1546

Sudden cardiac arrest (SCA) remains a leading cause of death in the industrialized world, representing ≈20%

of overall mortality.1 Except for a few recent encouraging reports,2–4 survival after SCA remains poor and relatively sta-ble over time, <10%, despite decades of research and major financial investments in resuscitation.5–7

Clinical Perspective on p 1554

Because SCA often presents with an initial ventricular fibrillation,8 external defibrillation has been incorporated as a key factor in the chain of survival since the early 1990s.9,10 Public-access defibrillation programs have thus been devel-oped to make automatic external defibrillators (AEDs)

publicly available, allowing rapid early defibrillation.11–17 Although the benefits of AED are undeniable, public use rates remain very low,4,15,18 and thus the effectiveness of such programs could be dramatically improved. One of the key issues is the disparity between the locations of AEDs and SCAs. In Paris, although AED deployment in the commu-nity, which started in 1993 with the progressive equipping of EMS ambulances, has been particularly active over the last decade, the choice of where to install AEDs in public places has been driven mainly by empirical considerations (local unguided initiatives). Recent studies have already pointed out the limitations of such an approach, calling for an evi-dence-based strategy.19

Background—Although the benefits of automatic external defibrillators are undeniable, their effectiveness could be dramatically improved. One of the key issues is the disparity between the locations of automatic external defibrillators and sudden cardiac arrests (SCAs).

Methods and Results—From emergency medical services and other Parisian agencies, data on all SCAs occurring in public places in Paris, France, were prospectively collected between 2000 and 2010 and recorded using 2020 grid areas. For each area, population density, population movements, and landmarks were analyzed. Of the 4176 SCAs, 1255 (30%) occurred in public areas, with a highly clustered distribution of SCAs, especially in areas containing major train stations (12% of SCAs in 0.75% of the Paris area). The association with population density was poor, with a nonsignificant increase in SCAs with population density (P=0.4). Occurrence of public SCAs was, in contrast, highly associated with population movements (P<0.001). In multivariate analysis including other landmarks in each grid cell in the model and demographic characteristics, population movement remained significantly associated with the occurrence of SCA (odds ratio, 1.48; 95% confidence interval, 1.34–1.63; P<0.0001), as well as grid cells containing train stations (odds ratio, 3.80; 95% confidence interval, 2.66–5.36; P<0.0001).

Conclusions—Using a systematic analysis of determinants of SCA in public places, we demonstrated the extent to which population movements influence SCA distribution. Our findings also suggested that beyond this key risk factor, some areas are dramatically associated with a higher risk of SCA. (Circulation. 2015;131:1546-1554. DOI: 10.1161/CIRCULATIONAHA.114.010498.)

Key Words: cardiopulmonary resuscitation ◼ death, sudden, cardiac ◼ defibrillators ◼ heart arrest

© 2015 American Heart Association, Inc.

Circulation is available at http://circ.ahajournals.org DOI: 10.1161/CIRCULATIONAHA.114.010498

Received April 7, 2014; accepted February 26, 2015.From Université Paris Descartes, France (E.M., W.B., M.T., N.K., L.L., F.B., G.B., A.H., J.-Y.L.H., M.D., F.D., C.S., A.C., X.J.); Paris Cardiovascular

Research Center (PARCC-Inserm U970), France (E.M., W.B., M.T., N.K., L.L., F.B., G.B., F.D., C.S., A.C., X.J.); Hôpital Européen Georges Pompidou, Département de Cardiologie, Paris, France (E.M., N.K., A.H., J.-Y.L.H., M.D., C.S., X.J.); Département de Réanimation Médicale, Hôpital Cochin, Paris, France (W.B., A.C.); Service Médical d’Urgence–Brigade de Sapeurs-Pompiers de Paris, France (D.J., H.D., J.-P.T.); Service d’Aide Médicale Urgente de Paris (SAMU 75), France (L.L.); Atelier d’Urbanisme Parisien, Paris, France (P.P.); Département des Urgences, Hôpital Cochin, Paris, France (F.D.); and Sydney Medical School, Australia (D.S.C.).

*Drs Marijon and Bougouin contributed equally.The online-only Data Supplement is available with this article at http://circ.ahajournals.org/lookup/suppl/doi:10.1161/CIRCULATIONAHA.

114.010498/-/DC1.Correspondence to Eloi Marijon, MD, PhD, European Georges Pompidou Hospital, Cardiology Department, Clinical Electrophysiology Unit, 20-40 Rue

Leblanc, 75908 Paris CEDEX 15, France. E-mail eloi_marijon@yahoo.fr

Population Movement and Sudden Cardiac Arrest LocationEloi Marijon, MD, PhD*; Wulfran Bougouin, MD, MPH*; Muriel Tafflet, PhD;

Nicole Karam, MD, MPH; Daniel Jost, MD; Lionel Lamhaut, MD; Frankie Beganton, MS; Patricia Pelloux, MS; Hervé Degrange, MD; Guillaume Béal, MS; Jean-Pierre Tourtier, MD, PhD;

Albert A. Hagège, MD, PhD; Jean-Yves Le Heuzey, MD; Michel Desnos, MD; Florence Dumas, MD, PhD; Christian Spaulding, MD, PhD; David S. Celermajer, PhD, FRACP;

Alain Cariou, MD, PhD; Xavier Jouven, MD, PhD

by guest on April 15, 2016http://circ.ahajournals.org/Downloaded from by guest on April 15, 2016http://circ.ahajournals.org/Downloaded from by guest on April 15, 2016http://circ.ahajournals.org/Downloaded from by guest on April 15, 2016http://circ.ahajournals.org/Downloaded from

Marijon et al High-Risk Public Areas for Sudden Cardiac Arrest 1547

Although previous studies have identified sites with a higher risk of SCAs,18–22 especially raising the association with train stations and airports, the reliability and determinants of such high-risk places have been only partially explored. In this study, we collected data on SCA locations in Paris. We hypothesized that population movement was a key factor for SCA distribution but could not explain by itself the overall heterogeneity of SCA distribution.

MethodsPopulation StudyFrom January 1, 2000, to July 31, 2010, using the 2 emergency medical services (EMS) systems existing in Paris, we prospectively collected all out-of-hospital SCA cases using a specific case report form.23 The Paris emergency medical system is a 2-tiered response system coordinated via a unique dispatcher center: a basic life sup-port tier provided by firefighters of the Brigade de Sapeurs Pompiers de Paris, who can apply an AED, and an advanced cardiac life support function provided by ambulance teams with an emergency physician, a nurse, and a paramedic (Service d’Aide Médicale Urgente).

SCA was defined according to the Utstein guidelines as a sudden cardiac death in which specific resuscitation records were available (deceased or successfully resuscitated) after exclusion of cases with obvious extracardiac causes (eg, drug overdose, suicide, drowning, hypoxia, exsanguination, cerebrovascular accident, subarachnoid hemorrhage, and trauma).24–26

In the present analysis, we considered SCA occurring in public (nonresidential) areas. We defined public areas as all areas accessible to the general public, including all outdoor locations, public transpor-tation sites, schools, and commercial and civic establishments.

