International Journal of Cardiology 179 (2015) 315–320

Contents lists available at ScienceDirect

International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard

Effect of weekend admission for acute myocardial infarction on in-hospital mortality: A retrospective cohort study Toshiaki Isogai a,b,1, Hideo Yasunaga a,⁎,1, Hiroki Matsui a,1, Hiroyuki Tanaka b,1, Tetsuro Ueda b,1, Hiromasa Horiguchi c,1, Kiyohide Fushimi d,1 a

Department of Clinical Epidemiology and Health Economics, School of Public Health, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan Department of Cardiology, Tokyo Metropolitan Tama Medical Center, 2-8-29 Musashidai, Fuchu, Tokyo 183-8524, Japan Department of Clinical Data Management and Research, Clinical Research Center, National Hospital Organization Headquarters, 2-5-21 Higashigaoka, Meguro-ku, Tokyo 152-0021, Japan d Department of Health Policy and Informatics, Graduate School of Medicine, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8510, Japan b c

a r t i c l e

i n f o

Article history: Received 25 March 2014 Received in revised form 8 July 2014 Accepted 5 November 2014 Available online 6 November 2014 Keywords: Acute myocardial infarction Weekend admission In-hospital mortality Reperfusion Revascularization

a b s t r a c t Background: No previous nationwide study has examined whether there is a ‘weekend effect’ of higher mortality after admission for acute myocardial infarction (AMI) when percutaneous coronary intervention (PCI) is easily accessible. Methods: Using the Diagnosis Procedure Combination inpatient database in Japan, we identified AMI patients aged ≥20 years who were admitted to acute care hospitals between July 1, 2010 and March 31, 2013. Multivariable regression models fitted with generalized estimating equations were used to determine the association between weekend admission and in-hospital mortality. Eligible patients were also classified into subgroups according to Killip class at admission. Results: Of 111,200 eligible patients, 30,847 patients were admitted on weekends and 80,353 patients on weekdays. Overall, the in-hospital mortality was significantly higher for weekend admission than for weekday admission (13.6% versus 11.4%; P b 0.001; unadjusted odds ratio [OR] 1.222; 95% confidence interval [CI] 1.172 to 1.274), despite the higher rate of PCI performed on the day of admission (68.9% versus 64.8%; P b 0.001). The association remained significant after adjusting for baseline characteristics, invasive procedures, and medications (adjusted OR 1.144, 95% CI 1.079 to 1.214). In subgroup analyses, the effect of weekend admission remained significant in the Killip II to IV subgroups, but became insignificant in the Killip I subgroup (adjusted OR 1.002, 95% CI 0.828 to 1.213). Conclusions: This study showed that weekend admission for AMI was significantly associated with higher inhospital mortality even in a setting where PCI was highly available. © 2014 Elsevier Ireland Ltd. All rights reserved.

1. Introduction A ‘weekend effect’ of hospital admission has been reported, whereby there is higher mortality in patients admitted on a weekend compared with admission on a weekday [1–3]. Hospitals generally provide comprehensive care on weekdays but have decreased staffing levels on weekends, with possible reductions in the quality of care [4–6]. The differences in the level of care between weekdays and weekends may cause ‘weekend effect’ in several diseases [1–3].

Abbreviations: DPC, Diagnosis Procedure Combination. ⁎ Corresponding author. E-mail addresses: [email protected] (T. Isogai), [email protected] (H. Yasunaga), [email protected] (H. Matsui), [email protected] (H. Tanaka), [email protected] (T. Ueda), [email protected] (H. Horiguchi), [email protected] (K. Fushimi). 1 All authors take responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation.

http://dx.doi.org/10.1016/j.ijcard.2014.11.070 0167-5273/© 2014 Elsevier Ireland Ltd. All rights reserved.

The management of acute myocardial infarction (AMI) generally requires emergency invasive cardiac procedures, such as percutaneous coronary intervention (PCI) [7,8], which are often less available during off-hours (weekends and nights) than during regular hours because of limited staff availability [4,5,9]. The effect of off-hour admission for AMI on mortality has been controversial because previous studies have shown inconsistent results [1,3–5,9–15]. Recently, a study using data from an administrative database in New Jersey reported that 30day mortality for AMI was higher in patients admitted on weekends than on weekdays [10]. On the other hand, a study from an American Heart Association database reported no significant difference in inhospital mortality between off-hour and regular hour admissions for AMI [9]. Most recently, a meta-analysis of 48 related studies suggested that patients with AMI admitted during off-hours had higher shortterm mortality, and patients with ST-elevation myocardial infarction (STEMI) had longer door-to-balloon times [16]. Off-hour admission for AMI, therefore, could affect the performance of invasive procedures and short-term mortality. In studies from Western countries, however,

