Perioperative myocardial injury: individual and population implications

Abstract

It is generally understood that patients who have cardiovascular disease are at risk of cardiac complications after surgery.33Mangano DT. Perioperative cardiac morbidity.Anesthesiology. 1990; 72: 153-184Crossref PubMed Scopus (1048) Google Scholar In patients undergoing major or vascular surgery, the risk of such complications can be significant. Eagle and colleagues studied patients enrolled on the CASS (Coronary Artery Surgery Study) registry who subsequently underwent non-cardiac surgery. They reported a perioperative myocardial infarction rate of 8.5% in patients with medically managed coronary artery disease who underwent vascular surgery.14Eagle KA Rihal CS Mickel MC Holmes DR Foster ED Gersh BJ. Cardiac risk of noncardiac surgery: influence of coronary disease and type of surgery in 3368 operations. CASS Investigators and University of Michigan Heart Care Program. Coronary Artery Surgery Study.Circulation. 1997; 96: 1882-1887Crossref PubMed Scopus (369) Google Scholar The authors defined high-risk surgery as surgery associated with a risk of perioperative death or myocardial infarction of greater than 4%. Abdominal surgery, thoracic surgery, and head and neck surgery fell into this category. For patients undergoing these types of surgery who had medically treated coronary artery disease, the overall perioperative myocardial infarction rate was 2.7% and the overall death rate 3.3%. This compared with rates of 0.8 and 1% respectively in patients undergoing similar surgery who did not have coronary artery disease. Data such as these are often considered with information on the prevalence of coronary artery disease to obtain some indication of the population burden of disease. For example, in 1990 Mangano stated that 25 million patients undergo major surgery in the USA each year and suggested that perhaps 6–7 million of these people are at risk of perioperative cardiac complications.33Mangano DT. Perioperative cardiac morbidity.Anesthesiology. 1990; 72: 153-184Crossref PubMed Scopus (1048) Google Scholar Such estimates suggest that perioperative cardiac complications are both a major issue in the care of individual surgical patients and a major public health issue. This review will examine in more detail the population impact of perioperative cardiac morbidity and the implications of this for the individual patient. Many of the studies that report the incidence of perioperative events have been conducted in patients undergoing major vascular surgery (the references given here identify a few of the best known studies).2Adams 3rd, JE Sicard GA Allen BT et al.Diagnosis of perioperative myocardial infarction with measurement of cardiac troponin I.N Engl J Med. 1994; 330: 670-674Crossref PubMed Scopus (498) Google Scholar 10Cutler BS Leppo JA. Dipyridamole thallium 201 scintigraphy to detect coronary artery disease before abdominal aortic surgery.J Vasc Surg. 1987; 5: 91-100Abstract Full Text Full Text PDF PubMed Scopus (174) Google Scholar 13Eagle KA Coley CM Newell JB et al.Combining clinical and thallium data optimizes preoperative assessment of cardiac risk before major vascular surgery.Ann Intern Med. 1989; 110: 859-866Crossref PubMed Scopus (628) Google Scholar 15Eagle KA Singer DE Brewster DC Darling RC Mulley AG Boucher CA. Dipyridamole‐thallium scanning in patients undergoing vascular surgery. Optimizing preoperative evaluation of cardiac risk.JAMA. 1987; 257: 2185-2189Crossref PubMed Scopus (271) Google Scholar 26L’Italien G Cambria RP Cutler BS et al.Comparative early and late cardiac morbidity among patients requiring different vascular surgery procedures.J Vasc Surg. 1995; 21: 935-944Abstract Full Text Full Text PDF PubMed Scopus (135) Google Scholar 27Landesberg G Luria MH Cotev S et al.Importance of long‐duration postoperative ST‐segment depression in cardiac morbidity after vascular surgery.Lancet. 1993; 341: 715-719Abstract PubMed Scopus (274) Google Scholar 31Leppo J Plaja J Gionet M Tumolo J Paraskos JA Cutler BS. Noninvasive evaluation of cardiac risk before elective vascular surgery.J Am Coll Cardiol. 1987; 9: 269-276Crossref PubMed Scopus (261) Google Scholar 37Pasternack PF Grossi EA Baumann FG et al.The value of silent myocardial ischemia monitoring in the prediction of perioperative myocardial infarction in patients undergoing peripheral vascular surgery.J Vasc Surg. 1989; 10: 617-625Abstract Full Text Full Text PDF PubMed Scopus (131) Google Scholar 43Poldermans D Rambaldi R Fioretti PM et al.Prognostic value of dobutamine‐atropine stress echocardiography for peri‐operative and late cardiac events in patients scheduled for vascular surgery.Eur Heart J. 1997; 18: D86-D96Crossref PubMed Google Scholar 44Raby KE Selwyn AP. Role of ambulatory ECG monitoring in estimating cardiac risk in vascular surgery patients.Isr J Med Sci. 1989; 25: 516-520PubMed Google Scholar 46vonKnorring J Lepantalo M. Prediction of perioperative cardiac complications by electrocardiographic monitoring during treadmill exercise testing before peripheral vascular surgery.Surgery. 