the concept of gene silencing by RNA interference. Alternative splicing, a wide variety of posttranslational modifications, and protein–protein interactions responsible for biologic function, would remain therefore undetected by gene expression profiling (Fig. 4-3). This has led to the emergence of a new field, proteomics, studying the sequence, modification, and function of many proteins in a biologic system at a given time. Rather than focusing on "static" DNA, proteomic studies examine dynamic protein products, with the goal of identifying proteins that undergo changes in abundance, modification, or localization in response to a particular disease state, trauma, stress, or therapeutic intervention (for a review, see 61). Thus proteomics offers a more global and integrated view of biology, complementing other functional genomic approaches. Several preclinical proteomic studies relevant to perioperative medicine have characterized the temporal changes in brain protein expression in response to various inhaled anesthetics62,63 or following cardiac surgery with hypothermic circulatory arrest.64 This may focus further studies aimed to identify new anesthetic binding sites and the development of neuroprotective strategies. Furthermore, detailed knowledge of the plasma proteome has profound implications in perioperative transfusion medicine,65 in particular related to peptide and protein changes that occur during storage of blood products. The development of protein arrays and real-time proteomic analysis technologies has the potential to allow the use of these versatile and rigorous high-throughput methods for clinical applications and is the object of intense investigation.

Emerging metabolomic tools have created the opportunity to establish metabolic signatures of myocardial injury. In a population of patients undergoing alcohol septal ablation for hypertrophic obstructive cardiomyopathy, a human model of planned (albeit chemical) MI that recapitulates spontaneous myocardial infarction targeted mass spectrometry-based metabolite profiling identified changes in circulating levels of metabolites participating in pyrimidine metabolism, the TCA cycle, and the pentose phosphate pathway as early as 10 minutes after MI in an initial derivation group and were validated in a second, independent group. Coronary sinus sampling distinguished cardiac-derived from peripheral metabolic changes. To assess generalizability, the planned MI-derived metabolic signature (consisting of aconitic acid, hypoxanthine, trimethylamine N-oxide, and threonine) differentiated with high-accuracy patients with spontaneous MI.66 We applied a similar approach to cardiac surgical patients undergoing planned global myocardial ischemia/reperfusion (I/R) and identified clear differences in metabolic fuel uptake based on the preexisting ventricular state (left ventricular dysfunction, CAD, or neither) as well as altered metabolic signatures predictive of postoperative hemodynamic course and perioperative MI.67 While simultaneous assessment of coronary sinus effluent in addition to the peripheral blood improves cardiac specificity of the observed signatures, direct measurements of metabolites in myocardial tissue allow marked enrichment and easier detection of potential biomarkers compared with plasma, as well as an assessment of how metabolic substrates are used in the tissue of interest. Such studies are possible in cardiac surgical patients where atrial tissues are routinely removed; for example, one study using high-resolution 1H-NMR spectroscopy identified alterations in myocardial ketone metabolism associated with persistent atrial fibrillation, and the ratio of glycolytic end products to end products of lipid metabolism correlated positively with time of onset of postoperative atrial fibrillation.68

More than 40 million patients undergo surgery annually in the United States at a cost of $450 billion. Each year approximately 1 million patients sustain medical complications after surgery, resulting in costs of $25 billion annually. The proportion of the US population older than 65 years is estimated to double in the next 2 decades, leading to a 25% increase in the number of surgeries, a 50% increase in surgery-related costs, and a 100% increase in complications from surgery. Recognizing the significant increase in surgical burden due to accelerated aging of the population and increased reliance on surgery for treatment of disease, the National Heart, Blood and Lung Institute recently convened a working group on perioperative medicine. The group concluded that perioperative complications are significant, costly, variably reported, and often imprecisely detected; and they identified a critical need for accurate comprehensive perioperative outcome databases. Furthermore, presurgical risk profiling is inconsistent and deserves further attention, especially for noncardiac, nonvascular surgery and older patients.69

Although many preoperative predictors have been identified and are constantly being refined, risk stratification based on clinical, procedural, and biologic markers explains only a small part of the variability in the incidence of perioperative complications. As mentioned earlier, it is becoming increasingly recognized that perioperative morbidity arises as a direct result of the environmental stress of surgery occurring on a landscape of susceptibility that is determined by an individual's clinical and genetic characteristics, and it may even occur in otherwise healthy individuals. Such adverse outcomes will develop only in patients whose combined burden of genetic and environmental risk factors exceeds a certain threshold, which may vary with age. Identification of such genetic contributions not only explains disease causation and susceptibility but also influences the response to disease and drug therapy; and incorporation of genetic risk information in clinical decision making may lead to improved health outcomes and reduced costs. For instance, understanding the gene–environment interactions involved in atherosclerotic cardiovascular disease and neurologic injury may facilitate preoperative patient optimization and resource utilization. Furthermore, understanding the role of allotypic variation in proinflammatory and prothrombotic pathways, the main pathophysiologic mechanisms responsible for perioperative complications, may contribute to the development of target-specific therapies, thereby limiting the incidence of adverse events in high-risk patients. To increase clinical relevance for the practicing perioperative physician, we summarize the existing evidence by specific outcome while highlighting candidate genes in relevant mechanistic pathways (Tables 4-4, 4-5, and 4-6).

Biomarkers of Adverse Perioperative Cardiovascular Outcomes

Perioperative Myocardial Infarction and Ventricular Dysfunction

Patients with underlying cardiovascular disease can be at increased risk for perioperative cardiac complications. Over the last few decades, several multifactorial risk indices have been developed and validated for both noncardiac (eg, Lee's Revised Cardiac Risk Index) and cardiac surgical patients (eg, Hannan scores), with the specific aim of stratifying risk for perioperative adverse events. However, these multifactorial risk indices have only limited predictive value for identifying patients at the highest risk of perioperative myocardial infarction (PMI).70 In this context, it has been proposed that genomic approaches could aid in refining an individual's risk profile. Several reports from animal models, linkage analysis, and family, twin, and population association studies have confirmed the role of genetic factors in the etiology and progression of CAD. Specifically, both deaths from CAD as well as hazardous patterns of angiographic CAD (left main and proximal disease), known to be major risk factors for adverse perioperative events, are highly heritable. Similarly, a number of linkage and association studies,10,16 including a well-powered and replicated GWAS,39 have identified genetic susceptibility to MI in ambulatory population-based cohorts. The collective evidence from these studies suggests a strong genetic contribution to the risk of adverse cardiovascular outcomes in general but do not directly address the heritability of adverse perioperative myocardial events.

The incidence of PMI following cardiovascular surgery remains between 7% and 19%,71,72 despite advances in surgical, cardioprotective, and anesthetic techniques, and it is consistently associated with reduced short- and long-term survival in these patients. The pathophysiology of PMI after cardiac surgery involves systemic and local inflammation, "vulnerable" blood, and neuroendocrine stress.3 In noncardiac surgery, PMI occurs as a result of two distinct mechanisms: (1) coronary plaque rupture and subsequent thrombosis triggered by a number of perioperative stressors including catecholamine surges, proinflammatory, and prothrombotic states; and (2) myocardial oxygen supply-and-demand imbalance.73 Interindividual genetic variability in these mechanistic pathways is extensive, which may combine to modulate overall susceptibility to perioperative stress and ultimately the magnitude of myocardial injury. Nevertheless, until recently, only a few studies have explored the role of genetic factors in the development of PMI,25,74,75 mainly conducted in patients undergoing CABG surgery (Table 4-4).

