Author + information
- Received April 4, 1996
- Revision received July 24, 1996
- Accepted October 28, 1996
- Published online February 1, 1997.
- Ralph A.H Stewart, MD, FRACPA,*,
- M.Clare Robertson, BSc(Hons), BComA,
- Gerard T Wilkins, MBChB, FRACPA,
- Clive J.S Low, MBChB, FRACPA and
- Norma J Restieaux, MD, FRACPA
- ↵*Dr. Ralph A. H. Stewart, Department of Medicine, University of Otago Medical School, P. O. Box 913, Dunedin, New Zealand.
Objectives. This study sought to compare the clinical features and outcome of a first myocardial infarction with onset of symptoms during or within 30 min of exercise, at rest and in bed.
Background. It is not known whether activity at onset influences outcome of acute myocardial infarction.
Methods. Information collected using a standard questionnaire was used to relate activity at the onset of symptoms to in-hospital outcome in 2,468 consecutive patients admitted to a coronary care unit with a first myocardial infarction between 1975 and 1993.
Results. Patients with exercise-related onset were more likely to be younger and male. Those with onset in bed were more likely to be older and have a history of stable or unstable angina. Compared with patients whose symptoms began at rest, those with exercise-related onset had a lower in-hospital mortality rate after adjusting for age, gender and year of admission (odds ratio [OR] 0.60, 95% confidence interval [CI] 0.40 to 0.89), and patients with onset in bed had a higher mortality rate (OR 1.38, 95% CI 1.03 to 1.85). The incidence of cardiac failure requiring diuretic therapy was also lower for exercise-related onset (OR 0.83, 95% CI 0.67 to 1.04) and higher when onset was in bed (OR 1.36, 95% CI 1.11 to 1.66).
Conclusions. There is an association between activity at onset and outcome of acute myocardial infarction. Differences in pathophysiology or in the population at risk could explain this observation.
(J Am Coll Cardiol 1997;29:250–3)
A circadian variation in the onset of myocardial infarction is well described with the highest risk during the first few hours after rising and a lower risk later in the day and at night while asleep ([1–3]). In addition to this circadian variation, it is known that strenuous exertion can trigger myocardial infarction ([4, 5]). However, it is not known whether the pathophysiology of exercise-, nonexercise- and sleep-related myocardial infarction is the same. Although they are likely to share similar mechanisms, it is possible that the relative importance of factors, such as the mode of plaque rupture, contribution of coronary vasoconstriction and coagulability, differ for exercise-, nonexercise- and sleep-related myocardial infarction. Differences in pathophysiology or in the population of patients at risk may influence the outcome of the coronary event. The present study investigates this possibility by comparing the clinical features and in-hospital outcome of patients whose symptoms of a first myocardial infarction began during exercise, at rest and in bed.
1.1 Study patients and data collection.
The study included 2,468 consecutive patients admitted to a coronary care unit from 1975 to 1993 with a first myocardial infarction. A standard questionnaire that contained information on activity at onset of symptoms, coronary risk factors and outcome was completed by the medical staff for each patient after admission. The diagnosis of myocardial infarction was based on the presence of at least two of the three following criteria: a typical history of chest discomfort lasting >15 min, an increase in cardiac enzyme levels to more than two times the upper limit of the normal range and the appearance of Q waves or evolution of characteristic ST segment or T waves on serial electrocardiograms (ECGs). Q wave myocardial infarctionwas defined as myocardial infarction with new Q waves >0.03 s or >0.1 mV on the ECG.
The patient or, when necessary, a relative or witness was asked what activity was undertaken when or during the 30 min before the symptoms of myocardial infarction began. Activity was documented as exercise related (yes/no) if the patient was walking at their normal brisk pace or undertaking some other activity equivalent to or more than ∼4 metabolic equivalents (). Other patients were classified as having symptom onset in bed (yes/no) or onset at rest if symptoms were not exercise related and did not begin in bed. Time of onset of symptoms was not recorded, nor whether the patient woke from sleep because of symptoms.
