https://orcid.org/0000-0002-8844-3230
Figures
Abstract
Introduction
Worldwide, the COVID-19 pandemic has been associated with an overall drop in acute coronary syndrome (ACS) hospitalizations. Additionally, there is a well-known association between ACS and socioeconomic status. This study aims to assess the COVID-19 effect on ACS admissions in France during the first national lockdown and investigate the factors associated with its spatial heterogeneity.
Materials and methods
In this retrospective study, we used the French hospital discharge database (PMSI) to estimate ACS admission rates in all public and private hospitals in 2019 and 2020. A negative binomial regression explored the nationwide change in ACS admissions during lockdown compared with 2019. A multivariate analysis explored the factors associated with the ACS admission incidence rate ratio (IRR, 2020 incidence rate/2019 incidence rate) variation at the county level.
Results
We found a significant but geographically heterogeneous nationwide reduction in ACS admissions during lockdown (IRR 0·70 [0·64–0·76]). After adjustment for cumulative COVID-19 admissions and the ageing index, a higher share of people on short-term working arrangements during lockdown at the county level was associated with a lower IRR, while a higher share of individuals with a high school degree and a higher density of acute care beds were associated with a higher ratio.
Citation: Plat R, Vasile M, Roubille F, Mercier G (2023) Relationships between the COVID-19 lockdown, socioeconomic factors and acute coronary syndrome hospitalisations in France. PLoS ONE 18(6): e0286700. https://doi.org/10.1371/journal.pone.0286700
Editor: Kjell Torén, University of Gothenburg: Goteborgs Universitet, SWEDEN
Received: February 16, 2023; Accepted: May 19, 2023; Published: June 7, 2023
Copyright: © 2023 Plat et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The complete dataset that supports the results cannot be shared publicly as it contains PMSI data. Due to the sensitive nature of the data, access to it is restricted by the french law on health data and the National Commission on Informatics and Liberty. However, a minimal dataset with aggregate data allowing to reproduce the results is available in a supplementary file. We are also willing to share our data in accordance with the law, upon reasonable request.
Funding: Support for this research was provided by the French Ministry of Health (sante.gouv.fr - grant number DGOS covid-19-20-0063). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
The current coronavirus 2019 (COVID-19) pandemic poses unprecedented challenges to health care systems. Transitions in the organisation of hospital services and in the use of health care by the population have had a major impact on hospital activities, particularly in the sectors managing noncommunicable diseases [1]. During the first wave of the pandemic, when hospital overload was at its highest, significant drops in the volume of admissions worldwide ranging from 20.2% to 73% were observed for acute cardiovascular diseases [2]. In France, a 30% decrease in the volume of admissions for myocardial infarction (MI) was reported during the first month of the national lockdown, which was implemented from 17 March to 11 May 2020 [3].
However, there are notable territorial disparities in the burden of acute coronary syndromes (ACSs) in France as well as in the impact of the epidemic on the health care system [4,5]. Moreover, the crisis has had a differentiated impact on individuals according to their socioeconomic status, affecting the most disadvantaged individuals more frequently and more severely [6,7]. Additionally, the independent link between social determinants and cardiovascular diseases is well established in developed countries [8–10]. These worrying findings highlight the collateral effects of the pandemic on hospital admissions for ACS and the role that health inequalities may have had. A better understanding of these events is essential to better anticipate them, enabling our health systems to adapt faster to future health crises and ensure the safety of individuals.
Objectives
We hypothesise that the first lockdown reduced hospital admissions for ACS in France and that this effect was partly dependent on contextual factors related to social inequalities in health. Our objectives were, first, to quantify the trend of admissions in 2019 and 2020 and more specifically during the first national lockdown period in comparison with the same period in 2019 and, second, to investigate the contribution of contextual factors to the variation of these admissions during this period.