Inflammatory Biomarkers and Perioperative Myocardial Outcomes

Although the role of inflammation in cardiovascular disease biology has long been established, we are just beginning to understand the relationship between genetically controlled variability in inflammatory responses to surgery and PMI pathogenesis. Recently, three inflammatory gene SNPs were described to have an independent predictive value for incident PMI after cardiac surgery performed with CPB.17 These include the proinflammatory cytokine interleukin-6 (IL6-572G>C) and two adhesion molecules: intercellular adhesion molecule-1 (ICAM1 Lys469Glu) and E-selectin (SELE 98G>T). Furthermore, the predictive ability of a PMI model based only on traditional risk factors was improved by the addition of genotypic information. Similarly, Collard et al identified a combined haplotype in the mannose-binding lectin gene (MBL2LYQA secretor haplotype), an important recognition molecule in the lectin complement pathway, to be independently associated with PMI in a cohort of white patients undergoing first-time CABG with CPB.18 Genetic variants in IL6 and TNFA have also been described in association with increased incidence of postoperative cardiovascular complications including PMI after lung surgery for cancer.76 Several polymorphisms in key proinflammatory genes have been associated with robust increases in perioperative inflammatory responses in patients undergoing cardiac surgery with CPB. These include the promoter SNPs in IL6 (-572G>C and -174G>C),77 also shown to prolong the hospital length of stay78; the apolipoprotein E genotype (the ϵ4 allele)79; SNPs in the tumor necrosis factor genes (TNFA-308G>A, LTA+250G>A),80 also associated with postoperative left ventricular dysfunction81; and a functional SNP in the macrophage migration inhibitory factor (MIF).82 Conversely, a genetic variant modulating the release of the anti-inflammatory cytokine interleukin-10 (IL10-1082G>A) in response to CPB has been described, with high levels of IL-10 surprisingly being associated with postoperative cardiovascular dysfunction.83 In patients undergoing elective surgical revascularization for peripheral vascular disease, several SNPs in IL6 (-174 G>C, nt565 G>A) and IL10 (-1082 G>A, -819 C>T, -592 C>A, and the ATA haplotype) were associated with endothelial dysfunction and an increased risk of a composite end point of acute postoperative cardiovascular events.84

C-reactive protein (CRP) is the prototypical acute-phase reactant and the most extensively studied inflammatory marker in clinical studies, and high-sensitivity CRP (hs-CRP) has emerged as a robust predictor of cardiovascular risk at all stages, from healthy subjects to patients with acute coronary syndromes and acute decompensated heart failure.85 Whether CRP is merely a marker or also a mediator of inflammatory processes is yet unclear, but several lines of evidence support the latter theory. In perioperative medicine, elevated preoperative CRP levels have been associated with increased short- and long-term morbidity and mortality in patients undergoing primary elective CABG (cut-off >3 mg/L)86 as well as in higher acuity CABG patients (cut-off >10 mg/L).87 Interestingly, in a retrospective analysis of patients with elevated baseline hs-CRP levels undergoing off-pump CABG surgery, preoperative statin therapy was associated with reduced postoperative myocardial injury and need for dialysis.88 In elective major noncardiac surgery patients, preoperative CRP levels (cut-off >3.4 mg/L) independently predicted perioperative major cardiovascular events (composite of MI, pulmonary edema, and cardiovascular death) and significantly improved the predictive power of the Revised Cardiac Risk Index (RCRI) in receiver operating characteristic analysis.89

In addition to the already established heritability of elevated baseline plasma CRP levels, recent reports indicate that the acute-phase rise in postoperative plasma CRP levels is also genetically determined. The CRP1059G>C polymorphism was associated with lower peak postoperative serum CRP following both elective CABG with CPB,90 as well as esophagectomy for thoracic esophageal cancer.91 Furthermore, CRP-717C>T polymorphism was associated with stress hyperglycemia in patients undergoing esophagectomy for cancer, leading to increased postoperative infectious complications and intensive care unit length of stay.92

Hemostatic Biomarkers and Perioperative Myocardial Outcomes

The host response to surgery is also characterized by alterations in the coagulation system, manifested as increased fibrinogen concentration, platelet adhesiveness, and plasminogen activator inhibitor (PAI)-1 production. These changes can be more pronounced after cardiac surgery, where the complex and multifactorial effects of hypothermia, hemodilution, and CPB-induced activation of coagulation, fibrinolytic, and inflammatory pathways are combined. Dysfunction of the coagulation system following cardiac surgery may manifest on a continuum ranging from increased thrombotic complications such as coronary graft thrombosis, PMI, stroke, pulmonary embolism at one end of the spectrum, to excessive bleeding as the other extreme. The balance between normal hemostasis, bleeding, and thrombosis is markedly influenced by the rate of thrombin formation and platelet activation, with genetic variability known to modulate each of these mechanistic pathways,93 suggesting significant heritability of the prothrombotic state (see Table 4-6 for an overview of genetic variants associated with postoperative bleeding). Several genotypes in hemostatic genes have been associated with increased risk of coronary graft thrombosis and myocardial injury following CABG. A genetic variant in the promoter of the PAI-1 gene, consisting of an insertion (5G)/deletion (4G) polymorphism at position -675 has been associated with changes in the plasma levels of PAI-1. Because PAI-1 is an important negative regulator of fibrinolytic activity, its polymorphism has been associated with increased risk of early graft thrombosis after CABG94 and, in a meta-analysis, with increased incidence of MI.95 Similarly, a polymorphism in the platelet glycoprotein IIIa gene (ITGB3), resulting in increased platelet aggregation (PlA2 polymorphism), has been associated with higher levels of postoperative troponin I release following CABG96 and with increased risk of thrombotic coronary graft occlusion, MI, and death 1 year following CABG.97 In the setting of noncardiac surgery, two polymorphisms in platelet glycoprotein receptors (ITGB3 and GP1BA) have been shown to be independent risk predictors of PMI in patients undergoing major vascular surgery and resulted in improved discrimination of an ischemia risk assessment tool when added to historic and procedural risk factors.98 Finally, a point mutation in coagulation factor V (1691G>A), resulting in resistance to activated protein C (factor V Leiden), was also associated with various postoperative thrombotic complications following noncardiac surgery.27 Conversely, in patients undergoing cardiac surgery, factor V Leiden was associated with significant reductions in postoperative blood loss and overall risk of transfusion.99 Nevertheless, in a prospective study of CABG patients with routine 3-month postoperative angiographic follow-up, carriers of factor V Leiden had a higher incidence of graft occlusion.100

Natriuretic Peptides and Perioperative Myocardial Outcomes

Circulating B-type natriuretic peptide (BNP) is a powerful biomarker of cardiovascular outcomes in many circumstances. Produced mainly in the ventricular myocardium, BNP is formed by cleavage of its prohormone by the enzyme corin into the biologically active C-terminal fragment (BNP) and an inactive N-terminal fragment (NT-proBNP). Known stimuli of BNP activation are myocardial mechanical stretch (from volume or pressure overload), acute ischemic injury, and a variety of other proinflammatory and neurohormonal stimuli inducing myocardial stress. Although secreted in a 1:1 ratio, circulating levels of BNP and NT-proBNP differ considerably due to different clearance characteristics.

