Risk Factors for Acute Myocardial Infarction in Our Patient Population A Retrospective Pilot Study

Introduction

Death from cardiac disease remains one of the leading causes of death in the U.S.  In our aging population where cardiac disease is common, one could reasonably expect an increases in this trend. Risks factors for AMI have been well established through numerous studies; the most well known of these being the Framingham Study. ¹   Also, many studies have evaluated risks factors for a small subset of post-menopausal women.  No study, to date, has specifically evaluated an inner city population, such as the population served by St. Barnabas Hospital, to see if these previously established risk factors are the same.

St. Barnabas is a busy, inner city hospital with approximately 93,000 to 95,000 emergency department visits per year.  Of those visits, there are a high number of patients complaining of chest pain.  Only a small number of these patients, approximately 180 patients per year, rule in for acute myocardial infarction.  Our goal is to investigate if the typical risk factors associated with acute myocardial infarction (AMI), are statistically significant in our patient population.

Recent studies suggest that there may be some disagreement regarding how significant cardiac risk factors are for predicting AMI.  “There has been controversy over whether the so-called “coronary risk factors” (i.e. diabetes, hypertension, smoking, hypercholesterolemia, age greater than 55) should be weighed significantly in the emergency physician’s triage decision.  An early report suggested that such factors, which were initially derived because of their ability to predict the development of coronary atherosclerosis and its complication over decades, have minimal predictive value as to whether a patient who comes to the emergency department will develop an acute myocardial infarction“. ²

Our goal is to see if the currently established risk factors for AMI are statistically significant in our population.  Upon completion of our study, we hope it will be used as a pilot study for a future prospective analysis of our patient population.  This, in theory, would lead to an improved predictive model of AMI with proactive endeavors to identify and educate those patients truly at risk for AMI. Hopefully, this study will serve as a channel to reach those patients who fit our predictive model and therefore work to prevent future cardiac events from occurring.

Method

A retrospective review was performed on all patients’ charts with the diagnosis of acute myocardial infarction, at St. Barnabas Hospital, from January 2000 to December 2001. A total of 358 charts were reviewed. Patients of all ethnic backgrounds, age, sex, and health status were included. The only inclusion criterion was the diagnosis of AMI.  The following was obtained from each patient’s chart; age greater than 55, sex, menopausal status of women, diabetes, hypertension,  dyslipidemia, previous MI, family history of MI (parent or immediate sibling),smoking history, and active cocaine use. The information extracted was obtained both from the emergency department, as well as, the patient’s discharge summary.  This information was then placed on a data form sheet.  If any information was unable to be obtained from the review of the chart it was given a designation of unknown.  Two patients were excluded from our studies that were inadvertently given the incorrect diagnosis of AMI.

A percentage was calculated for each of the above listed risk factors.  These percentages were then compared to percentages found in two large retrospective reviews published in the Journal of the American Medical Association in 2003.

The first review was “Major Risk Factors as Antecedents of Fatal and Nonfatal Coronary Heart Disease Events” by Greenland et al. ³   This article reviewed the risk factors for AMI from three prospective cohort studies.  These included the Chicago Heart Association Detection Project in Industry, with a population sample of 35,642 employed men and women aged 18-59 years.  Screenees for the Multiple Risk Factor Intervention Trial, included 347,978 men aged 35-57 years; and a population-based sample of 3,295 men and women aged 35-59 from the Framingham Heart Study.

The second review was “Prevalence of Conventional Risk Factors in Patients with Coronary Heart Disease” by Khot et al. ⁵  This article analyzed the risk factor data for 122,458 patients enrolled in 14 international randomized clinical trials of coronary heart disease (CHD) conducted during the prior decade. Patients included 76,716 with ST-elevated myocardial infarctions, 32,527 with unstable angina/non-ST-elevated myocardial infarctions, and 10,215 undergoing percutaneous coronary intervention.  The data from the 10,215 patients undergoing PCI was excluded from our comparison.