316

T. Isogai et al. / International Journal of Cardiology 179 (2015) 315–320

less than 50% of patients with AMI and less than 60% of patients with STEMI underwent PCI during hospitalization [4,5,9–12]. In the above-mentioned study from New Jersey, the difference in 30day mortality between weekend and weekday admissions became insignificant after adjustment for invasive cardiac procedures [10]. Thus, the higher mortality for weekend admissions could partially be a result of the decreased availability of PCI [4,10]. In Japan, many hospitals are equipped to perform PCI [17–19], and previous studies revealed that the proportion of patients undergoing primary PCI for AMI was much higher in Japan (75–97%) than in Western countries (6–60%) [4–6, 9–12,14,19,20]. It remains unclear whether weekend admission for AMI is associated with higher mortality when there are many hospitals equipped to perform PCI. It also remains unknown whether the effect of weekend admission for AMI is different according to the severity of AMI at admission. To our knowledge, no study has examined the effect of weekend admission in patients with different severities of AMI. The Killip classification is a classical but still useful predictor of short-term mortality in patients with AMI [21–23]. A higher Killip class is associated with a greater need for mechanical circulatory support such as intra-aortic balloon pumping (IABP) and extracorporeal membrane oxygenation (ECMO), which require high staff input for prompt administration in the acute phase. The present study therefore had two objectives: to determine whether weekend admission for AMI was associated with higher inhospital mortality compared with weekday admission in a nationwide setting in Japan, where PCI is easily accessible; and to assess whether the effect of weekend admission for AMI was different according to the Killip class at admission. 2. Methods 2.1. Data source Data for this study were extracted from the Diagnosis Procedure Combination (DPC) database in Japan, which includes hospital administrative claims data and discharge abstracts, and is described in detail elsewhere [24]. Since July 1, 2010, data have been continuously collected by the DPC Study Group, and in 2012 the database included data of approximately 7 million inpatients from over 1000 hospitals, representing approximately 50% of all inpatient admissions during the year to acute care hospitals in Japan. Attending physicians are obliged to record all discharge abstract data for each patient. Anonymized patient data are sent to the study group by the hospitals each month. The requirement for informed consent was waived because of the anonymous nature of the data. The Institutional Review Board at The University of Tokyo approved the present study. The DPC database includes the following data: unique identifiers of hospitals; patient age and sex; diagnoses at admission, comorbidities at admission, and complications after admission recorded with text data in the Japanese language and the International Classification of Diseases, Tenth Revision (ICD-10) codes; Killip class at admission; surgical and non-surgical procedures; drugs and devices used; length of stay; dead or alive within the first 24 h after admission; and discharge status. This database partially corresponds to the National Inpatient Sample in the United States [25], but has several unique advantages. First, in the DPC database, comorbidities already present at admission are clearly distinguished from complications that occur after admission. Second, the database contains data about not only the types of therapeutic approaches (drugs, devices, and procedures), but also the dates of starting and ending each treatment during hospitalization. The attending physicians are obliged to record the diagnoses with reference to the medical charts to optimize the accuracy of the recorded diagnoses. At discharge, diagnoses and comorbidities are registered in the database once per admission. Physicians and hospitals have a strong incentive for data compliance because it is mandatory to obtain DPCbased reimbursement of medical fees. 2.2. Study population We identified adult patients aged ≥20 years who were admitted with AMI (ICD10 code I21.x) and were discharged between July 1, 2010 and March 31, 2013. We excluded patients who were discharged alive on the day of admission because they were transferred to another hospital for treatment. The eligible patients were stratified according to admission on a weekday or weekend. In this study, weekdays included Monday through Friday, while weekends included Saturday, Sunday, Japanese statutory holidays, and New Year (from December 29 to January 3). 2.3. Baseline variables As covariates, we included age, sex, ambulance use, Killip class at admission, comorbidities at admission, type of hospital (academic or non-academic), and hospital volume.

