Resuscitation (2008) 79, 4145

available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/resuscitation

CLINICAL PAPER

Do risk factors for chronic coronary heart disease help diagnose acute myocardial infarction in the Emergency Department?
Richard Body , Garry McDowell, Simon Carley, Kevin Mackway-Jones
Emergency Medicine Research Group, Research Ofce, Emergency Department, Manchester Royal Inrmary, Oxford Road, Manchester M13 9WL, United Kingdom Received 29 November 2007; received in revised form 15 April 2008; accepted 8 June 2008

KEYWORDS
Myocardial infarction; Acute coronary syndromes; Diagnosis; Emergency Department

Summary Background: Hypertension, hyperlipidaemia, diabetes mellitus, tobacco smoking and a family history of premature coronary artery disease are known to be risk factors for the development of coronary artery disease. We sought to determine whether these traditional risk factors aid the diagnosis of acute myocardial infarction (AMI) in the Emergency Department (ED). Methods: We performed a prospective diagnostic cohort study within the ED at Manchester Royal Inrmary, a university-afliated teaching hospital with an annual ED census of approximately 145,000 patients. We recruited 804 patients who had presented to the ED with suspected cardiac chest pain. All patients had the presence or absence of traditional cardiac risk factors documented at the time of presentation using a custom-designed clinical report form. All patients subsequently underwent 12-h troponin T testing to provide a robust gold standard for the diagnosis of AMI according to revised World Health Organisation criteria. Results: The absence of any traditional cardiac risk factors carried a negative likelihood ratio of 0.61 for the diagnosis of AMI. 12.2% of patients with no cardiac risk factors had AMI, compared with 21.3% of patients with four or ve risk factors. The area under the receiver-operating characteristic curve was 0.49. Conclusions: Traditional cardiac risk factors are not helpful for the conrmation or exclusion of AMI within the ED. Future Emergency Medicine research should focus on those clinical and diagnostic features that are likely to alter during the acute phase of illness. 2008 Elsevier Ireland Ltd. All rights reserved.

A Spanish translated version of the summary of this article appears as Appendix in the online version at doi:10.1016/j.resuscitation.2008.06.009. Corresponding author. Tel.: +44 7880 712 929; fax: +44 1612 766 925. E-mail address: rbody@doctors.org.uk (R. Body). 0300-9572/$ see front matter 2008 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.resuscitation.2008.06.009

42

R. Body et al. idaemia or diabetes mellitus if they reported that they had the condition or that they were taking medication for the condition. Patients were considered to have a positive family history if they reported that a rst degree relative had ischaemic heart disease at 65 years of age. Patients were considered to be smokers if they reported current tobacco smoking or tobacco smoking within the past 6 weeks. All patients had blood taken for troponin T testing 12 h after onset of their most signicant symptoms (Roche Diagnostics, diagnostic cut-off 0.035 ng/ml). Any patient who had been discharged without an appropriately timed troponin test was recalled or visited the following day for venepuncture. All patients were followed up 6 months after presentation. At that time the patients outcome was checked by examining the National Health Strategic Tracing Service (NSTS) database for mortality data and hospital records for clinic letters, investigation reports and details of hospital admissions and/or ED attendances. Patients were then interviewed by telephone. If no contact could be established with the patient, their general practitioner was contacted.

Background
The Framingham Heart Study played a vital role in identifying important risk factors for the development of coronary heart disease.1 These included hypertension,2,3 hyperlipidaemia,2,4 tobacco smoking,36 diabetes mellitus79 and a family history of premature coronary artery disease.10 While it is apparent that each of these risk factors predisposes an individual to the development of coronary heart disease, the precise role of these risk factors in conrming or excluding acute myocardial infarction (AMI) among undifferentiated patients who present to the Emergency Department (ED) with suspected cardiac chest pain is not clear. Despite this many clinicians feel that a cardiac history is incomplete unless their presence or absence has been documented. We aimed to evaluate the usefulness of the presence or absence of each of the ve traditional cardiac risk factors in diagnosing or excluding AMI within the ED.

