We included 966 consecutive patients with AMI who were admitted to Kyungpook National University Hospital between November 2005 and December 2007. A diagnosis of AMI was made in patients having angina pain lasting more than 15 minutes that was associated with/without dyspnea or sweating, established/evolving ST-T wave changes on serial electrocardiograms, and an increase in levels of cardiac troponin I (cTnI). cTnI levels were measured on admission and every 6 hours to detect the presence of myocardial injury, using an enzyme immunoassay based on the sandwich principle (Dimension; Dade Behring, Deerfield, IL, USA); the lower detection limit of this assay was 0.04 ng/mL. Elevated cardiac biomarker levels were defined as having maximal values of cTnI that exceeded the cutoff (99th percentile of the values for a reference control group) on at least one occasion. The time of onset of AMI was determined by each patient's report of the chest pain that prompted hospital admission. Seventy-four patients with unclear onset of chest pain, including silent AMI, were excluded from further analysis. In total, 892 patients (mean age 67±12 years; 66.1% men) were included in this study. Informed consent was obtained from each patient.

Each day was divided into 12 equal parts of two hours each, and patients were grouped according to the time of symptom onset. For further analysis, a day was divided again into four equal 6-hour intervals: overnight (00:00 AM-05:59 AM); morning (06:00 AM-11:59 AM); afternoon (12:00 PM-17:59 PM) and evening (18:00 PM-23:59 PM).2)14) Routine duty hours were defined as Monday to Friday between 08:00 AM and 18:00 PM.13) Weekends were considered off-duty hours. We analyzed the baseline demographic characteristics, initial presentation, initial vital signs, results of laboratory tests, and discharge medications of the patients. Demographic and clinical characteristics were identified, including age, gender, body mass index (BMI), cardiovascular risk factors (hypertension, diabetes mellitus, hyperlipidemia, and current smoking), and co-morbidities, which indluded previous congestive heart failure and myocardial infarction (MI). Initial vital signs including the systolic pressure and Killip class were evaluated at admission. Initial blood samples collected at admission, except for those collected to determine peak cTnI, were used for baseline laboratory tests. The left ventricular ejection fraction (LVEF) was determined by 2-dimensional echocardiography during the index hospitalization. The mean symptom-to-door time (STD) and the proportion of patients with STD time ≥6 hours were assessed in patients with typical chest pain. In patients with STEMI, the median door-to-balloon (DTB) time and the proportion of patients with DTB time ≤90 minutes were also evaluated. Discharge medications assessed included aspirin, β-blockers, angiotensin converting enzyme inhibitors (ACE-I)/angiotensin receptor blockers (ARB), and diuretics. Mean follow-up duration was 357±141 days. The primary end-point was 12-month mortality, including both cardiac death and non-cardiac death. Cardiac death was defined as death from pump failure, arrhythmia, or mechanical complications, which included ventricular septal rupture and free wall rupture. During the follow-up period, data were obtained by reviewing medical records and by telephone interview with the patients.

Chi-squared goodness-of-fit was performed to test the uniformity of the distribution of patients among the time periods. Data are expressed as mean±SD for continuous variables and percentages for categorical variables. All comparisons between baseline variables were assessed with the analysis of variance for continuous variables, and with the Pearson's Chi-square test for categorical variables. A multivariate logistic regression model was used to determine independent predictors for evening-onset AMI. Twelve-month mortality according to the onset time of AMI was evaluated using Kaplan-Meier survival curve analysis. Univariate analyses were performed to determine the clinical predictors of 12-month mortality. The Cox proportional hazard model was used to calculate the risk of 12-month mortality related to the time of onset of AMI, and adjusted for baseline confounding characteristics. For all analyses, a two-sided p<0.05 was considered statistically significant. Statistical analysis was performed using SPSS version 15.0 for Windows (SPSS Inc., Chicago, IL, USA).

The frequency of the onset of AMI as determined by onset of chest pain for all 892 patients is shown in Fig. 1. The distribution is not uniform (p<0.001), and a peak was found between 08:00 and 09:59 hours (p<0.001). The distribution of AMI onset was obtained for each 6-hour interval, and a morning peak was found to be statistically significant (p<0.001) (Fig. 2). The numbers of patients grouped according to the AMI onset time were 184 (20.6%), 303 (34.0%), 187 (21.0%), and 218 (24.4%), corresponding to overnight, morning, afternoon and evening-onset groups, respectively. The incidence of morning-onset AMI was 54% higher than the average of the three other time periods. A morning peak of AMI was found irrespective of the presence or absence of ST-segment elevation (p=0.001 in STEMI and p<0.001 in NSTEMI).