Geolocalization of SCAWe used a 2-step approach to localize SCA. First, the exact loca-tions of SCAs were coded geographically and marked on the digi-tal city map. Automated geolocalization was performed first, and all addresses were then verified manually (Google Maps, API Web Services version 3, HTML 5, and JavaScript). Second, we merged geolocalized data with a Parisian grid system established with the Geographic Information System from the Atelier Parisien de l’Urbanisme (http://www.apur.org), including a defined and grid net with a standardized size of 2020 cells (200×200-m grid cells) over 101 km2. Of note, the 2 large public parks located in Paris (Bois de Vincennes and Bois de Boulogne) were not considered in this study, which aimed to evaluate an urban setting.

Population and Landmark DataWe obtained information for each of the grid cells on the density of the population, population movements, and landmarks in each grid. This involved data collection from different agencies that was coordi-nated by Atelier Parisien de l’Urbanisme.27 National and Parisian agen-cies enrolled in the evaluation of population densities and population movements in Paris were the Syndicat des transports d’Île-de-France, the Direction Régionale et Interdépartementale de l’Equipement et de l’Aménagement, the Observatoire de la Mobilité en Île-de-France, the Institut National de la Statistique et des Études Économiques, the Institut d’Aménagement et d’Urbanisme Ile de France, the Paris City Hall, the French Ministry of Transports, and the French Ministry of Education.

First, the density of the population was defined as the number of people living in the area (hundred inhabitants per hectare). Second, in terms of population movements, the flow of persons moving through the grid cell was determined by estimating the daily number of popu-lation movements in Paris for each grid cell. Those data were based on the 2001 Enquête Globale des Transports.28 This survey was car-ried out from October 2001 to April 2002 under the auspices of the French Ministry of Public Works, Transportation, and Housing. This survey, initiated in 1976, has been carried out by the Institut National de la Statistique et des Études Économiques every 7 to 10 years in

Paris (last survey in 2010).28 The methodology of the survey followed a standardized approach established by the Center d’Etudes sur les Réseaux, les Transports, l’Urbanisme et les Constructions Publiques. Briefly, a representative sample of 10 500 families were randomly selected in Paris. Nonpermanent residents and people living in hotels or other community locations were not enrolled in this evaluation. Each member of the family was interviewed face to face by an investi-gator at the family member’s home about all trips he or she had made the previous day through the use of a specific questionnaire (including 106 variables).28 Children <6 years old were not considered because they were considered to always be escorted by older people. Individual movement was defined as the mean of individual trips taken by living/local population per day between 1 origin and 1 destination, associ-ated with a specific intention or goal. Information on timing, specific itinerary, and mode of transportation (walking, cycling, car, public transportations) was also collected. Intentions were classified as being related to work, shopping, sports/exercise, education, tourism, health, administrative issues, or other activities (eg, other leisure-time activi-ties). For 1 given grid cell, the population movement (per grid cell) was the sum of individual movements of individuals entering into this specific grid cell per day either with the intention of participating in a specific activity or when commuting home. The mean accumulated number of trips taken was estimated at 8 259 134 per day in Paris. The results of the 2010 Enquête Global des Transports survey was very similar to those of the 2001 Enquête Global des Transports survey.28

Finally, we obtained information on the evaluation of different city landmarks in each cell. Train stations, exhibition centers, shows and convention centers, museums, public swimming pools or other sport facilities (public or private, outdoor or indoor), places of worship, primary to high schools, large shopping malls (>300 m2), hospitals, and tourist attractions (monuments) were considered.

In Paris, the major airports are outside the limit of the city district and could not be studied with the same methodology. To address this issue, we focused on Charles de Gaulle Airport, which is the most crowded airport in France and has a centralized systematic collection of all cases of SCA by EMS. We collected all cases of SCA in this airport from 2000 and 2010. For population movement, no detailed data comparable to Enquête Globale des Transports used in urban Paris were available. Consequently, we used the annual frequency of person visits of this airport (60.5 million passengers in 2010) as the reference value for the flow of population.

Statistical AnalysisThe subject characteristics were described as mean±SD, proportions, median, and interquartile ranges, as appropriate.

We first provided descriptive data on population density, popula-tion movements, and local landmarks in each of the 2020 grid cells. The associations between the number of SCAs and population den-sity and population movements were assessed by use of a general linear model over classes.

Second, to assess whether SCA incidence differed between 2 typi-cal places with the same population movement, we paired each grid cell containing the 6 major train stations (Austerlitz Station, Lyon Station, Nord Station, Est Station, Montparnasse Station, Saint Lazare Station) with the most populated grid cells containing a touristic place (Eiffel Tower, Georges Pompidou Beaubourg Center, Basilica of the Sacré Cœur–Montmartre, Louvre Museum, Notre-Dame of Paris, Paris Porte de Versailles Exhibition Center) on the closer population movement. We calculated for each pair the difference in the number of SCA between the 2 cells and tested the global difference to zero using the Wilcoxon signed-rank test to assess whether SCA incidence differed between 2 typical places with same population movement.

We then estimated the risk of having SCA in 1 of the 2020 grid cells containing a specific city landmark (listed above). We used a count model, the negative binomial regression. This model allowed better consideration of the overdispersion compared with the Poisson regression in our sample data. The dependent variable was the num-ber of public SCAs in each grid cell; the independent variable was the presence or absence of the landmarks in the grid cell. We normalized each cell by using the logarithm of population movements (per each unity, ie, the grid cell) as an offset. We provided first a univariate

by guest on April 15, 2016http://circ.ahajournals.org/Downloaded from

1548 Circulation May 5, 2015

model and then a multivariate model with adjustment for other land-marks in the grid cell. The relative risk of the cell containing the land-mark was the exponential of the regression coefficient, using SAS PROC GENMOD with negative binomial probability distribution and log link function.19,29 We performed a multinomial logistic regression explaining the number of SCAs in 3 categories—no SCA, 1 SCA, ≥2 SCAs—within a grid cell by landmarks in the grid cell and by population movement and population density (considered adjustment variables). We considered population movement and density in 6 cat-egories, as already described. No SCA was chosen as the reference category. We established univariate and multivariate regressions with logistic distribution and general logit link.

All tests were 2 tailed, and values of P<0.05 were considered to indicate statistical significance. All data were analyzed at INSERM, Unit 970, Cardiovascular Epidemiology and Sudden Death, Paris, with SAS version 9.3 (SAS Institute Inc, Cary, NC).

The authors had full access to data and designed the statistical analysis, had final responsibility for the decision to submit the manu-script for publication, and vouch for the accuracy and completeness of the data and the analyses. This prospective study was conducted according to the Declaration of Helsinki after institutional review board approval.