The presence or absence of the following risk factors was recorded: history of coronary artery disease in a first-degree relative; hypertension currently or previously requiring treatment; diabetes treated by diet, drugs or insulin; recognized hypercholesterolemia (previous total cholesterol >7.0 mmol/liter when known); and smoking during the preceding month. Time delay from onset of symptoms to hospital admission was classified as either <4 or ≥4 h. A previous history of stable angina, recent unstable angina and treatment with a beta-adrenergic blocking agent at the time of onset of myocardial infarction was recorded. Unstable anginawas defined as new-onset or increasing angina during the 2 weeks before the presenting myocardial infarction. Outcome measures were in-hospital mortality, cardiac arrest requiring cardioversion or cardiopulmonary resuscitation before admission to hospital, Q wave myocardial infarction and clinical or radiographic evidence of cardiac failure requiring diuretic treatment.
1.2 Analysis and statistics.
Patients with symptom onset during or after exercise were compared with those whose symptoms began at rest. Patients whose symptoms began in bed were also compared with those with symptom onset at rest. The relative likelihood of each patient characteristic and outcome measure was calculated as odds ratios with 95% confidence intervals after adjusting for age, gender and year of admission using logistic regression models (Statistical Package for Social Sciences Release 4.1 for VAX/VMX, SPPS, Inc., Chicago, Illinois). Age differences in activity at symptom onset are described for four age groups: <50, 50 to 60, 60 to 70 and >70 years old; p > 0.05 was regarded as not statistically significant.
Increased age was associated with a decrease in the proportion of patients with exercise-related onset of myocardial infarction and a greater proportion with symptom onset in bed. There was little age difference in the proportion of patients with symptom onset at rest (Table 1). The direction and size of these age-related changes were similar for men and women. However, after adjusting for age, women were relatively more likely to have onset of symptoms in bed and less likely to have exercise-related onset of symptoms (Table 2).
After adjusting for age, gender and year of admission, the proportion of patients who were current smokers or had recognized hypercholesterolemia, diabetes or a family history of cardiovascular disease was similar for those with symptom onset during or after exercise, at rest and in bed (Table 2). There was a longer time delay to hospital admission for patients with exercise-related onset of symptoms and a shorter time delay for those with onset of symptoms in bed, but these differences were not statistically significant (Table 2). Patients with symptom onset in bed were more likely to have a past history of stable angina or recent unstable angina but less likely to have a history of hypertension (Table 2). A similar proportion of patients with symptom onset on exercise and at rest had a history of stable or unstable angina. A lower proportion of patients with exercise-related onset were taking a beta-blocker, but this difference was not statistically significant.
After adjusting for age, gender and year of admission, patients with exercise-related onset of myocardial infarction had significantly lower in-hospital mortality than those with onset of symptoms at rest (Table 3). Patients with symptom onset in bed had the highest in-hospital mortality rate. The proportion of patients who developed heart failure needing diuretic treatment was lowest for exercise-related onset and highest when onset of symptoms was in bed. Q wave infarction was less frequent in patients with exercise-related onset than those with onset at rest or in bed. Prehospital cardiac arrests were increased in patients with onset related to exercise despite a better in-hospital outcome, but this difference was not statistically significant. Including previous beta-blockade and Q wave myocardial infarction in the regression models did not change the association between outcome and activity at the onset of symptoms. The odds ratio for in-hospital death for exercise-related onset compared with onset of rest was 0.62 (95% confidence interval [CI] 0.41 to 0.92) and for onset in bed compared with onset at rest was 1.36 (95% CI 1.01 to 1.83).
In the present study, patients whose symptoms of myocardial infarction began during or shortly after physical activity had both lower in-hospital mortality and less cardiac failure requiring diuretic therapy than those whose symptoms began at rest. The highest mortality and incidence of heart failure occurred in patients whose symptoms began in bed. The observation that the outcome of myocardial infarction is influenced by the activity at the onset of symptoms has not been made in earlier studies. Mortality was not reported in studies which demonstrated that exercise could trigger myocardial infarction ([4, 5]) or in studies which described a circadian variation in the onset of myocardial infarction ([1–3]). The reasons for the observed differences in outcome are uncertain.
