Increases in percentages of patients with preexisting comorbidities (A) and thromboembolic events (B) by year of surgery.
Independent risk factors for perioperative acute thromboembolism syndrome. Preexisting morbidities as risk factors for perioperative arterial or venous thrombosis and embolism are summarized, and cardiac and cerebrovascular thrombotic diseases and metabolic diseases produce a cross-linkage, causing perioperative acute thromboembolism syndrome and subsequent death.
Kikura M, Takada T, Sato S. Preexisting Morbidity as an Independent Risk Factor for Perioperative Acute Thromboembolism Syndrome. Arch Surg. 2005;140(12):1210-1217. doi:10.1001/archsurg.140.12.1210
Copyright 2005 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2005
Preexisting morbidities are risk factors for perioperative arterial or venous thromboembolic events and subsequent death within 30 postoperative days.
Prospective cohort study.
University-affiliated general hospital.
A total of 21 903 surgery patients treated from January 1, 1991, through December 31, 2002.
Main Outcome Measures
Independent risk factors for perioperative arterial or venous thromboembolic events.
History of atrial fibrillation and coronary artery disease increased the risk of myocardial infarction (odds ratio [95% confidence interval], 4.3 [2.8-6.7]). History of stroke increased the risk of stroke (2.4 [1.4-4.1]) and death (4.7 [1.3-17.3]). Diabetes mellitus increased the risk of myocardial infarction (2.1 [1.3-3.2]), and hyperuricemia increased the risk of stroke (3.5 [1.2-9.8]), and both increased the risk of death (4.3 [1.3-14.1] and 11.8 [2.2-63.5], respectively). History of myocardial infarction increased the risk of deep vein thrombosis (7.7 [1.7-34.7]). Cancer increased the risk of all thromboembolism (2.4 [1.9-3.2]). Trend analysis showed that preexisting morbidities will increase 1.5-fold and thromboembolic events will increase 3-fold during the next decade.
Cardiac and cerebrovascular diseases, metabolic diseases, and cancer are becoming increasingly high-risk comorbidities for perioperative acute thromboembolism syndrome.
Cardiac and cerebrovascular diseases are the leading causes of death worldwide and represent a critical health problem.1,2 They are accelerated by preexisting risk factors such as coronary artery disease, hypertension, diabetes mellitus, hyperlipidemia, and hyperuricemia, which exacerbate the progression of atherosclerotic disease, trigger atherothrombotic complications, and are associated with poor outcomes.3,4 Patients undergoing surgery are especially prone to perioperative myocardial infarction5,6 and stroke,7,8 because they tend to have atherosclerotic disease. Surgical patients are also vulnerable to perioperative deep vein thrombosis and pulmonary embolism, which are also major health problems and common causes of death.9,10 Although the incidences of ischemic heart disease, cerebrovascular disease, diabetes mellitus, and cancer are increasing worldwide,1,2,11,12 how these preexisting morbidities, as risk factors for perioperative arterial or venous thromboembolism, will influence future morbidity and mortality rates is unknown.
Perioperative myocardial infarction, ischemic stroke, deep vein thrombosis, and pulmonary embolism appear to share a common pathophysiology that includes perioperative hemodynamic disturbance,5- 8 bed rest,9,10 and a state described by a cell-based model of vascular thrombosis.13 This state is characterized by the Virchow triad: endothelial injury and dysfunction, disturbed blood flow, and hypercoagulability and reduced fibrinolysis. Thus, thrombin generation is induced, promoted by inflammatory cytokines and tissue factor, and results in thrombosis from adherence of activated neutrophils and platelets to activated endothelial cells.13- 16 Surgical patients are predisposed to these proinflammatory and prothrombotic processes and are particularly vulnerable to myocardial infarction and pulmonary embolism during week 1 after surgery, and to stroke, deep vein thrombosis, and cardiac or cerebrovascular death during weeks 1 and 2 after surgery. These events have been referred to collectively as a perioperative acute thromboembolism syndrome that is a multifactorial thromboembolic disorder rather than merely a combination of the separate individual diseases.17
Comprehensive study of preexisting morbidities as risk factors for perioperative thromboembolic events is necessary for informed patient consent, patient counseling, surgical treatment decisions, and prophylactic antithromboembolic medications.
