The BSR is a monocentric prospective registry of patients admitted with STEMI at the Bremen heart center. The Bremen heart center is exclusively responsible for 24-h-PCI-service in a large region in northwest Germany, serving a population of more than 800.000 residents from Bremen and in the northwestern part of Lower Saxony. Emergency services and regional hospitals are connected by telephone and fax for announcement of urgent coronary catheterization in case of STEMI. Since all STEMIs from the Bremen heart center are documented in the BSR it claims to give a complete statistical overview about the clinical course (short- and long-term) of STEMI-patients in the region of Bremen. The BSR was established in 2006 and is still running. Documentation is accomplished via data sheets completed by the responsible interventional cardiologist and/or through patient records after a physician has confirmed the exact diagnosis. Data about age, residence, sex, temporal delays of treatment, concomitant diseases, severity of STEMI and acute medical or interventional treatment are recorded. At discharge major adverse cardiac and cerebral events are documented. Follow-up examination is performed after 1 and 5 years by a telephone interview. Further details have been published previously [18].

STEMI was defined as persistent angina pectoris for ≥20 min in conjunction with a ST-segment elevation in two contigous leads of ≥0.25 mV in men below the age of 40 years, ≥0.2 mV in men over the age of 40 years, or ≥0.15 mV in women in leads V2-V3 and/or ≥0.1 mV in all other leads or new left bundle branch block [19].

Disease severity was evaluated using surrogate parameters, such as acute heart failure, pulmonary edema or cardiogenic shock at admission by the Killip-classification, multivessel disease with or without left-main-lesion and by measurement of the post-AMI ejection fraction. The peak creatine kinase (CK) was used to assess extension of myocardial infarction with laboratory controls routinely performed in STEMI-patients once to twice daily. To assess the quality of care, door-to-balloon time, interventional procedures and medication at hospital discharge were analysed.

The Bremen social deprivation index (SDI) was developed in 1993 to describe social inequalities in the city of Bremen in Germany. In its most recent update (2009) a ranking was established for 89 city districts, using census and public data on education, voting participation, unemployment, crime statistics and immigration flow. A positive composite index indicates a socially privileged, a negative composite index a socially disadvantaged city district with an overall mean SDI of 0 [20]. Since only postal codes were available in the BSR the SDIs of the 89 city districts were aggregated to 33 postal code regions of Bremen. Population data stem from municipial data of 2008.

Since a normal distribution of the social deprivation index across the postal code regions could be assumed (Kolgmorov-Smirnov-test of the 33 postal code regions: p = 0.15) distribution into groups was made according to the mean and standard deviation of the SDI: Within one standard deviation of the mean SDI postal codes were assigned to the groups G2 (social status: intermediate high) and G3 (social status: intermediate low). Postal codes with an SDI below −1 SD of the mean were assigned to G4 (social status: low), postal codes above +1 SD of the mean SDI were assigned to G1 (social status: high). Baseline characteristics of STEMI-patients were given as mean values ± standard deviation for continuous variables (age, left ventricular ejection fraction, door to balloon-time, number of coronary vessels affected) and in absolute numbers and percentages for categorical variables: (diabetes, obesity (body mass index > 30 kg/m²), smoking, multivessel-disease, prehospital cardiopulmonal resuscitation (CPR), peak CK > 3000 U/l, acute heart failure, positive family history (FH) for CAD/CVD, primary PCI, CABG, LVEF ≤ 40%, medication at admission and at discharge) by using SAS/STAT (version 9.3., 2011). Statistical significance was tested using Chi-square for categorical and t-tests for continuous variables. For comparison of the groups ANOVA tests were used for continuous variables, Chi-square for trend-test (Cochrane-Armitage) for categorical variables. A p-value <0.05 was considered statistically significant. Sex-, age-, and SES-group-specific incidence rates (per 100.000 persons/year) were estimated as the sum of of STEMI-events during 2006–2015 divided by the sum of persons at risk in the strata of sex, age- and SES-group. Incidence rate ratios (IRRs) and 95% confidence intervals were estimated by log-linear Poisson regression of sex-, age- and SES-group-specific number of STEMI-events with logarithm of persons at risk as offset. Multivariate logistic regression analysis were performed with SES-groups, age, sex and prior CVD as the independent variable and age, sex, smoking, obesity, diabetes mellitus, CK > 3000 U/l, primary PCI, successful PCI and LVEF < 40% as the dependent variable. Analysis of mortality and adverse-events-rates (major adverse cardiac and cerebrovascular events (MACCE): death, stroke, myocardial reinfarction) was performed after stratifying for age (<50 yrs., ≥50 yrs.).

The total study population comprised the city of Bremen with approximately 550,000 inhabitants. Between January 1st 2006 and December 31th 2015 a total of 3462 residents of the city of Bremen a STEMI was newly diagnosed. They were assigned to groups (G1 to G4) according to the social deprivation index of their home address, with G1 representing the most affluent neighbourhoods and G4 representing the most deprived ones (Table 1).