The percentage of patients from our population with each of the above mentioned risk factors are listed in the results section. The results from each of the above listed review articles is also summarized in the results section.

Results

Three hundred fifty eight charts were reviewed from January 2000 to December 2001. Two patients were excluded from our study who were inadvertently given the diagnosis of AMI, for a total of 356. Two hundred fifty eight patients (80%) were greater than 55 years old, and 71 patients (20%) were aged 55 or less.  One hundred eighty-seven were male (52%) and 169 (48%) were females. Only 9 females (5%) were pre-menopausal, while 160 females (95%) patients were either peri-menopausal or post-menopausal.  One hundred seventy-eight (50%) of patients had diabetes and 264 (74%) had hypertension.  One hundred twenty-three patients (35%) had dyslipidemia while 2 patients were unknown.  Our study demonstrated that 45 patients (13%) had a known family history of MI, while 69 patients had an unknown history.  Forty-eight percent of patients had a history of smoking, while this history was unknown for 22 patients.  Ninety-nine patients (28%) had a previous MI and 15 were unknown to have had a previous MI. Finally, 14 patients (4%) were active cocaine abusers, while this information was unknown in 13 patients. Refer to Table 1 and Table 2.

The data collected from our population listed above was then compared to the percentages obtained from the review article “Major Risk Factors as Antecedents of Fatal and Nonfatal Coronary Heart Disease Events” by Greenland et al. ³   This article looked at the percentage of patients who died from CHD who had a history of smoking, hypertension, diabetes, or dyslipidemia as a risk factor. They obtained their data from three cohort studies. These were the Chicago Heart Association Detection Project in Industry (CHA) 35,642 patients, the Multiple Risk Factor Intervention Trial (MRFIT) 347,978, and the Framingham Study (FHS) 3,295 patients.  The percentage of patients, with any of the above listed risk factors from each cohort is summarized below.  See Table 3.

The data from our population was also compared to the percentages obtained from the second review article “Sex, Age, and Clinical Presentation of Acute Coronary Syndromes” by Rosengren et al ⁴ printed in the European Heart Journal. This article looked at the percentages of patients with any of the four previously listed risk factors.  Of the 10,253 patients studied, during September 4, 2000 to May 15, 2001, presenting to 1 of 6 European hospitals, received the discharge diagnosis of acute coronary syndrome. Table 4 summarizes and compares risk factor percentages with the aforementioned studies.

Summary/Conclusion

Comparison of the data shows the following. In regards to dyslipidemia, the St. Barnabas Hospital (SBH) population showed 35% of patients presenting with AMI had dyslipidemia by history or found on admission.  This compares to 72% (CHA), 76 % (MRFIT), 87 % (FHS), and 47 % (EHS).  Therefore, there is a statistically significant difference in hyperlipidemia between the studies analyzed (p &ln; 0.01).  Also of note, it appears dyslipidemia was not demonstrated to be one of the top two risk factors in our patient population.

The SBH population had 74% of patients presenting with AMI and a history of hypertension. This compared to 88% (CHA), 82% (MRFIT), 83% (FHS), and 57.8% (EHS).  Our study demonstrated that HTN and DM disproportionately make up the top two risk factors for our patient population.

In regards to diabetes, we see a dramatic difference in the SBH population. In our population, 50% of patients presenting with an AMI had a history of, or were diagnosed with diabetes prior to or during their hospitalization.  This compares to 6% (CHA), 5% (MRFIT), 4% (FHS), and 22.9% (EHS).  Therefore, with respect to DM, there is a statistically significant difference between the studies analyzed with a p < 0.01.

Another important risk factor is a patient’s smoking history.  The SBH population had a 48% history compared to 55% (CHA), 50% (MRFIT), 64% (FHS), and 30.2% (EHS).  As with hypertension, it appears that a history of smoking, in our study, is comparable to other studied populations.