For Killip class in the DPC database, attending physicians could record an ‘unclassified class’ if they could not choose any specific class (I to IV). We identified comorbidities at admission using ICD-10 codes based on the Ontario AMI mortality predictive rule, which was validated for predicting death within 30 days and 1 year after AMI [26,27]. We defined hospital volume as the annual number of AMI patients admitted to each hospital, and categorized it into quartiles with approximately equal numbers of patients in each group. We also identified invasive procedures and medications as potential confounding factors. Reperfusion therapies included PCI and fibrinolytic therapy; revascularization procedures included PCI and coronary artery bypass graft; mechanical support included IABP, ECMO, mechanical ventilation, and renal replacement therapy; and medications included aspirin, thienopyridinic derivatives (clopidogrel or ticlopidine), renin-angiotensin system inhibitors (angiotensin converting enzyme-inhibitors [ACE-I] or angiotensin II receptor blockers [ARB]), β-blockers, and hydroxymethylglutaryl-CoA reductase inhibitors (statins). 2.4. Outcome measures The primary outcome was in-hospital mortality. The secondary outcomes were mortality within the first 24 h after admission, 7-day mortality, and 30-day mortality. 2.5. Statistical analyses Categorical variables are presented as numbers and proportions. Continuous variables are presented as mean and standard deviation if they followed a normal distribution, or otherwise as median and interquartile range (IQR). The chi-square test was used to compare proportions in baseline characteristics, invasive procedures, and drugs between weekend and weekday admissions. We used the Fisher exact test to compare mortality rates. When comparing continuous variable data, we used the Student t-test for normally distributed data, and the non-parametric Mann–Whitney U test for skewed data. In the DPC database, data are derived from multiple hospitals and are structured by two strata (patients and hospitals). We therefore accounted for clustering within hospitals using generalized estimating equations [28]. Instead of using basic regression models, we used generalized estimating equations with a unique hospital identifier as a subject variable because the outcomes of patients with AMI in the same hospital would be correlated. An adjusted model for outcomes was created using the following measured variables as covariates: baseline characteristics; reperfusion therapies on the day of admission; mechanical support on the day of admission; revascularization procedures during hospitalization; and drugs administered during hospitalization. The odds ratios (ORs) and 95% confidence intervals (CIs) were determined. The model included reperfusion therapies and mechanical support performed on the day of admission as covariates, because the therapies would have potential to modulate the association between weekend and weekday admissions. Furthermore, to assess whether any effect of weekend admission for AMI was different according to Killip class, patients were divided into five subgroups (Killip I to IV and unclassified) and the ORs and 95% CIs were estimated for each subgroup with multivariable adjustment. All hypothesis tests had a two-sided significance level of 0.05. All statistical analyses were conducted using IBM SPSS Statistics, version 22 (IBM SPSS, Armonk, NY, USA).

3. Results 3.1. Study population Among approximately 18 million inpatients in the database between July 2010 and March 2013, 111,885 patients aged ≥ 20 years were admitted with AMI. After exclusion of 685 patients who were discharged alive on the day of admission, there were 111,200 patients eligible for inclusion in the study, 30,847 admitted during weekends and 80,353 admitted during weekdays. 3.2. Patient characteristics Table 1 shows the baseline characteristics of the study population. Fewer hospitals were available on weekends than weekdays. Patients admitted on weekends were slightly younger and more often male, tended to have higher Killip class, used an ambulance more frequently, and were treated in hospitals with higher hospital volume. Patients admitted on weekends also showed higher rates of shock, pulmonary edema, and cardiac arrhythmias, but showed lower rates of diabetes mellitus with complications, neoplasms, and chronic renal failure. Table 2 shows invasive procedures and drugs used during hospitalization. Patients admitted on weekends were more likely to undergo coronary angiography, PCI, and mechanical support on the day of admission. They were also more likely to undergo revascularization procedures and to receive clopidogrel/ticlopidine during hospitalization.

T. Isogai et al. / International Journal of Cardiology 179 (2015) 315–320

317

Table 1 Baseline characteristics of patients admitted on weekends and on weekdays.

Total No. of hospitals Age, years, mean (SD) Sex Ambulance use Killip class at admission

Comorbidities present at admission Shock DM with complications Heart failure Neoplasms Cerebrovascular disease Pulmonary edema Acute renal failure Chronic renal failure Cardiac arrhythmias Type of hospital Hospital volume (case/year)

Overall (n = 111,200)

Weekday admission (n = 80,353)

Weekend admission (n = 30,847)

P-value

1112 69.7 (13.0) 57,296 (71.3) 23,057 (28.7) 45,643 (56.8) 34,286 (42.7) 21,510 (26.8) 6731 (8.4) 9539 (11.9) 8287 (10.3)

1002 69.1 (13.3) 22,287 (72.3) 8560 (27.7) 20,462 (66.3) 12,703 (41.2) 8008 (26.0) 2718 (8.8) 4434 (14.4) 2984 (9.7)

– b0.001 0.002

I II III IV Unclassified

1123 69.5 (13.1) 79,583 (71.6) 31,617 (28.4) 66,105 (59.4) 46,989 (42.3) 29,518 (26.5) 9449 (8.5) 13,973 (12.6) 11,271 (10.1)

Academic Non-academic Quartile 1 (≤43) Quartile 2 (44–68) Quartile 3 (69–106) Quartile 4 (≥107)

8625 (7.8) 7519 (6.8) 33,421 (30.1) 4234 (3.8) 4356 (3.9) 381 (0.3) 1547 (1.4) 4924 (4.4) 13,432 (12.1) 20,802 (18.7) 90,398 (81.3) 26,869 (24.2) 28,023 (25.2) 28,114 (25.3) 28,194 (25.4)