Materials and methods
Study design and population
In this French retrospective nationwide study, we used data from the French hospital discharge database (PMSI) to estimate weekly rates of admission for ACS. The PMSI is a national database that includes standardised medical information on hospital admissions, such as diagnoses, medical procedures and health care use. We included every admission for adult patients to any public or private hospital for ACS in metropolitan France from 1 January 2019 to 31 December 2020.
Procedures
ACS admissions included hospitalisations for unstable angina (ICD-10 code I.20), ST segment elevation myocardial infarction (STEMI, ICD-10 code I.21 except I21.4), non–ST-segment elevation myocardial infarction (NSTEMI, ICD-10 code I.21.4), and recurrent myocardial infarction (ICD-10 code I.22) [11]. To take into account some iterative revisions of the PMSI coding rules published during the crisis and the coding errors that may have resulted, especially in the case of hospitalisation for an ACS associated with a COVID-19 infection, we also included stays with an associated secondary diagnosis code related to COVID-19 (ICD-10 codes U.07.10 to U.07.15) if the ACS ICD code was found at the onset of the hospitalisation. To avoid iatrogenic events and MI occurring during hospitalisation, we excluded patients with a main diagnosis of ACS transferred from another hospital or department (Fig 1).
To investigate the contribution of the factors involved in the variation of admissions during the first lockdown, the multivariate analyses used environmental data aggregated at the county level. The socioeconomic level was measured by the average income, education level, labourers and unemployment rates at the county level provided by the National Institute of Statistics and Economic Studies (INSEE) [12]. Social isolation was measured by the share of single-person households [12]. Socioeconomic determinants also included data on living conditions during the lockdown (decline of financial situation, short time working, teleworking) from the national Epicov study [13].
The supply of inpatient and outpatient care was measured by the density of general practitioners and the density of acute care beds in medicine. Hospital accessibility was also measured by the average travel time between the ZIP code of residence and the ZIP code of the nearest hospital for the patients included in this study, available in the PMSI. To adjust for the local burden of the epidemic, we used the cumulative incidence of admissions for COVID-19 during the first lockdown at the county level. We also included the precrisis global and cardiovascular health condition of populations with the ageing index (i.e., the ratio of the population aged 65 and over to the population under 20), the standardised mortality rate of people aged over 65 years, and the standardised rate of people hospitalised for ACS in 2018 at the county level.
Statistical analysis
The weekly incidences of admissions for the different types of ACS, as well as for hospital mortality, were estimated at the national level and over the years 2019 and 2020. We conducted a univariate analysis to quantify the first lockdown’s effect on admissions for all types of ACS. Their national cumulative incidence were compared over identical periods, including the first French national lockdown, corresponding to weeks 12 to 19 of 2019 and 2020 (i.e., 18 March to 12 May 2019 and 16 March to 10 May 2020). To address the overdispersion found in the data, a negative binomial regression was used to assess the change in the number of cumulative admissions for the different types of ACS over these two periods (except for recurrent MI, for which there was insufficient data). This regression quantifies the impact of the first confinement on ACS admissions, which is represented by a national cumulative incidence rate ratio (IRR) [3]:
Using ZIP codes attached to each stay in the PMSI, the IRR for any type of ACS was mapped at the county level, leading to a local cumulative incidence ratio that allowed us to report the impact of lockdown on ACS hospitalisations in each county:
Using linear regression, we conducted univariate and multivariate analyses using counties as statistical units to explore the factors associated with the variation in the local cumulative IRR, which was our dependant variable. Each county-level aggregate environmental data listed in the procedure section was tested univariately with our dependant variable. Multi-collinearity issues were investigated, and each variable associated with the local IRR with a p-value < 0.20 was inserted into the multivariate model. Of the included variables that were no longer significant in the multivariate analysis, the ageing index and the cumulative county incidence of admissions for COVID-19 during lockdown were maintained in the model as they are adjustment variables, reflecting the local burden of COVID-19, as well as the overall and cardiovascular health condition of populations prior to the crisis. Descriptive analyses were performed, reporting continuous data as mean values with standard deviation (SD) and median values with range (quartile 1 and quartile 3).