1986; 99: 610-613PubMed Google Scholar This is appropriate in that coronary artery disease and perioperative myocardial infarction are both common in these patients.14Eagle KA Rihal CS Mickel MC Holmes DR Foster ED Gersh BJ. Cardiac risk of noncardiac surgery: influence of coronary disease and type of surgery in 3368 operations. CASS Investigators and University of Michigan Heart Care Program. Coronary Artery Surgery Study.Circulation. 1997; 96: 1882-1887Crossref PubMed Scopus (369) Google Scholar 20Hertzer NR. Basic data concerning associated coronary disease in peripheral vascular patients.Ann Vasc Surg. 1987; 1: 616-620Abstract Full Text PDF PubMed Scopus (149) Google Scholar 21Hertzer NR Beven EG Young JR et al.Coronary artery disease in peripheral vascular patients. A classification of 1000 coronary angiograms and results of surgical management.Ann Surg. 1984; 199: 223-233Crossref PubMed Scopus (1196) Google Scholar However, it may not be valid to extrapolate from this population to patients undergoing other types of surgery. In a study by Borsma and colleagues of 1351 patients presenting for major vascular surgery, 976 (72%) of patients had at least one of the risk factors for cardiovascular complications included in the Lee risk score (see below). Four hundred and ninety-one (36%) of the patients had suffered a previous myocardial infarction.5Boersma E Poldermans D Bax JJ et al.Predictors of cardiac events after major vascular surgery: role of clinical characteristics, dobutamine echocardiography, and beta‐blocker therapy.JAMA. 2001; 285: 1865-1873Crossref PubMed Scopus (485) Google Scholar In contrast, among 1066 patients enrolled into a study of fluids and vasopressors in spinal and general surgery, 94 (9%) had evidence of coronary artery disease.4Arndt JO Bomer W Krauth J Marquardt B. Incidence and time course of cardiovascular side effects during spinal anesthesia after prophylactic administration of intravenous fluids or vasoconstrictors.Anesth Analg. 1998; 87: 347-354PubMed Google Scholar Hertzer and colleagues describe a series of 1000 patients presenting for major vascular surgery, all of whom were subjected to coronary angiography as part of their preoperative assessment. Sixty per cent of these had serious coronary artery disease.21Hertzer NR Beven EG Young JR et al.Coronary artery disease in peripheral vascular patients. A classification of 1000 coronary angiograms and results of surgical management.Ann Surg. 1984; 199: 223-233Crossref PubMed Scopus (1196) Google Scholar Given the high prevalence of coronary artery disease in patients presenting for vascular surgery, it is likely that estimates of the population burden of disease based on the incidence of cardiac events in this group may be misleading. However, although the incidence of perioperative cardiac complications is almost certainly lower in patients undergoing major surgery other than vascular (or cardiac) surgery than in those undergoing cardiovascular operations, these patients may contribute much the greater population burden of perioperative cardiac events. The Hospital Episode Statistics for England and Wales indicate that in the year 2002–2003 there were 65 567 operations with codes OPCS4 in the range L16 to L31 and L37 to L75.11Department of Health, UK Table 4, Main operations.http://www.dh.gov.uk/PublicationsAndStatistics/Statistics/HospitalEpisodeStatisticsDate: 2003Google Scholar These codes broadly cover major vascular surgical operations on arteries, but exclude operations on the cerebral arteries with codes in the range L33 to L35. In the same period 3 135 784 operations in people aged 60–74 yr and 2 323 902 operations in people aged 75 yr or more were recorded. Thus, in total 5 459 686 operations were reported in patients aged over 59 yr. The Hospital Episode Statistics report crude data and include many minor and diagnostic procedures. Perhaps only 5–10% of the procedures recorded were major operations. However, this still represents between a quarter and half a million people aged over 60 yr in England and Wales undergoing major surgery each year. It is clear that vascular surgery patients represent only a fraction of the surgical population. Several million major and minor procedures are undertaken on people aged 60 yr or over in England and Wales each year. It seems likely that, although the cardiac event rate may be higher in patients undergoing major vascular surgery, in any given year, more patients will suffer a perioperative infarction after other types of major non-cardiac surgery. If this is the case, there are considerable implications for the benefits to be gained from strategies to reduce perioperative cardiac events. An intervention that reduces risk by half in a vascular