A large number of studies have reported consistent associations of baseline plasma BNP or NT-proBNP levels with a variety of postoperative short- and long-term morbidity and mortality end points, independent of the traditional risk factors. For noncardiac surgery, these were summarized in two meta-analyses that overall indicate an approximately 20-fold increase in risk of adverse perioperative cardiovascular outcomes.101,102 Similarly, for cardiac surgery patients, preoperative BNP was a strong independent predictor of in-hospital postoperative ventricular dysfunction, hospital length of stay, and 5-year mortality following primary CABG,103 performing better than peak postoperative BNP.104 The current guidelines for preoperative cardiac risk assessment in noncardiac surgery list BNP and NT-proBNP measurements as class IIa/level B indications.105 However, despite the large number of studies conducted in both cardiac and noncardiac surgery, precise cut-off levels for BNP still need to be determined and adjusted for age, gender, and renal function. Similarly, no BNP-based goal directed therapies have been reported in the perioperative period. However, a role for BNP assays in monitoring aortic valve disease for optimal timing of surgery has been described.106

Furthermore, a recent study by Fox et al identified genetic variation in natriuretic peptide precursor genes (NPPA/NPPB) to be independently associated with a decreased risk of postoperative ventricular dysfunction following primary CABG, whereas variants in natriuretic peptide receptor NPR3 were associated with an increased risk (Table 4-4),107 offering additional clues into the molecular mechanisms underlying postoperative ventricular dysfunction.

The Role of Genetic Variability in Perioperative Vascular Reactivity

The perioperative period is characterized by robust activation of the sympathetic nervous system, which plays an important role in the pathophysiology of PMI. Thus patients with CAD who carry specific polymorphisms in adrenergic receptor (AR) genes can be at high risk for catecholamine toxicity and cardiovascular complications. Several functionally important SNPs modulating the AR pathways have been described.108 One of them is the Arg389Gly polymorphism in β1-AR gene (ADRB1), a SNP associated with increased risk of composite cardiovascular morbidity at 1 year after noncardiac surgery under spinal anesthesia.109 Of note, perioperative β-blockade had no effect. These findings prompted the investigators to suggest that stratification on AR genotype in future trials may help identify patients likely to benefit from perioperative β-blocker therapy. Significantly increased vascular responsiveness to α-adrenergic stimulation (phenylephrine) has been observed in carriers of the endothelial nitric oxide synthase (NOS3) 894>T polymorphism110 and angiotensin-converting enzyme (ACE) insertion/deletion (I/D) polymorphism28,111 undergoing cardiac surgery with CPB. Differences in perioperative vascular reactivity in relation to genetic variants of the β2-AR (ADRB2) have also been noted in patients undergoing noncardiac surgery. In patients with a common functional ADRB2 SNP (Glu27), increased blood pressure responses to endotracheal intubation were observed in one study.112 In a different study, in obstetric patients who had spinal anesthesia for cesarean delivery, the incidence and severity of maternal hypotension and response to treatment was affected by ADRB2 genotype (Gly16 and/or Glu27 led to lower vasopressor use for the treatment of hypotension).113 In patients undergoing cardiac surgery, the frequently observed vasoplegic syndrome and vasopressor requirements have been associated with a common polymorphism in the dimethylarginine dimethyl-aminohydrolase II (DDAH II) gene, an important regulator of nitric oxide synthase activity.114

Two recent studies have revealed the association of a common SNP at the 9p21 locus with both perioperative myocardial injury115 and all-cause mortality after primary CABG.116 This SNP was previously correlated with a wide array of vascular phenotypes in ambulatory populations (including CAD, MI, carotid atherosclerosis, abdominal aortic aneurysms, intracranial aneurysms) in replicated GWAS analyses. The mechanism of action of this SNP in the development of PMI and mortality is not completely understood, but it involves altered regulation of cell proliferation, senescence, and apoptosis. It seems that cardiac surgery with CPB may trigger the effects of the 9p21 gene variant leading to accumulation of senescent cells or cells that show evidence of necrotic death with cellular edema and lysis.

Perioperative Atrial Fibrillation

Perioperative atrial fibrillation (PoAF) remains a significant clinical problem after cardiac and noncardiac thoracic procedures. With an incidence of 27% to 40%, PoAF is associated with increased morbidity, hospital length of stay, rehospitalization, health care costs, and reduced survival. The high incidence of PoAF has prompted several investigators to develop comprehensive risk indices for the prediction of PoAF based on demographic, clinical, electrocardiographic, and procedural risk factors. Nevertheless, the predictive accuracy of these risk indices remains limited,117 suggesting that genetic variation may play a significant role in the occurrence of PoAF. Heritable forms of AF have been described in the ambulatory nonsurgical population, and it appears both monogenic forms like "lone" AF as well as polygenic predisposition to more common acquired forms like PoAF do exist.118 A recent GWAS for AF found two polymorphisms on chromosome 4q25 to be significantly associated with AF,42 findings replicated in other patient groups from Sweden, the United States, and Hong Kong. Recently, this locus was also associated with new-onset PoAF after cardiac surgery with CPB (CABG with or without concurrent valve surgery).119 The mechanism of action of the genetic locus identified by the 2 noncoding SNPs is unknown, but it lies close to several genes involved in the development of the pulmonary myocardium, or the sleeve of cardiomyocytes extending from the left atrium into the initial portion of the pulmonary veins. Clinical studies have demonstrated that ectopic foci of electric activity arising from within the pulmonary veins and posterior left atrium play a substantial role in initiating and maintaining AF.

Other candidate susceptibility genes for PoAF include those determining the duration of action potential (voltage-gated ion channels, ion transporters), responses to extracellular factors (adrenergic and other hormone receptors, heat shock proteins), remodeling processes, and magnitude of inflammatory and oxidative stress. It has been described that inflammation, reflected by elevated baseline CRP or IL6 levels and exaggerated postoperative leukocytosis, predicts the occurrence of PoAF. A link between inflammation and the development of PoAF is also supported by evidence that postoperative administration of nonsteroidal anti-inflammatory drugs may reduce the incidence of PoAF. Several recent studies have found that a functional SNP in the IL6 promoter (-174G>C) is associated with higher perioperative plasma IL6 levels and several adverse outcomes after CABG, including PoAF.120-122 In noncardiac surgery, polymorphisms in IL6 and TNFA genes have been shown to be associated with an increased risk of postoperative morbidity, including new-onset arrhythmias.76 There is, however, a contradictory lack of association between CRP levels (strongly regulated by IL6) and PoAF in women undergoing cardiac surgery,123 which may reflect gender-related differences. However, a recent study reported that both pre- and postoperative PAI-1 levels were independently associated with development of PoAF following cardiac surgery.124

Several investigations in the transcriptional responses to AF in human atrial appendage myocardium collected at the time of cardiac surgery or in preclinical models (Table 4-3) have identified a ventricular-like genomic signature in fibrillating atria, with increased ratios of ventricular to atrial isoforms, suggesting dedifferentiation.125 It remains unclear whether this "ventricularization" of atrial gene expression reflects cause or effect of AF, but it likely represents an adaptive energy-saving process to the high metabolic demand of fibrillating atrial myocardium, akin to chronic hibernation. Recently, a different mechanism was proposed as being involved in PoAF; it has been found that patients who exhibit PoAF after cardiac surgery display a differential genomic response to CPB in their peripheral blood leukocytes, characterized by upregulation of oxidative stress genes, which correlated with a significantly larger increase in oxidant stress both systemically (as measured by total peroxide levels) as well as at the myocardial level (as measured in the right atrium).126