6034 (7.5) 5553 (6.9) 24,095 (30.0) 3129 (3.9) 3180 (4.0) 256 (0.3) 1107 (1.4) 3632 (4.5) 9448 (11.8) 15,094 (18.8) 65,259 (81.2) 20,266 (25.2) 20,186 (25.1) 19,941 (24.8) 19,960 (24.8)

2591 (8.4) 1966 (6.4) 9326 (30.2) 1105 (3.6) 1176 (3.8) 125 (0.4) 440 (1.4) 1292 (4.2) 3984 (12.9) 5708 (18.5) 25,139 (81.5) 6603 (21.4) 7837 (25.4) 8173 (26.5) 8234 (26.7)

b0.001 0.001 0.422 0.015 0.264 0.027 0.535 0.016 b0.001 0.283

Male Female

b0.001 b0.001

b0.001

Numbers are n (%), unless otherwise stated. SD: standard deviation, DM: diabetes mellitus.

There was no significant difference in fibrinolytic therapy, aspirin, ACEI/ARB, β-blocker, and statin administration between weekend and weekday admissions. 3.3. In-hospital mortality Fig. 1 shows in-hospital mortality and mortality within the first 24 h after admission according to the day of the week. Patients admitted on Sunday showed the highest in-hospital mortality and mortality within the first 24 h after admission, followed by those admitted on Saturday. The proportion of deaths occurring within the first 24 h after admission among the total in-hospital deaths for AMI according to admission on a

specific day was also highest for Sunday admission (53.1%), followed by Saturday admission (48.1%), compared with 44.9% to 47.7% for weekday admission. Table 3 shows the comparison of outcomes between weekend and weekday admissions. Patients admitted on weekends showed significantly higher in-hospital mortality than those admitted on weekdays (13.6% versus 11.4%; P b 0.001; unadjusted OR 1.222, 95% CI 1.172 to 1.274), although there was no significant difference in median length of stay between weekend and weekday admissions (14 days [IQR, 10–21 days] versus 14 days [IQR, 9–21 days]; P = 0.600). In the multivariable analysis, weekend admission remained significantly associated with higher in-hospital mortality after adjustment for the measured

Table 2 Invasive procedures and medications during hospitalization.

Coronary angiography on the day of admission Reperfusion therapy on the day of admission PCI Fibrinolytic therapy Any reperfusion therapy Mechanical support on the day of admission IABP ECMO Mechanical ventilation Renal replacement therapy Coronary angiography during hospitalization Revascularization procedure during hospitalization PCI CABG Any revascularization procedure Drugs during hospitalization Aspirin Clopidogrel/ticlopidine ACE-I/ARB β-blocker Statin

Overall (n = 111,200)

Weekday admission (n = 80,353)

Weekend admission (n = 30,847)

P-value

80,611 (72.5)

57,649 (71.7)

22,962 (74.4)

b0.001

73,292 (65.9) 1747 (1.6) 73,677 (66.3)

52,030 (64.8) 1246 (1.6) 52,305 (65.1)

21,262 (68.9) 501 (1.6) 21,372 (69.3)

b0.001 0.378 b0.001

11,493 (10.3) 1601 (1.4) 11,531 (10.4) 8643 (7.8) 97,249 (87.5)

8021 (10.0) 1084 (1.3) 7856 (9.8) 5902 (7.3) 70,409 (87.6)

3472 (11.3) 517 (1.7) 3675 (11.9) 2741 (8.9) 26,840 (87.0)

b0.001 b0.001 b0.001 b0.001 0.006

86,548 (77.8) 2380 (2.1) 87,928 (79.1)

62,029 (77.2) 1775 (2.2) 63,083 (78.5)

24,519 (79.5) 605 (2.0) 24,845 (80.5)

b0.001 0.011 b0.001

96,623 (86.9) 86,997 (78.2) 75,265 (67.7) 59,365 (53.4) 78,040 (70.2)

69,770 (86.8) 62,583 (77.9) 54,252 (67.5) 42,843 (53.3) 56,395 (70.2)

26,853 (87.1) 24,414 (79.1) 21,013 (68.1) 16,522 (53.6) 21,645 (70.2)

0.324 b0.001 0.054 0.468 0.961

Numbers are n (%). PCI: percutaneous coronary intervention, CABG: coronary artery bypass graft, IABP: intra-aortic balloon pumping, ECMO: extracorporeal membrane oxygenation, ACEI: angiotensin converting enzyme-inhibitors, ARB: angiotensin II receptor blockers.