We conducted a prospective cohort study during a 12-year period to identify which preexisting morbidities are risk factors for perioperative acute thromboembolism syndrome, including myocardial infarction, ischemic stroke, deep vein thrombosis, pulmonary embolism, and subsequent cardiac or cerebrovascular death within 30 postoperative days, and to test the hypothesis that the incidence of this syndrome will increase because of an increase in the incidence of preexisting morbidities during the next decade.
The study protocol was approved by the institutional review board (Sei-rei Mikatabara General Hospital, Hamamatsu, Japan), and consecutive patients who underwent elective general surgery, orthopedic surgery, or thoracic or peripheral vascular surgery between January 1, 1991, and December 31, 2002, were enrolled in the study. Written informed consent for medical care and procedures in the clinic and hospital was obtained from all patients. Patients undergoing surgery on the head, neck, or thoracic aorta were not included to exclude thromboembolic accidents that are directly related to surgical procedures. All patients were monitored for new onset of myocardial infarction, pulmonary embolism, deep vein thrombosis, ischemic stroke, and death from cardiovascular causes for 30 days postoperatively. Diagnosis of myocardial infarction was confirmed if symptoms met the World Heath Organization criteria and the event was associated with abnormal levels of cardiac enzymes and met diagnostic electrocardiographic criteria. Ischemic stroke was diagnosed if the patient showed new neurologic deficits that persisted for more than 24 hours. Computed tomographic scans or magnetic resonance images were available for most events and were used to distinguish hemorrhagic from ischemic events. Pulmonary embolism was diagnosed if a pulmonary angiogram, computed tomographic scan, magnetic resonance image, echocardiogram, or pathologic characteristics of a thrombus removed at surgery or autopsy confirmed pulmonary embolism. Deep vein thrombosis was diagnosed if a duplex ultrasonogram, venous angiogram, computed tomographic scan, magnetic resonance image, or pathologic characteristics of a thrombus removed at surgery or autopsy confirmed deep vein thrombosis. Deaths from cardiovascular causes were confirmed by review of the death certificates, medical records, and autopsy reports.
All patients in this study underwent preoperative evaluation of preexisting morbidities. Preexisting morbidities in surgical patients recorded in the hospital database included hypertension (defined as a systolic blood pressure of ≥140 mm Hg, a diastolic blood pressure of ≥90 mm Hg, or both), diabetes mellitus (defined as a blood glucose level of ≥126 mg/dL [6.93 mmol/L] after an overnight fast, a glycosylated hemoglobin value of ≥6.5%, or both), hyperlipidemia (serum total cholesterol level of ≥220 mg/dL [5.72 mmol/L], low-density lipoprotein cholesterol level of ≥140 mg/dL [2.8 mmol/L], or both), and hyperuricemia (serum uric acid level of ≥7.0 mg/dL [0.42 mmol/L]). Preexisting morbidities also included cardiovascular disease diagnosed by board-certified cardiologists, including atrial fibrillation confirmed by electrocardiography, stable angina (symptoms of frequent chest pain, dyspnea on exertion, or nuclear imaging abnormalities), unstable angina (increased frequency, intensity, or duration of chest pain, or decreased response to nitrates in the previous 2 months), myocardial infarction (as described in the previous section), valvular disease (diagnosed on the basis of symptoms, echocardiographic findings, cardiac catheterization findings, or angiographic findings), pulmonary hypertension (diagnosed via cardiac catheterization, mean pulmonary arterial pressure ≥25 mm Hg, or echocardiographic or angiographic findings), ischemic stroke diagnosed by board-certified neurologists (as described in the previous section), and cancer. The computerized medical database file of each patient in the study was prospectively collected and secured in the Division of Medical Records and Information of our institution.