ResultsCharacteristics and Clustered Distribution of SCAs in ParisOverall, of the 8234 out-of-hospital cardiac arrests in Paris, 5296 received EMS resuscitation attempts (Figure 1) and 4176 were presumed to have a cardiac cause (eventually con-sidered SCA): 2921 (69.9%) occurring in a residential loca-tion and 1255 (30.1%) occurring in a public location. Baseline characteristics of public SCAs are described in Table 1. Among all cardiac arrest occurrences in Paris, those occur-ring in public places had more attempts compared with those

occurring at home (85.7% versus 55.4%; P<0.0001). Among those SCAs that occurred in public places, the proportion con-sidered for resuscitation was very similar across different pub-lic locations: 83.7% in train stations compared with 86.5% in others areas (P=0.41). The proportion of SCA was observed to be higher in the morning and afternoons (representing 56% of the overall SCAs occurring in 30% of the day).

Figure 2 shows the locations of all 1255 SCAs in public places over the 2020 Parisian grid cells. The mean number of SCA occurring in 1 grid cell was 0.62, varying from 0.56 in a grid cell including a school to 9.93 for a grid cell including a large train station. Among public SCAs, 478 cases (38.1%) occurred in public facilities such as train stations, malls, or museums, whereas 777 (61.9%) took place outside of such facilities, mostly on the sidewalk. The SCAs were scattered around the city with a highly heterogeneous frequency dis-tribution, with many SCAs clustered in areas: 50% of SCAs occurred in 9% of the Paris total area. We noted a particularly high density of SCAs near the 6 major train stations where 12% of SCAs occurred (0.75% of the total Paris area). This distribution was stable over 3 sequential periods of 40 months.

Population Density and Population MovementThe incidence of SCA was evaluated across the classes of population density and population movements (Figure 3). The mean number of public SCAs by grid cell was not sig-nificantly associated with population density, varying from 0.33 to 0.78 SCA per grid cell for the first and the sixth class, respectively (P for trend=0.4). In contrast, when considering population movements, we observed a highly significant asso-ciation between the number of public SCAs and the different classes of population movements, with a 6-fold difference: the mean number of public SCA by grid cell varying from 0.21 to 1.11 SCAs per grid cell for the lowest and highest classes of population movement, respectively (P for trend <0.001).

Population Movement and Specific LocationsWe analyzed the 6 most visited places in Paris, which accounted for a very similar population movement number compared with the 6 major train stations (median, 26 930.0 versus 24 591.5 per day and grid cell; P=0.06). The median number of SCAs per grid cell was almost 5-fold lower in the tourist sites (4.5 versus 22.5; P=0.03; Figure 4).

Using a multinomial logistic regression model (Table 2), we explored the extent to which SCA occurrence was influ-enced by landmarks in cells after adjustment for the popula-tion movement and the population density of each grid cell. In univariate modeling, we observed that the odds ratio for hav-ing ≥2 SCAs in grid cells containing a train station was 5.76 (95% confidence interval [CI], 4.14–8.01; P<0.0001) higher than in grid cells without this landmark. Grid cells contain-ing large shopping malls or exhibitions were also significantly associated with the occurrence of ≥2 SCAs compared with grid cells without those landmarks. In terms of demographic characteristics, in univariate analysis, categories of population density and categories of population movement were associ-ated with the occurrence of ≥2 SCAs.

In multivariate analysis, with other landmarks in each grid cell and demographic characteristics included in the model,

Out-of-Hospital Cardiac ArrestsConsidered for Resuscitation

From 01/2000 to 07/2010Paris City

N=5,296

Home LocationN=2,921

Admitted to HospitalN=615 (NA=52)

Sudden Cardiac ArrestsN=4,176

Evidence of Extra-Cardiac EtiologyN= 1,120

Public LocationN=1,255

Figure 1. Flowchart.

by guest on April 15, 2016http://circ.ahajournals.org/Downloaded from

Marijon et al High-Risk Public Areas for Sudden Cardiac Arrest 1549

categories of population movement remained significantly associated with the occurrence of ≥2 SCAs (odds ratio, 1.48; 95% CI, 1.34–1.63; P<0.0001), whereas categories of population density were no longer associated. Grid cells containing train stations remained significantly associated with the occurrence of ≥2 SCAs (odds ratio, 3.80; 95% CI, 2.66–5.36; P<0.0001), whereas shopping malls or exhibitions were no longer significantly associated with the occurrence of ≥2 SCAs. In contrast, in multivariate analysis, grid cells

containing primary to high schools were negatively associated with the occurrence of ≥2 SCAs (odds ratio, 0.65; 95% CI, 0.48–0.89; P=0.01).

As a sensitivity analysis, we also performed a negative binomial regression model (Table I in the online-only Data Supplement) to explore the extent to which SCA occurrence was influenced by landmarks in cells normalized by the popu-lation movement of each grid cell (population movement used as an offset). In a multivariate model accounting for land-marks and population movements, large train stations were significantly associated with the occurrence of SCA (relative risk, 2.60; 95% CI, 1.69–4.01; P<0.0001).

Overall, 67 SCAs occurred in Paris Charles de Gaulle Airport during the same period (2000–2010), giving a fre-quency of occurrence of 0.12 per million in Charles de Gaulle Airport.

DiscussionOur findings provide a better understanding of the extent to which SCAs occurring in public places may be clustered in a large city, with 12% of SCAs occurring in 0.75% of the total Paris area, emphasizing why population movement is crucial to consider, with very few events occurring in less frequented areas. To the best of our knowledge, our study is the first to provide information on population movements in addition to population density. We demonstrated that the population movements was a major determining factor of such heterogeneity, much more than variations in population density. Finally, our results strongly suggest that beyond sim-ply population movements, some areas (major train stations) remain at particularly high risk for SCA occurrence. These data should help optimize AED placement in the community by concentrating AEDs in high-risk areas for SCAs such as train stations.

Figure 2. Location of sudden cardiac arrests occurring in public places in Paris (2000–2010).

Table 1. Characteristics of Subjects and Selected Features of the 1255 SCAs Occurring in Public Areas in Paris

Demographic data

Age, mean±SD, y 62.5±16

Male sex, n (%) 1036 (82.5)

History of cardiovascular disease, n (%) 441 (38.4)

Management of SCA, n (%)

Time of occurrence, h

6–9 188 (15.0)

10–13 351 (28.0)

14–17 350 (27.9)

18–21 238 (19.0)

22–5 128 (10.1)

Witnessed SCA 996 (85.4)

Bystander CPR 389 (35.4)

Initial ventricular fibrillation 446 (37.3)

Shock by external defibrillator 658 (53.9)

Time from call to AED placement, min 10.9±4.9

Alive at hospital admission, n (%) 625 (50.6)

Percentages were calculated on the basis of the total number of known events. n=1255. AED indicates automatic external defibrillator; CPR, cardiopulmonary resuscitation; and SCA, sudden cardiac arrest.

by guest on April 15, 2016http://circ.ahajournals.org/Downloaded from

1550 Circulation May 5, 2015

Stability in the incidence of SCA within geographic areas has been demonstrated.30 Accordingly, identifying high-risk places for SCA should be considered a reliable method to guide

optimal AED location to eventually increase the cost-effective-ness of such programs. For this, a better understanding of high-density SCA areas is needed to provide evidence to the medical community and policy makers to guide strategies for optimal AED deployment, especially in concentrating AEDs in high-risk areas for SCA and acknowledging that some large areas do not need to be so well equipped. The hypothesis that population movement was a highly influential factor, driving “hot spots” for SCA, was strong, although this has not yet been formally tested, mainly because of the lack of reliable data on popula-tion movements. Data from different urban Parisian agencies gave us the unique opportunity to explore this question; we demonstrate the extent to which SCA density was associated with population movements and that this association remained significant after adjustment for different landmarks in each grid cell (including train stations) and demographic characteristics. Further attention should be given in the future to those locations with the highest population movement.

