Impact of genetic variation on perioperative bleeding

Abstract

Variation in bleeding in the perioperative period is a complex and multifactorial event associated with immediate and delayed consequences for the patient and health care resources. Little is known about the complex genetic influences on perioperative bleeding. With the discovery of multiple variations in the human genome and ever-growing databases of well-phenotyped surgical patients, better identification of patients at risk of bleeding is becoming a reality. In this review, polymorphisms in the platelet receptor genes, plasminogen activator inhibitor, and angiotensin genes among others will be discussed. We will explore the nature, effects, and implications of the genetics that influence perioperative bleeding above and beyond surgical bleeding, particularly in cardiac surgery.

Introduction

The number of patients undergoing surgery has increased in recent years; today, 40 million surgical procedures are performed annually, in the United States alone [1]. Of the numerous complications that can arise, perioperative bleeding is one of the most serious. Patients with excessive bleeding require blood product transfusion and reoperation. These interventions have a large financial impact on the health care system, in which the health care costs for surgical patients now total over $450 billion per year [1]. In addition, perioperative bleeding and related interventions exact a substantial human toll, conferring elevated risks of other morbidity and mortality.

Approximately 20% of the 13 million units of blood products transfused in the United States annually are consumed during the 750,000 cardiac surgical procedures performed on a mere 0.3% of the US population [2]. Transfusion requirements vary considerably among cardiac surgical patients, with prediction indices rarely able to predict more than 20% of variability in transfusion or concomitant excessive bleeding [3,4].

The need for transfusion increases risk of mortality and other morbidities in the cardiac surgical population. Patients undergoing CABG who require perioperative blood transfusions have significantly elevated risk of mortality (odds ratio [OR], 1.77; 95% confidence interval [CI] 1.7–1.9), prolonged ventilatory support (OR, 1.79; 95% CI 1.7–1.9), renal failure (OR, 2.06, 95% CI 1.9–2.3), serious infections (OR, 1.76; 95% CI 1.7–1.8), cardiac complications (OR, 1.55; 95% CI 1.5–1.6), and neurologic events (OR, 1.37; 95% CI 1.3–1.4) [5], independent of other clinical predictors of these adverse events. Moreover, perioperative red blood cell transfusion is the single most reliable factor associated with adverse postoperative events after isolated CABG [5] and confers decreased long-term survival [6]. In addition, the risk of postoperative arrhythmias, especially atrial fibrillation (AF), is equally increased with transfusions [7]. However, it is not clear whether patients experience these complications as a result of blood product transfusion or as a result of their excessive bleeding (and any associated comorbidities). Randomized trials assessing the safety and efficacy of blood products do not exist.

The etiology of perioperative bleeding is multifactorial and includes nongenetic, notably and most important surgical technique and concurrent drug therapies, as well as genetic influences, some of which can result in a catastrophic bleeding diathesis. The relatively new field of perioperative genomics seeks to elucidate the different outcomes of patients with similar traditional risk factors based on their genes. Most common medical diseases, including excessive bleeding, are genetically complex. Usually, these conditions involve multiple genes (polygenic) that individually have a small effect, show incomplete penetrance and variable expression, and are dependent on gene–gene and gene-environment interaction.

This review focuses on the group of complex genetic variants that induce changes in the activity of gene products affecting perioperative bleeding. Although their existence in a single patient is usually unknown, as they do not manifest until later in life or on exposure to an environmental stressor, they are very common in the population. Often, many different genetic susceptibility variants interact with environmental factors to incrementally increase risk of bleeding [8]. The combination of environmental factors and surgical stress can unmask these defects in the perioperative period.

Etiology of Perioperative Bleeding

Environmental factors contributing to surgical bleeding include the type, length, complexity, and location of the surgery performed, and pharmacologic anticoagulation, among others (Table I) [9]. Bleeding also tends to be greater when surgery involves highly vascular tissues that are not easily cauterized (e.g., bone, liver). Although multiple studies have documented these risk factors for bleeding and defined a high-risk patient group, there still exists wide interpatient variability of bleeding and transfusion, which is not identifiable with conventional preoperative assessment.

TABLE I.