318

T. Isogai et al. / International Journal of Cardiology 179 (2015) 315–320

Fig. 1. Mortality according to day of week. The bars, including white and black parts, show in-hospital mortality on each day of week. The black parts show the mortality within the first 24 h after admission.

confounding factors (OR 1.144, 95% CI 1.079 to 1.214). We obtained consistent results for mortality within the first 24 h after admission, 7day mortality, and 30-day mortality. 3.4. Subgroup analyses Table 4 shows subgroup analyses according to Killip class at admission. These analyses with multivariable adjustment showed that weekend admission was significantly associated with higher in-hospital mortality among Killip II to IV subgroups. In Killip I or unclassified subgroups, however, there was no significant association between weekend admission and in-hospital mortality. 4. Discussion This study showed that in-hospital mortality of AMI patients was significantly higher for weekend admission than weekday admission, despite the higher rate of PCI performed on the day of admission. In multivariable analysis, weekend admission was significantly associated with higher in-hospital mortality, even after adjustment for confounding factors. In subgroup analyses, weekend admission was associated with higher mortality in the Killip II to IV subgroups, but not in the Killip I subgroup. Although previous studies have examined the effect of off-hour admission for AMI on mortality, the results were inconsistent [1,3–5, 9–14]. In 2014, a meta-analysis which included 1,896,859 patients in 48 studies, reported that off-hour admission of patients with AMI was associated with higher short term mortality (OR 1.06, 95% CI 1.04 to 1.09), and patients with STEMI admitted during off-hours had longer door-to-balloon time (mean difference 14.8 min, 95% CI 10.7 to 19.0 min) [16]. However, there was high heterogeneity in the studies

included in the meta-analysis. From 2005 to 2008, three large-scale studies from the United States examined the effect of off-hour admission for AMI [4,9,10]. Using the data from the National Registry of Myocardial Infarction database, Magid et al. reported that patients with STEMI admitted during off-hours had significantly higher adjusted inhospital mortality and longer door-to-balloon times than patients admitted during regular hours, although the difference in mortality became insignificant after adjustment for reperfusion treatment time [4]. Using data from an administrative database in New Jersey, Kostis et al. found that mortality for AMI was higher in patients admitted on weekends than on weekdays [10], although the difference became insignificant after adjustment for invasive cardiac procedures. Both studies suggested that the higher mortality in patients admitted during offhours could be explained partially by the decreased availability of PCI or fibrinolytic therapy [4,10]. Contrary to the two studies, using data from the Get With the Guidelines-Coronary Artery Disease database, Jneid et al. showed that patients presenting with AMI during off-hours had in-hospital mortality similar to those presenting during regular hours despite significantly longer door-to-balloon times [9]. In Europe, several recent studies showed that there was no significant difference in short-term mortality between off-hour and regular hour admissions [5,11,12], but a study from Germany showed a significantly higher inhospital mortality after weekend admission for STEMI [14]. The abovementioned results from Western nations may not be applicable to the Japanese population because the availability of PCI for AMI is higher in Japan [19]. The proportion of patients undergoing PCI for AMI was low at 32–45% in Western nations [4,5,9–12], whereas our study showed that 66% of the study patients underwent PCI on the day of admission and 78% during hospitalization. We speculated that one of the reasons for the relatively low rate of PCI in Western nations might be an insufficient number of hospitals equipped to perform PCI for AMI. For

Table 3 Mortality in patients admitted on weekends and on weekdays. Overall (n = 111,200)

In-hospital mortality Mortality within the first 24 h after admission 7-day mortality 30-day mortality

Weekday admission (n = 80,353)

Weekend admission (n = 30,847)

P-value

Univariable analysis

Multivariable analysis*

No. deaths (%)

No. deaths (%)

No. deaths (%)

Odds ratio (95%CI)

Odds ratio (95%CI)

13,322 (12.0) 6327 (5.7)

9140 (11.4) 4217 (5.2)

4182 (13.6) 2110 (6.8)

b0.001 b0.001

1.222 (1.172 to 1.274) 1.326 (1.254 to 1.402)

1.144 (1.079 to 1.214) 1.158 (1.049 to 1.278)

9385 (8.4) 11,938 (10.7)

6343 (7.9) 8170 (10.2)

3042 (9.9) 3768 (12.2)

b0.001 b0.001

1.277 (1.218 to 1.338) 1.299 (1.178 to 1.283)

1.172 (1.088 to 1.262) 1.124 (1.055 to 1.198)

Odds ratios of weekend admission in reference to weekday admission are shown. *In multivariable analysis, odds ratios were adjusted for age, sex, ambulance use, Killip class at admission, nine comorbidities present at admission, type of hospital, hospital volume, reperfusion therapies on the day of admission, mechanical support on the day of admission, revascularization procedures during hospitalization, and drugs during hospitalization. CI: confidence interval.

