Temporal trends in clinical characteristics of patients without known coronary
artery disease presenting with a first episode of myocardial infarction
Binita Shah, MD, MS; Sripal Bangalore, MD, MHA; Eugenia Gianos, MD; Li Liang, PhD; W. Frank Peacock, MD;
Gregg C. Fonarow, MD; Warren K. Laskey, MD, MPH; Adrian F. Hernandez, MD; Deepak L. Bhatt, MD, MPH
Background
Measures
Results
•  Recently, greater focus has been
placed on primary prevention of
cardiovascular disease
•  National guidelines aim at improving
modifiable risk factors with lifestyle
changes and pharmacologic therapy
•  Increased public policy efforts further
aim to delay a 1st cardiovascular event
•  We utilized the Get With The
Guidelines (GWTG)-Coronary Artery
Disease (CAD) registry to evaluate
change in risk factor profile over time of
patients without known CAD presenting
with their first episode of myocardial
infarction (MI)
Primary Measure
•  Composite of 6 cardiovascular risk
factors
•  Age >45 years (men), >55 years
(women)
•  Male sex
•  Diabetes Mellitus (DM)
•  Hypertension
•  Hyperlipidemia
•  Tobacco use
•  Defined as a high risk factor profile if
>3 risk factors
Secondary Measures
•  Composite of 4 modifiable risk factors
•  Each individual risk factor
Time Trend of >3 Factors
Study Population
Statistical Analyses
•  276,540 patients in 435 hospitals in
the United States between January 1,
2002 and December 31, 2008 in
GWTG-CAD registry
•  Exclusion criteria:
•  Known CAD, MI, percutaneous
coronary intervention, coronary artery
bypass graft surgery, stroke, transient
ischemic attack, or peripheral artery
disease (n=106,051)
•  Presentation other than an acute MI
(n=68,880)
•  Inconsistent data on diagnosis of
non-ST-segment elevation MI
(NSTEMI) versus ST-segment
elevation MI (STEMI) (n=725)
•  Final study cohort of 100,884 patients
without known CAD presenting with first
episode of MI at 408 hospitals
•  Cochran-Mantel-Haenszel test for
time trend
•  Non-zero correlation statistic for
continuous variables
•  Row-mean score statistic for
categorical variables
•  Logistic regression analyses of risk
factors
•  Generalized estimating equations
used to account for within-hospital
clustering
•  Variables included in the model are
calendar year, NSTEMI versus STEMI
groups, interaction between calendar
years and NSTEMI vs STEMI groups
•  For modifiable risk factors, age and
gender adjusted multivariable logistic
regression analyses performed
Time Trend (per 1 year) for
modifiable risk factors after
adjusting for age and gender
Time Trend of Each Risk Factor
!
!
!
!
!
!
!
Limitations
•  Retrospective analysis
•  Includes only patients
hospitalized with AMI and hospitals
participating in GWTG
•  Selection bias and survival time
bias
•  No data on family history of
premature CAD, impaired glucose
tolerance, and pre-hospital
medications
Conclusions
Funding Sources
•  While risk factor profiles in patients presenting with first
MI have shown slight improvements over time, the
changes are modest.
•  More data is needed to understand where the disparities
in risk factor modification and prevention of cardiovascular
disease lie
•  B. Shah partially funded by NIH/
NHLBI grant (T32HL098129-2012;
UL1 TR000038-2013)
•  Data analysis performed by the
GWTG Data Coordinating Center at
DCRI and supported by AHA Young
Investigator Database Seed Grant