Environmental Factors Associated With Increased Perioperative Bleeding or Blood Transfusion

Type	Examples	Mechanism
Inherent factors	Advanced age
	Small body size	Low preop red cell volume =>   more blood transfusions
Comorbidities	Hemorrhagic shock
	Renal failure	Can lead to hemolytic uremic   syndrome (see below)
	Liver disease	Decrease in all factors   except factor VIII
	Preoperative anemia
	Prosthetic cardiac valves	Hemolysis
	Nutritional deficit	Deficiency of vitamin   K–dependent factors
Difficult surgical/   technical   hemostasis	Transplantation	Liver, kidney dx leading to   coagulation abnormalities
	Extracorporeal circulation
Acquired factors	Sepsis	Excessive fibrinolysis   counterbalanced   by increased levels of PAI-1,   antithrombin   deficiency, reduced activity   of APC
	DIC	Excessive consumption
	Hemolytic Uremic   Syndrome	Thrombocytopenia   and mechanical   hemolytic anemia
	Massive transfusion	Dilutional
	Thrombocytopenia
Pharmacologic   anticoagulation	Warfarin, aspirin, clopidogrel,   heparin, bivalirudin, …
	Chemotherapy, antibiotics	Can cause thrombocytopenia
Procedure-related   factors	Reduced body temperature	Decrease in protein function
	Urgent operation	Patient not medically optimized
	Non-CABG cardiac surgery	Longer time on CPB =>   dilution of clotting factors

PAI-1, plasminogen activator inhibitor-1; DIC, disseminated intravascular coagulation; APC, activated protein C; CPB, cardiopulmonary bypass; CABG, coronary artery bypass grafting.

Among surgical patients undergoing coronary artery bypass grafting (CABG) surgery, the factors affecting the risk of perioperative bleeding include the extent of the procedure (single-vessel primary operation having the lowest risk and reoperative multivessel or valve/CABG procedures having the highest risk), duration of cardiopulmonary bypass (CPB) [10], increasing age, preoperative renal insufficiency, greater volume of salvaged red cell infusion, and lower core body temperature [11]. Other reasons for excessive bleeding unrelated to surgical factors are reductions in platelet number, size, and mass; increased fibrinolytic activity; and receipt of a high dose of heparin and/or protamine [12,13].

The singular process that separates cardiac from non-cardiac surgery is the use of CPB, a physiological and biomaterial insult that increases the likelihood of bleeding through activation of the coagulation, inflammation, and fibrinolysis pathways, by contact of the cellular elements of blood with CPB circuit materials and air, reperfusion of ischemic tissues induced hypothermia and the use of heparin and protamine. The most notable defect in hemostasis after CPB is platelet dysfunction due to hemodilution, adherence of platelets to the extracorporeal circuit material with associated activation, degranulation, and desensitization [14] and the use of platelet inhibitors such as inhibitors of adenosine diphosphate (ADP) or glycoprotein (GP) IIB/IIIA pathways, further increasing the likelihood of bleeding and transfusion.

Genetic Influences on Perioperative Bleeding

Genetic variation falls into two broad groups: the rare, so-called Mendelian traits with high penetrance, that usually arise from one or more single defects in a gene (monogenic), and with little or no modification by environmental factors; and complex factors, which often involve multiple polymorphisms in multiple genes (polygenic), with considerable modification of genetic influence by nongenetic (environmental) factors.

Monogenic risk factors of bleeding

Von Willebrand disease (vWD) and the hemophilias make up 95–97% of all inherited major deficiencies of coagulation proteins [15] exhibiting strong familial inheritance, but have little overall impact in the entire cohort of patients undergoing cardiac surgery because of their very low frequency. Most patients present with excessive bleeding and/or abnormal laboratory testing early in life and have a well established diagnosis prior to surgery. Their management during cardiac surgery, usually by coagulation factor transfusion or pharmacologic therapy, is well delineated [16–20].

Complex risk factors of bleeding

In contrast to highly penetrant monogenic disease, complex risk factors involve multiple genes that individually have low (<2) incremental risk ratios, show incomplete penetrance and variable expression, and are dependent on gene–gene and gene-environment interaction.