The unadjusted incidence-rate (per 100.000 persons/year) of STEMI in the city of Bremen between 2006 and 2015 was 63 per 100,000 inhabitants/year or 74 per 100,000 inhabitants/year when excluding inhabitants <18 yrs. from the calculation model (Table 2). The hypothesis, that the number of STEMIs was equally distributed over time could not be rejected for any of the four SDI-Groups (p-values calculated by means of a chi-square goodness-of-fit test): G1 (p = 0.45), G2 (p = 0.14), G3 (p = 0.14), G4 (p = 0.10). A rising incidence of STEMI could be observed with declining SDI category. Adjusted relative risk (for age and sex) for STEMI showed a 56% increase of STEMI-incidence among pts. from low SES neighbourhoods as compared to the more affluent city regions. The association of rising STEMI incidence with increased social deprivation could be demonstrated in both sexes (Fig. 1a). The socioeconomic gradient was most prominent in young patients (<50 years of age) showing a 2.18 fold elevation of STEMI-incidence when comparing G4 to G1 in women (Fig. 1b) and a 2.17 fold elevation in men (Fig. 1c). In the older ages groups this gradient consistently decreased until being no longer significant in pts. ≥ 80 years for men and women alike (Fig. 1b and c).

The 3462 patients admitted with STEMI between 2006 and 2015 were on average 64.8+/−13 years old, 71% were male, 44% were active smokers, 21% had diabetes and 23% were obese (BMI > 30 kg/m²). Patients from neighbourhoods with a lower SDI were on average younger at time of STEMI, more likely to smoke and more likely to be obese than those from more affluent city districts. In contrast the prevalence of a positive family history for CAD was significantly lower in patients from low-SES city districts. Gender distribution as well as prevalence of diabetes in STEMI patients did not differ substantially between SDI categories (Table 3). A multivariate analysis confirmed these results: higher degrees of social disadvantage was positively associated with higher likelihood of suffering the STEMI at young age, to smoke or to be obese at time of STEMI. In contrast social deprivation and the prevalence of positive family history for CAD were inversely related while gender and diabetes mellitus did not show significant differences in prevalence between SDI-categories (Fig. 2).

While the prevalence of multi-vessel-disease did not differ substantially between SES-classes, pts. from low-SES city districts showed greater extension of STEMI, measured by the peak creatine kinase (peak CK) and had on average lower left-ventricular-ejection-fraction (LVEF) post-STEMI (Table 4). The relation between social deprivation and infarct size and subsequent LV-EF-reduction post STEMI remained significant in a multivariate model (Fig. 3). Rates of prehospital resuscitation as well as the rate of cardiogenic shock in STEMI patients did not differ between SES-classes (Table 4).

The proportion of patients receiving a successful primary PCI was >89%, with no significant disparities when comparing the socioeconomic groups (Fig. 3). Moreover, comparable proportions of STEMI-patients had early surgical revascularisation procedures (CABG) or a conservative therapy regime. Prior to the coronary angiography pts. from all SES-groups showed similar average door-to-balloon-times. Patients from all quartiles received additional Gp2b3A-antagonist-medication (Tirofiban or Abciximab) at a similar frequency. At discharge from hospital there was no difference in standard medication (ASA, ADP-antagonists, beta blockers, statins, ACE/ATR-Inhibitors) between the groups (Table 4).

While SES had no significant impact on inhospital-mortality, 1-year-mortality, 1-year-MACCE-rates and 5-year-mortality-rates, lower SES was associated with a significant increase in MACCE-events within 5 years after STEMI, showing a 34% increase in event rates when comparing pts. from the most socially deprived city areas to those from the city areas with the highest SES (Table 4). When stratifying follow-up-data by age (<50 yrs. of age and ≥50 yrs. of age) a negative impact of social deprivation on long-term-outcome could be shown for the young STEMI-cohort with elevated 1-year mortality-, 1-year-MACCE-rates as well as 5-year-mortality- and 5-year-MACCE-rates for the socially deprived (Fig. 4a). In contrast short-term- (<72 h) (G1: 1.3%, G2: 0.9%, G3 1.0%, G4:1.6%; p (for trend) = 0.8) and in-hospital mortality-rates (G1: 2.6% G2: 2.4%, G3: 3.9%, G4: 3.6%, p (for trend) = 0.1) were not significantly affected by SES-class. For pts. ≥ 50 yrs. SES did not show an impact on 1- and 5-year mortality-and MACCE-rates, while overall event-rates were generally higher due to the advanced patients' age (Fig. 4b).

An association between socioeconomic deprivation and increased risk for suffering from CAD and AMI has been described by prior investigations [10, 21]. Besides the observation of an increased risk for coronary events in socioeconomic disadvantaged neighbourhoods in four US study sites by Diez Roux [21], a social down-gradient of STEMI incidence could be observed in both sex groups and most significantly in young inhabitants in a Dutch study (relative risk in women 1.44, in men 1.34, decreasing social gradient with increasing age) [10]. These results are in good accordance to our data.