Cardiac Allograft Rejection

Identification of peripheral blood gene- and protein-based biomarkers to noninvasively monitor, diagnose, and predict perioperative cardiac allograft rejection is an area of rapid scientific growth. Although several polymorphisms in genes involved in alloimmune interactions, the renin-angiotensin-aldosterone system, and the transforming growth factor-β superfamily have been associated with cardiac transplant outcomes, their relevance as useful clinical monitoring tools remains uncertain. However, peripheral blood mononuclear cell–based molecular assays have shown much promise for monitoring the dynamic responses of the immune system to the transplanted heart, discriminating immunologic allograft quiescence and predicting future rejection.127 A noninvasive molecular test to identify patients at risk for acute cellular rejection is commercially available (Allomap, XDx), in which the expression levels of 11 genes are measured by quantitative real-time polymerase chain reaction (qRT-PCR) and translated using a mathematical algorithm into a clinically actionable AlloMap score that enhances the ability to deliver personalized monitoring and treatment to heart transplant patients.128 Furthermore, several clinically available protein-based biomarkers of alloimmune activation, microvascular injury (troponins), systemic inflammation (CRP), and wall stress and remodeling (BNP) correlate well with allograft failure and vasculopathy and have good negative predictive values, but they require additional studies to guide their clinical use. Similarly, molecular signatures of functional recovery in end-stage heart failure following left ventricular assist device (LVAD) support using gene expression profiling have been reported using only mRNA129 or combined microRNA and mRNA profiling,130 and they could be used to monitor patients who received an LVAD as destination therapy or assess the timing of potential device explantation.

Genetic Variability and Postoperative Event-Free Survival

Several large randomized clinical trials examining the benefits of CABG surgery and percutaneous coronary interventions relative to medical therapy and/or to one another have refined our knowledge of early and long-term survival after CABG. These studies have helped define the subgroups of patients who benefit from surgical revascularization, and they also demonstrated a substantial variability in long-term survival after CABG, altered by important demographic and environmental risk factors. Increasing evidence suggests that the ACE gene indel polymorphism may influence post-CABG complications, with carriers of the D allele having higher mortality and restenosis rates after CABG surgery compared with the I allele.75 As discussed earlier, a prothrombotic amino acid alteration in the β3-integrin chain of the glycoprotein IIb/IIIa platelet receptor (the PlA2 polymorphism) is associated with an increased risk for major adverse cardiac events (a composite of MI, coronary bypass graft occlusion, or death) following CABG surgery (Table 4-4).97 We found preliminary evidence for association of two functional SNPs modulating β2-adrenergic receptor activity (Arg16Gly and Gln27Glu) with incidence of death or major adverse cardiac events following cardiac surgery,131 and recently identified a functional polymorphism in thrombomodulin (THBD Ala455Val; OR: 2.64) gene associated with increased 5-year mortality after CABG independent of EuroSCORE.132

Genetic Susceptibility to Adverse Perioperative Neurologic Outcomes

Despite advances in surgical and anesthetic techniques, significant neurologic morbidity continues to occur following cardiac surgery, ranging in severity from coma and focal stroke (incidence: 1%-3%) to more subtle cognitive deficits (incidence up to 69%), with a substantial impact on the risk of perioperative death, quality of life, and resource utilization. Variability in the reported incidence of both early and late neurologic deficits remains poorly explained by procedural risk factors, suggesting that environmental (operative) and genetic factors may interact to determine disease onset, progression, and recovery. The pathophysiology of perioperative neurologic injury is thought to involve complex interactions between primary pathways associated with atherosclerosis and thrombosis, and secondary response pathways like inflammation, vascular reactivity, and direct cellular injury. Many functional genetic variants have been reported in each of these mechanistic pathways involved in modulating the magnitude and the response to neurologic injury, which may have implications in chronic as well as acute perioperative neurocognitive outcomes. For example, Grocott at al examined 26 SNPs in relationship to the incidence of acute postoperative ischemic stroke in 1635 patients undergoing cardiac surgery, and they found that the interaction of minor alleles of the CRP (1846C>T) and IL6 promoter SNP -174G>C significantly increases the risk of acute stroke.133 Similarly, a recent study suggests that P-selectin and CRP genes both contribute to modulating the susceptibility to postoperative cognitive decline (POCD) following cardiac surgery.21 Specifically, the loss-of-function minor alleles of CRP 1059G>C and SELP 1087G>A are independently associated with a reduction in the observed incidence of POCD after adjustment for known clinical and demographic covariates (Table 4-5).

Our group has demonstrated a significant association between the apolipoprotein E (APOE) E4 genotype and adverse cerebral outcomes in cardiac surgery patients.19,134 This is consistent with the role of the APOE genotype in recovery from acute brain injury, such as intracranial hemorrhage,135 closed-head injury,136 and stroke,137 as well as experimental models of cerebral I/R injury138; two subsequent studies in CABG patients, however, have not replicated these initial findings. Furthermore, the incidence of postoperative delirium following major noncardiac surgery in the elderly139 and in critically ill patients140 is increased in carriers of the APOE ϵ4 allele. Unlike adult cardiac surgery patients, infants possessing the APOE ϵ2 allele are at increased risk for developing adverse neurodevelopmental sequelae following cardiac surgery.141,142 The mechanisms by which the APOE genotypes might influence neurologic outcomes are yet to be determined, but they do not seem to be related to alterations in global cerebral blood flow of oxygen metabolism during CPB143; however, genotypic effects in modulating the inflammatory response,79 extent of aortic atheroma burden,144 and risk for premature coronary atherosclerosis145 may play a role.

Recent studies have suggested a role for platelet activation in the pathophysiology of adverse neurologic sequelae. Genetic variants in surface platelet membrane glycoproteins, important mediators of platelet adhesion and platelet–platelet interactions, have been shown to increase the susceptibility to prothrombotic events. Among these, the PlA2 polymorphism in glycoprotein IIb/IIIa has been related to various adverse thrombotic outcomes, including acute coronary thrombosis146 and atherothrombotic stroke.147 We found the PlA2 allele to be associated with more severe neurocognitive decline after CPB,20 which could represent exacerbation of platelet-dependent thrombotic processes associated with plaque embolism.

Cardiac surgical patients who develop POCD demonstrate inherently different genetic responses to cardiopulmonary bypass from those without POCD, as evidenced by acute deregulation in peripheral blood leukocytes of gene expression pathways involving inflammation, antigen presentation, and cellular adhesion.148 These findings corroborate with proteomic changes, in which patients with POCD similarly have significantly higher serologic inflammatory indices compared with those patients without POCD149,150 and add to the increasing level of evidence that CPB does not cause an indiscriminate variation in gene expression but rather distinct patterns in specific pathways that are highly associated with the development of postoperative complications such as POCD. The implications for perioperative medicine include identifying populations at risk who might benefit not only from an improved informed consent, stratification, and resource allocation, but also from targeted anti-inflammatory strategies.