Genetic variation of coagulation has been studied extensively in the nonoperative setting. In several ambulatory populations, levels of circulating coagulation proteins have been shown to be highly heritable (40–70% of variation in level is explained by parentage) [21–25], notably for plasminogen activator inhibitor-1 (PAI-1), factor XIII (SNPs only account for 2% of variance), factor VII (promoter SNPs accounted for 12–19% of variation), fibrinogen, tissue plasminogen activator (tPA) and vWF. In addition, the response to anticoagulant medications commonly used in cardiovascular disease, such as aspirin, warfarin, the GPIIb/IIIa inhibitors, and other platelet inhibitors, varies with genetic influences upon metabolism and effector mechanisms [26–31].

Genetic Variation Associated with Perioperative Bleeding

As surgery—especially cardiac surgery—involves multiple environmental stressors and blood loss, this is an ideal environment in which to examine genetic variations in bleeding. The clinical phenotype that emerges during this period reflects a combination of the patient’s genotype and the superimposed stressors of surgery and environmental influences, whereby the surgical stressor still maintains the strongest influence on perioperative bleeding (see Fig. 1) [32]. Because genetic databases initially focused on the ambulatory population and perioperative genomics is a relatively new field of study, few polymorphisms contributing to bleeding in the surgical setting have been identified thus far.

Figure 1.

Factors influencing clinical phenotype in the perioperative period. A patient’s genotype along with gene–gene and gene-environment interactions determines the phenotype at baseline. Depending on this genotype and the addition of stressors in the operating room, the clinical phenotype in the perioperative period can range from thrombosis to a normal response to abnormal bleeding. IABP indicates intra-aortic balloon pump. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Noncardiac surgery

In humans, the level and activity of the angiotensin-converting enzyme (ACE) is associated with the insertion/deletion (in/del) polymorphism which has been associated with a wide spectrum of cardiovascular diseases and noncardiovascular characteristics [33]. The in/del is unlikely to be the causative polymorphism, as it has no identified functional significance despite considerable investigation, and is in strong linkage disequilibrium with several other more-likely candidate polymorphisms that alter ACE activity. Nevertheless, homozygosity of the deletion genotype (DD) is associated with plasminogen activator inhibitor-1 (PAI-1) plasma activity, and has been implicated as a risk factor for thrombophilia by altering vascular tone and blood flow [34,35].

In a markedly underpowered study of 105 geriatric patients undergoing total hip arthroplasty (THA), those with the ACE deletion allele had increased blood loss, whereas those homozygous for the insertion allele had reduced blood loss [36]. In contrast, in a better-powered study of 780 cardiac surgical patients, the deletion allele of the ACE polymorphism was associated with decreased bleeding [37]. This was supported by another study, again examining patients undergoing THA, demonstrating decreased bleeding. These contradictory findings do not stand on their own and are additionally doubtful because of a lack of biological plausibility.

Cardiac surgery

Several studies have examined perioperative bleeding during and after cardiac surgery performed with CPB and have identified SNPs associated with bleeding in the genes for the platelet glycoproteins and circulating coagulation factors [37–40].

PAI-1 Polymorphism

PAI-1 is the primary inhibitor of tissue plasminogen activator (t-PA). Both circulating t-PA levels and the activity of its inhibitor, PAI-1, determine t-PA activity. Deficiencies in PAI-1 are associated with an increased likelihood of hemorrhagic events due to increased t-PA dependant plasmin production [41–43]. Although CPB causes a marked increase in t-PA levels within minutes of initiation of CPB, due to contact activation of coagulation proteins with the oxygenator surface, this response is not ubiquitous, with one-third of patients having little change in t-PA activity [44,45]. Over the 2–24 hr after initiation of CPB, PAI-1 levels increase, decreasing t-PA activity and fibrinolysis favoring a procoagulant state. PAI-1 levels are determined by a common single base G in/del of the promoter region of the PAI-1 gene, and also by circadian variation in PAI-1 release [46,47], lipid profile (cholesterol and lipoprotein(a) levels) [48,49], and the acute-phase responses to CPB/CABG. The so-called 4G allele (the deletion allele of the above-mentioned promoter polymorphism) has been associated with higher plasma PAI-1 activity as the genetic sequence does not favor binding of a transcription repressor protein (Factor B), thus increasing transcription of PAI-1 messenger RNA (mRNA) [50]. When the 5G allele (the insertion allele) is present, the transcription repressor protein can bind to the promoter, thus reducing transcription of PAI-1 gene mRNA. Among patients undergoing elective CABG and valve surgery, the homozygous 5G genotype is associated with lower PAI-1 mRNA production, lower circulating PAI-1 levels, and increased transfusion of coagulation factors [12,38]. This finding was confirmed in a similar, albeit smaller, cohort of CPB patients showing that patients with the 5G/5G genotype had significantly higher amounts of chest tube drainage (CTD) compared with patients with other genotypes at all study time-points [51].