T. Isogai et al. / International Journal of Cardiology 179 (2015) 315–320

319

Table 4 In-hospital mortality according to Killip class at admission. Killip class

I II III IV Unclassified

Overall

Weekday admission

Weekend admission

No. of deaths/Total No. (%)

No. of deaths/Total No. (%)

No. of deaths/Total No. (%)

815/46,989 (1.7) 1308/29,518 (4.4) 1373/9449 (14.5) 6081/13,973 (43.5) 3745/11,271 (33.2)

600/34,286 (1.7) 925/21,510 (4.3) 939/6731 (14.0) 4074/9539 (42.7) 2602/8287 (31.4)

215/12,703 (1.7) 383/8008 (4.8) 434/2718 (16.0) 2007/4434 (45.3) 1143/2984 (38.3)

P-value

0.689 0.074 0.013 0.005 b0.001

Univariable analysis

Multivariable analysis*

Odds ratio (95%CI)

Odds ratio (95%CI)

0.967 (0.820 to 1.139) 1.118 (0.985 to 1.268) 1.172 (1.040 to 1.321) 1.109 (1.031 to 1.193) 1.356 (1.240 to 1.484)

1.002 (0.828 to 1.213) 1.155 (1.006 to 1.325) 1.328 (1.143 to 1.544) 1.140 (1.016 to 1.279) 1.071 (0.934 to 1.227)

Odds ratios of weekend admission in reference to weekday admission are shown. *In multivariable analysis, odds ratios were adjusted for age, sex, ambulance use, nine comorbidities present at admission, type of hospital, hospital volume, reperfusion therapies on the day of admission, mechanical support on the day of admission, revascularization procedures during hospitalization, and drugs during hospitalization.

example, in the study from New Jersey, only one quarter of the study patients were admitted to hospitals equipped to perform PCI [10]. In Japan, many hospitals are equipped to perform PCI, and primary PCI is widely established as a first choice for AMI, leading to a higher proportion of AMI patients receiving PCI on admission [19]. Although two previous studies from Japan reported no weekend effect on mortality of AMI patients [29,30], these were limited because of the small sample sizes (n = 307 to 4805) and the limited area. In contrast, the present study has the advantage of being based on a large number of patients with AMI in a nationwide database. Regarding baseline characteristics, differences of ≤0.5% in diabetes mellitus with complications, neoplasms, pulmonary edema, and chronic renal failure were statistically significant (P b 0.05), but they may not be important from a clinical standpoint. However, the differences in the outcomes between weekend and weekday admissions showed both statistical and clinical significance. Our study described that invasive procedures and medications were equally available for weekday and weekend admissions in Japan, suggesting that the quality of the Japanese healthcare system should be similar throughout the whole week. In such a preferable setting for AMI, however, the present study showed that weekend admission for AMI was significantly associated with higher mortality. This increased mortality of patients admitted on weekends could lead to a substantial number of deaths which might be preventable. One reason for the higher mortality might be that more severely ill patients with high Killip class or requiring invasive procedures were admitted on weekends. However, the association between weekend admission and higher mortality remained significant after multivariable adjustment. We speculate that two reasons could explain this weekend effect. First, a potential delay in time from symptom onset to PCI on weekends may have influenced the in-hospital outcomes in patients with AMI. Previous studies showed that presentation during off-hours was associated with longer prehospital and in-hospital delays to primary PCI for patients with STEMI [4,14]. In our study, 121 hospitals (10.8%) did not receive patients with AMI on weekends. Such a situation may have changed the nearest PCI-equipped hospital for many patients to take longer time from symptom onset to arrival at hospital. Additionally, it may have taken a longer door-to-balloon time on weekends, as in previous studies [4,9], because many institutions need to gather on-call staff to activate the cardiac catheterization laboratory during off-hours. Thus, total ischemia time during weekends might have been prolonged and led to increased inhospital mortality. Second, there may be insufficient staff to provide prompt treatment for more serious cases on weekends. Weekend admission was significantly associated with mortality within the first 24 h as well as in-hospital mortality. The patients admitted on Sunday showed the highest mortality within the first 24 h (7.5%) and the highest proportion of deaths occurring within the first 24 h among inhospital deaths of patients admitted on a specific day (53.1%), followed by those admitted on Saturday. To rescue patients with severe AMI, we often need to perform not only PCI but also other intensive care promptly to prevent rapidly deterioration [31,32]. In Japanese PCI-equipped hospitals, the staff available on weekends might be able to perform invasive procedures, but there might not be enough staff in immediate attendance to provide the level of care required to rescue patients with