surgery patient with a 7% risk of perioperative infarction offers an absolute risk reduction to that patient of 3.5%. The same relative risk reduction in a patient undergoing orthopaedic surgery who has a 1% risk of perioperative infarction produces an absolute risk reduction of 0.5%, but may prevent many more events in the population as a whole. There are few more startling results than the reduction in cardiac death and perioperative myocardial infarction from 34 to 3.4% produced by perioperative beta-adrenergic blockade in the study conducted by Poldermans and colleagues.42Poldermans D Boersma E Bax JJ et al.The effect of bisoprolol on perioperative mortality and myocardial infarction in high‐risk patients undergoing vascular surgery. Dutch Echocardiographic Cardiac Risk Evaluation Applying Stress Echocardiography Study Group.N Engl J Med. 1999; 341: 1789-1794Crossref PubMed Scopus (1364) Google Scholar This translates into an absolute risk reduction of 30.6% and a crude odds ratio for the benefit of perioperative beta-blockade of 0.07 (0–0.21). However, this result was obtained in a group of 112 of the highest-risk patients selected from a total population of 1351 patients. The data on all 1351 patients are reported in a separate paper. A total of 45 patients suffered perioperative cardiac death or myocardial infarction. Sixteen of these events occurred in the highest-risk patients, whose risk of a perioperative cardiac event was of the order of 30%.5Boersma E Poldermans D Bax JJ et al.Predictors of cardiac events after major vascular surgery: role of clinical characteristics, dobutamine echocardiography, and beta‐blocker therapy.JAMA. 2001; 285: 1865-1873Crossref PubMed Scopus (485) Google Scholar These patients had three or more risk factors for cardiac disease and inducible wall motion abnormalities on dobutamine stress echocardiography. However, 29 of the events occurred in the much larger group of intermediate- and low-risk patients. In these patients the perioperative event rate ranged between 1.2 and 5.8% in patients not taking beta-blockers. The benefits of perioperative beta-blockade have not been proven in intermediate- or low-risk patients. The case for perioperative beta-blockade rests primarily on two randomized controlled trials. That by Poldermans and colleagues, described above, randomized only very high-risk patients.42Poldermans D Boersma E Bax JJ et al.The effect of bisoprolol on perioperative mortality and myocardial infarction in high‐risk patients undergoing vascular surgery. Dutch Echocardiographic Cardiac Risk Evaluation Applying Stress Echocardiography Study Group.N Engl J Med. 1999; 341: 1789-1794Crossref PubMed Scopus (1364) Google Scholar The other, by Mangano and colleagues, was underpowered and excluded patients who died in hospital.34Mangano DT Layug EL Wallace A Tateo I. Effect of atenolol on mortality and cardiovascular morbidity after noncardiac surgery. Multicenter Study of Perioperative Ischemia Research Group.N Engl J Med. 1996; 335: 1713-1720Crossref PubMed Scopus (1558) Google Scholar It remains to be established that the benefits of perioperative beta-blockade outweigh the risks of adverse effects in patients at relatively low risk of perioperative complications.22Howell SJ Sear JW Foex P. Peri‐operative beta‐blockade: a useful treatment that should be greeted with cautious enthusiasm.Br J Anaesth. 2001; 86: 161-164Crossref PubMed Scopus (39) Google Scholar However, the underlying message is clear. While the individual high-risk patient may gain a considerable absolute reduction in risk from an effective intervention, a greater reduction in the population burden of disease may be achieved by a more modest risk reduction in intermediate- and low-risk patients. While an intervention may yield a several-fold reduction in perioperative risk, if the baseline risk is relatively small the absolute risk reduction will also be small. In the randomized study of Oliver and colleagues comparing mivazerol and placebo, patients undergoing vascular surgery who were randomized to mivazerol had a relative risk of perioperative cardiac death of 0.33 (95% confidence interval 0.18–0.91).36Oliver MF Goldman L Julian DG Holme I. Effect of mivazerol on perioperative cardiac complications during non‐cardiac surgery in patients with coronary heart disease: the European Mivazerol Trial (EMIT).Anesthesiology. 