Platelet genetic variation

At the site of endothelial injury, the extracellular matrix proteins von Willebrand Factor (vWF), collagen, fibronectin, laminin, elastin, thrombospondin, and fibrin are exposed to platelet surface membrane receptors. The first contact between platelets and the lesion on the vessel wall is established by an interaction of the platelet receptor glycoprotein (GP)Ib-V-IX and vWF (see Fig. 2) [52]. The GPIb-V-IX complex is composed of four polypeptide subunits: GPIbα, GPIbβ, GPV, and GPIX. Binding of vWF and GPIb-V-IX, which occurs at the GPIbα subunit, tethers platelets to the endothelium at high shear rates, but is insufficient for stable adhesion and is rapidly reversible. For stable adhesion, the principal collagen-binding receptor, GPVI, is required [53]. In the absence of this GP receptor, other surface receptors are inadequate at adhering to the subendothelium [52]. GPVI binding and activation stimulates the expression of other glycoprotein receptors notably GPIIb/IIIa (comprised of the integrins α2b and β3), binding fibrinogen [52,54,55]; and GPIaIIa (comprised of the integrins α2 and β1), also binding collagen [56].

Figure 2.

Platelet dependent thrombus formation. At site of injury, extracellular matrix proteins von Willebrand factor (vWF) and collagen (Col) are exposed to the blood. Platelets loosely adhere to the subendothelium via the membrane adhesion receptors GPIb and GPVI. This results in GPIIbIIIa (fibrinogen receptor) and GP IaIIa (collagen receptor) activation. This interaction of with extracellular matrix proteins stabilizes platelet adhesion. Platelets bind via fibrinogen between two GPIIbIIIa receptors. Microparticles catalyze thrombin generation and fibrin formation that stabilizes the platelet thrombus. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Platelet glycoprotein polymorphisms

The majority of studies of platelet GP receptors in the surgical population have been associated with thrombotic outcomes. Few studies have examined for association with increased perioperative bleeding.

In nonsurgical populations, expression of GPIaIIa receptors on the platelet surface varies significantly [57] and the rate of platelet attachment to collagen under conditions of high shear, is highly correlated to the expression of GPIaIIa receptors [58]. Hereditary variation of the number of alleles is associated with receptor density and has a direct impact on coagulation. The −92C/G promoter variant of GPIa (ITGA2) is strongly associated with expression of the GPIaIIa receptor [59]. Platelet expression of GPIaIIa and genetic variation that modifies such expression have been strongly associated with retinal vein thrombosis and less well associated with myocardial infarction [60–63].

Numerous studies have examined the effect of integrin β3 (ITGB3; GPIIIa) genetic variation, principally to determine responsiveness to GPIIb/IIIa receptor antagonists used in the treatment of myocardial ischemia and infarction. The nonsynonymous coding SNP rs5918 (1565C/T) encodes a common leucine–proline substitution at codon 33 (Leu33Pro), commonly known as the human platelet antigen (HPA)-1 or Pl^A1/^A2 [64]. This polymorphism, with a prevalence of 20% in Caucasians, increases α-granule release and fibrinogen binding. It is speculated that the increased affinity of fibrinogen to the GPIIb/IIIa receptor in Pl^A2 patients is a result of conformational changes at the fibrinogen binding site. This hyperreactivity is manifested by a lower ADP activation threshold and increased GPIIb/IIIa - fibrinogen binding leading to increased platelet activation and aggregation [65]. Therefore, platelet glycoprotein polymorphisms can contribute to an increased risk of thrombosis or bleeding in relevant disease states [37].

In a study of 780 CABG patients, 19 functional polymorphisms in 13 candidate genes were examined for their role in postoperative bleeding defined as 12-hr CTD [37]. Apart from the ACE in/del polymorphism, which was associated with decreased bleeding, six polymorphisms showed an association with increased CTD. These included the GPIaIIa (α2β1 integrin) −52C/T and 807C/T polymorphisms, the GPIbα 524C/T polymorphism, the tissue factor −603A/G polymorphism, the prothrombin 20210G/A polymorphism, and the tissue factor pathway inhibitor (TFPI) −399C/T polymorphism. In a multivariable model in interaction with each other, they accounted for 9% of the variability of CTD (P < 0.01). The exact mechanisms for increased bleeding are not known. It is speculated, since these genetic variants modulate thrombin generation, platelet surface receptor density, clotting factor and platelet consumption on CPB, they each contribute a risk for perioperative bleeding.