As the data were spatially distributed, local and global spatial dependence phenomena were screened using Moran’s I and LISA tests. Residual spatial interactions remaining in the linear regression model were searched using Moran’s I test on residuals. Given the persistence of spatial autocorrelation in the errors of the model, a spatial error model (SEM) using queen contiguity was estimated [14]. Its validity was verified by the Hausman test. The analyses were carried out using R studio v4.0.3.
Ethics statement
This study was approved by the Institutional Review Board of the Montpellier university hospital (IORG number 0009525—I.R.B approval number: 202000600). Since all patient-level data in the PMSI database are anonymized before third parties are given access, informed consent at an individual patient level was not required. The study posed no potential risks to individuals or individual privacy, and was conducted in accordance with relevant international and French regulatory requirements.
Results
Admissions and in-hospital mortality rates for ACS in 2019 and 2020
Fig 2 shows the annual trend of the respective incidences measured. A total of 304,467 stays were included, of which 42,706 occurred during the periods of interest.
During 2019, the average weekly incidence of admissions in France for all types of ACS was 4.68/100,000 inhabitants; 2.03/100,000 inhabitants for unstable angina; 1.30, 1.26 and 0.10/100,000 inhabitants for STEMI, NSTEMI and recurrent MI, respectively; and 0.13/100,000 inhabitants for in-hospital mortality. In 2020, they were equal to 4.34/100,000 inhabitants for all types of ACS; 1.79/100,000 inhabitants for unstable angina; 1.25, 1.22 and 0.09/100,000 inhabitants for STEMI, NSTEMI and recurrent MI, respectively; and 0.12/100,000 inhabitants for in-hospital mortality.
For all types of ACS, weekly admissions peaked in early and late 2019 and 2020, with weekly of approximately 5.0/100,000 inhabitants. During the first lockdown, the incidence ranged from 2.89 to 3.92 hospitalisations/100,000 inhabitants, while the corresponding period in 2019 recorded an incidence ranging from 4.55 to 5.11 hospitalisations/100,000 inhabitants. In the latter part of 2020, the weekly incidence remained, on average, lower than those observed in 2019. These trends are found for all types of ACS, except for recurrent MI, as well as for in-hospital mortality, which appears to be stable overall during 2019 and 2020. The data additionally highlights the presence of a second drop in incidences from weeks 25 to 40 in 2019 and 2020, which could be linked to the well-known summer decrease in incident cases of ACS [15].
The first national lockdown impact
In France, there was a significant drop in admissions for all types of ACS during the lockdown, accounting for 30% of the volume of hospitalisations recorded over the same time period in 2019 (IRR 0.70 [0.64–0.76], p value < 0.0001). This decline was most noticeable for unstable angina hospitalisations, with a 39% reduction in the number of admissions in 2020 (IRR 0.61 [0.54–0.70], p value < 0.0001). For myocardial infarction, hospitalisations were more modestly affected, with a similar decrease in volumes for NSTEMI (IRR 0.79 [0.71–0.88], p value < 0.0001) and STEMI (IRR 0.80 [0.74–0.86], p value < 0.0001). In-hospital mortality did not vary significantly during the first lockdown (IRR 0.91 [0.79–1.06], p value 0.22).
Fig 3 shows the county variation of the national admission IRR for all types of ACS, depicting variation in volumes from a decrease of 39% to an increase of 61%. Despite a national trend of a 30% drop in the total number of admissions, four counties recorded a higher number of admissions for ACS in 2020. In contrast, fourteen departments were the most severely affected by the crisis, with relative decreases exceeding 40% over the course of the lockdown.