The preexisting specific incidences and risks of myocardial infarction, pulmonary embolism, deep vein thrombosis, ischemic stroke, and death were calculated. Correlation between the year of surgery and the percentage of patients with a preexisting morbidity or with a perioperative arterial or venous thromboembolic event within 30 postoperative days was determined by linear regression analysis.
The potential association of age, sex, and histories of cardiac and cerebrovascular disease, metabolic disease, and cancer with the occurrence of any perioperative thromboembolic event was analyzed by χ2 test and Fisher exact test for trend. To estimate odds ratios and 95% confidence intervals by multivariate logistic regression analysis, age category (<50 years, 50-69 years, and ≥70 years), sex, and history of cardiac or cerebrovascular disease, metabolic disease, and cancer were treated as independent categorical variables. Type of thromboembolic event (myocardial infarction, pulmonary embolism, deep vein thrombosis, ischemic stroke, or death) was treated as a dependent categorical variable. All independent variables that were significant (2-tailed nominal P<.1) in univariate analyses were entered into a multivariate logistic analysis by the proportional odds model. Stepwise logistic regression was performed, and variables that were significant (2-tailed nominal P<.05) were retained. Statistical analysis was performed with SAS statistical software (SAS Institute Inc, Cary, NC).
The study population consisted of 21 903 surgical patients (Table 1). General surgery included lower (19.1% of patients) or upper (16.0%) abdominal surgery and other general procedures (4.4%); orthopedic surgery included total knee (4.7%) or hip (4.0%) replacement, spinal surgery (10.4%), and other orthopedic procedures (29.1%); thoracic or vascular surgery included lung and mediastinal surgery (10.6%), peripheral vascular surgery (1.5%), and other related operations (0.2%). A total of 255 patients (1.16%) experienced thromboembolic complications within 30 postoperative days, 105 patients (0.48%) had myocardial infarction, 30 (0.14%) had pulmonary embolism, 23 (0.11%) had deep vein thrombosis, 97 (0.44%) had ischemic stroke, and 13 (0.06%) died of cardiac or cerebrovascular causes. The mortality was 13 patients (5.1%) in the 255 patients with any thromboembolic complications, 7 patients (6.7%) in the 105 patients with myocardial infarction, 5 patients (5.2%) in the 97 patients with ischemic stroke, and 1 patient (3.3%) in the 30 patients with pulmonary embolism.
The incidences of all thromboembolic events and of myocardial infarction, ischemic stroke, pulmonary embolism, deep vein thrombosis, and subsequent death are shown in Tables 1, 2, and 3 in relation to age, sex, and preexisting morbidities. The incidence of perioperative thromboembolism was higher in patients who had a history of atrial fibrillation, angina pectoris, myocardial infarction, ischemic stroke, diabetes mellitus, hyperuricemia, and cancer.
The risk and incidence of all arterial or venous thromboembolic events increased with age, particularly for patients 70 years or older (P<.001), with a history of atrial fibrillation (P = .005), myocardial infarction (P<.001), ischemic stroke (P = .001), diabetes mellitus (P<.001), or cancer (P<.001) as determined by multivariate logistic regression analysis and summarized in Table 4 and Table 5.
Men had an increased risk of myocardial infarction (P = .03), and women had an increased risk of pulmonary embolism (P = .01) and deep vein thrombosis (P = .004). The risk and incidence of myocardial infarction increased with age (P<.001), and with a history of atrial fibrillation (P = .02), angina pectoris (P = .01), myocardial infarction (P<.001), or diabetes mellitus (P = .001). The risk and incidence of ischemic stroke increased with age (P<.001) and with a history of ischemic stroke (P = .002) or hyperuricemia (P = .01). The risk and incidence of deep vein thrombosis increased with age (P<.001) and with a history of myocardial infarction (P = .008). The risk and incidence of postoperative death increased with a history of ischemic stroke (P = .02), diabetes mellitus (P = .01), or hyperuricemia (P = .004). Patients with cancer had an increased risk of myocardial infarction (P<.001), ischemic stroke (P = .03), pulmonary embolism (P = .001), deep vein thrombosis (P = .02), and subsequent death (P = .01).