As a result of our main finding, that population movement is a major determinant of SCA distribution, we observed that large train stations are particularly high-risk areas for SCAs and confirm that in the setting of a large urban European area, clustered dis-tribution is observed. These data are highly consistent with previ-ous publications that demonstrated an association between major train stations and SCA.18,19,31–35 A differential proportion in resus-citation attempted by EMS (according to the location of SCA) may have led to a significant reporting bias. This has not been observed, however, with a very similar proportion of SCA with resuscitation attempted being observed in different public places (train stations versus areas other than train stations). Accordingly, in public places, population movement appears unlikely to be a surrogate for an increased likelihood of having a SCA that is wit-nessed and thus more likely to undergo resuscitation.

Our data add to this prior literature by providing the extent to which those associations were driven by population movements. Consequently, our results are in line with prior literature on the location of SCA. Associations between SCA incidence and population density have been tested in the setting of residential

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

0 <100 100-200 200-300 300-400 >400

Mea

nSC

Ape

rGrid

-Cel

l

Population density (per hectare)

Population Movement <2 2-3 3-4 4-5 5-6 ≥6

Number of Grid-Cells 424 223 291 265 241 576

Number of SCA 89 83 125 143 176 639Mean SCA per Grid-Cell 0.21 0.37 0.43 0.54 0.73 1.11

Population Density 0 <100 100-200 200-300 300-400 ≥400

Number of Grid-Cell 93 506 495 526 284 116

Number of SCA 31 335 302 324 173 90

Mean SCA per Grid-Cell 0.33 0.66 0.61 0.62 0.61 0.78

Figure 3. Association between classes of population density and population movements in regard to sudden cardiac arrest (SCA) occurrence.

Lyon Station

Parc des expositions

Montparnasse Station

Notre-Dame

Saint Lazare Station

Beaubourg Center

Nord Station

Louvre Museum

Est Station

Montmartre

Austerlitz Station

Eiffel Tower

Mean movement per day (thousand) Total cases of SCA

Figure 4. Descriptive approach aiming to compare the population movements and occurrence of sudden cardiac arrest (SCA) in the 6 major train stations (red) vs the 6 most touristic sites (green) in Paris.

by guest on April 15, 2016http://circ.ahajournals.org/Downloaded from

Marijon et al High-Risk Public Areas for Sudden Cardiac Arrest 1551

(nonpublic) areas, demonstrating the poor predictive relevance of this factor.36 In quantifying the association between SCA occurrence and the level of population movements, our findings may be particularly useful to provide a strong rationale for AED deployment in the community, targeting high-risk areas.

In 2005, North American guidelines recommended implan-tation of AEDs in sites “where the probability of cardiac arrest occurring is at least once in every 5 years.” European guide-lines considered areas with at least 1 SCA every 2 years can-didate sites for AED placement.37 However, these guidelines were based on limited evidence.24–26 Recent data, available since 2009, also elegantly demonstrated the limitation of using such an approach,19 and these recommendations were finally deleted in the recent updated guidelines.38 When we look at the history of AED deployment in Paris, it is interesting to observe that this has been carried out independently of any evidence-based approach. For example, the 20 city halls of the different districts were the first public places equipped with AEDs; during our study period, no SCAs at all occurred in these locations. Overall, a relatively consistent pattern of deployment of almost 1000 AEDs in public places from 2000 to 2010 has been observed. For example, the first AEDs were installed in 2005 in major train stations, eventu-ally providing a very similar level of coverage compared with well-known tourist sites in Paris such as the Eiffel Tower and the

Musée du Louvre. This contrasts with the actual pattern of SCA distribution seen in our study. This finding emphasizes the point that AED placement may be substantially improved, emphasiz-ing the importance for policy makers and the medical community to consider that AED distribution should followed the SCA distri-bution on the basis of evidence such as that provided in this study.

Our study also reports additional and interesting findings, suggested by our descriptive approach and confirmed by our modeling. Population movements cannot be the only explana-tion for the high rate of SCAs in some public areas. Compared with the most tourist-visited sites in Paris, large train stations (with, however, very similar population movements) had a much higher frequency of SCAs. Our approach, based on a count modeling (negative binomial regression), showed a highly significant association between SCA and major train stations, which is attenuated only when considering population movements in the model. Possible explanations include particu-lar physical activity at train stations, which is known to favor ventricular arrhythmias, especially among untrained middle-age men.39 Another hypothesis may be postulated concerning the associated psychological stress associated with public trans-portation areas (train stations or airports), mainly related to tim-ing departure issues.40 Third, particulate pollution in these areas may possibly play a role.41 Finally, populations in train stations

Table 2. Relative Risk for SCA of Grid Cells Containing Specific Landmarks: Univariate and Multivariate Analyses (Multinomial Logistic Regression Model)

FeaturesSCA

Category

Univariate Multivariate

OR (95% CI) P Value OR (95% CI) P Value

Population density (per 1 category)

1 1.30 (1.20–1.41) <0.0001 1.08 (0.97–1.19) 0.17

2 1.20 (1.08–1.34) 0.95 (0.84–1.08)

Population movement (per 1 category)

1 1.32 (1.24–1.39) <0.0001 1.19 (1.11–1.29) <0.0001

2 1.55 (1.43–1.68) 1.48 (1.34–1.63)

Primary to high schools

1 1.41 (1.14–1.73) 0.003 1.06 (0.84–1.33) 0.01

2 0.91 (0.70–1.20) 0.65 (0.48–0.89)

Large shopping malls 1 2.15 (1.73–2.68) <0.0001 1.49 (1.17–1.89) 0.05

2 2.30 (1.75–3.03) 1.21 (0.88–1.65)

Exhibitions and shows 1 1.70 (1.31–2.21) <0.0001 1.22 (0.92–1.61) 0.21

2 2.12 (1.56–2.90) 1.29 (0.91–1.82)

Places of worship 1 1.34 (1.01–1.78) 0.05 1.08 (0.81–1.45) 0.80

2 1.38 (0.97–1.96) 1.11 (0.76–1.62)

Sports facilities 1 0.98 (0.72–1.35) 0.9 1.02 (0.73–1.43) 0.95

2 0.92 (0.61–1.39) 1.07 (0.68–1.67)