1999; 91: 951-961Crossref PubMed Scopus (182) Google Scholar The absolute risk of cardiac death among vascular surgery patients randomized to placebo was 4% and mivazerol reduction reduced this to 1.3%, an absolute risk reduction of 2.7%. This small benefit is reflected in the number needed to treat (NNT) of 37. That is to say, 37 patients undergoing major vascular surgery need to be treated with mivazerol in order to prevent one cardiac event. In fact, this NNT is not dissimilar to that seen for some well-established interventions in other fields. For example, amongst patients aged between 65 and 74 yr presenting with myocardial infarction, 37 need to be treated with thrombolysis to prevent one death between presentation and 35 days. For patients aged between 55 and 64 yr, 56 patients need to be treated, while for patients aged less than 55 yr 91 patients need to be treated.19Heller RF Dobson AJ. Disease impact number and population impact number: population perspectives to measures of risk and benefit.BMJ. 2000; 321: 950-953Crossref PubMed Google Scholar Thus, it may be argued that the therapies that may be used to prevent perioperative infarction and those used to treat patients presenting with a myocardial infarction offer similar levels of benefit. From a public health perspective, the difference between patients presenting with myocardial infarction and those suffering perioperative myocardial infarction is that, while very many people undergo surgery, very few of these suffer a perioperative infarction, so the population burden of disease from perioperative infarction is less than that from acute non-operative infarction. In the 12-month period covered by the hospital episode statistics for 2002 to 2003 there were 105 476 consultant episodes and 69 116 admissions for acute myocardial infarction.11Department of Health, UK Table 4, Main operations.http://www.dh.gov.uk/PublicationsAndStatistics/Statistics/HospitalEpisodeStatisticsDate: 2003Google Scholar Lee and colleagues, in their study of patients undergoing major surgery, reported an overall in-hospital perioperative myocardial infarction rate in their derivation cohort of 1%.30Lee TH Marcantonio ER Mangione CM et al.Derivation and prospective validation of a simple index for prediction of cardiac risk of major noncardiac surgery.Circulation. 1999; 100: 1043-1049Crossref PubMed Scopus (2645) Google Scholar A total of 5 459 686 operations in patients aged 60 yr or over are recorded as having taken place in England and Wales in the period 2002–2003 (see above). However, this included a large number of diagnostic procedures and minor operations. One might estimate that perhaps only 10% of the operations were major procedures. If this is the case, one can estimate that perhaps five and a half thousand patients suffered a perioperative infarction in the study period. Thus, it seems likely at first sight that the absolute number of patients likely to suffer perioperative infarction is considerably lower than the number of patients presenting with infarction in the non-operative setting. The implication is that greater population benefit will be obtained from aggressive management of non-operative infarction than of perioperative infarction. This assumption will be challenged below. It is very much part of anaesthetic practice to identify patients at increased risk of perioperative complications. A number of risk scores, such as the Goldman Multifactorial Cardiac Risk Index, have been developed to assist the anaesthetist in doing this. There are a number of potential uses of information on perioperative cardiac risk. It may be used to counsel the patient on the risks that they face; to inform the planning of anaesthesia and surgery; to guide preoperative preparation; and as an epidemiological tool to examine the actual incidence of morbidity and mortality in a surgical population against a calculated expected outcome. To be of value for any of these applications, a risk score should be robust, have been tested in a number of settings, and produce reliable estimates of risk in a range of operative settings. It is not clear that the tools currently available to us fulfil these requirements. The earliest, and perhaps still the best known, cardiac risk score is the Goldman Multifactorial Cardiac Risk Index.16Goldman L Caldera DL Nussbaum SR et al.Multifactorial index of cardiac risk in noncardiac surgical procedures.N Engl J Med. 1977; 297: 845-850Crossref PubMed Scopus (2072) Google Scholar This was derived in a prospective study of 1001 patients who between them suffered 58 cardiac events (19 patients died from a cardiac cause and 39 suffered one or more life-threatening cardiac complications, but did not die). Approximately 45 possible predictors of perioperative cardiac morbidity were examined. An unspecified number of further predictors were created by the combination of raw factors. All of these potential predictors were then subjected to discriminant function analysis. Nine predictors of perioperative cardiac events were identified and given numerical weights for inclusion in the final model. Depending on their total risk score derived using these weights, patients were divided into four categories of increasing risk. The Goldman index is important both because it is still cited and used and because it led the way for subsequent work. However, it suffers from a number of weaknesses. Almost every conceivable cardiac symptom and sign was included in the initial