AMI who rapidly progress to a fatal condition. This might also explain the absence of a weekend effect in the Killip I subgroup in the present study: less severe cases with AMI did not require as many staff for survival in the acute phase, while greater staff input was required for more severe emergency events. Thus, the decreased staffing levels on weekends may have contributed to higher mortality. Further studies are needed to explore these possibilities. To reduce the unfavorable effect of weekend admission, a healthcare system should provide consistent quality of care throughout the week, including weekends. Our study has several limitations. First, it was a retrospective design using a secondary database, in which recorded diagnoses are less well validated than those in planned prospective cohorts or registries. Second, the DPC database does not include detailed data about the patients' medical history and laboratory findings, e.g., with respect to the time course of AMI, cardiac biomarkers, electrocardiogram, echocardiogram, and coronary angiography. In this study, therefore, some baseline characteristics about AMI were unavailable, e.g., ST-elevation or non-STelevation AMI, time from symptom onset to PCI, infarct size, and infarct location. These unmeasured confounding factors may have contributed to the differences in mortality between weekend and weekday admissions found in this study [7,8]. Third, the DPC database does not include AMI patients who died outside hospital. We therefore could not assess the difference in prehospital mortality between patients developing AMI on weekdays and those on weekends. Finally, the present study covered only a Japanese population, reflecting Japanese situations in terms of geography, infrastructure and healthcare services. Our results therefore might not be generalizable to other countries, but can be generalized to the whole population of patients with AMI in Japan because the study was based on a nationwide database. 5. Conclusions The present study showed that weekend admission for AMI was significantly associated with higher in-hospital mortality even in a setting where PCI was readily available. The effect of weekend admission on inhospital mortality remained significant after adjusting for baseline characteristics, invasive procedures, and medication. In a subgroup analysis, the effect of weekend admission remained significant in the Killip II to IV subgroups, but was not significant in the Killip I subgroup. Funding sources This work was supported by grants from the Ministry of Health, Labour and Welfare, Japan (Research on Policy Planning and Evaluation grant number: H25-Policy-010), and the Council for Science and Technology Policy, Japan (Funding Program for World-Leading Innovative R&D on Science and Technology, FIRST program grant number: 0301002001001). Conflict of interest The authors report no relationships that could be construed as a conflict of interest.

320

T. Isogai et al. / International Journal of Cardiology 179 (2015) 315–320

References [1] C.M. Bell, D.A. Redelmeier, Mortality among patients admitted to hospitals on weekends as compared with weekdays, N. Engl. J. Med. 345 (2001) 663–668. [2] P. Cram, S.L. Hillis, M. Barnett, G.E. Rosenthal, Effects of weekend admission and hospital teaching status on in-hospital mortality, Am. J. Med. 117 (2004) 151–157. [3] P. Aylin, A. Yunus, A. Bottle, A. Majeed, D. Bell, Weekend mortality for emergency admissions. A large, multicentre study, Qual. Saf. Health Care 19 (2010) 213–217. [4] D.J. Magid, Y. Wang, J. Herrin, et al., Relationship between time of day, day of week, timeliness of reperfusion, and in-hospital mortality for patients with acute STsegment elevation myocardial infarction, JAMA 294 (2005) 803–812. [5] A. Berger, J.C. Stauffer, D. Radovanovic, P. Urban, O. Bertel, P. Erne, Comparison of inhospital mortality for acute myocardial infarction in Switzerland with admission during routine duty hours versus admission during out of hours (Insight into the AMIS Plus Registry), Am. J. Cardiol. 101 (2008) 422–427. [6] C.M. Bell, D.A. Redelmeier, Waiting for urgent procedures on the weekend among emergently hospitalized patients, Am. J. Med. 117 (2004) 175–181. [7] H. Jneid, J.L. Anderson, R.S. Wright, et al., 2012 ACCF/AHA focused update of the Guidelines for the Management of Patients with Unstable Angina/non-STElevation Myocardial Infarction (updating the 2007 Guidelines and replacing the 2011 focused update): a report of the American College of Cardiology Foundation/ American Heart Association Task Force on Practice Guidelines, Circulation 126 (2012) 875–910. [8] P.T. O'Gara, F.G. Kushner FG, D.D. Ascheim, et al., 2013 ACCF/AHA Guideline for the Management of ST-Elevation Myocardial Infarction: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines, Circulation 127 (2013) e362–e425. [9] H. Jneid, G.C. Fonarow, C.P. Cannon, et al., Impact of time of presentation on the care and outcomes of acute myocardial infarction, Circulation 117 (2008) 2502–2509. [10] W.J. Kostis, K. Demissie, S.W. Marcella, Y.H. Shao, A.C. Wilson, A.E. Moreyra, Weekend versus weekday admission and mortality from myocardial infarction, N. Engl. J. Med. 356 (2007) 1099–1109. [11] S.M. Snelder, S.T. Nauta, K.M. Akkerhuis, J.W. Deckers, R.T. van Domburg, Weekend versus weekday mortality in ST-segment elevation acute myocardial infarction patients between 1985 and 2008, Int. J. Cardiol. 168 (2013) 1576–1577. [12] K.W. Hansen, A. Hvelplund, S.Z. Abildstrom, et al., Prognosis and treatment in patients admitted with acute myocardial infarction on weekends and weekdays from 1997 to 2009, Int. J. Cardiol. 168 (2013) 1167–1173. [13] J.W. Ryan, E.D. Peterson, A.Y. Chen, et al., Optimal timing of intervention in non-STsegment elevation acute coronary syndromes: Insights from the CRUSADE (can rapid risk stratification of unstable angina patients suppress adverse outcomes with early implementation of the ACC/AHA Guidelines) Registry, Circulation 112 (2005) 3049–3057. [14] P. Kruth, U. Zeymer, A. Gitt, et al., Influence of presentation at the weekend on treatment and outcome in ST-elevation myocardial infarction in hospitals with catheterization laboratories, Clin. Res. Cardiol. 97 (2008) 742–747. [15] J.S. Hong, H.C. Kang, S.H. Lee, Comparison of case fatality rates for acute myocardial infarction in weekday vs weekend admissions in South Korea, Circ. J. 74 (2010) 496–502.