This figure displays the values of the county cumulative admission Incidence Rate Ratio (Local IRR, which is the ratio of the local cumulative incidence for all types of ACS admissions between weeks 12 and 19 of 2020, to the same cumulative incidence in 2019) for all types of ACS (Background map source: IGN 2016 –Etalab Open license).
Moran’s I test revealed a positive spatial autocorrelation of county IRRs (Moran index 0.24, p value < 0.001). In addition, the local spatial autocorrelation tests LISA found statistically significant clusters of greater decrease or increase in hospital admissions for ACS in 2020. Fig 4 shows their distribution.
This figure highlights the countries for which LISA local spatial autocorrelation tests revealed statistically significant clusters within the local IRRs. Significant LISA clusters are combinations of similar or dissimilar values more marked than what might have been observed based on random spatial distribution. These groupings can match up with four types of spatial groupings (high-high, low-low, high-low or low-low). The left panel displays the type of the LISA test clusters. The right panel shows the significance (p-value) of the LISA test clusters. (Background map source: IGN 2016 –Etalab Open license).
Multivariate analysis
Table 1 presents the mean value and distribution by county of the variables used in this study. Each variable univariably associated with the variation of the IRR with a p value < 0.20 was inserted in the multivariate model. No interactions for the inserted variables were found. The assumptions of the linear model were respected, but a spatial interaction phenomenon was detected in the residuals of the linear model (Moran’s test on the residuals, p value 0.001). The use of an SEM was considered valid (Hausman test p value 0.93) [14].
Tables 2 and 3 presents the results of the univariate and multivariate analyses. At the county level, the regression models indicate an association between, on the one hand, the rate of individuals whose highest degree is a high school diploma (p 0.003) and the density of acute care beds in medicine (p 0.013) and, on the other hand, the positive variation in the IRR of admissions for ACS. Thus, for each 1% increase in the high school graduate rate, the IRR increased by 4.1%, while the provision of one additional acute care bed per 100,000 inhabitants was associated with a 0.07% ratio increase. Conversely, we found an association between, on the one hand, the rate of individuals who were on short time working during the crisis and, on the other hand, the negative variation in the IRR of hospitalisations for ACS (p 0.008). As a result, for each 1% increase in the short-term working individual rate, the IRR decreased by 1.5%. These results are adjusted by the cumulative incidences of hospital admissions for COVID-19 and the ageing index.
Discussion
Admission trends in 2019–2020 and the impact of the lockdown
During the first lockdown, our study highlights a substantial drop for all types of ACS admissions in France. This decrease was heterogeneous over the French metropolitan territory and seemed to intensify following a line connecting southwestern France to northeastern France. It was also more pronounced for given types of ACS, such as unstable angina, and in the first half of the lockdown, as the rates then started to rise again within its second half. As the incidences of weekly hospitalisations registered in the latter part of 2020 did not exceed those of 2019, the curves do not highlight the presence of a compensatory rebound in hospitalisations.
Worldwide, several studies have explored the decrease in ACS admissions during the lockdown and have suggested possible hypothesis and explanations. The first hypothesis is that many ACS did not reach the hospital, a first explanation being that many patients may have forgone hospital care for various reasons, including the fear of being infected by COVID-19 at the hospital, the concern about overloading the health services, or the willingness to avoid long waiting times. In a crisis context where the messages spread by the authorities urged people to stay home [2]. In France, during the first month of lockdown, this hypothesis can be supported by a larger reduction in hospital admissions in elderly patients or patients with STEMI; the first being fragile and therefore cautious, the second experiencing less severe symptoms and therefore being more likely to minimise them [3]. Our data do not highlight differences in this reduction according to the type of myocardial infarction, this difference possibly having been rebalanced during the catch-up phenomenon observed in the second half of the lockdown. However our study reveals a larger reduction in hospitalisations for unstable angina, which is the least severe and least symptomatic ACS. This could be related to a phenomenon of avoidance of care, which is easier in the case of a minor ACS, than in the case of a severe presentation. A second explanation for this decrease in admissions could be linked to the overload and disorganisation of the health care system, leading to dysfunctions, particularly in the prehospital stage, and missed opportunities to provide care. For example, in Gironde county, emergency calls for chest pain increased fivefold during the lockdown, while our data revealed a 26% drop in admissions for ACS in this county [16]. This is echoed in one region of England where, although admissions for ACS were also significantly impacted, emergency calls for MI remained unchanged [17]. A French study also reported a doubling of the average time between the onset of ACS symptoms and the first medical contact, with similar findings in other countries [2,18]. Overall, whether related to care avoidance events or pre-hospital care overload issues, some circumstances may have led to a substantial decrease in ACS admissions during the lockdown period. Thus, a decline in admission does not automatically imply a decrease in ACS occurrence, and it can be hypothesised that a share of the missing patients in hospitals experienced an ACS, but were not willing or able to reach the hospital and receive the needed care for their condition.