Methods
Study design and setting
This is a substudy of the Early Vascular Markers of Acute Coronary Syndromes (EVaMACS) study, a single-centre prospective diagnostic cohort study. The study was approved by the Local Research Ethics Committee. All patients provided written informed consent. We recruited patients at Manchester Royal Inrmary, a university-afliated teaching hospital. The annual ED census is approximately 145,000 (comprising approximately 39,000 cases triaged to Majors including the Resuscitation Room and treatment cubicles, 43,000 minor injuries, 19,000 ophthalmological emergencies, 24,000 primary care emergencies (patients who are seen by a general (family) practitioner or equivalent), 13,000 presentations to the Walk in Centre (patients with minor complaints who are seen by a nurse practitioner) and 7000 others).

Outcome measures
The primary outcome was a diagnosis of acute myocardial infarction (AMI), dened according to recent American Heart Association and European Society of Cardiology guidelines.11 Patients fullled the diagnosis of AMI if they had a troponin T elevation 0.035 ng/ml (i.e. above the 99th percentile of the upper reference limit with a co-efcient of variation 10%) with at least one of the following: symptoms of ischaemia, ECG evidence of AMI (acute ischaemic changes or development of pathological Q waves) or imaging evidence of new loss of viable myocardium. If any patient should have died before troponin testing, diagnosis was to be assigned according to post-mortem ndings.12 The principal outcomes at 6-month follow up were death, AMI (excluding the index event) and urgent coronary revascularisation. Coronary revascularisation included the detection of a new angiographic coronary stenosis of 50% where revascularisation could not be achieved, as reported by the patients interventional cardiologist.

Selection of participants
Consecutive patients aged over 25 years presenting to the ED with chest pain occurring within the past 24 h that the treating physician suspected to be cardiac in origin were eligible for inclusion. Patients were excluded if they had another medical condition mandating hospital admission, renal failure needing dialysis, signicant chest trauma with suspicion of myocardial contusion, if they were pregnant, did not speak English, were prisoners or if follow-up would be impossible.

Statistical methods
To assess for clinical utility sensitivities, specicities and likelihood ratios were calculated. Receiver-operating characteristic curves were created from the sensitivities and specicities obtained when using 0, 1, 2, 3, 4 and 5 risk factors as cut-offs. Logistic regression was used to determine whether each individual risk factor helped to predict the diagnosis of AMI. Odds ratios with 95% condence intervals are reported. All statistical analyses were performed using SPSS version 12.0. Score condence intervals, calculated using accepted methodology, are reported.13

Data collection and processing

Clinical data was recorded by the initial treating physician at the time of ED presentation using a custom-designed clinical report form. The presence or absence of each cardiac risk factor was recorded using tick boxes. If no boxes were ticked, the data was considered to be missing. Patients were considered to have a history of hypertension, hyperlip-

Results
804 patients were recruited to the study between January 2006 and February 2007. Eight patients were excluded

Risk factors and myocardial infarction

43

Table 1 Baseline characteristics according to diagnosis (AMI, met diagnostic criteria for acute myocardial infarction; non-AMI, did not meet diagnostic criteria for acute myocardial infarction) Variable Age in years (S.D.) Men (%) Previous angina (%) Previous myocardial infarction (%) Hypertension (%) Hyperlipidaemia (%) Diabetes mellitus (%) Smoking (%) Family history (%) Number of risk factors (%) 0 1 2 3 4 or 5 Grade of attending physician (%) Senior House Ofcer Registrar Consultant Total (N = 796) 58.9 (14.2) 481 (60.4) 258 (32.4) 194 (24.4) 399 (50.1) 379 (47.6) 141 (17.7) 247 (31.0) 379 (47.6) 90 (11.3) 212 (26.6) 237 (29.8) 177 (22.2) 80 (10.1) 71 (8.9) 724 (91.0) 1 (0.1) AMI (N = 148) 63.1 (13.2) 104 (70.3) 35 (23.6) 32 (21.6) 73 (49.3) 57 (38.5) 23 (15.5) 69 (46.6) 63 (42.6) 11 (7.4) 51 (34.5) 45 (30.4) 24 (16.2) 17 (11.0) 13 (8.8) 135 (91.2) 0 Non-AMI (N = 648) 57.9 (14.3) 377 (58.2) 223 (34.4) 163 (25.2) 326 (50.3) 322 (49.7) 118 (18.2) 178 (27.5) 316 (48.8) 79 (12.2) 171 (24.8) 192 (29.6) 153 (23.6) 63 (9.8) 58 (9.0) 589 (90.9) 1 (0.2)