analysis. Thus, the factors relevant to heart failure that were examined included dyspnoea, orthopnoea, oedema, New York Heart Association Score, jugular venous distension, the presence of a third or fourth heart sound, rales, pulmonary vasculature appearance on X-ray, heart size and QRS axis. No account is taken of the fact that these risk factors are not independent. That is to say, the presence of one of these factors is likely to be reflected by the presence of several others. The result is that rather than the clinical condition of heart failure being included in the risk score, it is represented by one or more of its symptoms or signs. Heart failure is represented in the final index by an elevated jugular venous pressure and/or the presence of a third heart sound. These signs may or may not be present in any individual patient with heart failure. Nine risk factors emerged from the discriminant function analysis as significant, with weights ranging from 0.123 to 0.451. These were translated into scores of between 3 and 11 points. However, there has to be some suspicion about the robustness of the weights. It is difficult, if not impossible, to perform a formal power calculation for a multivariate analysis because one cannot anticipate the relative weightings that will be afforded to individual risk factors. In the Goldman risk index the presence of a third heart sound or jugular venous distension is given a weight of 11 points and the presence of important valvular aortic stenosis a weight of 3 points. These values cannot be known in advance, but the precision with which they are determined will have a considerable impact on the final risk model. In a series of computer simulation studies of logistic regression and proportional hazards analysis, Concato and colleagues demonstrated that, if the weighting afforded to a given variable is to be determined with reasonable precision, then 10 subjects who suffered an adverse event are required for each variable in the final model.9Concato J Peduzzi P Holford TR Feinstein AR. Importance of events per independent variable in proportional hazards analysis. I. Background, goals, and general strategy.J Clin Epidemiol. 1995; 48: 1495-1501Abstract Full Text PDF PubMed Scopus (536) Google Scholar 38Peduzzi P Concato J Feinstein AR Holford TR. Importance of events per independent variable in proportional hazards regression analysis. II. Accuracy and precision of regression estimates.J Clin Epidemiol. 1995; 48: 1503-1510Abstract Full Text PDF PubMed Scopus (1326) Google Scholar 39Peduzzi P Concato J Kemper E Holford TR Feinstein AR. A simulation study of the number of events per variable in logistic regression analysis.J Clin Epidemiol. 1996; 49: 1373-1379Abstract Full Text PDF PubMed Scopus (5052) Google Scholar Thus, for the nine factors in the Multifactorial risk index, 90 adverse events would be required rather than the 58 actually observed. When tested in populations other than that in which it was derived, the index does not perform especially well.8Charlson ME Ales KL Simon R MacKenzie CR. Why predictive indexes perform less well in validation studies. Is it magic or methods?.Arch Intern Med. 1987; 147: 2155-2161Crossref PubMed Scopus (180) Google Scholar 30Lee TH Marcantonio ER Mangione CM et al.Derivation and prospective validation of a simple index for prediction of cardiac risk of major noncardiac surgery.Circulation. 1999; 100: 1043-1049Crossref PubMed Scopus (2645) Google Scholar It appears to be poor at identifying high-risk patients. These formed only a small part of the derivation data set, so relatively few data were available on these patients. At the same time, they are perhaps the group of greatest interest to the clinician. In the validation study conducted by Lee and colleagues, the performance of a modified risk index, derived by these authors and described below, was compared with that of the Goldman risk index.30Lee TH Marcantonio ER Mangione CM et al.Derivation and prospective validation of a simple index for prediction of cardiac risk of major noncardiac surgery.Circulation. 1999; 100: 1043-1049Crossref PubMed Scopus (2645) Google Scholar The area under the receiver operating characteristic (ROC) curve for the Goldman index was 0.701 (se 0.043), while that under the ROC curve for the index derived by Lee and colleagues was 0.806 (se 0.034). While the areas under these two curves were significantly different (P=0.021), the difference does not appear startlingly great. Indeed, the performance of the indices was very similar for the lower risk categories of each index. This was not the case for the highest risk categories of each index, class IV in each case. The Goldman index predicts a 78% cardiac event rate in this group. Neither of the two Goldman class IV patients from the validation cohort, nor the seven Goldman class IV patients in the study population as a whole, suffered a cardiac event. In contrast, 