The clinical characteristics of the patients are shown in Table 1. The patients mean age was 67±12 years, and 590 (66.1%) were men. There were no significant differences in age, sex, BMI, systolic blood pressure, preinfarct angina pectoris, location of AMI, previous history of diabetes mellitus, hypertension, hyperlipidemia, MI, use of tobacco, percutaneous coronary intervention (PCI) at index hospitalization, peak cTnI, or pro-brain natriuretic peptide (proBNP) levels among the four groups. In the evening-onset AMI group serum creatinine levels (p=0.001) and the frequency of Killip class ≥3 (p=0.027) were significantly higher, while the LVEF (p=0.011) was significantly lower. The mean STD time was 564±1,157 minutes, with no significant differences in the mean STD time (p=0.746) or the proportion of patients with an STD time ≥6 hours (p=0.384) among the four groups. The proportion of patients who arrived in the emergency room (ER) during routine duty hours was significantly higher in the morning-onset AMI group (p<0.001). The discharge prescription rates of aspirin, β-receptor blockers, ACE-I/ARBs, and diuretics were not significantly different among the four groups.

Compared with the morning-onset AMI group, serum creatinine levels (p=0.002),the frequency of Killip class ≥3 (p=0.004), and the prescription rate of diuretics (p=0.011) were significantly higher in the evening-onset AMI group, while LVEF (p=0.012) was significantly lower. The proportion of patients who arrived in the ER during routine duty hours was significantly lower in the evening-onset group compared to the morning-onset group (p<0.001).

In the STEMI group (Table 2), the numbers of patients with a previous history of diabetes mellitus was significantly higher within the evening-onset group (p=0.049). The median DTB time was 80 minutes, and was significantly longer in the overnight-onset and evening-onset groups than in the morning-onset and afternoon-onset groups (p=0.039). The proportion of all patients with a DTB time ≤90 minutes was not significantly different among the four groups. However, compared with the morning-onset STEMI group, the proportion of patients with a DTB time ≤90 minutes was significantly lower for the evening-onset STEMI group (p=0.039). Serum creatinine levels, frequency of Killip class ≥3, and LVEF were not different among four STEMI groups.

In the NSTEMI group (Table 3), the frequency of Killip class ≥3 (p=0.013), levels of serum creatinine (p=0.001), and peak cTnI (p=0.004) were significantly higher in the evening-onset NSTEMI group, while the LVEF (p=0.017) was significantly lower. Compared with the morning-onset NSTEMI group, the log proBNP level (p=0.048), prescription rate of diuretics (p=0.027), frequency of Killip class ≥3, levels of serum creatinine, and peak cTnI were significantly different in the evening-onset NSTEMI group. The proportion of patients who arrived in the ER during routine duty hours was significantly lower in the evening-onset group, irrespective of the presence or absence of ST-segment elevation (p<0.001).

During the study period, 48 (5.4%) patients were lost during follow-up. The overall mortality rate was 11.7% (n=99), with 10.5% (n=89) from cardiac deaths and 1.2% (n=10) from non-cardiac deaths. No significant difference in the 12-month mortality rate was observed between STEMI and NSTEMI groups (11.6% vs. 11.8%, p=NS). Kaplan-Meier survival curves showed 12-month mortality rates of 9.6% (n=17), 9.1% (n= 26), 12.1% (n=21) and 16.7% (n=35) in the overnight, morning, afternoon and evening-onset groups, respectively (p= 0.012) (Figs. 2 and 3). Although there were no significant differences in inhospital mortality among the four groups according to univariate analysis, 6-month (p=0.006) and 12-month mortality rates (p=0.012) were significantly higher in the evening-onset group than for the three other groups. Compared with the morning onset-AMI group, 6-month {hazard ratio (HR) 2.040, 95% CI 1.181 to 3.523, p=0.011} and 12-month mortality rates [HR 1.998, 95% confidence interval (CI) 1.196 to 3.338, p= 0.008] were significantly higher in the evening onset-AMI group (Table 4). A similar result was obtained for 12-month mortality in patients with NSTEMI. However there were no significant differences between morning- and evening-onset STEMI groups in in-hospital, 6-month, and 12-month mortality rates. According to the multivariate Cox proportional-hazards model, evening-onset AMI (HR 0.518, 95% Cl 0.168 to 1.601, p=0.253) was not an independent predictor for 12-month mortality after adjusting for age, gender, ER arrival during off-duty hours, LVEF, Killip class ≥3, prescription of diuretics, levels of serum creatinine, peak cTnI, and log proBNP.