Museums 1 1.02 (0.66–1.56) 0.8 0.94 (0.60–1.48) 0.88

2 1.17 (0.69–1.96) 0.87 (0.49–1.54)

Hospitals 1 1.42 (0.90–2.24) 0.2 1.24 (0.77–1.97) 0.34

2 0.85 (0.43–1.69) 0.72 (0.35–1.48)

Monuments, tourist settings

1 1.51 (0.69–3.29) 0.15 1.33 (0.58–3.03) 0.73

2 2.25 (0.97–5.24) 1.35 (0.53–3.44)

Stations 1 2.08 (1.51–2.88) <0.0001 1.65 (1.18 -2.32) <0.0001

2 5.76 (4.14–8.01) 3.80 (2.66–5.36)

SCA categories: 1=1 SCA in the grid cell; 2=≥2 SCAs in the grid cell (cells without SCA used as reference). CI indicates confidence interval; OR, odds ratio; and SCA, sudden cardiac arrest.

by guest on April 15, 2016http://circ.ahajournals.org/Downloaded from

1552 Circulation May 5, 2015

are relatively “unselected” in contrast to users of sport facilities who are likely to be younger and in better physical shape and to have been screened by physicians. This may contribute to this higher frequency of events. Of interest, when we look at the number of SCAs in Paris Charles de Gaulle Airport (60.5 mil-lion passengers in 2010) during the same period (2000–2010), we observe a frequency of SCAs very similar to that of the train stations: 0.12 per million in Charles de Gaulle Airport compared with 0.08 per million observed in the 6 major train stations in Paris. The very similar findings observed in major train stations and airports reinforce the idea that physical and psychological stress may be a trigger of SCA in those high-risk areas; other socioeconomic factors might also account for higher rates at those areas and need further investigation. Considering that exposure time (waiting time) may be larger in airports com-pared with those expected in train stations, this could explain (at least in part) the absolute lower rate of SCAs observed in train stations compared with the airport.

Our study contributes to the existing literature on the local-ization of SCA,19,30,42 with an innovative approach based on population movement. However, we must acknowledge some limitations. First, we studied only SCA with presumed cardiac cause, which could represent a selection bias. However, this pop-ulation is the likeliest to benefit from AED and represents the main target of public access to defibrillation programs. Second, the low density of SCAs in the majority of grid cells could distort (to a minor extent) the binomial negative model used but with an unlikely overall impact on the results and model fit. Third, according to the Utstein guidelines, we did not collect informa-tion on out-of-hospital cardiac arrests not related to a primary cardiac cause.8,25,26 The extent to which those data (additional 20% of cases, in agreement with prior literature) would change our findings seems limited because those noncardiac cases are known to present a very small proportion with shockable rhythm and thus would be less likely to benefit from public AEDs. Fourth, although the classification status of SCA cases (tour-ists or permanent resident) was not collected during the study, the extent to which movements related to tourism may influ-ence the distribution of SCA in Paris was of particular interest. It could be possible that population movement may have been underestimated in train stations on the basis of a survey of Paris residents alone. However, the 2005 evaluation carried out by the Institut National de la Statistique et des Études Économiques emphasized that tourism frequency was equivalent to 224 400 permanent residents (≈11% of the total population of Paris) and that the number of those additional subjects was similar to the number of permanent residents from outside the city (mainly commuting for business purposes).43 Overall, even in one of the most tourist-visited cities worldwide, the influence of tourism on SCA distribution appears limited. Fifth, interesting existing data have demonstrated the influence of neighborhood socio-economic status on the incidence and management of SCA.44 However, in terms of SCA occurring in public places, this “time of exposition” (influence of neighborhood socioeconomic sta-tus) to neighborhood appears of limited value, especially in a city like Paris with very high population movement and conse-quently considerable heterogeneity in socioeconomic status. The extent to which the association between SCA and major train station may be driven by different (lower) socioeconomic status

needs to be examined in further studies. Sixth, spatial autocor-relation is an important issue that may be considered through a variety of complex analytic approaches guided by novel specific software (eg, SaTScan).45 Although crucial to consider in the field of infectious diseases and pollution effects, in the setting of SCA, this approach remains more debatable. Finally, our study is limited to Paris, and high-risk areas may differ according to the preferred public transportation system of the city/country (mainly major train stations and airports). Although our descrip-tive results fit extremely well with data from Folke and col-leagues19 (in Copenhagen, 19.5% of SCAs occurred in 1.2% of the city area compared with 1.4% in Paris), the extent to which the present results may be generalized to other non-European cities requires further study. The urban and demographic data available for Paris, the population density, and the large number of tourists provided a unique opportunity to assess the relation-ship between SCA and population movements. This model may not be easy to replicate in other cities.

ConclusionsIn providing a systematic analysis of determinants of SCA within a public urban setting, we demonstrated the extent to which population movements may account for the hot spots of SCA observed. Our results also suggested that beyond this key risk factor, some areas show a dramatically higher risk of SCA. This should provide evidence to the medical com-munity and policy makers to guide strategies for optimal AED deployment.

AcknowledgmentsWe thank Vincent Lanoe, Caroline Barnes, Aurelien Bidot, Antoine Labouze, and Guillaume Prang for their valuable help during the study, as well as the Association pour le Recensement et la Localisation des Défibrillateurs Committee (http://arlod.e-monsite.com, ARLoD, 15 Rue St Charles, 75015 Paris): Dr Tomislav Petrovic (Bobigny, France), Dr Louis Soulat (Châteauroux, France), Dr Sybille Goddet (Chalons sur Saône, France), Dr Bruno Thomas-Lamotte (Paris, France), and Marcel Mazeas (Paris, France).

Sources of Funding This work was supported by the French Institute of Health and Medical Research, French Society of Cardiology, and French Federation of Cardiology

DisclosuresNone.

References 1. Deo R, Albert CM. Epidemiology and genetics of sudden car-

diac death. Circulation. 2012;125:620–637. doi: 10.1161/CIRCULATIONAHA.111.023838.

2. Hollenberg J, Herlitz J, Lindqvist J, Riva G, Bohm K, Rosenqvist M, Svensson L. Improved survival after out-of-hospital cardiac arrest is asso-ciated with an increase in proportion of emergency crew–witnessed cases and bystander cardiopulmonary resuscitation. Circulation. 2008;118:389–396. doi: 10.1161/CIRCULATIONAHA.107.734137.

3. Iwami T, Nichol G, Hiraide A, Hayashi Y, Nishiuchi T, Kajino K, Morita H, Yukioka H, Ikeuchi H, Sugimoto H, Nonogi H, Kawamura T. Continuous improvements in “chain of survival” increased survival after out-of-hos-pital cardiac arrests: a large-scale population-based study. Circulation. 2009;119:728–734. doi: 10.1161/CIRCULATIONAHA.108.802058.