[16] A. Sorita, A. Ahmed, S.R. Starr, et al., Off-hour presentation and outcomes in patients with acute myocardial infarction: systematic review and meta-analysis, BMJ 348 (2014) f7393. [17] K. Sakurai, J. Watanabe, K. Iwabuchi, et al., Comparison of the efficacy of reperfusion therapies for early mortality from acute myocardial infarction in Japan: registry of Miyagi Study Group for AMI (MSAMI), Circ. J. 67 (2003) 209–214. [18] T. Takii, S. Yasuda, J. Takahashi, et al., Trends in acute myocardial infarction incidence and mortality over 30 years in Japan: report from the Miyagi-AMI registry study, Circ. J. 74 (2010) 93–100. [19] S. Ui, M. Chino, T. Isshiki, Rates of primary percutaneous coronary intervention worldwide, Circ. J. 69 (2005) 95–100. [20] M. Nakamura, M. Yamagishi, T. Ueno, et al., Current treatment of st elevation acute myocardial infarction in Japan: door-to-balloon time and total ischemic time from the J-AMI registry, Cardiovasc. Interv. Ther. 28 (2013) 30–36. [21] T. Killip III, J.T. Kimball, Treatment of myocardial infarction in a coronary care unit. A two year experience with 250 patients, Am. J. Cardiol. 20 (1967) 457–464. [22] V.S. DeGeare, J.A. Boura, L.L. Grines, W.W. O'Neill, C.L. Grines, Predictive value of the Killip classification in patients undergoing primary percutaneous coronary intervention for acute myocardial infarction, Am. J. Cardiol. 87 (2001) 1035–1038. [23] T.H. Tsai, S. Chua, H. Hussein, et al., Outcomes of patients with Killip class III acute myocardial infarction after primary percutaneous coronary intervention, Crit. Care Med. 39 (2011) 436–442. [24] M. Sumitani, K. Uchida, H. Yasunaga, et al., Prevalence of malignant hyperthermia and relationship with anesthetics in Japan: data from the diagnosis procedure combination database, Anesthesiology 114 (2011) 84–90. [25] A.B. Jena, E.C. Sun, J.A. Romley, Mortality among high-risk patients with acute myocardial infarction admitted to US teaching-intensive hospitals in July: a retrospective observational study, Circulation 128 (2013) 2754–2763. [26] J.V. Tu, P.C. Austin, R. Walld, L. Roos, J. Agras, K.M. McDonald, Development and validation of the Ontario acute myocardial infarction mortality prediction rules, J. Am. Coll. Cardiol. 37 (2001) 992–997. [27] M.J. Vermeulen, J.V. Tu, M.J. Schull, ICD-10 adaptations of the Ontario acute myocardial infarction mortality prediction rules performed as well as the original versions, J. Clin. Epidemiol. 60 (2007) 971–974. [28] A.E. Hubbard, J. Ahern, N.L. Fleischer, et al., To GEE or not to GEE: comparing population average and mixed models for estimating the associations between neighborhood risk factors and health, Epidemiology 21 (2010) 467–474. [29] K. Matsui, S. Kojima, T. Sakamoto, et al., Weekend onset of acute myocardial infarction does not have a negative impact on outcome in Japan, Circ. J. 71 (2007) 1841–1844. [30] T.C. Turin, Y. Kita, N. Rumana, et al., Incidence, admission and case–fatality of acute myocardial infarction: weekend versus weekday in a Japanese population: 16-year results from Takashima AMI Registry (1988–2003), Eur. J. Epidemiol. 24 (2009) 93–100. [31] D.M. Henkel, B.J. Witt, B.J. Gersh, et al., Ventricular arrhythmias after acute myocardial infarction: a 20-year community study, Am. Heart J. 151 (2006) 806–812. [32] M.P. Donataccio, E. Puymirat, C. Vassanelli, et al., Presentation and revascularization patterns of patients admitted for acute coronary syndromes in France between 2004 and 2008 (from the National Observational Study of Diagnostic and Interventional Cardiac Catheterization [ONACI]), Am. J. Cardiol. 113 (2014) 243–248.