The second hypothesis to explain the decrease in ACS admissions is that there was a real decrease in the incidence of ACS during lockdown [19]. The absence of a rebound in hospitalisations following the lockdown seems to support this hypothesis. This could be the consequence of a significant change in living conditions and behaviours influencing cardiovascular risk (e.g., professional and physical activity, diet and consumption, sleep, anxiety); or of a change in environmental conditions with, notably, a reduction in air pollution connected with the reduction in activity during confinement. Indeed, air pollution is a recognized risk factor and a known trigger for ACS, and numerous studies highlights the relationship between transient elevations in ambient air pollutants and an increased risk of ischemic myocardial infarction [20]. In addition, the concentration of most air pollutants decreased transiently during the lockdown in France, to a greater or lesser extent depending on the geographical area and its baseline pollution level [21]. But interestingly, a study specifically investigating the relation between the hospital management of ACS during the lockdown and air pollutants found no association between an short term air cleaning effect and a concomitant reduction in invasive procedures for ACS [22]. These facts raise questions about the impact of a transient and geographically heterogeneous decrease in air pollutants on the occurrence of ACS, and how it may have interacted with the other factors considered. Finally, a last assumption to explain a potential decrease in the incidence of ACS would be the existence of a competitive risk on mortality between COVID-19 and MI [23].
Regarding the impact of the crisis on ACS management once patients were admitted to an hospital, our study found no significant variation in inpatient mortality during the lockdown.
The role of determinants linked to social and territorial inequalities in health
Regarding the determinants of this decrease in admissions, our results suggest that, at the county level, a higher share of individuals on short-term working arrangements resulted in a larger decrease in admissions for ACS in 2020, i.e., a larger impact of the first lockdown on admissions. Some employment categories were more directly affected by part-time working than others; 24% of labourers, 18% of employees and 20% of shop keepers became part-time workers during the first lockdown, while the national mean was 15% [13]. Since we did not find other associations between the IRR variation and other determinants linked to occupational status or financial condition, we hypothesise that this result reflects a lower incidence of ACS among people placed on part-time working during the first lockdown. This could be explained first by the eviction from the workplace, as several studies documented the independent relationship between a high-demand occupation coupled with a low level of autonomy and the occurrence of MI [24,25]. In addition, intensive outdoor physical activity is associated with an increased risk of atheromatous plaque rupture leading to ACS [26]. A prolonged eviction from this environment could, thus, have reduced the excess risk connected with this exposure, as well as exposures to other occupational risks impacting health (e.g., tedious work, nightshifts, toxic substances). Once at home, partially employed individuals could have seized this opportunity to modify their lifestyles and improve their overall cardiovascular risk. However, this second interpretation must be taken with caution, as evolutions in modifiable cardiovascular risk factors during the lockdown have been observed in both favourable and unfavourable directions [27,28]. In contrast, at the county level, a high share of individuals with a high school degree increased ACS admissions in 2020, thus reducing the impact of confinement. This outcome can also be linked to occupation-related exposure at the time of the lockdown. In effect, 62.8% of individuals with this degree are working as employees or in intermediate occupations [29]. During the first lockdown, a majority of individuals in intermediate occupations continued to work for the most part at the worksite, as teleworking was uncommon. This was also the case for 73% of the employees who kept working [30]. These individuals thus continued to be exposed to occupational risks on cardiovascular health; in working conditions that likely worsened, increasing the burden of these exposures. In addition, a French study found an association between a job involving contact with the public, which is fairly common in these professions, and a deterioration of cardiovascular risk during the lockdown [31].