There was a correlation between the year of surgery (x variable) and the percentages of patients with preexisting comorbidities (y variables) (Figure 1A). We estimate from these trends that the percentage of surgical patients with any preexisting cardiac or cerebrovascular disease, metabolic disease, or cancer will increase to 44.0% in 2015, a 1.5-fold increase from the mean percentage during the 12 years of the present study.
There was a correlation between the year of surgery (x variable) and the percentages of patients experiencing a thromboembolic event, myocardial infarction, ischemic stroke, deep vein thrombosis, pulmonary embolism, and subsequent death (y variables) (Figure 1B). We estimate from these trends that the percentage of surgical patients experiencing any thromboembolic event will increase to 4.0% in 2015, a 3-fold increase from the mean percentage during the 12 years of our study.
The relationship of preexisting morbidities to perioperative risk of myocardial infarction, ischemic stroke, pulmonary embolism, deep vein thrombosis, and subsequent cardiac and cerebrovascular death is summarized in Figure 2.
This large, contemporary, prospective 12-year study of consecutive surgical patients offers physicians important, clinically relevant information about the relationship of patients’ preexisting morbidity to perioperative morbidity, mortality, and risk of a perioperative acute thromboembolism syndrome (Figure 2). Our data on arterial and venous thromboembolic events have serious implications. The incidence of these events (1.2%) is projected to increase more than 3-fold, accelerated by a predicted 1.5-fold increase in preexisting morbidity during the next decade, creating a more severe health care problem.
The present study showed atrial fibrillation and cancer to be independent risk factors for perioperative myocardial infarction, in addition to traditional risk factors such as aging or a history of myocardial infarction, angina pectoris, and diabetes mellitus.5,6 Atrial fibrillation, the most common sustained arrhythmia seen in clinical practice, is associated with a 2-fold increase in total and cardiovascular mortality.18 C-reactive protein level, an indicator of systemic inflammation and coronary risk,16 increases in patients with atrial fibrillation,19 and increased levels of C-reactive protein predict increased risk for future myocardial infarction and thromboembolic stroke.20 In addition to an irregular heart rate and the resulting turbulent flow in the systemic vessels, proinflammatory and prothrombotic status could be related to the association between atrial fibrillation and the risk of perioperative myocardial infarction.16,18- 20 A relatively lower risk observed in our study compared with the previous studies5,6 is probably a result of the noncardiac surgery and Asian population of the present study.
A history of hyperuricemia was shown to be an independent risk factor for perioperative stroke, as are aging and history of stroke.7,8 An increased level of serum uric acid has been identified as an independent risk factor for stroke in the general medical population.21,22 Increased uric acid level may play a pathogenetic role in stroke by promoting hypertension, atherosclerotic change and subsequent endothelial dysfunction,23 or activation of platelets and inflammatory cytokines.24 The uric acid level reflects the degree of circulating xanthine oxidase activity, an important source of oxygen free radicals.25
A history of myocardial infarction was also shown to be an independent risk factor for perioperative deep vein thrombosis, as are aging and a history of varix and deep vein thrombosis.9,10 The risk of deep vein thrombosis in patients hospitalized with acute myocardial infarction is moderate and is comparable to that in general surgical patients.26 Recent study indicates that atherosclerosis can induce deep vein thromboembolism, and atherosclerosis and deep vein thrombosis share a common prothrombotic pathophysiology, such as activated platelets and up-regulated blood coagulation.27 Our results support the relationship between history of coronary atherosclerotic disease and perioperative deep vein thrombosis. A relatively lower perioperative incidence in our study compared with the other studies10,28 might be due to lower prevalence of deep venous thromboembolism in the Asian population.
Diabetes mellitus, hyperuricemia, stroke, cancer, and age of 70 years and older were also shown to be independent risk factors for perioperative cardiac or cerebrovascular death. Diabetes mellitus and hyperuricemia increase cardiac and cerebrovascular mortality in the general population by promoting atherosclerosis, endothelial dysfunction, and activation of platelets and inflammatory cytokines.4,21,22,29,30 The present study findings indicate that surgical stress enhances the prothrombotic pathophysiology and inhibits the fibrinolytic pathophysiology in diabetes mellitus and hyperuricemia, thus increasing the risk of perioperative cardiac or cerebrovascular death.