11% of the 109 patients place in class IV by the Lee index suffered a cardiac event. This is very similar to the predicted event rate of 9.1% (95% confidence interval 5.5–13.8%). For all its shortcomings, the Goldman index is important because it laid the groundwork for further endeavours in this field.12Detsky AS Abrams HB Forbath N Scott JG Hilliard JR. Cardiac assessment for patients undergoing noncardiac surgery. A multifactorial clinical risk index.Arch Intern Med. 1986; 146: 2131-2134Crossref PubMed Scopus (413) Google Scholar 29Larsen SF Olesen KH Jacobsen E et al.Prediction of cardiac risk in non‐cardiac surgery.Eur Heart J. 1987; 8: 179-185Crossref PubMed Scopus (120) Google Scholar 30Lee TH Marcantonio ER Mangione CM et al.Derivation and prospective validation of a simple index for prediction of cardiac risk of major noncardiac surgery.Circulation. 1999; 100: 1043-1049Crossref PubMed Scopus (2645) Google Scholar The most current scoring system for cardiac risk in non-cardiac surgery is probably that of Lee and colleagues.30Lee TH Marcantonio ER Mangione CM et al.Derivation and prospective validation of a simple index for prediction of cardiac risk of major noncardiac surgery.Circulation. 1999; 100: 1043-1049Crossref PubMed Scopus (2645) Google Scholar They studied 4315 patients, dividing them into a derivation cohort of 2893 patients from which a risk score was derived and a validation cohort of 1422 patients on which the performance of the score was examined. This includes six risk factors, as listed in Table 1. It will be seen that these represent clinical conditions such as ischaemic heart disease rather than the signs and symptoms that are in the Goldman risk index. The odds ratios for the individual risk factors are all of the same order of magnitude and in the risk score all the factors are given the same weight. Patients with zero, one, two and three or more risk factors are assigned to classes I, II, III and IV respectively. The score was derived from a data set that included 2893 patients who suffered 56 adverse events, so that there were 9.3 adverse events for each risk factor in the final model.Table 1Risk factors included in the cardiac risk index developed by Lee and colleagues30High-risk type of surgeryIschaemic heart diseaseHistory of congestive heart failureHistory of cerebrovascular diseaseInsulin therapy for diabetesPreoperative serum creatinine >2.0 mg dl−1 Open table in a new tab The Lee score divides patients into four categories in which the risk of cardiovascular complications ranges from 0.5 to 9.1%. Even in the highest risk category the risk of complications is less than 10%. This and other risk scoring systems are specific but not especially sensitive. A patient with a low score who is placed in the lowest risk category is very unlikely to suffer cardiovascular complications. The opposite is not true, however. Patients in the highest risk category are not extremely likely to suffer a complication. In an unselected population, in which the overall risk of perioperative cardiac complications is low, the distinction between sensitivity and specificity and positive and negative predictive value is thrown into sharp relief. Even using a highly specific scoring system, the large group of patients who will not suffer a perioperative cardiac complication will generate many false-positive results. The false-positive results will outnumber the true positives, leading to a low positive predictive value. The available risk scoring systems do not dichotomize patients into high- and low-risk groups. A risk of complications of 10 or 15% is worrying but would rarely constitute grounds to cancel surgery. Risk scores are not tools to direct management. The risk scores are tools derived from population studies, and their strength lies at this level. It is not appropriate (from an epidemiological perspective) to ask ‘why did this patient, who had a Lee score of 1, suffer a perioperative myocardial infarction?’ It is reasonable to ask why the perioperative infarction rate in patients with a Lee score of 1 is observed to be considerably higher than the expected 1.3%. The cardiac risk scores are based on perioperative outcome. Both the Goldman Multifactorial Risk Index and the Lee cardiac risk index were derived using cardiac events that occurred during the patients’ hospital stay.16Goldman L Caldera DL Nussbaum SR et al.Multifactorial index of cardiac risk in noncardiac surgical procedures.N Engl J Med. 1977; 297: 845-850Crossref PubMed Scopus (2072) Google Scholar 30Lee TH Marcantonio ER Mangione CM et al.Derivation and prospective validation of a simple index for prediction of cardiac risk of major noncardiac surgery.Circulation. 