4. Wissenberg M, Lippert FK, Folke F, Weeke P, Hansen CM, Christensen EF, Jans H, Hansen PA, Lang-Jensen T, Olesen JB, Lindhardsen J, Fosbol

by guest on April 15, 2016http://circ.ahajournals.org/Downloaded from

Marijon et al High-Risk Public Areas for Sudden Cardiac Arrest 1553

EL, Nielsen SL, Gislason GH, Kober L, Torp-Pedersen C. Association of national initiatives to improve cardiac arrest management with rates of bystander intervention and patient survival after out-of-hospital cardiac arrest. JAMA. 2013;310:1377–1384. doi: 10.1001/jama.2013.278483.

5. Sasson C, Rogers MA, Dahl J, Kellermann AL. Predictors of survival from out-of-hospital cardiac arrest: a systematic review and meta-anal-ysis. Circ Cardiovasc Qual Outcomes. 2010;3:63–81. doi: 10.1161/CIRCOUTCOMES.109.889576.

6. Sanders AB, Kern KB. Surviving cardiac arrest: location, location, loca-tion. JAMA. 2008;300:1462–1463. doi: 10.1001/jama.300.12.1462.

7. Nichol G, Thomas E, Callaway CW, Hedges J, Powell JL, Aufderheide TP, Rea T, Lowe R, Brown T, Dreyer J, Davis D, Idris A, Stiell I; Resuscitation Outcomes Consortium Investigators. Regional variation in out-of-hospital cardiac arrest incidence and outcome. JAMA. 2008;300:1423–1431. doi: 10.1001/jama.300.12.1423.

8. Fishman GI, Chugh SS, Dimarco JP, Albert CM, Anderson ME, Bonow RO, Buxton AE, Chen PS, Estes M, Jouven X, Kwong R, Lathrop DA, Mascette AM, Nerbonne JM, O’Rourke B, Page RL, Roden DM, Rosenbaum DS, Sotoodehnia N, Trayanova NA, Zheng ZJ. Sudden cardiac death prediction and prevention: report from a National Heart, Lung, and Blood Institute and Heart Rhythm Society Workshop. Circulation. 2010;122:2335–2348. doi: 10.1161/CIRCULATIONAHA.110.976092.

9. Cummins RO, Ornato JP, Thies WH, Pepe PE. Improving survival from sudden cardiac arrest: the “chain of survival” concept: a state-ment for health professionals from the Advanced Cardiac Life Support Subcommittee and the Emergency Cardiac Care Committee, American Heart Association. Circulation. 1991;83:1832–1847.

10. Eisenberg MS, Hallstrom AP, Copass MK, Bergner L, Short F, Pierce J. Treatment of ventricular fibrillation: emergency medical technician defi-brillation and paramedic services. JAMA. 1984;251:1723–1726.

11. Weisfeldt ML, Kerber RE, McGoldrick RP, Moss AJ, Nichol G, Ornato JP, Palmer DG, Riegel B, Smith SC Jr. Public access defibrillation: a statement for healthcare professionals from the American Heart Association Task Force on Automatic External Defibrillation. Circulation. 1995;92:2763.

12. Caffrey SL, Willoughby PJ, Pepe PE, Becker LB. Public use of automated external defibrillators. N Engl J Med. 2002;347:1242–1247. doi: 10.1056/NEJMoa020932.

13. Groeneveld PW, Kwong JL, Liu Y, Rodriguez AJ, Jones MP, Sanders GD, Garber AM. Cost-effectiveness of automated external defibrillators on air-lines. JAMA. 2001;286:1482–1489.

14. Capucci A, Aschieri D, Piepoli MF, Bardy GH, Iconomu E, Arvedi M. Tripling survival from sudden cardiac arrest via early defibrillation with-out traditional education in cardiopulmonary resuscitation. Circulation. 2002;106:1065–1070.

15. Kitamura T, Iwami T, Kawamura T, Nagao K, Tanaka H, Hiraide A; Implementation Working Group for the All-Japan Utstein Registry of the Fire and Disaster Management Agency. Nationwide public-access defibrillation in Japan. N Engl J Med. 2010;362:994–1004. doi: 10.1056/NEJMoa0906644.

16. Hallstrom AP, Ornato JP, Weisfeldt M, Travers A, Christenson J, McBurnie MA, Zalenski R, Becker LB, Schron EB, Proschan M; Public Access Defibrillation Trial Investigators. Public-access defibrillation and survival after out-of-hospital cardiac arrest. N Engl J Med. 2004;351:637–646. doi: 10.1056/NEJMoa040566.

17. Berdowski J, Blom MT, Bardai A, Tan HL, Tijssen JG, Koster RW. Impact of onsite or dispatched automated external defibrillator use on survival after out-of-hospital cardiac arrest. Circulation. 2011;124:2225–2232. doi: 10.1161/CIRCULATIONAHA.110.015545.

18. Hansen CM, Wissenberg M, Weeke P, Ruwald MH, Lamberts M, Lippert F, Gislason GH, Nielsen SL, Kober L, Torp-Pedersen C, Folke F. Automated external defibrillators inaccessible to more than half of nearby cardiac arrests in public locations during evening, nighttime and week-ends. Circulation. 2013;128:2224–2231.

19. Folke F, Lippert FK, Nielsen SL, Gislason GH, Hansen ML, Schramm TK, Sørensen R, Fosbøl EL, Andersen SS, Rasmussen S, Køber L, Torp-Pedersen C. Location of cardiac arrest in a city center: strategic place-ment of automated external defibrillators in public locations. Circulation. 2009;120:510–517. doi: 10.1161/CIRCULATIONAHA.108.843755.

20. Eckstein M. The Los Angeles public access defibrillator (PAD) pro-gram: ten years after. Resuscitation. 2012;83:1411–1412. doi: 10.1016/j.resuscitation.2012.03.029.

21. Reed DB, Birnbaum A, Brown LH, O’Connor RE, Fleg JL, Peberdy MA, Van Ottingham L, Hallstrom AP; PAD Trial Investigators. Location of

cardiac arrests in the public access defibrillation trial. Prehosp Emerg Care. 2006;10:61–76.

22. Becker LB, Aufderheide TP, Geocadin RG, Callaway CW, Lazar RM, Donnino MW, Nadkarni VM, Abella BS, Adrie C, Berg RA, Merchant RM, O’Connor RE, Meltzer DO, Holm MB, Longstreth WT, Halperin HR; American Heart Association Emergency Cardiovascular Care Committee; Council on Cardiopulmonary, Critical Care, Perioperative and Resuscitation. Primary outcomes for resuscitation science studies: a consensus statement from the American Heart Association. Circulation. 2011;124:2158–2177. doi: 10.1161/CIR.0b013e3182340239.

23. Jost D, Degrange H, Verret C, Hersan O, Banville IL, Chapman FW, Lank P, Petit JL, Fuilla C, Migliani R, Carpentier JP; DEFI 2005 Work Group. DEFI 2005: a randomized controlled trial of the effect of automated exter-nal defibrillator cardiopulmonary resuscitation protocol on outcome from out-of-hospital cardiac arrest. Circulation. 2010;121:1614–1622. doi: 10.1161/CIRCULATIONAHA.109.878389.

