Summary
Background
While the death rate from acute coronary syndromes (ACS) has been in decline for more than 50 years, out-of-hospital mortality remains high despite improvements in care.
Aim
To evaluate the importance of out-of-hospital mortality and identify the main predictors of in-hospital and 1-year mortality in France.
Methods
Analyses were based on data from the French MONICA population-based registry, which included all cases of ACS occurring in people aged 35–74 years during 2006 in three geographic areas in France. We first evaluated out-of-hospital mortality; then, using data from patients with incident ACS who reached hospital alive, Cox models were performed to determine the main predictors of 1-year mortality. The number of attributable deaths was assessed for variables of interest.
Results
After 1-year follow-up, case-fatality was 29.3% for incident events ( n = 2547); the proportion of out-of-hospital deaths was 70.3%, and 91.5% of deaths occurred in the 28 days following the ACS. On multivariable analysis, the number of attributable deaths associated with three scenarios (out-of-hospital life-and-death emergency, hospitalization before ACS occurrence, and lack of coronary angiography) was 130 (accounting for 59% of deaths occurring after reaching the hospital) during 1-year follow-up. These scenarios corresponded to patients with an initial severe clinical presentation in whom rates of use of specific treatments and invasive procedures were very low.
Conclusion
A large proportion of fatalities after an ACS occurs in the out-of-hospital phase. Moreover, the major component of 1-year mortality is associated with a poor prognosis at initial presentation. This finding highlights the importance of cardiovascular prevention, population education and better out-of-hospital emergency management in improving prognosis after an ACS.
Résumé
Rationnel
La mortalité des syndromes coronariens aigus diminue depuis 50 ans. La mortalité extrahospitalière reste élevée malgré les améliorations de soins.
Objectif
Évaluer l’importance de la mortalité extrahospitalière et les principaux facteurs prédictifs de la mortalité à un an en France.
Méthodes
Données de l’année 2006 des registres français MONICA incluant exhaustivement tous les syndromes coronariens aigus entre 35 et 74 ans dans trois bassins géographiques. La mortalité extrahospitalière a été évaluée, les facteurs prédictifs de la mortalité à un an des SCA inauguraux hospitalisés ont été analysés par modèles de Cox. Le nombre de décès attribuables a ensuite été calculé pour des variables d’intérêt.
Résultats
La létalité à un an des 2547 épisodes inauguraux était de 29,3 %. La proportion de décès extrahospitaliers était de 70,3 % ; 91,5 % des décès survenant dans les 28 jours suivant l’hospitalisation. Le nombre de décès attribuables à trois situations identifiées à partir de l’analyse multivariée (menaces vitales pré-hospitalières, sujets déjà hospitalisés et patients sans coronarographie) était de 130 à un an, soit 59 % des décès survenant chez les sujets arrivés vivants à l’hôpital. Ces situations correspondaient à des patients déjà très graves initialement et ne bénéficiant pas des traitements habituels.
Conclusion
La majorité des décès surviennent à la phase pré-hospitalière. De plus, la majorité des décès à un an est associée à un tableau déjà grave avant la médicalisation. Cela démontre l’importance de la prévention, de l’éducation de la population et d’une amélioration de la prise en charge pré-hospitalière.
Introduction
Management of acute coronary syndromes (ACS) has improved greatly over the past few decades, with important consequences in terms of the epidemiology of coronary heart disease. Mortality is determined by incidence (incident events or recurrences) and by lethality (at the acute phase and over the long term). The mortality rate after an ACS has been in decline for more than 50 years throughout the world, but some geographic differences have been reported . In the United States, the rate of death from coronary heart disease decreased by 59% between 1950 and 1999 . This improvement was associated primarily with a decrease in the case-fatality of acute events, and in a more inconsistent way with a decrease in incidence of ACS . However, recent data from France suggest that overall mortality has decreased slowly, mainly due to a decrease in incidence of coronary heart disease, whereas short-term case-fatality appears to have plateaued . This pattern may reflect the impact of deaths occurring outside of hospital , which appear to be almost non-modifiable.
The aim of our study was first to estimate and describe the distribution of in-hospital and out-of hospital deaths following ACS in a population-based registry, and then to identify the main predictors of mortality occurring during the out-of-hospital phase, and post-discharge up to 1 year.