The last association in our study indicates that, at the county level, a higher density of acute care hospital equipment increased ACS admissions in 2020, thus reducing the impact of the first lockdown. Since our model is adjusted for the cumulative incidence of hospitalisations for COVID-19, this finding may reflect the effect of the resulting hospital overload. Indeed, counties with a lower level of acute care hospital resources were more likely to be overloaded with an equivalent number of COVID-19 patients, increasing dysfunctions in the prehospital care phase and complicating the admission of patients with ACS to the hospital. In a given county, greater hospital pressure may have also discouraged some patients from seeking care if they experienced ACS symptoms. This could have had the effect of exacerbating the foregoing of care, which was preexisting and probably more common in these territories due to the poorer availability of care.
Strengths and limits
This study has various strengths. First, it includes data from all public and private hospitals in metropolitan France before, during and after the lockdown. To our knowledge, this is the first study to examine the socioeconomic and health care supply-related determinants of the impact of lockdown on ACS admissions in France. The seasonality of ACS was considered, and spatial approaches allowed us to capture the processes involved more accurately. Our study also presents limitations. First, the study is a retrospective, descriptive, observational study of hospitalised ACS only, with a comparison to the year 2019 only. Prehospital care pathways and individual patient characteristics were not investigated, and the multivariate analysis of the impact of the health crisis focused only on the first lockdown. Another limitation is that the inpatient admission data may present coding heterogeneities according to the practices of the hospitals, inducing potential biases [11]. Finally, as the data are aggregated at the county level, the results are subject to the modifiable area unit problem of scale and their interpretation to an ecological bias [14,30].
Conclusions
This study highlights the significant impact of lockdown on ACS admissions in France, as well as the spatial variation of this impact. This variation was explained by socioeconomic and health care determinants, highlighting the impact of occupational status and hospital care supply on the incidence of ACS admissions during the lockdown. This is likely to be the result of a real change in the incidence of ACS, as well as a decrease in the number of patients reaching the hospital, due to dysfunctions in prehospital care and the fact that some patients did not seek care. Consequently, we believe that clear and targeted messages from health authorities to high-risk communities, urging them to seek care when necessary, and increased attention to the integrity of all stages of the management of emergency care, will be essential to ensure that the health of individuals is better advised and protected and to limit the collateral damages of potential future public health crises.
Supporting information
S1 Dataset. Minimal dataset.
Aggregate data used to perform figures and statistical analysis.
https://doi.org/10.1371/journal.pone.0286700.s001
(XLSX)
References
- 1. Modesti PA, Wang J, Damasceno A, Agyemang C, Van Bortel L, Persu A, et al. Indirect implications of COVID-19 prevention strategies on non-communicable diseases. BMC Med. 2020 Aug 14;18(1):256.
- 2. Kiss P, Carcel C, Hockham C, Peters SAE. The impact of the COVID-19 pandemic on the care and management of patients with acute cardiovascular disease: a systematic review. Eur Heart J Qual Care Clin Outcomes. 2021 Jan 25;7(1):18–27. pmid:33151274
- 3. Mesnier J, Cottin Y, Coste P, Ferrari E, Schiele F, Lemesle G, et al. Hospital admissions for acute myocardial infarction before and after lockdown according to regional prevalence of COVID-19 and patient profile in France: a registry study. Lancet Public Health. 2020 Oct 1;5(10):e536–42. pmid:32950075
- 4.