In noncardiac surgery, a study conducted in patients undergoing carotid endarterectomy demonstrated that preoperative plasma levels of fibrinogen and hs-CRP were independently associated with new periprocedural cerebral ischemic lesions caused by microembolic events, as determined by MRI diffusion-weighted imaging.151

Genetic Susceptibility to Adverse Perioperative Renal Outcomes

Acute renal dysfunction is a common, serious complication of cardiac surgery; about 8% to 15% of patients develop moderate renal injury (>1.0 mg/dL peak creatinine rise), and up to 5% of them develop renal failure requiring dialysis.152 Acute renal failure is independently associated with in-hospital mortality rates of greater than 60% in patients requiring dialysis.152 Several studies have demonstrated that inheritance of genetic polymorphisms in the APOE gene (ϵ4 allele)24 and in the promoter region of the IL6 gene (−174C allele)121 are associated with acute kidney injury following CABG surgery (Table 4-6). Stafford-Smith et al reported that major differences in peak postoperative serum creatinine rise after CABG are predicted by possession of combinations of polymorphisms that interestingly differ by race: the angiotensinogen (AGT) 842T>C and IL6 -572G>C variants in whites and the endothelial nitric oxide synthase (NOS3) 894G>T and angiotensin-converting enzyme (ACE) insertion/deletion in African Americans are associated with >50% reduction in the postoperative glomerular filtration rate.22 Further identification of genotypes predictive of adverse perioperative renal outcomes may facilitate individually tailored therapy, risk stratify the patients for interventional trials targeting the gene product itself, and aid in medical decision making (eg, selecting medical over surgical management).

Genetic Variants and Risk for Prolonged Postoperative Mechanical Ventilation

Prolonged mechanical ventilation (inability to extubate patient by 24 hours postoperatively) is a significant complication following cardiac surgery, occurring in 5.6% and 10.5% of patients undergoing first and repeat CABG surgery, respectively.153 Several pulmonary and nonpulmonary causes have been identified, and scoring systems based on preoperative and procedural risk factors have been proposed and validated. Recently, genetic variants in the renin-angiotensin pathway and in proinflammatory cytokine genes have been associated with respiratory complications post-CPB. The D allele of a common functional insertion/deletion polymorphism in the angiotensin-converting enzyme (ACE) gene, accounting for 47% of variance in circulating ACE levels,154 is associated with prolonged mechanical ventilation following CABG155 and with susceptibility to and prognosis of acute respiratory disease syndrome (ARDS).156 Furthermore, a hyposecretor haplotype in the neighboring genes tumor necrosis factor α (TNFA) and lymphotoxin α (LTA) on chromosome 6 (TNFA-308G/LTA+250G haplotype)157 and a functional polymorphism modulating postoperative IL6 levels (IL6-174G>C)121 are independently associated with higher risk of prolonged mechanical ventilation post-CABG. The association is more dramatic in patients undergoing conventional CABG than in those undergoing off-pump CABG (OPCAB), suggesting that in high-risk patients identified by preoperative genetic screening, OPCAB may be the optimal surgical procedure.

A next crucial step in understanding the complexity of adverse perioperative outcomes is to assess the contribution of variations in many genes simultaneously and their interaction with traditional risk factors to the longitudinal prediction of outcomes in individual patients. The use of such outcome predictive models incorporating genetic information may help stratify mortality and morbidity in surgical patients, improve prognostication, direct medical decision making both intraoperatively and during postoperative follow-up, and even suggest novel targets for therapeutic intervention in the perioperative period.

Interindividual variability in response to drug therapy, both in terms of efficacy and safety, is a rule by which anesthesiologists live. In fact, much of the art of anesthesiology is the astute clinician being prepared to deal with outliers. The term pharmacogenomics is used to describe how inherited variations in genes modulating drug actions are related to interindividual variability in drug response. Such variability in drug action may be pharmacokinetic or pharmacodynamic (Fig. 4-4). Pharmacokinetic variability refers to variability in a drug's absorption, distribution, metabolism, and excretion that mediates its efficacy and/or toxicity. The molecules involved in these processes include drug-metabolizing enzymes (such as members of the cytochrome P450, or CYP, superfamily) and transport molecules that mediate drug uptake into, and efflux from, intracellular sites. Pharmacodynamic variability refers to variable drug effects despite equivalent drug delivery to molecular sites of action. This may reflect variability in the function of the molecular target of the drug or in the pathophysiologic context in which the drug interacts with its receptor target (eg, affinity, coupling, expression).158 Thus pharmacogenomics investigates complex, polygenically determined phenotypes of drug efficacy or toxicity, with the goal of identifying novel therapeutic targets and customizing drug therapy.

Pseudocholinesterase Deficiency

Historically, characterization of the genetic basis for plasma pseudocholinesterase deficiency in 1956 was of fundamental importance to anesthesia and the further development and understanding of genetically determined differences in drug response.159 Individuals with an atypical form of pseudocholinesterase resulting in a markedly reduced rate of drug metabolism are at risk for excessive neuromuscular blockade and prolonged apnea. More than 20 variants have since been identified in the butyrylcholinesterase gene (BCHE), the most common of which are the A-variant (209A>G) and the K-variant (1615G>A), with various and somewhat poorly defined phenotypic consequences on prolonged neuromuscular blockade. Therefore, pharmacogenetic testing is currently not recommended in the population at large but only as an explanation for an adverse event.160

Genetics of Malignant Hyperthermia

Malignant hyperthermia (MH) is a rare autosomal dominant genetic disease of skeletal muscle calcium metabolism, triggered by administration of general anesthesia with volatile anesthetic agents or succinylcholine in susceptible individuals. The clinical MH syndrome is characterized by skeletal muscle hypermetabolism and manifested as skeletal muscle rigidity, tachycardia, tachypnea, hemodynamic instability, increased oxygen consumption and carbon dioxide production, lactic acidosis and fever, progressing to malignant ventricular arrhythmias, disseminated intravascular coagulation, and myoglobinuric renal failure. MH susceptibility has been initially linked to the ryanodine receptor (RYRI) gene locus on chromosome 19q.161 However, subsequent studies have shown that MH may represent a common severe phenotype that originates not only from point mutations in the RYRI gene (Arg614Cys), but also within its functionally and/or structurally associated proteins regulating excitation-contraction coupling (such as α1DHPR and FKBP12). It is becoming increasingly apparent that MH susceptibility results from a complex interaction between multiple genes and environment (such as environmental toxins), suggested by the heterogeneity observed in the clinical MH syndrome and the variable penetrance of the MH phenotype.162 Current diagnostic methods (the caffeine-halothane contracture test) are invasive and potentially nonspecific. Unfortunately, because of the polygenic determinism and variable penetrance, direct DNA testing in the general population for susceptibility to MH is currently not recommended; in contrast, testing in individuals from families with affected individuals has the potential to greatly reduce mortality and morbidity.160 Furthermore, genomic approaches may help elucidate the molecular mechanisms involved in altered RYRI-mediated calcium signaling and identify novel, more specific therapeutic targets.