1999; 100: 1043-1049Crossref PubMed Scopus (2645) Google Scholar The anaesthetist's perception of operative risk tends to be based on how the patient fares whilst in hospital. It is becoming clear that the story is not so simple. The cardiac troponin assays are described in a review elsewhere in this Postgraduate Issue. The troponins are cardiac-specific proteins released during myocardial injury. Studies using the cardiac troponins in patients presenting with chest pain have led to a revision of our understanding of myocardial infarction. Infarction is no longer viewed as a binary event, such that patients have either suffered a myocardial infarction or they have not. Rather, it is now held that there is a spectrum of myocardial injury characterized by increasing troponin release and ranging from stable angina, through unstable angina to full-blown myocardial infarction.1Adams 3rd, JE Clinical application of markers of cardiac injury: basic concepts and new considerations.Clin Chim Acta. 1999; 284: 127-134Crossref PubMed Scopus (22) Google Scholar Amongst patients presenting with chest pain, some troponin release is seen in many who are not subsequently diagnosed as having suffered a myocardial infarction. The greater the magnitude of this troponin release, the greater the likelihood of a further cardiac event in the ensuing weeks and months.3Antman EM Tanasijevic MJ Thompson B et al.Cardiac‐specific troponin I levels to predict the risk of mortality in patients with acute coronary syndromes.N Engl J Med. 1996; 335: 1342-1349Crossref PubMed Scopus (1545) Google Scholar 18Hamm CW Goldmann BU Heeschen C Kreymann G Berger J Meinertz T. Emergency room triage of patients with acute chest pain by means of rapid testing for cardiac troponin T or troponin I.N Engl J Med. 1997; 337: 1648-1653Crossref PubMed Scopus (951) Google Scholar 32Lindahl B Venge P Wallentin L. Relation between troponin T and the risk of subsequent cardiac events in unstable coronary artery disease. The FRISC study group.Circulation. 1996; 93: 1651-1657Crossref PubMed Scopus (553) Google Scholar 35Ohman EM Armstrong PW Christenson RH et al.Cardiac troponin T levels for risk stratification in acute myocardial ischemia. GUSTO IIA Investigators.N Engl J Med. 1996; 335: 1333-1341Crossref PubMed Scopus (1010) Google Scholar 45Ravkilde J Nissen H Horder M Thygesen K. Independent prognostic value of serum creatine kinase isoenzyme MB mass, cardiac troponin T and myosin light chain levels in suspected acute myocardial infarction. Analysis of 28 months of follow‐up in 196 patients.Am J Cardiol. 1995; 25: 574-581Crossref Scopus (231) Google Scholar It is becoming clear that similar considerations apply to myocardial injury in the context of non-cardiac surgery. Several studies, such as that by Haggart and colleagues and the Vascular Anaesthesia Society Study by Howell and colleagues, have shown a continuum of cardiac troponin release in patients undergoing vascular surgery.17Haggart PC Adam DJ Ludman PF Bradbury AW. Comparison of cardiac troponin I and creatine kinase ratios in the detection of myocardial injury after aortic surgery.Br J Surg. 2001; 88: 1196-1200Crossref PubMed Scopus (44) Google Scholar 23Howell SJ Thompson J Nimmo A et al.Subclinical perioperative myocardial injury in patients undergoing major vascular surgery.Br J Anaesth. 2001; 87: 661P-662PGoogle Scholar Just as increasing troponin release is associated with increasing risk in patients presenting with chest pain, a number of studies suggest that the same is true for non-cardiac surgery patients. Kim and colleagues studied 229 patients undergoing aortic or infra-inguinal surgery or lower extremity amputation.25Kim LJ Martinez EA Faraday N et al.Cardiac troponin I predicts short‐term mortality in vascular surgery patients.Circulation. 2002; 106: 2366-2371Crossref PubMed Scopus (251) Google Scholar Cardiac troponin I (cTnI) levels were measured immediately after surgery and on postoperative days 1, 2 and 3. Ninety-eight patients had cTnI levels above 0.35 ng/ml (the lower detection limit of the assay used). There was a clear dose–response relationship between the serum cTnI level and the risk of death in the 6 months after surgery. Taking the group of patients with cTnI levels less than 0.35 ng/l as the reference group, the odds ratio for death (95% confidence interval) in the 6 months after surgery was 1.3 (0.4–4.4) for patients with a troponin level between 0.4 and 1.5 ng/ml. It was 4.3 (0.8–24.3) for patients with a cTnI level between 1.6 and 3.0 mg/ml, and 4.9 (1.9–19.0) for patients with cTnI levels greater than 3.0 ng/ml. Landesberg and colleagues obtained similar results in a study of 447 patients undergoing major vascular surgery. Cardiac troponin levels were measured in all patients for the first 3 days after surgery. The odds ratio for death during long-term follow-up (between 1 and 5 yr) increased steadily with increasing postoperative cardiac troponin levels.28Landesberg G Shatz V Akopnik I et al.Association of cardiac troponin, CK‐MB, and postoperative myocardial ischemia with long‐term survival after major vascular surgery.J Am Coll Cardiol. 