Santos F, Gabet A, Carcaillon-Bentata L, Grave C, Olié V. Departmental disparities in potential years of life lost due to premature cardiovascular diseases in France (2013–2015) [Internet]. [cited 2021 Aug 19]. https://www.santepubliquefrance.fr/import/disparites-departementales-d-annees-potentielles-de-vie-perdues-prematurement-par-maladies-cardiovasculaires-en-france-2013-2015.
- 5. Gaudart J, Landier J, Huiart L, Legendre E, Lehot L, Bendiane MK, et al. Factors associated with the spatial heterogeneity of the first wave of COVID-19 in France: a nationwide geo-epidemiological study. Lancet Public Health. 2021 Apr 1;6(4):e222–31. pmid:33556327
- 6. Hawkins RB, Charles EJ, Mehaffey JH. Socio-economic status and COVID-19–related cases and fatalities. Public Health. 2020 Dec;189:129–34. pmid:33227595
- 7. Ginsburgh V, Magerman G, Natali I. COVID-19 and the role of inequality in French regional departments. Eur J Health Econ. 2021 Mar;22(2):311–27. pmid:33387139
- 8. Rosengren A, Smyth A, Rangarajan S, Ramasundarahettige C, Bangdiwala SI, AlHabib KF, et al. Socioeconomic status and risk of cardiovascular disease in 20 low-income, middle-income, and high-income countries: the Prospective Urban Rural Epidemiologic (PURE) study. Lancet Glob Health. 2019 Jun;7(6):e748–60. pmid:31028013
- 9. Manrique-Garcia E, Sidorchuk A, Hallqvist J, Moradi T. Socioeconomic position and incidence of acute myocardial infarction: a meta-analysis. J Epidemiol Community Health. 2011 Apr 1;65(4):301–9. pmid:20841371
- 10. Lecoffre C, Decool E, Olié V. Hospitalizations for neuro-cardiovascular diseases and social deprivation in France in 2013. 2016;8.
- 11. Bezin J, Girodet P-O, Rambelomanana S, Touya M, Ferreira P, Gilleron V, et al. Choice of ICD-10 codes for the identification of acute coronary syndrome in the French hospitalization database. Fundam Clin Pharmacol. 2015 Dec;29(6):586–91. pmid:26301735
- 12.
INSEE. Local statistics [Internet]. [cited 2021 Aug 23]. https://statistiques-locales.insee.fr/#c=home.
- 13.
Givord P, Silhol J. INSEE Première—Lockdown: unequal economic consequences for households [Internet]. [cited 2021 Aug 20]. https://www.insee.fr/fr/statistiques/4801313.
- 14.
INSEE. Handbook of Spatial Analysis. Theory and Application with R [Internet]. 2018 [cited 2022 Mar 30]. https://www.insee.fr/en/information/3635545.
