Since the end of the MONICA study [
12,
13], few studies have assessed the management and outcome of AMI in Switzerland. Our results show an increase in most revascularization procedures, namely PCI, while CABG tended to stabilize or even to decrease in the most recent years. Moreover, these trends strongly differ according to region. Overall, the changes in revascularization procedure in Switzerland follow the ones observed in other countries, [
6,
7,
14,
15]. These findings are also in agreement with a previous study conducted on 19,500 patients of the AMIS Plus registry [
8], although the increase in PCI procedures observed [
16] was considerably stronger than in ours. This difference is most likely due to the overrepresentation of PCI procedures among the AMIS Plus registry hospitals as the Registry includes only selected voluntary teams.
Between 1998 and 2008, the number of hospital discharges for AMI increased over two-fold in Switzerland. This finding is in contradiction with some previous studies [
17,
18], but it should be noted that these studies showed declining rates of AMI rather than number of discharges. Part of this increase might be due to population aging [
2], although the observed increase in population age (from 66.5 years in 1998 to 67.6 years in 2008) would not lead to a doubling of the number of hospitalizations for AMI. Further, all analyses were adjusted for age. Another possibility would be a change in disease coding, but again this is unlikely as hospital discharge data has been shown to be reliable regarding the diagnosis of AMI [
19] and as this study focused on a specific code. It is also possible that more people are hospitalized for a second or even a third AMI (partly related to the decrease in case fatality rate of the first AMI, leading to an increased incidence of 2
nd or 3
rd AMI). Another possibility is the change in criteria defining AMI [
20], which has been shown to increase the incidence of AMI [
21]. The most likely explanation for the considerable increase in the number of hospital discharges for AMI in Switzerland is the steep rise in hospital transfers, which accounted for almost half of hospital discharges for AMI in 2008. The increase in hospital transfers might be due to the limited number of hospitals capable of performing revascularization procedures, prompting smaller hospitals to transfer their patients as recommended by European guidelines [
22,
23]. Still, considerable differences in the percentages of hospital discharges due to transfers were noted: in 2008 the values ranged from 40% in Leman to 68% in Ticino. These differences might be due to local policies. As AMI management is expensive [
24], the increase in hospital discharges for AMI would add a considerable pressure in Swiss health expenditures, which already increased from 9.6% in 1995 to 11.4% of the Swiss GNP in 2009 [
25].
Intensive care unit
The percentage of hospitalizations with ICU stay remained stable between 1998 and 2008 in Switzerland. This stabilization might be the consequence of improved management, requiring less ICU admissions before or after a revascularization intervention. Another likely explanation is the increasing number of transfers, the patients being stabilized before being transferred to another hospital for revascularization. Still, considerable differences in the proportion of hospitalizations with ICU and corresponding trends were found between regions, suggesting that other factors such as the number of ICU beds available or local options for AMI management might be at play. Unfortunately there is no available information regarding the number of ICU beds in each region, so the precise reasons for the regional differences in ICU admissions for AMI remain to be further assessed.
Revascularisation procedures
Stent use increased steeply from slightly over 1% in 1998 to nearly 40% in 2008, a finding in agreement with the literature [
26]. DES appeared in 2004 and rose steadily until 2006, when several reports questioning their costs [
27,
28] and long term complications [
29,
30] were published. These reports slowed down DES use, but some Swiss regions were more responsive than others. For instance, in the Leman area DES use equaled but never exceeded bare stent use, while in South Switzerland (Ticino) DES represented nearly all stents implanted since 2005. The reasons for such regional differences are more likely due to local preferences or to local agreements regarding DES cost than to decisions based on scientific evidence.
The rate of CABG remained stable at 3% throughout the study period, a finding also reported in other countries [
3,
6]. As for stents, very different trends according to the region were found. For instance, South Switzerland (Ticino) showed a steep rise in CABG use, starting in 1999 to reach an average rate of 7%, well above the Swiss average. This trend might partly be explained by the creation in 1999 of new medical centers with the capacity to carry out these interventions. Conversely, in Eastern and Central Switzerland, hardly any CABG was performed until 2007, likely the result of the lack of required infrastructure. Overall, our results show that CABG rates evolved differently between Swiss regions, the most likely explanations being the existence of infrastructures and local preferences for certain types of revascularization procedures.
In-hospital mortality
An overall decrease in seven-day in-hospital mortality was found for Switzerland, a finding also reported elsewhere [
8,
16]. This trend was no longer significant after multivariate adjustment, suggesting that most of the decrease in mortality is due to revascularization interventions and the improved quality of treatment according to the newest guidelines [
9,
10]. Restricting the analysis to patients who survived at least three days led to similar findings, although some regions presented a slight increase in mortality (Additional file
5: Table S1). The rising prevalence of CV risk factors in Switzerland [
31] could also contribute to lessen the decrease in AMI mortality [
32], but it was not possible to adequately consider them in the multivariate model as they are not reported systematically in the database. Similarly, it was not possible to assess whether the patients presented with a STEMI or a non-STEMI, and which drug treatments were provided during hospital stay, all factors associated with in-hospital mortality [
22,
23,
33]. Overall, our data suggest that the increasing number of revascularization interventions performed among patients discharged with a main diagnosis of AMI do not suffice to decrease in-hospital mortality in Switzerland, a country presenting the lowest CVD mortality within Europe [
1]. Further studies are mandatory to better assess the factors associated with in-hospital mortality from AMI in order to optimize patient management.
Study limitations
This study has some limitations. Medicines are not included in the hospital discharge database, thus precluding the assessment of their trends. This absence might also explain the very low rates for thrombolysis, which are likely due to the absence of coding rather than a true absence of use. Furthermore, we lack data to assess whether the changes in thrombolytic therapy are related to changes in reporting levels or are actually real changes. The ICD10 coding does not distinguish STEMI from NSTEMI; hence, it is possible that part of the increase in discharges from AMI during the study period might be related to the change in the definition of NSTEMI populations due to the generalization of the use of troponin measurements. Indeed, introduction of troponin measurements led to the reclassification of many unstable angina patients into NSTEMI patients, with a 33% increase of the population of NSTEMIs from before to after the introduction of troponin measurements [
34]. Still, including unstable angina (ICD10 code I200) in the analysis led to similar findings, with the exception that the increase in CABG was no longer significant and even decreased in some regions (Additional file
5: Table S1). The strong increase in the number of between-hospital transfers and the anonymization of the data (which precluded the follow-up of the patients) complicated the interpretation of the trends. Still, restricting the analysis to patients fully managed in a single hospital led to similar results. It was not possible to assess “true” seven- and 30-day mortality rates as no follow-up data was available for the patients discharged. The AMIS registry, which collects vital status data for AMI patients hospitalized in the participating institutions, will be able to provide such information [
35]. Finally, there is no published information on the validity of hospital discharge data in Switzerland regarding the diagnosis of AMI. Still, using the results from an ongoing study (CoLaus), of 159 hospitalizations for AMI, 145 (91%) were confirmed as such by an independent panel of cardiologists, a value in agreement with the literature [
36,
37].