Genetic Variability and Response to Anesthetic Agents

Anesthetic potency, defined by the minimum alveolar concentration (MAC) of an inhaled anesthetic that abolishes purposeful movement in response to a noxious stimulus, varies among individuals, with a coefficient of variation (the ratio of standard deviation to the mean) of approximately 10%.163 This observed variability may be explained by interindividual differences in multiple genes that underlie responsiveness to anesthetics, by environmental or physiologic factors (eg, brain temperature, age), or by measurement errors. With growing public concern over intraoperative awareness, understanding the mechanisms responsible for this variability may facilitate implementation of patient-specific preventive strategies. Evidence of a genetic basis for increased anesthetic requirements is beginning to emerge, suggested, for instance, by the observation that desflurane requirements are increased in subjects with red hair versus dark hair,164 and by recently reported variability in the immobilizing dose of sevoflurane (as much as 24%) in populations with different ethnic (and thus genetic) backgrounds.165 Several studies evaluating the genetic control of anesthetic responses, coupled with molecular modeling, proteomic, neurophysiology, and pharmacologic approaches have provided important developments in our understanding of general anesthetic mechanisms. Triggered by the seminal work of Franks and Lieb,166 research shifted from the membrane lipid bilayer to protein receptors (specifically ligand- and voltage-gated ion channels) as potential anesthetic targets, ending a few decades of stagnation that were primarily due to an almost universal acceptance of the dogma of nonspecific anesthetic action (the so-called lipid theory). Some of the genes responsible for phenotypic differences in anesthetic effects have been mapped in various animal models, and following genomic manipulation of plausible candidate receptors to investigate their function in vitro, they were evaluated in genetically engineered animals for their relationship to various anesthetic end points, such as immobility (ie, MAC), hypnosis, amnesia, and analgesia (for reviews, see Sonner et al167). Several thousand different strains of knockout mice have been created and are used to investigate specific functions of particular genes and mechanisms of drug action, including the sensitivity to general anesthetic in animals lacking the β3 subunit168 or the α6 subunit169 of the GABAA receptor. In contrast, knock-in animals express a site-directed mutation in the targeted gene that remains under the control of endogenous regulatory elements, allowing the mutated gene to be expressed in the same amount, at the same time, and in the same tissues as the normal gene. This method has provided remarkable insight into the mechanisms of action of benzodiazepines170 and IV anesthetics. In a seminal study by Jurd et al, a point mutation in the gene encoding the β3 subunit of the GABAA receptor, previously known to render the receptor insensitive to etomidate and propofol in vitro,171 was validated in vivo by creating a knock-in mouse strain that proved also essentially insensitive to the immobilizing actions of etomidate and propofol.172 A point mutation in the β2 subunit of the GABAA receptor results in a knock-in mouse with reduced sensitivity to the sedative173 and hypothermic effects174 of etomidate. Knock-in mice harboring point mutations in the α2A-adrenergic receptor have enabled the elucidation of the role of this receptor in anesthetic-sparing, analgesic, and sedative responses to dexmedetomidine.175

The situation is far more complex for inhaled anesthetics, which appear to mediate their effects by acting on several receptor targets. Based on combined pharmacologic and genetic in vivo studies to date, several receptors are unlikely to be direct mediators of MAC, including the GABAA (despite their compelling role in IV anesthetic-induced immobility), 5-HT3, AMPA, kainate, acetylcholine and α2-adrenergic receptors, and potassium channels.176 Glycine, NMDA receptors, and sodium channels remain likely candidates.167 These conclusions, however, do not apply to other anesthetic end points, such as hypnosis, amnesia, and analgesia. Several preclinical proteomic analyses have identified in a more unbiased way a group of potential anesthetic targets for halothane,61 desflurane,62 and sevoflurane,63 which should provide the basis for more focused studies of anesthetic binding sites. Such "omic" approaches have the potential to evolve into preoperative screening profiles useful in guiding individualized therapeutic decisions, such as prevention of anesthetic awareness in patients with a genetic predisposition to increased anesthetic requirements.

Genetic Variability and Response to Pain

Similar to the observed variability in anesthetic potency, the response to painful stimuli and analgesic manipulations varies among individuals. The sources of variability in the report and experience of pain and analgesia (ie, the so-called pain threshold) are multifactorial, including factors extrinsic to the organism (such as cultural factors or circadian rhythms) and intrinsic factors (such as age, gender, hormonal status, or genetic makeup). Increasing evidence suggests that pain behavior in response to noxious stimuli, its modulation by the central nervous system in response to drug administration or environmental stress, as well as the development of persistent pain conditions through pain amplification are strongly influenced by genetic factors.177-179

Results from studies in twins180 and inbred mouse strains181 indicate a moderate heritability for chronic pain syndromes and nociceptive sensitivity, which appears to be mediated by multiple genes. Various strains of knockout mice lacking target genes like neurotrophins and their receptors (eg, nerve growth factor), peripheral mediators of nociception, and hyperalgesia (eg, substance P), opioid and nonopioid transmitters and their receptors, and intracellular signaling molecules have significantly contributed to the understanding of pain processing mechanisms.182 A locus responsible for 28% of phenotypic variance in magnitude of systemic morphine analgesia in mouse has been mapped to chromosome 10, in or near the OPRM (μ-opioid receptor) gene. The μ-opioid receptor is also subject to pharmacodynamic variability; polymorphisms in the promoter region of the OPRM gene modulating interleukin 4–mediated gene expression have been correlated with morphine antinociception. The much quoted OPRM 188A>G polymorphism is associated with decreased responses to morphine-6-glucuronide, resulting in altered analgesic requirements but also reduced incidence of postoperative nausea and vomiting, and reduced risks of toxicity in renal failure patients. Conversely, variants of the melanocortin 1 receptor (MC1R) gene, which produce a red hair/fair skin phenotype, are associated with increased analgesic responses to κ-opioid agonists in women but not men, providing evidence for a gene-by-gender interaction in regulating analgesic response (for a review, see Somogyi et al183). Very recent reports suggest that peripherally located β2-adrenergic receptors (ADRB2) also contribute to basal pain sensitivity, the development of chronic pain states, as well as opioid-induced hyperalgesia.179 Functionally important haplotypes in the ADRB2178 and catechol-O-methyltransferase (COMT)184 genes are associated with enhanced pain sensitivity in humans.

In addition to the genetic control of peripheral nociceptive pathways, considerable evidence exists for genetic variability in the descending central pain modulatory pathways, further explaining the interindividual variability in analgesic responsiveness. One good example relevant to analgesic efficacy is cytochrome P450D6 (CYP2D6), a member of the superfamily of microsomal enzymes that catalyze phase I drug metabolism and are responsible for the metabolism of a large number of therapeutic compounds. The relationship between the CYP2D6 genotype and the enzyme metabolic rate has been extensively characterized, with at least 12 known mutations leading to a tetramodal distribution CYP2D6 activity: ultrarapid metabolizers (5% to 7% of the population), extensive metabolizers (60%), intermediate metabolizers (25%), and poor metabolizers (10%). Currently, pharmacogenomic screening tests predict CYP2D6 phenotype with more than 95% reliability. The consequences of inheriting an allele that compromises CYP2D6 function include the inability to metabolize codeine (a prodrug) to morphine by O-demethylation, leading to lack of analgesia but increased side effects from the parent drug (eg, fatigue) in poor metabolizers.160,177

Genetic Variability in Response to Other Drugs Used Perioperatively

A wide variety of drugs used in the perioperative period display significant pharmacokinetic or pharmacodynamic variability that is genetically modulated (Table 4-7). Although genetic variation in drug-metabolizing enzymes or targets usually results in hypervariable drug response, genetic markers associated with rare but life-threatening side effects have also been described. Of note, the most commonly cited categories of drugs involved in adverse drug reactions include cardiovascular, antibiotic, psychiatric, and analgesic medications, and interestingly, each category has a known genetic basis for an increased risk of adverse reactions.