because they were found to meet pre-dened exclusion criteria, meaning that 796 patients were suitable for nal analysis. 148 (18.6%) patients had AMI. There was no missing data regarding the presence or absence of cardiac risk factors in any patient. All patients underwent troponin T testing 12 h after the onset of their most signicant symptoms. No patients were lost to follow up at 6 months. Table 1 shows the baseline characteristics of the included patients stratied by the presence of AMI. The incidence of AMI according to the number of risk factors present is demonstrated in Figure 1. There was no trend towards increasing incidence of AMI with increasing number of risk factors. The receiver-operating characteristic (ROC) curve demonstrating the value of cardiac risk factors for the diagnosis of AMI is shown in Figure 2. The area under the ROC curve was 0.49 (95% condence intervals 0.440.54). When tested against the null hypothesis that the true area under the curve is 0.50 this yielded a p-value of 0.59, suggesting

that cardiac risk factor burden is not useful in the diagnosis or exclusion of AMI. The sensitivities, specicities and likelihood ratios at each cut-off are demonstrated in Table 2. The absence of any cardiac risk factors carried a negative likelihood ratio of 0.61 for the diagnosis of AMI, a sensitivity of 92.6% (95% condence intervals 87.295.8%) and a negative predictive value of 87.8% (79.493.0%). The results of the logistic regression analysis are shown in Table 3. Only smoking carried signicant positive predictive

Figure 1

Incidence of AMI according to risk factor burden.

Figure 2 Receiver-operating characteristic curve for risk factors and acute myocardial infarction.

44
Table 2 Sensitivities, specicities and predictive values for each risk factor burden cut-off for diagnosis of AMI Sensitivity, % (95% CI) 92.6 58.1 27.7 11.5 (87.295.8) (50.165.8) (21.335.4) (7.317.6) Specicity, % (95% CI) 12.2 37.0 66.7 90.3 (9.914.9) (33.440.8) (62.970.2) (87.892.3) LR+ 1.05 0.92 0.83 1.18

R. Body et al.

Number of risk factors 1 2 3 4

CI: condence intervals.

LR 0.61 1.12 1.08 0.98

value for the diagnosis of AMI. Interestingly, the presence of hyperlipidaemia was a weak negative predictor of AMI. As only one cardiac risk factor was a signicant positive predictor on univariate analysis a multivariate analysis was not undertaken. The incidence of adverse events at 6-month follow up is demonstrated in Table 4. With increasing number of risk factors there was a weak trend towards increasing incidence of death or AMI and a stronger trend towards increasing incidence of death, AMI or urgent revascularisation.

Discussion
Our results indicate that traditional cardiac risk factors do not assist in the diagnosis or exclusion of AMI within the ED. Of the ve traditional risk factors only smoking was a signicant positive predictor of the diagnosis of AMI. While the absence of any cardiac risk factors has slight negative predictive value, a likelihood ratio of 0.61 means that in our population with a 18.6% incidence of AMI (a similar incidence to other similar studies14 ) the post-test probability of AMI is shifted only to 12.2%. Further, the presence at least four of the ve traditional cardiac risk factors does not help to conrm a diagnosis of AMI, as the post-test probability is shifted only as far as 21.3%. While we found that the number of traditional risk factors did not correlate with the incidence of AMI, we did demonstrate a correlation with the incidence of death, AMI or urgent coronary revascularisation within 6 months. Thus, it may be that while cardiac risk factors are useful for predicting prognosis with regard to coronary heart disease, they do little to tell us what is actually happening now. Interestingly we found that the presence of hyperlipidaemia carried weak but statistically signicant negative predictive value for the diagnosis of AMI. This may be a statistical anomaly. However, there may be a rational explanation for this observation. Patients who have been told that they have hyperlipidaemia may have a heightened awareness of their cardiac risk and therefore more likely to attend