Methods
Population and registry
Our work was based on 2006 data from the French MONICA registries, in which information was collected in all cases of ACS occurring in people aged 35–74 years in three areas in North, North-East and South-West France. These exhaustive, population-based registries collected data from the case files of all patients hospitalized for an ACS (multiple sources were cross-checked), but also from all patients who died out-of-hospital from a suspected ACS, in which case the general practitioner or the physician who notified the death was interviewed . Out-of-hospital deaths corresponded to patients dying after an out-of-hospital ACS diagnosis (five of the incident events), those dying after symptoms suggestive of an ACS (67 of the incident events), sudden deaths occurring within 24 hours of the index event (422 of the incident events), and other deaths with insufficient data but without any other suspected cause or medical history (30 of the incident events).
We first analysed the data collected over 1 year from 4220 ACS patients to estimate and describe the distribution of in-hospital and out-of hospital mortality. We then conducted an analysis to identify predictors of 1-year mortality using data from the 2023 patients with an incident event who reached hospital alive ( Fig. 1 ). The rate of complete follow-up (follow-up to ≥ 1 year or death during follow-up) among these 2023 patients was 95.9%.
Determinants and outcomes
The main outcome was mortality at 1 year after the index event. Out-of-hospital deaths were recorded from the patient’s medical history, but only age, gender and previous coronary heart disease were recorded for those who died before reaching the hospital.
More data were collected for hospitalized patients, including age, gender and chronic treatments for cardiovascular diseases (anticoagulants, medications for diabetes, antihypertensive drugs, lipid-lowering drugs, antiarrhythmic drugs and antiplatelets).
The type of ACS was determined by troponin concentration and electrocardiographic data, and was categorized as unstable angina or non-ST-elevation myocardial infarction (UA/NSTEMI), ST-elevation myocardial infarction (STEMI) and unclassified (non-interpretable electrocardiogram; corresponding to 4% of cases). Early complications were described as acute heart failure (dyspnoea, acute pulmonary oedema and out-of-hospital diuretic use) and life-threatening complications (resuscitated cardiac arrests, shocks and out-of-hospital use of amines). Life-threatening complications were divided into those occurring out-of-hospital (associated with out-of-hospital explicit treatments for cardiac arrest or shock: xylocaine, amines, sodium bicarbonate or external electrical defibrillation) and those occurring in hospital (all others; defined only for patients reaching the hospital alive). Coronary anatomical status was categorized as one-vessel disease, multivessel disease, left main coronary artery lesion and unknown.
Out-of-hospital management was described in terms of delays between symptom onset and hospitalization (patients already hospitalized before the event were considered as a specific category), management in a mobile intensive care unit and out-of-hospital specific treatment (e.g. aspirin, clopidogrel or anticoagulant).
In-hospital management was described in terms of prescription of recommended treatments (these were combined into a single variable corresponding to the prescription of dual antiplatelet therapy, anticoagulants, beta-blockers, statins and angiotensin-converting enzyme [ACE] inhibitors), and by revascularization strategy (no invasive exploration, coronary angiography without revascularization, percutaneous coronary intervention [PCI] and coronary artery bypass graft [CABG]).
Statistical analyses
The overall population was described according to case-fatality, corresponding to the proportion of deaths in the study groups (4220 incident and recurrent cases of ACS; 2547 incident ACS and 2023 incident hospitalized ACS). Cox modelling was then use to identify predictors of 1-year mortality. Before introducing covariates in the model, the assumption of proportional risk was checked for all variables and log-linearity was assessed for quantitative variables. Missing data for both the type of ACS and the delay to hospitalization were analysed as a specific category because they were more frequent, but for all other missing data were switched to a negative response. Sensitivity analyses were performed, replacing missing data with positive responses. Predictors of 1-year mortality were identified by creating a model of causation in which explanatory covariates were introduced successively and considered in the model according to their distance from the outcome variable (1-year mortality). More precisely, variables considered as the most distant from the outcome were first analysed (age, gender and previous medical treatments). Then, intermediate and proximal explanatory variables were added to the model in the following order: hospitalization before ACS occurrence, type of ACS, early complications, out-of-hospital management and in-hospital management and complications. This approach is more appropriate than stepwise regression to study determinants that are not equally distant to the outcome in the causal chain. This process was inspired by that used more frequently in social epidemiology . Variables interpreted as clinically significant markers of severity of initial presentation and with a strong statistical weight were analysed in greater detail.
The different subgroups are described using counts and percentages or as medians. The χ 2 test or Fisher’s exact test were used to test for differences in qualitative variables and Student’s t test or the Mann-Whitney test for differences in quantitative variables.
We calculated the number of attributable deaths for each variable of interest with a similar method used for estimating attributable risk . Variables of interest were considered as the ‘exposure’, and death was considered as the ‘disease’. The population-attributable fraction for each exposure was estimated as:
p × a d j u s t e d H a z a r d r a t i o − 1 / a d j u s t e d H a z a r d r a t i o
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