There are more than 30 families of drug-metabolizing enzymes in humans, most with genetic polymorphisms shown to influence enzymatic activity. Of special importance to the anesthesiologists is the CYP2D6, one of the most intensively studied and best understood examples of pharmacogenetic variation, involved in the metabolism of several drugs including analgesics (codeine, dextromethorphan), β-blockers, antiarrhythmics (flecainide, propafenone, quinidine), and diltiazem. Another important pharmacogenetic variation has been described in cytochrome P450C9 (CYP2C9), involved in metabolizing anticoagulants (warfarin), anticonvulsants (phenytoin), antidiabetic agents (glipizide, tolbutamide), and nonsteroidal anti-inflammatory drugs (celecoxib, ibuprofen), among others. Three known CYP2C9 variant alleles result in different enzyme activities (extensive, intermediate, and slow metabolizer phenotypes) and have clinical implications in the increased risk of life-threatening bleeding complications in slow metabolizers during standard warfarin therapy. This illustrates the concept of "high-risk pharmacokinetics," which applies to drugs with low therapeutic ratios eliminated by a single pathway (in this case CYP2C9-mediated oxidation); genetic variation in that pathway may lead to large changes in drug clearance, concentrations, and effects.158 Dose adjustments based on the pharmacogenetic phenotype have been proposed for drugs metabolized via both CYP2D6 and CYP2C9 pathways,160 and a commercially available approved test by the Food and Drug Administration (FDA) (CYP450 AmpliChip, Roche Molecular Diagnostics) allows clinicians for the first time to test patients for a wide spectrum of genetic variation in drug-metabolizing enzymes. Using this technology, Candiotti et al showed that patients carrying either 3 copies of the CYP2D6 gene, a genotype consistent with ultrarapid metabolism, or both have an increased risk of ondansetron failure for the prevention of postoperative vomiting but not nausea.185 The strongest evidence to date for use of pharmacogenomic testing is to aid in the determination of warfarin dosage by using genotypes in the CYP2C9 and vitamin K epoxide reductase complex 1 (VKORC1) genes, and at least four FDA-approved tests are now commercially available.

Genetic variation in drug targets (receptors) can have a profound effect on drug efficacy, and more than 25 examples have already been identified. For example, functional polymorphisms in the β2-AR (Arg16Gly, Gln27Glu) influence the bronchodilator and vascular responses to β-agonists, and β1-AR variants (Arg389Gly) modulate responses to β-blockers and may have an impact on postoperative cardiovascular-adverse events.108,109

Finally, clinically important genetic polymorphisms with indirect effects on drug response have been described. These include variants in candidate genes like sodium (SCN5A) and potassium ion channels (KCNH2, KCNE2, KCNQ1), which alter susceptibility to drug-induced long-QT syndrome and ventricular arrhythmias (torsade de pointes) associated with the use of drugs like erythromycin, terfenadine, disopyramide, sotalol, cisapride, or quinidine. Carriers of such susceptibility alleles have no manifest QT-interval prolongation or family history of sudden death until QT-prolonging drug challenge is superimposed.158 Predisposition to QT-interval prolongation (considered a surrogate for risk of life-threatening ventricular arrhythmias) has been responsible for more drug withdrawals from the market than any other category of adverse event in recent times, so understanding genetic predisposing factors constitutes one of the highest priorities of current pharmacogenomic efforts.

Pharmacogenomics is emerging as an additional modifying component to anesthesia along with age, gender, comorbidities, and medication usage. Specific testing and treatment guidelines allowing clinicians to modify drug use appropriately (eg, adjust dose or change drug) already exist for a few compounds160 and will likely be expanded to all relevant therapeutic compounds, together with identification of novel therapeutic targets.

Systems Biology Approach to Perioperative Medicine: The "Perioptome"

Systems biology is a conceptual framework within which scientists attempt to correlate massive amounts of apparently unrelated data into a single unifying explanation of how biologic processes occur.186 This evolving discipline that merges experimental and computational approaches to observe, record, and integrate information from the molecular, cellular, tissue, and whole organism level into testable models of a dynamic biologic process can be applied to understand the way patients respond to a multidimensional stimulus such as a surgical procedure and the mechanistic basis of perioperative morbidity. Such an approach involves multiple levels of data integration. First, delineating the composition of the perioperative phenome (the representation of all perioperative phenotypes expressed by a given patient) requires standardized definitions, controlled vocabularies, and data dictionaries (a perioperative phenotype ontology), new (molecular) imaging technologies, and the availability of comprehensive data warehousing capabilities that will allow cataloging of individual perioperative phenotypes as well as correlations between combinations of phenotypes (organ cross-talk, multiple organ failure). Second, the orthogonal integration of whole-genome genotypic, transcriptomic, proteomic, and metabolomic data, augmented by more recent functional genomic and proteomic approaches including protein–protein, protein–DNA, or other "component–component" interaction mapping (interactome), transcript, or protein three-dimensional localization mapping (localizome),187 and literature data within individual biologic systems involved in perioperative morbidity. This highest level of data integration is the mapping of the integrated high-throughput static and dynamic genomic data into regulatory networks to model interactions of the different components of the system and identify modules of highly interconnected genes and hub points that can be prioritized as therapeutic targets. Ultimately, mathematical models require experimental validation in animal models of disease or tissue culture, in an iterative process that is one of the core characteristics of systems biology.188 Such integrative approaches to study cardiovascular function (the Cardiome Project), but also perioperative morbidity (the Perioptome)189 have already been outlined, and they promise to increase the identification of key drivers of perioperative adverse events beyond what could be achieved by genetic associations alone.

Targeted Therapeutic Applications: The "5 P's" of Perioperative Medicine and Pain Management

Genomic and proteomic approaches are rapidly becoming platforms for all aspects of drug discovery and development, from target identification and validation to individualization of drug therapy. As mentioned, the human genome contains about 26,000 genes encoding for approximately 200,000 proteins that represent potential drug targets. However, only about 120 drug targets are currently being marketed, thus making identification of novel therapeutic targets an area of intense research. Following gene identification, its therapeutic potential needs to be validated by defining the sequence function, its role in disease, and demonstrating that the gene product can be manipulated with beneficial effect and no toxic effects. A developing field, toxicogenomics, studies the influence of toxic or potentially toxic substances on different model organisms by evaluating the gene expression changes induced by novel drugs in a given tissue. Sponsored by the National Institutes of Health, a nationwide collaborative effort called the Pharmacogenetics Research Network (http://www.nigms.nih.gov/pharmacogenetics/) is aiming to establish a strong pharmacogenomics knowledge base (http://www.pharmgkb.org/) as well as create a shared computational and experimental infrastructure required to connect human sequence variation with drug responses and translate information into novel therapeutics.