Table 3 Logistic regression analyses of risk factors for diagnosis of AMI (univariate) Variable Hyperlipidaemia Hypertension Diabetes mellitus Smoking Family history Odds ratio (95% CI) 0.63 0.96 0.83 2.31 0.78 (0.440.91) (0.671.37) (0.511.35) (1.603.23) (0.541.12) p-Value 0.014 0.829 0.443 <0.0001 0.174

the ED with chest pain. Further, diagnostic thresholds for hyperlipidaemia are currently much lower than they were when the original Framingham study data was derived. It is possible that only severe hyperlipidaemia is predictive of AMI within the ED. Finally, as it is only possible to assess self-reported risk factors at the point of presentation to the ED it is not possible for us to know whether patients who reported a history of hyperlipidaemia had a raised lowdensity lipoprotein concentration, which is known to be a strong risk factor for coronary heart disease or (for example) hypertriglyceridaemia, which is a weak one. This study was appropriately designed to achieve the stated objectives. It is a prospective diagnostic cohort study set within the ED. We recruited undifferentiated patients with chest pain that the initial treating physician felt may represent an acute coronary syndrome at the time of ED presentation. This is the precise group of patients for whom the study question is relevant. Further, all patients underwent robust gold standard investigations for the diagnosis of AMI. Finally, our custom-designed clinical report form documented either the presence or absence of each risk factor and there was no missing data. The study has several limitations, however. We documented the presence of hypertension, hyperlipidaemia or diabetes mellitus if the patient reported that they had the condition or were taking medication for that condition. Medical records were not consulted as these are not immediately available within the ED. While it is therefore possible that these conditions may have been over or under reported, the history obtained in this study is that which would be available to clinicians. Further, it is not always easy for patients to remember whether a rst degree relative developed coronary heart disease before 65 years of age. Where there remained some uncertainty regarding the precise age at which the disease was diagnosed, patients were considered to have a positive family history. Finally, this study included a relatively small cohort of patients and was undertaken in a single inner city ED, which is an important limitation. It is not possible for us to know whether these results can be generalised to other centres with different patient populations. The results of this study do agree with previous data from a United States registry, which suggested that cardiac risk factor burden has limited clinical value for diagnosing ACS in the ED15 and with a cohort study undertaken in 1979 in the United States.16 However, studies that have derived and validated the Thrombolysis in Myocardial Infarction (TIMI) risk score have found that the presence of at least three of the ve traditional cardiac risk factors helps to predict the development of adverse cardiac events in the near future.17,18 In the undifferentiated ED population, the pres-

Risk factors and myocardial infarction

Table 4 Risk factors and incidence of adverse events at 6-month follow up Number of risk factors 0 Death/AMI, number (%) Death/AMI/UR, number (%) 3(3.3) 14(15.6) 1 10 (4.7) 44 (20.8) 2 8 (3.4) 54 (22.8) 3 11 (6.2) 41 (23.2) 4 or 5

45

5 (6.9) 27 (33.8)

Abbreviations: AMI, acute myocardial infarction; UR, urgent coronary.

ence of at least three risk factors was associated with a relative risk of 2.13 for death, AMI or coronary revascularisation within 30 days.18 These ndings are also consistent with our results. While cardiac risk factors are not helpful in the diagnosis of AMI they may help to identify populations who are at higher risk of events in the near future. It is also possible that the presence of risk factors increases the probability that patients will undergo diagnostic coronary angiography and thus a revascularisation procedure. There is now a growing amount of evidence to support the theory that traditional cardiac risk factors are unhelpful in the ED diagnosis or exclusion of AMI. Risk factor burden does not alter the post-test probability of AMI to an extent that approaches clinical signicance. These results should be considered by practitioners, policy makers and guideline developers to ensure that the presence or absence of cardiac risk factors does not inuence the decision to investigate patients with possible cardiac chest pain.