Polymorphism of other proteins

Genetic variants of other proteins, including some less centrally involved in coagulation, may also influence perioperative bleeding. Notable examples are ACE—discussed in the section on non-cardiac surgery—and the selectins, principally E-selectin, which mediates the adhesion of blood leukocytes to vascular endothelium at sites of inflammation. In a cardiac surgical population, the 98G/T polymorphism in the 5′ UTR of the E-selectin gene was associated with increased bleeding in a gene-dose effect, independent and additive to a clinical risk model [66]. It is thought to be responsible for the accumulation of blood leukocytes at sites of inflammation by mediating the adhesion of cells to the vascular lining.

ABO blood group

A relationship between the ABO blood groups and coagulation disturbances has been described, mainly attributed to its association with lower vWF levels. Both differences in in vitro bleeding times and vWF activity have been found [67,68]. Patients with blood group O may have an increased risk of bleeding due to lower vWF activity. In the nonsurgical population, the non-OO blood group decreased the risk of bleeding from vitamin-K antagonist treatment [69]. On the other hand, this variability in bleeding linked to ABO blood group could not be reproduced in a large study of cardiac surgical patients [70].

Diagnosis of Risk Factors of Perioperative Bleeding

The underlying risk of bleeding is often difficult to estimate preoperatively. In patients with rare highly penetrant disorders such as the hemophilias, the childhood presentation of these disorders means that treatment options are well-established prior to surgery. However, in patients with complex risk factors, the disorder is generally undiagnosed because the genetic variation is not sufficient, by itself, to cause symptoms in the absence of surgery, trauma or other initiating event. Consequently, presentation occurs during or after surgery and a definitive cause is impossible to establish at that time [71]. Treatment is often empiric and nonspecific.

Routine preoperative laboratory tests are rarely helpful to help identify patients at elevated risk for bleeding. The specific tests obtained vary depending on the patient’s age, comorbidities, and type of surgery. The American Society of Anesthesiologists Task Force on Perioperative Blood Transfusion and Adjuvant Therapies has examined the usefulness of preoperative coagulation studies [71,72]. Routine preoperative coagulation studies identified abnormalities in bleeding time, prothrombin time (PT), activated partial prothrombin time (PTT), or platelet count in an excessively high 22.0% of patients which only led to changes in clinical management in up to 4.0% of cases. Thus the current testing regimens are too sensitive and the abnormal results are routinely ignored [73–75].

Whole blood coagulation assessment

For cardiac surgery, a common standard of care for perioperative coagulation monitoring consists of a platelet count, PT, and PTT [71]. However, results of these tests, as well as other indices of coagulation such as bleeding time, fibrinogen and D-dimer levels, have not been shown to correlate well with bleeding or arterial thrombotic events after CABG surgery [11,76,77]. Recently, functional assays that measure hemostasis in vitro, such as whole blood aggregometry (WBA) and thromboelastography (TEG), have shown good correlation with both intraoperative and postoperative bleeding in patients undergoing cardiac surgery [11,76–78]. However, their cost and complexity often reduce their utilization.

The TEG measures viscoelastic properties of nonanticoagulated or anticoagulated blood after initiation of coagulation under low-shear conditions, resembling venous rheologic conditions. The patterns of change in the viscoelastic properties reflect the kinetics of each stage of thrombus formation, including fibrinolysis. A TEG-guided transfusion algorithm has been used in cardiac surgery to effectively reduce transfusion of coagulation factors and platelets, by diagnosis of the specific functional deficit, or perhaps more importantly, to diagnose the absence of a specific functional deficit, thus empowering greater surgical diligence in hemostasis [79]. In this prospective, randomized, controlled and blinded study of 105 patients, the patients in the TEG group received less fresh frozen plasma and platelet transfusions and exhibited less chest tube drainage. A meta-analysis evaluated 170 publications, of which 14 represented the best studies on the intraoperative use of TEG, and came to the conclusion that although it does predict bleeding and guide therapy, it needs to be validated [80].