3.1 Possible reasons for outcome differences.
It is possible that exercise is the trigger for myocardial infarction in patients with less extensive vascular disease and therefore a better prognosis. This conclusion is consistent with the observation from the Thrombolysis in Myocardial Infarction (TIMI) II study () that patients with exercise-related onset myocardial infarction were less likely to have multivessel disease and more likely to have a single occluded coronary artery after thrombolysis. Conversely patients with onset of symptoms in bed are more likely to be older and have a history of stable or unstable angina, suggesting more extensive vascular disease or a preexisting unstable lesion and therefore a lower threshold for abrupt coronary occlusion. Information on the routine amount of physical activity of participants was not available. Patients with exercise-related onset may be habitually more active and therefore have a greater exposure to both the risks and benefits of exercise ([4, 8, 9]). Randomized trials of cardiac rehabilitation suggest that regular exercise reduces the risk of fatal more than nonfatal myocardial infarction ([10, 11]).
Differences in the pathophysiology of exercise- and nonexercise-related myocardial infarction are possible. The lower incidence of Q wave myocardial infarction in patients with exercise-related onset would be consistent with more rapid coronary reperfusion because of spontaneous thrombolysis or relief of coronary vasoconstriction induced by exercise (). Other potentially protective mechanisms, such as ischemic preconditioning () or the “warm-up” effect (), could be influenced by activity before coronary occlusion.
3.2 Study limitations.
Assessment of the level of activity at the onset of symptoms by necessity relied on patient recall; no distinction was made between light, moderate and strenuous exercise, and the type of activity was not recorded. This relatively crude assessment of activity may reduce the accuracy of the study but is unlikely to bias the results. The time of symptom onset was not recorded, so analysis of circadian variation in outcome is not possible. In other studies (), the morning increase in risk of myocardial infarction begins after getting out of bed. For the majority of patients whose symptoms began in bed, onset would be between 11 pmand 7 amand occur while the patient was asleep.
Data were collected over many years during which the treatment of myocardial infarction changed. However, the relation between activity at the onset of symptoms and outcome remained after adjusting for study year, and a standard format for data collection was used throughout. Prehospital treatment other than beta-blockers was not recorded, and data on use of thrombolytic and other therapy that may influence outcome was not available. A systematic difference in treatment by activity at symptom onset cannot be excluded. However, it is unlikely that treatment decisions were directly influenced by knowledge of whether symptom onset was related to exercise, rest or sleep.
The “two of three” criteria used for diagnosis of myocardial infarction may miss some patients with small myocardial infarction and include others with unstable angina. Because standard criteria for inclusion were used for all patients, the possibility that a small proportion of the study cohort had unstable angina rather than myocardial infarction should not bias the results. Peak cardiac enzyme levels were not documented, so estimates of infarct size are limited to the presence or absence of Q waves on the ECG and clinical or radiographic evidence of cardiac failure. The study was limited to patients with a first myocardial infarction because of the significant impact of previous infarction on outcome.
The study does not include subjects not admitted to the coronary care unit or those who died before hospital admission. This exclusion could explain the trend for more prehospital cardiac arrests in patients with exercise-related onset. A witnessed arrest and therefore a successful resuscitation out of the hospital may be more likely for persons exercising than for those in bed. Differences between groups in time delay from onset of symptoms to hospital admission were small compared with the outcome differences, and this alone is unlikely to explain the observed differences in mortality.
The reasons for the relatively large differences in outcome of myocardial infarction by activity at onset are uncertain and cannot be determined from this study. It is possible that the pathophysiology of myocardial infarction is influenced by the circumstances in which it occurs. Further studies are needed to gain a better understanding of the reasons for the differences observed and their potential relevance to the prevention and management of ischemic heart disease.
We thank our coronary care medical and nursing staff for their assistance over many years with the questionnaires used for this study. We are particularly grateful to Deborah Scott and Mary Blok, MD for their help with the management of questionnaire data. We thank Shiela Williams, PhD for statistical advice.
☆ The study was supported in part by the National Heart Foundation of New Zealand, Auckland.
- confidence interval
- odds ratio
- Received April 4, 1996.
- Revision received July 24, 1996.
- Accepted October 28, 1996.
- The American College of Cardiology
- Willich SN,
- Linderer T,
- Wegscheider K,
- et al.
- Muller JE,
- Toller GH,
- Stone PH
- Astrand PO,
- Rodahl K
- Tofler GH,
- Muller JE,
- Stone PH,
- et al.
- O’Connor GT,
- Buring JE,
- Yusuf S,
- et al.
- Gage JE,
- Hess OM,
- Murakami T,
- Ritter M,
- Grimm J,
- Krayenbuehl HP
- Kloner RA,
- Yellon D