Conclusions
Risk factors for chronic coronary heart disease are not clinically helpful in the diagnosis or exclusion of AMI in the ED setting. Future Emergency Medicine research should focus upon factors that are likely to alter during the acute phase of illness.

Conict of interest
The authors conrm that they have no conicts of interest to declare with regard to this study.

Acknowledgements
The authors would like to acknowledge the contribution of Dr. Jamie Ferguson, who assisted with patient recruitment and follow up and research administration. We would also like to acknowledge the assistance given by all of the nursing and medical staff in the Emergency Department at Manchester Royal Inrmary. Study sponsors: The study was sponsored by Central Manchester & Manchester Childrens NHS Trust, who played no role in data analysis or the decision to submit the manuscript for publication.

References
1. Kannel WB, McGee D, Gordon T. A general cardiovascular risk prole: the Framingham study. American Journal of Cardiology 1978;38:4651.

2. Csastelli WP, Anderson K. A population at risk: prevalence of high cholesterol levels in hypertensive patients in the Framingham Study. American Journal of Medicine 1986;80:2332. 3. Kannel WB. Hypertension, blood lipids, and cigarette smoking as co-risk factors for coronary heart disease. Annals of the New York Academy of Science 1978;304:12839. 4. Anderson KM, Castelli WP, Levy D. Cholesterol and mortality. 30 years of follow-up from the Framingham study. JAMA: The Journal of the American Medical Association 1987;25716:217680. 5. Doyle JT, Dawber TR, Kannel WB, Heslin AS, Kahn HA. Cigarette smoking and coronary heart disease: combined experience of the Albany and Framingham studies. New England Journal of Medicine 1988;266:796801. 6. Kannel WB. Cigarette smoking and coronary heart disease. Annals of Internal Medicine 1964;60:11036. 7. Kannel WB, McGee DL. Diabetes and cardiovascular disease. The Framingham study. JAMA: The Journal of the American Medical Association 1979;24119:20358. 8. Kannel WB, McGee DL. Diabetes and cardiovascular risk factors: the Framingham study. Circulation 1979;591:813. 9. Kannel WB, DAgostino RB, Wilson PW, Belanger AJ, Gagnon DR. Diabetes, brinogen, and risk of cardiovascular disease: the Framingham experience. American Heart Journal 1990;1203:6726. 10. Myers RH, Kiely DK, Cupples A, Kannel WB. Parental history is an independent risk factor for coronary artery disease: the Framingham study. American Heart 1990;1204:9639. 11. Thygesen K, et al. Universal Denition of Myocardial Infarction. Circulation 2007;11622:263453. 12. Apple FS, Wu AHB, Jaffe AS. European Society of Cardiology and American College of Cardiology guidelines for redenition of myocardial infarction: how to use existing assays clinically and for clinical trials. American Heart Journal 2002;144:9816. 13. Zhou XH, Obuchowski NA, McClish DK. Statistical methods in diagnostic medicine. 1st ed. New York: Wiley Interscience; 2002. 14. Carley SD, Jenkins M, Mackway-Jones K. Body surface mapping versus the standard 12 lead ECG in the detection of myocardial infarction amongst Emergency Department patients: a Bayesian approach. Resuscitation 2005;64:30914. 15. Han JH, Lindsell CJ, Storrow AB, et al. The role of cardiac risk factor burden in diagnosing acute coronary syndromes in the Emergency Department setting. Annals of Emergency Medicine 2007;49:14552. 16. Jayes Jr RL, Beshansky JR, DAgostino RB, Selker HP. Do patients coronary risk factor reports predict acute cardiac ischemia in the Emergency Department? A multicenter study. Journal of Clinical Epidemiology 1992;456:6216. 17. Antman EM, Cohen M, Bernink PJLM, et al. The TIMI risk score for unstable angina/non-ST elevation MI: a method for prognostication and therapeutic decision making. JAMA 2000;2847: 83542. 18. Chase M, Robey JL, Zogby KE, Sease KL, Shofer FS, Hollander JE. Prospective validation of the Thrombolysis in Myocardial Infarction Risk score in the Emergency Department chest pain population. Annals of Emergency Medicine 2006;483:2529.