24. Fishman GI, Chugh SS, Dimarco JP, Albert CM, Anderson ME, Bonow RO, Buxton AE, Chen PS, Estes M, Jouven X, Kwong R, Lathrop DA, Mascette AM, Nerbonne JM, O’Rourke B, Page RL, Roden DM, Rosenbaum DS, Sotoodehnia N, Trayanova NA, Zheng ZJ. Sudden cardiac death prediction and prevention: report from a National Heart, Lung, and Blood Institute and Heart Rhythm Society Workshop. Circulation. 2010;122:2335–2348. doi: 10.1161/CIRCULATIONAHA.110.976092.

25. Cummins RO, Chamberlain DA, Abramson NS, Allen M, Baskett PJ, Becker L, Bossaert L, Delooz HH, Dick WF, Eisenberg MS. Recommended guidelines for uniform reporting of data from out-of-hospital cardiac arrest: the Utstein style: a statement for health pro-fessionals from a task force of the American Heart Association, the European Resuscitation Council, the Heart and Stroke Foundation of Canada, and the Australian Resuscitation Council. Circulation. 1991;84:960–975.

26. Jacobs I, Nadkarni V, Bahr J, Berg RA, Billi JE, Bossaert L, Cassan P, Coovadia A, D’Este K, Finn J, Halperin H, Handley A, Herlitz J, Hickey R, Idris A, Kloeck W, Larkin GL, Mancini ME, Mason P, Mears G, Monsieurs K, Montgomery W, Morley P, Nichol G, Nolan J, Okada K, Perlman J, Shuster M, Steen PA, Sterz F, Tibballs J, Timerman S, Truitt T, Zideman D. Cardiac arrest and cardiopulmonary resuscitation outcome reports: update and simplification of the Utstein templates for resuscita-tion registries: a statement for healthcare professionals from a task force of the International Liaison Committee on Resuscitation (American Heart Association, European Resuscitation Council, Australian Resuscitation Council, New Zealand Resuscitation Council, Heart and Stroke Foundation of Canada, InterAmerican Heart Foundation, Resuscitation Councils of Southern Africa). Circulation. 2004;110:3385–97.

27. Atelier d’Urbanisme Parisien. Etude de la Localisation des stations de velos en libre service. http://www.apur.org/sites/default/files/docu-ments/216.pdf. 2006. Accessed July 11, 2014.

28. Omnil.Fr. http://www.omnil.fr/spip.php?article13. Accessed July 11, 2014.

29. Cameron AC, Trivedi PK. Regression Analysis of Count Data (Econometric Society Monographs). Cambridge: Cambridge University Press; 1998.

30. Sasson C, Keirns CC, Smith D, Sayre M, Macy M, Meurer W, McNally BF, Kellermann AL, Iwashyna TJ; CARES (Cardiac Arrest Registry to Enhance Survival) Study Group. Small area variations in out-of-hos-pital cardiac arrest: does the neighborhood matter? Ann Intern Med. 2010;153:19–22. doi: 10.7326/0003-4819-153-1-201007060-00255.

31. Sasaki M, Iwami T, Kitamura T, Nomoto S, Nishiyama C, Sakai T, Tanigawa K, Kajino K, Irisawa T, Nishiuchi T, Hayashida S, Hiraide A, Kawamura T. Incidence and outcome of out-of-hospital cardiac arrest with public-access defibrillation: a descriptive epidemiological study in a large urban community. Circ J. 2011;75:2821–2826.

32. Engdahl J, Herlitz J. Localization of out-of-hospital cardiac arrest in Goteborg 1994–2002 and implications for public access defibrillation. Resuscitation. 2005;64:171–175.

33. Becker L, Eisenberg M, Fahrenbruch C, Cobb L. Public locations of cardiac arrest: implications for public access defibrillation. Circulation. 1998;97:2106–2109.

34. Folke F, Lippert FK, Nielsen SL, Gislason GH, Hansen ML, Schramm TK, Sørensen R, Fosbøl EL, Andersen SS, Rasmussen S, Køber L, Torp-Pedersen C. Location of cardiac arrest in a city center: strategic place-ment of automated external defibrillators in public locations. Circulation. 2009;120:510–517. doi: 10.1161/CIRCULATIONAHA.108.843755.

35. Ong ME, Tan EH, Yan X, Anushia P, Lim SH, Leong BS, Ong VY, Tiah L, Yap S, Overton J, Anantharaman V. An observational study describing

by guest on April 15, 2016http://circ.ahajournals.org/Downloaded from

1554 Circulation May 5, 2015

the geographic-time distribution of cardiac arrests in Singapore: what is the utility of geographic information systems for planning public access defibrillation? (PADS Phase I). Resuscitation. 2008;76:388–396. doi: 10.1016/j.resuscitation.2007.09.006.

36. Folke F, Gislason GH, Lippert FK, Nielsen SL, Weeke P, Hansen ML, Fosbøl EL, Andersen SS, Rasmussen S, Schramm TK, Køber L, Torp-Pedersen C. Differences between out-of-hospital cardiac arrest in residential and pub-lic locations and implications for public-access defibrillation. Circulation. 2010;122:623–630. doi: 10.1161/CIRCULATIONAHA.109.924423.

37. Handley AJ, Koster R, Monsieurs K, Perkins GD, Davies S, Bossaert L; European Resuscitation Council. European Resuscitation Council guide-lines for resuscitation 2005, section 2: adult basic life support and use of automated external defibrillators. Resuscitation. 2005;67(suppl 1):S7–S23. doi: 10.1016/j.resuscitation.2005.10.007.

38. Koster RW, Baubin MA, Bossaert LL, Caballero A, Cassan P, Castrén M, Granja C, Handley AJ, Monsieurs KG, Perkins GD, Raffay V, Sandroni C. European Resuscitation Council guidelines for resuscitation 2010, sec-tion 2: adult basic life support and use of automated external defibrillators. Resuscitation. 2010;81:1277–1292. doi: 10.1016/j.resuscitation.2010.08.009.

39. Albert CM, Mittleman MA, Chae CU, Lee IM, Hennekens CH, Manson JE. Triggering of sudden death from cardiac causes by vigorous exertion. N Engl J Med. 2000;343:1355–1361. doi: 10.1056/NEJM200011093431902.

40. Lampert R, Joska T, Burg MM, Batsford WP, McPherson CA, Jain D. Emotional and physical precipitants of ventricular arrhythmia. Circulation. 2002;106:1800–1805.

41. Mustafic H, Jabre P, Caussin C, Murad MH, Escolano S, Tafflet M, Périer MC, Marijon E, Vernerey D, Empana JP, Jouven X. Main air pollutants and myocardial infarction: a systematic review and meta-analysis. JAMA. 2012;307:713–721. doi: 10.1001/jama.2012.126.