Lastly, we examined patient profiles for family history of MI 13% (SBH), previous MI 28% (SBH), and active cocaine use 4% (SBH). No studies were found by these authors that listed the percentage of patients with these risk factors presenting with AMI. The article published in the European Heart Journal in 2004, “Sex, Age, and Clinical Presentation of Acute Coronary Syndrome” by Rosengren et al ⁴ did study history of MI.  Overall, 30% of these patients were found to have a prior history of MI.  This compares equally to the 28% found in the SBH population. However, it is difficult to give an overall assessment of these three risk factors without more national data to compare to.

While the reason for such a dramatic increase in the SBH population is unclear, several reasons can be inferred.  First, SBH serves a very large Hispanic population.  It has been proposed that there is a higher incidence of diabetes in this population as compared to others.  Second, many of these patients are immigrants and under-served.  It can be theorized that many do not obtain routine medical checkups or annual physicals.  Therefore, the diagnosis of diabetes is often not known and not treated until later in the disease process.  Also, many patients with a known history of diabetes are poorly controlled leading to increased complications of the disease, such as AMI.  Lastly, the prevalence of diabetes has increased over the past decade.  Much of the data presented in the review articles was collected over the past decade.  The SBH data was collected from 2000 to 2001.  While this may explain a small increase in the number of patients with diabetes in the SBH population, it is surely not the main contributing factor.

It would appear, based on this data that three conclusions can be made regarding the population presenting to SBH with AMI.  First, a comparable percentage of these patients present with a history of hypertension, smoking, and/or dyslipidemia as compared to other patient populations.  Second, there is a significant increase in the number of patients presenting to SBH for AMI found to have diabetes as compared to other populations.  Lastly, more national data needs to be collected regarding family history of MI, previous MI, and current cocaine use, before comparison can be made to the population presenting to SBH.

Comments

As with any retrospective review of data, there are limitations to the information gathered.  Some of these limitations include, but are not limited to the following.  In several of the risk factor categories there was a significant amount of unknown presence of a risk factor.  An example is the presence of a family history of MI.  In our patient population there were 69 patients in whom this information was unknown.  This is approximately 19% of our population sample.  If this information was known it could have significantly altered the 13% of known family history of MI in this study.  Similar, but smaller groups of unknowns were found in the risk factor categories of smoking, previous MI, and active cocaine use.

Another limitation to this study is its relatively small sample size.  There were 356 charts reviewed over a two year period.  Although small, this sample is probably an accurate representation of the population presenting to SBH.  This sample did not allow significant time for change in the population which one would expect to occur over a longer time period.  Also limiting, in our investigation was the fact that all our patients had the same outcome.  This is an issue that can be addressed in the prospective study with comparing chest pain AMI to chest pain without AMI. With such a comparison, other statistically valuable information may be obtained such as a linear regression analysis which could not be done with this study.

In light of these limitations, one must keep in mind the goal of this investigation.  This study was set up to be a pilot study.  The goal of the authors was to establish a relatively accurate picture of the patient population presenting to SBH for AMI.  The next goal would be to use this information in preparation of conducting a prospective study.  This would, then, allow for further gathering of information regarding MI risk factors for all patients presenting to SBH with AMI.  In addition, it would increase our data pool, sample size and hopefully decrease the amount of unknown information, which, in turn would increase accuracy.

As mentioned earlier in the introduction, death from cardiac disease remains the number one cause of death in the U.S.  An aging population will only continue to increase this trend.  It would seem prudent for emergency medicine physicians to become knowledgeable about the population that they serve.  This will allow us to improve on treating patients who present to our emergency department and more opportunity to educate those patients who are at risk for complications, such as AMI.  Information obtained in this study, along with information that could be gained in a future prospective study, will hopefully continue to challenge us to identify risk factors of AMI in an urban environment and configure early screening programs specifically tailored to decrease complications leading to adverse cardiac events.