Utility of History, Physical Examination, Electrocardiogram, and ...

from the attending physician’s note (internist or cardiologist)
from hospital day 1 or 2; whether the attending
physicians were aware of a patient’s systolic function
when a note was written is unknown. Recorded symptoms
included dyspnea at rest, dyspnea with exertion,
orthopnea, paroxysmal nocturnal dyspnea, angina, and
nonanginal chest pain. Physical examination findings included
height, weight, heart rate, systolic and diastolic
blood pressure, jugular venous distention, hepatojugular
reflux, rales, wheezing, S 3 and S 4 gallops, and pedal
edema. Pulse pressure was calculated as systolic minus
diastolic blood pressure, and body mass index was calculated
as weight (in kg) divided by height (in m 2 ). Most
parameters (80%) were recorded in 97% of the patients.
We assessed whether patients met the Framingham
criteria for heart failure (17).
Admission electrocardiograms (ECGs) were assessed
by a cardiologist who was unaware of other patient data
for rhythm (recorded as sinus, atrial fibrillation, or other),
presence of abnormal Q waves, left ventricular hypertrophy
by the Estes criteria (18), and left atrial abnormality
(defined as P terminal force in V 1 more negative than
0.04 ms, or a notched P wave 0.12 s) (19). The admission
chest radiograph was evaluated by an attending radiologist
for the presence of cardiomegaly, flow cephalization,
pleural effusion, and pulmonary edema.
Left ventricular systolic function was determined by
echocardiography. All ECGs were performed at Cook
County Hospital and were interpreted by experienced
cardiologists who were unaware of the specific clinical
findings in the patients. Normal systolic function was defined
as an ejection fraction 45% (20) as assessed by
visual inspection. Decreased systolic function was defined
as ejection fraction 45%. Patients were excluded if
left ventricular function was not measured, and also if
primary valvular disease was present. Patients with mitral
or tricuspid regurgitation judged to be a result of heart
failure, rather than the cause of the condition, were not
excluded.
Statistical Analysis
Chi-squared tests were used to compare categorical variables,
and two-sided t tests were used for continuous variables.
The sensitivity, specificity, and positive and negative
predictive values for normal systolic function were
calculated for each clinical variable. Multivariate logistic
regression analysis was also performed, including variables
with univariate P values 0.10, to identify independent
predictors of decreased systolic function. Receiver
operating characteristic curves were constructed for these
predictors. A sub-analysis was performed in which patients
with markedly depressed systolic function (ejection
fraction 30%) were compared with those who had an
ejection fraction 30%. A P value 0.05 (two sided) was
Clinical Findings in Heart Failure/Thomas et al
438 April 15, 2002 THE AMERICAN JOURNAL OF MEDICINE Volume 112
considered significant. This study was approved by the
Scientific Committee of Cook County Hospital.
RESULTS
A total of 282 patients were admitted with a primary diagnosis
of congestive heart failure. Left ventricular function
was not documented during the index admission in
43 patients (15%), and 14 patients (5%) had primary valvular
disease; these patients were excluded. The remaining
225 patients comprised the study sample, of whom
121 (54%) had decreased left ventricular systolic function.
Of the 225 patients, 75% (n 169) were black, 10%
(n 22) were white, 11% (n 25) were Hispanic, and
4% (n 9) were of other races. All but 2 of these patients
met the Framingham criteria for congestive heart failure
(17).
Patients with normal systolic function were older and
more likely to be female than were those with decreased
systolic function (Table 1). There were no significant differences
between the groups with respect to coronary risk
factors (including hypertension), history of coronary disease,
or history of chronic renal failure. A history of alcohol
use was significantly more common in patients with
decreased systolic function. More patients with decreased
function were taking angiotensin-converting enzyme
(ACE) inhibitors and digoxin (both P 0.001), whereas
use of calcium antagonists (P 0.001) and beta-blockers
(P 0.08) tended to be higher in patients with normal
function. Symptoms were similar in the patients with
normal or decreased systolic function, except that angina
was more common in patients with decreased systolic
function. Those with normal function tended to have a
higher body mass index (P 0.06), and significantly
more patients with normal function had a body mass index
30 kg/m 2 (P 0.04).
Mean heart rate was significantly higher in patients
with decreased systolic function, and tachycardia (heart
rate 100 beats per minute) was more common (Table
1). Systolic blood pressure and pulse pressure were significantly
higher in those with normal function, who were
also significantly more likely to have a systolic blood pressure
160 mm Hg or a pulse pressure 60 mm Hg. Although
there was no difference in mean diastolic blood
pressure between the groups, patients with normal function
were significantly more likely to have a diastolic pressure
110 mm Hg. There were no significant differences
in the presence of jugular venous distention, pedal
edema, or S 4 gallop sounds. An S 3 gallop tended to be
more common in patients with decreased function (P
0.07), whereas rales were significantly more common in
those with normal function (P 0.05).
There were no significant differences in heart rhythm
or the prevalence of abnormal Q waves. Left ventricular