42. Ong ME, Earnest A, Shahidah N, Ng WM, Foo C, Nott DJ. Spatial varia-tion and geographic-demographic determinants of out-of-hospital cardiac arrests in the city-state of Singapore. Ann Emerg Med. 2011;58:343–351. doi: 10.1016/j.annemergmed.2010.12.014.

43. Population présente: méthodes de détermination à partir des enquêtes sur les touristes. http://www.veilleinfotourisme.fr/population-presente-meth-odes-de-determination-a-partir-des-enquetes-sur-les-touristes–100583.kjsp. Accessed July 11, 2014.

44. Sasson C, Magid DJ, Chan P, Root ED, McNally BF, Kellermann AL, Haukoos JS; CARES Surveillance Group. Association of neighbor-hood characteristics with bystander-initiated CPR. N Engl J Med. 2012;367:1607–1615. doi: 10.1056/NEJMoa1110700.

45. Kleinman K, Abrams A, Katherine Yih W, Platt R, Kulldorff M. Evaluating spatial surveillance: detection of known outbreaks in real data. Stat Med. 2006;25:755–769. doi: 10.1002/sim.2402.

CLINICAL PeRSPeCTIveSudden cardiac arrest (SCA) remains a leading cause of death in the industrialized world. Survival remains poor despite decades of research and major financial investments in resuscitation. Because SCA most often presents with initial ventricu-lar fibrillation, external defibrillation has been incorporated as a key factor in the chain of survival since the early 1990s. Public-access defibrillation programs have thus been developed to make automatic external defibrillators (AEDs) more widely available. Although the benefits of AEDs are undeniable, the effectiveness of such programs could be dramatically improved. One of the key issues is the disparity between the locations of AEDs and SCAs. Although preliminary data have identified sites with a higher risk of SCAs, the determinants of those areas have not yet been identified and quantified. In this study, we collected data on SCA locations in Paris over a 10-year period. We emphasized the extent to which, in the setting of a large urban area, the distribution of SCA was clustered in major “hot spots.” We demonstrated that the number of population movements was a major determining factor of SCA location. However, beyond population movements, our results strongly suggest that some areas (especially major train stations) remain at particularly high risk for SCA occurrence (5-fold higher risk compared with tourist areas with similar population movements). This should provide evidence to the medical community and policy makers to guide strategies for optimal AED deployment, especially in concentrating AEDs in high-risk areas for SCA.

by guest on April 15, 2016http://circ.ahajournals.org/Downloaded from

David S. Celermajer, Alain Cariou and Xavier JouvenAlbert A. Hagège, Jean-Yves Le Heuzey, Michel Desnos, Florence Dumas, Christian Spaulding,

Frankie Beganton, Patricia Pelloux, Hervé Degrange, Guillaume Béal, Jean-Pierre Tourtier, Eloi Marijon, Wulfran Bougouin, Muriel Tafflet, Nicole Karam, Daniel Jost, Lionel Lamhaut,

Population Movement and Sudden Cardiac Arrest Location

Print ISSN: 0009-7322. Online ISSN: 1524-4539 Copyright © 2015 American Heart Association, Inc. All rights reserved.

is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Circulation doi: 10.1161/CIRCULATIONAHA.114.010498

2015;131:1546-1554; originally published online March 11, 2015;Circulation. 

http://circ.ahajournals.org/content/131/18/1546World Wide Web at:

The online version of this article, along with updated information and services, is located on the

http://circ.ahajournals.org/content/suppl/2015/03/11/CIRCULATIONAHA.114.010498.DC1.htmlData Supplement (unedited) at:

  http://circ.ahajournals.org//subscriptions/

is online at: Circulation Information about subscribing to Subscriptions: 

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

  document. Permissions and Rights Question and Answer this process is available in the

click Request Permissions in the middle column of the Web page under Services. Further information aboutOffice. Once the online version of the published article for which permission is being requested is located,

can be obtained via RightsLink, a service of the Copyright Clearance Center, not the EditorialCirculationin Requests for permissions to reproduce figures, tables, or portions of articles originally publishedPermissions:

by guest on April 15, 2016http://circ.ahajournals.org/Downloaded from

Supplemental Material

Methods

We estimated the risk of having SCA in one of the 2,020 grid-cells containing

specific city landmark. We used a count model, the negative binomial

regression. This model allowed better consideration of the over-dispersion

compared to the Poisson regression in our sample data. The dependent

variable was the number of public SCA in each grid-cell; the independent

variable was the presence or absence of the landmarks in the grid-cell. We

normalized each cell by using the logarithm of population movements (per

each unity i.e. the grid-cell) as an offset. We provided first a univariate model

and then a multivariate model considering adjustment on other landmarks in

the grid-cell. The relative risk (RR) of the cell containing the landmark was the

exponential of the regression coefficient, using SAS PROC GENMOD with

negative binomial probability distribution and log link function.19, 31

Table Legend Relative Risk for SCA of Grid-Cells Containing Specific Landmark. Univariate and Multivariate Analyses (Negative Binomial Regression Model). The RR represents the risk of SCA in a grid-cell with specific landmark compared to grid-cell without the considered landmark. Table

Grid-Cell Characteristics†

Grid-Cell, N

SCA, N‡

Univariable RR (95%CI)

P Value

Multivariable RR (95%CI) *

P Value

Primary to high schools 893 496 0.69 (0.6 - 0.8) <0.0001 0.73 (0.63 - 0.84) <0.0001 Large shopping mall 646 531 0.95 (0.82 - 1.1) 0.49 0.93 (0.8 - 1.07) 0.49 Exhibitions and shows 383 309 0.93 (0.79 - 1.1) 0.41 0.95 (0.8 - 1.12) 0.29 Place of worship 311 222 0.89 (0.74 - 1.08) 0.23 0.94 (0.78 - 1.13) 0.52 Train station 258 418 1.82 (1.54 - 2.15) <0.0001 1.77 (1.5 - 2.09) <0.0001 Sports facilities 250 161 1.12 (0.9 - 1.39) 0.32 1.16 (0.93 - 1.43) 0.19 Museum 132 115 0.97 (0.74 - 1.28) 0.84 0.91 (0.69 - 1.19) 0.49 Hospital 98 57 0.86 (0.62 - 1.19) 0.37 0.93 (0.68 - 1.28) 0.66 Touristic sites 36 37 1.05 (0.65 - 1.69) 0.84 0.98 (0.63 - 1.55) 0.94

CI denotes confidence interval, RR, relative risk, SCA, sudden cardiac arrest. †Grid cells are defined as 200*200-m areas.

‡Denotes the total number of SCA within a grid cell containing a given landmark. The number of SCA can overlap between the different grid-cell characteristics because one grid-cell can contain several landmarks. Overall, 1,255 SCA occur over 2,020 grid-cells, including 1,555 grid-cells with at least one characteristics/landmark (1,132 SCA) and 465 grid-cells with none of the characteristics mentioned on the table (123 SCA). *Multivariable model considering adjustment on other landmarks of the grid-cell.