The elevated cardiovascular risk profile in patients with a low SES, in particular the higher rates of smokers and obese pts., might partly explain the elevated STEMI incidence. The association of active smoking and obesity with CAD and CAD-related mortality has been proven [22‐25] and the influence of tobacco smoking in public areas on non-smokers` risk for AMI has been demonstrated by recent investigations [18]. Additionally there is some evidence, that beyond the classic risk factors for CAD the SES may have biological, behavioral and psychosocial components affecting the deprived [7, 26, 27]. Chetty et al. found that life expectancy more likely depends on health behaviors (i.e. smoking, obesity, exercise) rather than access to medical supplies or environmental factors (i.e. air pollution) [13]. In concern of CVD and MI disease determinants are varied. One aspect could be the neighbourhood in which people live, but social influences on health are complex and seem to operate through many different processes. Worldwide debate about to what extend contextual and/or compositional effects account for higher morbidity and mortality of the underprivileged is ongoing [28].

With respect to quality of care (door-to-balloon-times, PCI rates and medication) as well as with respect to short-term outcome there were no significant differences between the four SES groups. These results are in contrast to findings from Finland and the US, where low-income pts. received fewer therapy with antiplatelets, beta-blockers, lipid-lowering medication [13, 29] and reperfusion therapy [13]. In addition obtaining similar invasive procedures in treatment of STEMI was unlikely for ethnic minority members in the US, where racial disparities led to lower rates of primary PCI and longer door-to-balloon times in Afro-Americans [30, 31]. Alter et al. reported a negative association of SES and access to cardiac services, leading to elevated long -term mortality for patients from low SES areas even for Canada with its universal health care system [32]. Differences in findings about treatment strategies within different countries might still be attributed to altering health care systems. The fact, that there is no striking inequity in acute therapy and medication between the social groups in Bremen suggests, that the German health care system provides a guideline conform treatment for STEMI pts. independent of their social background.

Data from the BSR showed that living in low SES areas was associated with more severe myocardial infarctions, measured by the peak creatine kinase (CK) and higher rates of severe impairment of left ventricular ejection fraction post-STEMI. While Gerber et al. [33] showed higher rates of infarction-associated left-heart-failure for patients with low SES the impact of socioeconomic status on extent of MI has, to our knowledge, not been described before. However our data does not offer an explanation for the greater extent of MIs for the socially deprived since pre- as well as in-hospital-treatment modalities were similar across all SES-groups. While the rate of subacute MIs was similar in all SES-groups it however cannot be excluded, that the pain-to-alarm-interval, which is not coherently registered in this database, might have been longer for patients from socially deprived neighbourhoods.

Although acute treatment strategies were comparable between the different SES groups in the BSR, 5-year-MACCE-rates were significantly higher for STEMI-patients from low SES city areas. Stratification for age revealed, that the worse prognosis after STEMI was evident for young-STEMI patients showing a steep elevation in 1-year-mortality and 5-year mortality and MACCE-rates. This disadvantage could no longer be shown for elder patients. Similar findings with increased long-term mortality after AMI in pts. with lower SES have been described by prior studies [9, 32‐38]. Winkleby et al. [38] demonstrated that 1-year case fatality from CAD was 1.6 times higher for women and 1.7 times higher for men in high versus low deprivation neighbourhoods. In Israel low income was shown to be associated with an increase in mortality risk when accompanied by low education among STEMI pts. (HR 2.64, 95% CI: 1.92–3.63) [34]. This disadvantage for STEMI-patients from underprivileged city areas regarding long-term outcome may be the result of poor treatment compliance especially in young STEMI patients, as previously low SES was found to be a predictor for non-adherence to post-discharge medication and lifestyle recommendations [39, 40]. Patients from low SES city areas were more likely to smoke at the index event. Since previous investigations have shown that the rate of smoking cessation after AMI is lower in the less educated [40], higher rates of remaining smokers even after STEMI should be assumed among the socially deprived patients contributing to their worse long-term-prognosis.

In total the elevated STEMI incidence and the higher long-term mortality in low SES city areas of Bremen may indicate a lack of efficacy of primary and secondary prevention strategies in those city districts. Consequently prevention programs should focus on socially disadvantaged neighbourhoods to further improve primary and secondary prevention strategies.

Limited data on long-term medical adherence and risk-factor-adjustment of pts. after STEMI was available in the BSR, so that further studies are required to complete knowledge about adherence to secondary prevention strategies.

Due to the catchment area of the Bremen heart center, we are limited to a relative small study population, but on the other hand for the same reason we can give a comprehensive view of the clinical care of STEMI in Bremen, which minimizes bias due to regional differences in treatment strategies. Although the Bremen heart center is the only regional provider for emergency PCI in case of STEMI, we cannot rule out that single cases of STEMI were treated elsewhere, were not presented to receive emergency-PCI or died before admission to the cath-lab. This might have generated a selection bias.

Our database did not provide individual level SES. The analysis therefore relied on area level SES. This and the pooling from city district-SES to postal-code district SES to ultimately 4 SES-groups probably weakened the effect of social inequalities. However both, area level SES and individual level SES are valid for investigating the impact of socioeconomic status, only that the effects of area level SES are supposed to be weaker. Despite this, our study demonstrates a clear effect of SES on incidence and long-term-outcome, which might have been even more prominent if data on individual level SES had been available.