The epidemiologic framework for assessing the applicability of previously identified biomarkers of perioperative morbidity and the successful implementation of molecular diagnostics in perioperative medicine is contingent on demonstrating their clinical validity, analytical validity, and clinical utility.190 Perioperative genomic investigators are currently conducting replication studies in different surgical patient populations to formally assess the clinical validity of the markers reported so far. For genomic classifiers, the emphasis during external validation is placed on prospectively testing the accuracy of the entire molecular fingerprint in a new patient population rather than corroborating results in individual genes. In perioperative and critical care settings, it is vital to have fast turnaround time (several hours) and easy-to-use testing capabilities, so that meaningful therapeutic interventions can take place. In this regard, new molecular diagnostic systems based on the random access technology such as the GeneXpert (Cepheid), eSensor (Osmetech), and Liat Analyzer (Iquum) are already becoming available. Clinical utility (targeted interventions to reduce perioperative morbidity among patients with a certain genomic profile) remains to be evaluated in future genomically stratified perioperative trials. Indeed, a landmark study on the effects of a 5-lipoxygenase–activating protein (FLAP) inhibitor on biomarkers associated with the risk of MI demonstrates that by defining at-risk patients for two genes in the leukotriene pathway, one can predict who will respond to targeted drug therapy. Specifically, in patients carrying the at-risk variants in FLAP and in the leukotriene A4 hydrolase genes, use of a FLAP inhibitor in a randomized controlled trial resulted in significant and dose-dependent suppression of biomarkers associated with increased risk of MI.191 It is expected that similar principles of targeted therapeutics could be operational in the perioperative period, thus beginning to fulfill the 5 P's of modern medicine (personalized, preventive, predictive, participatory, and prospective).

Ethical Considerations

Although one of the aims of the Human Genome Project is to improve therapy through genome-based prediction, the birth of personal genomics opens up a Pandora's box of ethical issues, including privacy and the risk for discrimination against individuals who are genetically predisposed for a medical disorder. Such discrimination may include barriers to obtaining health, life, or long-term care insurance or obtaining employment. Thus extensive efforts are made to protect patients participating in genetic research from prejudice, discrimination, or uses of genetic information that will adversely affect them. To address the concerns of both biomedical research and health communities, the U.S. Senate approved in 2003 the Genetic Information and Nondiscrimination Act, which provides the strong safeguards required to protect the public participating in human genome research.

Another ethical concern is the transferability of genetic tests across ethnic groups, particularly in the prediction of adverse drug responses. It is known that most polymorphisms associated with variability in drug response show significant differences in allele frequencies among populations and racial groups. Furthermore, the patterns of linkage disequilibrium are markedly different between ethnic groups, which may lead to spurious findings when markers, instead of causal variants, are used in diagnostic tests extrapolated across populations. In exploring racial disparities in health and disease outcomes, considerable debate has focused on whether race and ethnic identity are primarily social or biologic constructs, and the contribution of genetic variability in explaining observed differences in the rates of disease between racial groups. With the goal of personalized medicine the prediction of risk and treatment of disease on the basis of an individual's genetic profile, some have argued that biologic consideration of race will become obsolete. However, in this discovery phase of the postgenome era, continuing to incorporate racial information in genetic studies should improve our understanding of the architecture of the human genome and its implications for novel strategies aiming at identifying variants protecting against, or conferring susceptibility to, common diseases and modulating drug effects.192

The Human Genome Project has revolutionized all aspects of medicine, allowing us to assess the impact of genetic variability on disease taxonomy, characterization, and outcome, and individual responses to various drugs and injuries. Mechanistically, information gleaned through genomic approaches is already unraveling long-standing mysteries behind general anesthetic action and adverse responses to drugs used perioperatively. However, a strong need remains for prospective, well-powered genetic studies in highly phenotyped surgical populations, which require the development of multidimensional perioperative databases. For the anesthesiologist, this may soon translate into prospective risk assessment incorporating genetic profiling of markers important in thrombotic, inflammatory, vascular, and neurologic responses to perioperative stress, with implications ranging from individualized additional preoperative testing and physiologic optimization, to choice of perioperative monitoring strategies and critical care resource utilization. Furthermore, genetic profiling of drug metabolizing enzymes, carrier proteins, and receptors, using currently available high-throughput molecular technologies, will enable personalized choice of drugs and dosage regimens tailored to suit a patient's pharmacogenetic profile. At that point, perioperative physicians will have far more robust information to use in designing the most appropriate and safest anesthetic plan for a given patient.

Future trends and challenges in perioperative genomics are still being defined, but they mainly concern interdisciplinary studies designed to combine an analytical system approach, mathematical modeling, and engineering principles with the multiple molecular and genetic factors and stimuli, and the macroscale interactions that determine the pathophysiologic response to surgery.

Acknowledgments: Supported in part by NIH grants HL075273 and HL092071 (to MVP).

• Ashley EA, Butte AJ, Wheeler MT, et al. Clinical assessment incorporating a personal genome. Lancet. 2010;375:1525-1535.
• Body SC, Collard CD, Shernan SK, et al. Variation in the 4q25 chromosomal locus predicts atrial fibrillation after coronary artery bypass graft surgery. Circ Cardiovasc Genet. 2009;2:499-506.
• Diatchenko L, Nackley AG, Tchivileva IE, et al. Genetic architecture of human pain perception. Trends Genet. 2007;23:605-613.
• Feero WG, Guttmacher AE, Collins FS. Genomic medicine—an updated primer. N Engl J Med. 2010;362:2001-2011.
• Fox AA, Collard CD, Shernan SK, et al. Natriuretic peptide system gene variants are associated with ventricular dysfunction after coronary artery bypass grafting. Anesthesiology. 2009;110:738-747.
• Fox AA, Muehlschlegel JD, Body SC, et al. Comparison of the utility of preoperative versus postoperative B-type natriuretic peptide for predicting hospital length of stay and mortality after primary coronary artery bypass grafting. Anesthesiology. 2010;112:842-851.
• Mathew JP, Podgoreanu MV, Grocott HP, et al. Genetic variants in P-selectin and C-reactive protein influence susceptibility to cognitive decline after cardiac surgery. J Am Coll Cardiol. 2007;49:1934-1942.
• Muehlschlegel JD, Liu KY, Perry TE, et al. Chromosome 9p21 variant predicts mortality after coronary artery bypass graft surgery. Circulation. 2010;122:S60-S65.
• Podgoreanu MV, Schwinn DA. New paradigms in cardiovascular medicine: emerging technologies and practices: perioperative genomics. J Am Coll Cardiol. 2005;46:1965-1977.
• Podgoreanu MV, White WD, Morris RW, et al. Inflammatory gene polymorphisms and risk of postoperative myocardial infarction after cardiac surgery. Circulation. 2006;114:1275-1281.
• Rosenberg S, Elashoff MR, Beineke P, et al. Multicenter validation of the diagnostic accuracy of a blood-based gene expression test for assessing obstructive coronary artery disease in nondiabetic patients. Ann Intern Med. 2010;153:425-434.
• Turer AT, Stevens RD, Bain JR, et al. Metabolomic profiling reveals distinct patterns of myocardial substrate use in humans with coronary artery disease or left ventricular dysfunction during surgical ischemia/reperfusion. Circulation. 2009;119:1736-1746.
• Zaugg M, Bestmann L, Wacker J, et al. Adrenergic receptor genotype but not perioperative bisoprolol therapy may determine cardiovascular outcome in at-risk patients undergoing surgery with spinal block: the Swiss Beta Blocker in Spinal Anesthesia (BBSA) study: a double-blinded, placebo-controlled, multicenter trial with 1-year follow-up. Anesthesiology. 2007;107:33-44.