Platelet function assessment

As cardiac surgery and CPB particularly induce thrombocytopenia and a platelet function deficit, monitors of platelet function per se have been developed. Classic assessment of platelet aggregation by platelet aggregometry is labor intensive, expensive, and time-consuming, and not suited for real-time platelet function analysis in the operating room. Newer platelet function analyzers circumvent most of these shortcomings by simulating in vivo conditions of high shear stress and with platelet receptor agonists such as collagen [81]. These devices are often designed to assess efficacy of antiplatelet therapy. such as GPIIb/IIIa inhibitors, adenosine diphosphate and aspirin, and have been used successfully to predict major adverse cardiac events [82].

Value of coagulation monitors in identifying complex genetic bleeding risk factors

So how does this apply to genomics? Presently, the phenotype to diagnose a complex bleeding risk factor is the measured amount of perioperative blood loss, or in cardiac surgery, the amount of chest tube drainage. This is a relatively crude measure of effect of the genotype, and its association does not take into account the multiple steps between genotype and phenotype (see Fig. 1). An intermediate phenotype that is closer to the patient’s genotype, such as a functional assay might be more likely to identify a risk of bleeding. Our present use of such tests is rudimentary but some studies have shown reasonable ability to predict patients with increased perioperative bleeding [79].

Diagnosis of complex bleeding risk factors by genotyping—Present and future

The promise of perioperative genetic testing lies in the ability to predict which patients have a propensity to excessive bleeding from surgery. The rationale for the routine preoperative work-up is to risk-stratify patients, guide clinical management and decision making, predict postoperative morbidity and mortality, and ultimately tailor the perioperative therapy to the individual patient. However, perioperative stressors can unmask deficiencies otherwise unrecognized by this work-up. Therefore, in the large majority of patients, there is no reliable way to determine which patients will have surgical bleeding without performing genotyping for known polymorphisms.

Routine genotyping of surgical patients is still in the future. The cost of genotyping has dropped dramatically within the last 2 years, but even if certain polymorphisms are discovered on preoperative genotypic screening, in complex disease they usually only account for 5–15% of the variability of the disease. This leaves over 80% of the variability that is unaccounted for. Additionally, with over 20 million major surgeries performed annually in the United States alone, even a 99.9% positive predictive value will result in a large number of patients with false-positive results. For now, until the accuracy and reliability of POCMs have been established, the most logical approach to assessing the risk of perioperative bleeding would be routine patient assessment for conventional risk factors combined with targeted genotyping of those patients deemed at increased risk.

The analytical validity of genotyping platforms is very high [83]; however, the clinical implications of the genotyping results are poorly delineated for several reasons. We are still a long ways away from identifying all susceptibility-associated variants, especially for such complex diseases such as perioperative bleeding. Genome wide associations studies (GWAS) performing characterization and analysis of very-high-resolution SNP genotype data for thousands of individuals has not been performed for this phenotype. Additionally, the observed relative risks of ~1.5, as we have been finding in most of the complex diseases discovered to date, have mediocre performance in predicting who will develop a phenotype, versus those who will not, especially with the amount of intraoperative variables present.

Lastly, what are we to do if a patient truly has a polymorphism of a gene variant known to influence perioperative bleeding? Patients with genotypes conferring an increased bleeding tendency could receive higher doses of antifibrinolytics, smaller incisions (e.g., ministernotomy vs. full sternotomy), or lower doses of anticoagulants (e.g., low-dose heparin with a heparin-bonded circuit). These interventions are unproven in clinical trials and potentially carry significant risks. Large scale studies comparing patients with genetic variation to healthy controls would be needed before these interventions can go into mainstream practice.

Conclusions

Perioperative bleeding is a common and serious—yet still largely unpredictable—complication of surgery. Although there is evidence that genetic factors are associated with excessive bleeding in cardiac surgery, more studies with larger patient populations are needed. The polymorphisms found in this population with upregulated coagulation and fibrinolytic pathways will likely apply to other surgical patients as well, albeit to a lesser degree. In addition, whole-genome association studies are needed to better elucidate gene–gene and gene-environment interactions. Perioperative genomics holds potential for improving clinical outcomes in the surgical population, to the benefit of patients and the health care system alike.