- 15. Spencer FA, Goldberg RJ, Becker RC, Gore JM. Seasonal distribution of acute myocardial infarction in the second National Registry of Myocardial Infarction. J Am Coll Cardiol. 1998 May;31(6):1226–33. pmid:9581712
- 16. Gil-Jardiné C, Chenais G, Pradeau C, Tentillier E, Revel P, Combes X, et al. Trends in reasons for emergency calls during the COVID-19 crisis in the department of Gironde, France using artificial neural network for natural language classification. Scand J Trauma Resusc Emerg Med. 2021 Mar 31;29(1):55. pmid:33789721
- 17. Mafham MM, Spata E, Goldacre R, Gair D, Curnow P, Bray M, et al. COVID-19 pandemic and admission rates for and management of acute coronary syndromes in England. Lancet Lond Engl. 2020;396(10248):381–9. pmid:32679111
- 18. Rangé G, Hakim R, Beygui F, Angoulvant D, Marcollet P, Godin M, et al. Incidence, delays, and outcomes of STEMI during COVID‐19 outbreak: Analysis from the France PCI registry. J Am Coll Emerg Physicians Open. 2020 Nov 23;1(6):1168–76. pmid:33363285
- 19. Huet F, Prieur C, Schurtz G, Gerbaud E, Manzo-Silberman S, Vanzetto G, et al. One train may hide another: Acute cardiovascular diseases could be neglected because of the COVID-19 pandemic. Arch Cardiovasc Dis. 2020 May;113(5):303–7. pmid:32362433
- 20. Claeys MJ, Rajagopalan S, Nawrot TS, Brook RD. Climate and environmental triggers of acute myocardial infarction. Eur Heart J. 2017 Apr 1;38(13):955–60. pmid:27106953
- 21. Ikhlasse H., Benjamin D., Vincent C. et al. Environmental impacts of pre/during and post-lockdown periods on prominent air pollutants in France. Environ Dev Sustain 23, 14140–14161 (2021). pmid:33519298
- 22. Arrivi A, Dominici M, Bier N, Truglio M, Vaudo G, Pucci G. Association Between Air Pollution and Acute Coronary Syndromes During Lockdown for COVID-19: Results From the Terni Hub Center. Front Public Health. 2021 Jun 24;9:683683. pmid:34249847.
- 23. Lantelme P, Couray Targe S, Metral P, Bochaton T, Ranc S, Le Bourhis Zaimi M, et al. Worrying decrease in hospital admissions for myocardial infarction during the COVID-19 pandemic. Arch Cardiovasc Dis. 2020 Jul;113(6–7):443–7. pmid:32636131
- 24. Netterstrom B, Nielsen FE, Kristensen TS, Bach E, Moller L. Relation between job strain and myocardial infarction: a case-control study. Occup Environ Med. 1999 May;56(5):339–42. pmid:10472309
- 25. Kornitzer M, deSmet P, Sans S, Dramaix M, Boulenguez C, DeBacker G, et al. Job stress and major coronary events: results from the Job Stress, Absenteeism and Coronary Heart Disease in Europe study. Eur J Cardiovasc Prev Rehabil. 2006 Oct 1;13(5):695–704. pmid:17001207
- 26. Kato A, Minami Y, Katsura A, Muramatsu Y, Sato T, Kakizaki R, et al. Physical exertion as a trigger of acute coronary syndrome caused by plaque erosion. J Thromb Thrombolysis. 2020 Apr;49(3):377–85. pmid:32144654
- 27. Guignard R, Andler R, Quatremère G, Pasquereau A, du Roscoät E, Arwidson P, et al. Changes in smoking and alcohol consumption during COVID-19-related lockdown: A cross-sectional study in France. Eur J Public Health. 2021 Apr 7;ckab054. pmid:33826721
- 28. Deschasaux-Tanguy M, Druesne-Pecollo N, Esseddik Y, de Edelenyi FS, Allès B, Andreeva VA, et al. Diet and physical activity during the coronavirus disease 2019 (COVID-19) lockdown (March-May 2020): results from the French NutriNet-Santé cohort study. Am J Clin Nutr. 2021 Apr 6;113(4):924–38.
- 29.
INSEE. Fact sheets: The integration of young people, Training and employment [Internet]. 2018 [cited 2022 Mar 30]. https://www.insee.fr/fr/statistiques/3526080?sommaire=3526086.
- 30. Lambert A, Cayouette-Remblière J, Guéraut É, Roux GL, Bonvalet C, Girard V, et al. How the COVID-19 epidemic changed working conditions in France. Popul Soc. 2020 Jun 17;579(7):1–4.
- 31. Bérard E, Huo Yung Kai S, Coley N, Bongard V, Ferrières J. Lockdown-related factors associated with the worsening of cardiovascular risk and anxiety or depression during the COVID-19 pandemic. Prev Med Rep. 2020 Dec 28;21:101300. pmid:33457196