Cancer was shown to be an independent risk factor for perioperative arterial and venous thromboembolism. A critical component of cancerous tumor progression is inflammation, in which abundant proinflammatory cytokines such as tumor necrosis factor α and interleukin 1 are produced.31 The mechanisms explaining the association between preexisting cancer and perioperative myocardial infarction and stroke are unknown. Perhaps the elevated levels of D dimers and serum angiogenic factors that promote the proinflammatory and prothrombotic state increase the risk of arterial thrombosis.31- 33 Cancer increases the risk of deep vein thrombosis and pulmonary embolism in the general medical population.10,34 Endothelial dysfunction, alterations in procoagulant and anticoagulant protein levels, and deregulation of cytokine activities have been proposed as underlying mechanisms.31,33,34
Perioperative administration of aspirin reduces the incidence of myocardial infarction and stroke by one half35 and the risk of pulmonary embolism and deep vein thrombosis by one third and two thirds, respectively.28 An orally active direct thrombin inhibitor, ximelagatran, has been shown to prevent stroke in patients with atrial fibrillation,36 major cardiovascular events following acute myocardial infarction,37 and venous thromboembolism in patients undergoing elective hip or knee replacement surgery.38 Its further usefulness in perioperative prophylaxis is anticipated, in addition to the traditional antithromboembolic prophylaxis by warfarin, low-dose heparin, and low-molecular-weight heparin.
Recent evidence supports a role of inflammatory stimulation in the pathogenesis of atherosclerosis.16 Serum markers of inflammation such as C-reactive protein have proven remarkably robust as markers of cardiac and cerebrovascular risk, and targeting inflammation has become a focal point for understanding and treating atherosclerosis.16,20,39 Aspirin, β-adrenergic blocking agents, statins (inhibitors of 3-hydroxy-3-methylglutaryl coenzyme A reductase), angiotensin-converting enzyme inhibitors, angiotensin-receptor blockers, and fibric acid derivatives have emerged as new pharmacologic strategies for preventing perioperative myocardial infarction, stroke, and subsequent death, thus complementing traditional antiplatelet and anticoagulant therapies.16,28,35,39,40
We should address the limitations of the present study. First, we did not identify the patients with asymptomatic arterial and venous thromboembolic events such as silent myocardial ischemia, cerebrovascular transient ischemic attack, or asymptomatic deep vein thromboembolism. Therefore, the numbers of surgical patients exposed to the risk of perioperative arterial and venous thromboembolic events should be greater than those we estimated in this study. Second, hypertension and hyperlipidemia, traditional atherothrombotic risk factors, were not identified as risk factors for any thromboembolic events in this study. The reason might be that the Asian population has less severe atherosclerotic disease. In addition, perioperative surgical stress itself is a more influential and detrimental factor for thromboembolic events than hypertension and hyperlipidemia in surgical patients, and the risk of hypertension and hyperlipidemia might not be demonstrated in this study. Third, we estimated an increase in a perioperative arterial or venous thromboembolic events by linear regression analysis. However, the increases in comorbidities associated with the expanding elderly population might increase the incidence of a perioperative acute thromboembolism syndrome more than we expected in the present study.
The results of the present study permit identification of patients at risk of perioperative acute thromboembolism syndrome, thus indicating the need for prophylactic antithromboembolic treatment. Our findings underscore the importance of minimizing risk of perioperative acute thromboembolism syndrome arising from comorbid diseases. We hope to reduce its incidence and thereby reduce associated costs as well as improve patient outcomes.
Correspondence: Mutsuhito Kikura, MD, PhD, Department of Anesthesia and Intensive Care, Hamamatsu Medical Center, 328 Tomizuka-cho, Hamamatsu, Japan 432-8580 (email@example.com).
Accepted for Publication: March 23, 2005.