2003; 42: 1547-1554Crossref PubMed Scopus (403) Google Scholar Fewer data exist on the incidence and implications of troponin release in patients undergoing non-vascular surgery. Further data are needed both to determine the incidence of cardiac troponin release in patients undergoing non-cardiac, non-vascular surgery and to determine the long-term implications of such release. One would certainly expect the incidence of myocardial injury to be lower in patients not undergoing vascular surgery but, as already discussed, there are many more of these patients compared with the vascular surgery population. Potentially, perioperative myocardial injury going undetected at the time of operation, but associated with subsequent cardiac events, carries with it a very considerable population burden of disease. If this is indeed the case, the focus of perioperative medicine needs to change from the prevention of perioperative myocardial infarction to the prevention of any perioperative myocardial injury. In many respects the approach will not differ from that to the prevention of perioperative infarction. It remains appropriate for anaesthetists to take due care over identifying at-risk patients and to plan their preoperative assessment and perioperative care appropriately. However, the issue of secondary prevention of cardiovascular disease is thrown into sharper perspective. In the ideal world, all patients with a history of cardiovascular disease who arrive at hospital for surgery and who are found not to be receiving secondary prevention would have this rectified before discharge. It is already clear that patients presenting with peripheral vascular disease benefit from treatment with statins.24Kertai MD Boersma E Westerhout CM et al.Association between long‐term statin use and mortality after successful abdominal aortic aneurysm surgery.Am J Med. 2004; 116: 96-103Abstract Full Text Full Text PDF PubMed Scopus (195) Google Scholar 41Poldermans D Bax JJ Kertai MD et al.Statins are associated with a reduced incidence of perioperative mortality in patients undergoing major noncardiac vascular surgery.Circulation. 2003; 107 (Epub 2003 Apr 14): 1848-1851Crossref PubMed Scopus (424) Google Scholar It may be argued that patients who present for vascular surgery should receive cardiovascular secondary prevention in the same way as patients with cardiac disease. A consensus statement on the use of antiplatelet therapy in patients with peripheral vascular disease was published in 2003.40Peripheral PADACG. Antiplatelet therapy in peripheral arterial disease. Consensus statement.Eur J Vasc Endovasc Surg. 2003; 26: 1-16Abstract Full Text Full Text PDF PubMed Scopus (67) Google Scholar National cardiac guidelines in the UK suggest that patients who have intermittent claudication should be managed in the same way as patients with established coronary artery disease, although we are still some way from achieving this ideal.6Cassar K Belch JJ Brittenden J. Are national cardiac guidelines being applied by vascular surgeons?.Eur J Vasc Endovasc Surg. 2003; 26: 623-628Abstract Full Text Full Text PDF PubMed Scopus (28) Google Scholar 7Cassar K Coull R Bachoo P Macaulay E Brittenden J. Management of secondary risk factors in patients with intermittent claudication.Eur J Vasc Endovasc Surg. 2003; 26: 262-266Abstract Full Text Full Text PDF PubMed Scopus (30) Google Scholar It now seems reasonable to ask if patients with one or more risk factors for cardiovascular disease, who undergo surgery, should have pre- and postoperative measurements of cardiac troponin and should receive cardiovascular secondary prevention if any postoperative elevation is detected. This relatively cheap intervention might have a significant impact on the long-term outcome in this group. In summary, it is clear that perioperative myocardial infarction is a significant issue in patients undergoing vascular surgery. While the incidence of perioperative cardiac complications may be lower in patients undergoing other types of non-cardiac surgery, there are many more of these patients and their impact on the population burden of disease may be greater. Identifying patients at the highest risk of perioperative infarction can be difficult, and risk scores, while potentially of value for population studies, are not an ideal tool for directing care in the individual patient. To some extent this is due to the relatively low incidence of perioperative cardiac events and, thus, the problem in identifying a person likely to suffer such an event. However, studies with cardiac troponins make it clear that the incidence may not be as low as first thought and imply that subclinical perioperative myocardial injury may, in fact, be quite common. Further studies are needed in this field, but if this is the case the development of formal strategies for preventing, identifying and managing subclinical preoperative myocardial injury may be appropriate.