Typical Symptoms Are Predictive of Acute Coronary Syndromes in Women


Am Heart J. 2002;143(2) 

In This Article


We conducted this study in the ED of Yale-New Haven Hospital, a 900-bed university teaching hospital and a regional cardiac referral center. Patients seen in the ED between September 1995 and August 1997 were considered for participation if they were 45 years old and reported 1 of a prespecified set of typical or atypical symptoms suggestive of ACS (Table I). Additionally, patients 18 to 44 years old were asked to participate if they had diabetes mellitus or 2 cardiac risk factors (history of coronary heart disease, systemic hypertension, hypercholesterolemia, smoking, obesity, family history of premature coronary heart disease, and postmenopausal status), and 1 typical or atypical symptoms suggestive of ACS. We included these additional criteria for younger patients to exclude patients at very low risk for ACS. Patients were approached for study participation after the ED staff had completed initial assessment. In the event that patients were too sick to provide oral consent in the ED, they were approached when they were considered stable, within 24 hours of hospital admission. No patient died before we were able to obtain consent. A total of 527 patients met study criteria and were approached for participation. Of these, 522 patients agreed to participate (<1% refusal rate).

During shifts of 2 to 5 hours across all 24 hours and all 7 days, nurse data collectors unobtrusively observed all the patients who met the study criteria as they came to the ED. Data on presenting symptoms were obtained by observing the patient-physician or patient-nurse interview, and symptoms were documented verbatim. Data on demographics and cardiac risk factors were gathered from the patient and the medical record. These data elements are defined in detail elsewhere.[8,11,12]

ACS included either acute ischemia or acute myocardial infarction (AMI). Acute ischemia was determined by electrocardiogram evidence of ST-segment depression, ST-segment elevation, or T-wave inversion or T-wave abnormalities different from the last electrocardiogram in at least 2 consecutive leads and by lack of cardiac enzyme elevation. Elevated cardiac enzymes (creatine kinase-MB [CK-MB] >5% of total CK in at least one set) indicated AMI.

All analyses were done using SAS 8.1 statistical software (SAS Institute, Cary, NC). We used X[2] and t tests to examine the associations between ACS and baseline characteristics and symptoms in both women and men. Typical symptoms included chest pain or discomfort as a single variable (either chest pressure, heaviness, tightness/squeezing, or center or left chest pain), neck/jaw pain, arm or shoulder pain, diaphoresis, and dyspnea. Atypical symptoms included atypical chest pain or discomfort as a single variable (either chest fullness, stabbing, burning, or right chest pain), indigestion, nausea or vomiting, upper extremity numbness or tingling, cough, pain with deep breath, palpitations, midback pain, dizziness/faintness, and fatigue. We also created an interval-level variable for the "number of presenting symptoms." Only those symptoms that were reported by at least 5% of women and men were included in the analyses.

To determine symptom predictors in women and men, we performed logistic regression analyses stratified by sex, with an a level of .20 for entering or staying in the model. In these models, typical and atypical symptoms reported by >5% women and men were included to identify the most important symptom predictors in the presence of many typical and atypical presenting symptoms. Age, diabetes, and number of presenting symptoms were forced in as control factors. Adjusted relative risks were estimated from generalized linear models for each of the symptom predictors. Last, we calculated individual relative risks of ACS for typical symptoms in women relative to men. In these analyses, we controlled for age, diabetes, and number of presenting symptoms. The adequacy of fit and the discriminatory power of all models were assessed according to standard methods.[13,14]


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.