Resuscitation (2006) 71, 180—187

CLINICAL PAPER

Monitoring intervention programmes forout-of-hospital cardiac arrest in a mixedurban and rural setting�

Andrea Fabbri a,∗, Giulio Marchesinib, Marco Spadaa, Tiziana Iervesea,Massimo Dentea, Marcello Galvanic, Alberto Vandelli a

a Dipartimento dell’Emergenza, Azienda USL Forlı, Italyb Alma Mater Studiorum, Universita of Bologna, Italy

c Unita di Terapia Intensiva Cardiologica, Azienda USL di Forlı, Italy

Received 17 January 2006; received in revised form 29 March 2006; accepted 7 April 2006

KEYWORDSWitnessed cardiacarrest;Resuscitation;Outcome predictors;Survival

SummaryBackground: Only a few data are available on the survival rate following out-of-hospital cardiac arrest in different Italian settings. We report an analysis of a 10-year experience in a mixed rural/urban setting, the main variables associated withsurvival, and the preliminary results of the implementation of an automated externaldefibrillator (AED) programme operated by lay volunteers on the effectiveness of theexisting Emergency Medical Service (EMS).Methods: We report data from an observational cohort study on all adults, resus-citated from witnessed cardiac arrest between 1994 and 2004 in the district areaof Forlı (Italy). The AED programme was introduced in 2002. Entry variables, timeintervals and nodal events were tested according to Utstein recommendations. Thepredictors of favourable outcomes (Overall Performance Category 1—2) were iden-tified by logistic regression analysis.Results: The witnessed cardiac arrest rate was 27/100,000 population per year(95% confidence interval, 18—38). The initial rhythm was shockable in 241/479cases (50.3%). After resuscitation, 55 (11.5%) subjects had a favourable outcomeat discharge and 38 (7.9%) at 1 year. Time-to-treatment was longer for EMS thanfor AED-equipped units (median, 8 min interquartile range, 6—10 (467 cases) ver-sus 6 min interquartile range, 4—8 (13 cases); P < 0.013), but the final results of

the AED programme were poor, with only 1 subject saved/171,000 inhabitants in 2years. Positive outcome predictors were male sex, younger age, shockable rhythms,low number of defibrillations, and no history of hypertension, diabetes, myocardialinfarction or congestive heart failure.

� A Spanish translated version of the summary of this article appears as Appendix in the online version atdoi:10.1016/j.resuscitation.2006.04.003.

∗ Correspondence to: Dipartimento dell’Emergenza, Ospedale Morgagni-Pierantoni, Azienda USL di Forlı, Via Forlanini 34, I-47100Forlı, Italy. Tel.: +39 0543 735151; fax: +39 0543 735163.

E-mail address: andfabbri@libero.it (A. Fabbri).

0300-9572/$ — see front matter © 2006 Elsevier Ireland Ltd. All rights reserved.doi:10.1016/j.resuscitation.2006.04.003

saved by an AED programme is limited when accompanied by an efficient traditionalEMS. The allocation of resources to an AED programme should be reconsidered in a

g.td.

I

SrtRtipElCiea

dslivcotfgtAdc

efimpisE

M

S

Tht

tri1iAespaenjwtR

alciwiaDcc

bccavEmtatc

fbcentral computerized system.

From 2002, a defibrillation programme by lay

mixed rural/urban settin© 2006 Elsevier Ireland L

ntroduction

urvival following out-of-hospital cardiac arrestemains poor, one of the major determinants beinghe immediate availability of a defibrillator.1,2.ecovery rate decreases to 5—10% when defibrilla-ion is delayed over 10 min, as frequently observedn the events occurring in the community.3 Thisrompted the development of rapid and efficientmergency Medical Services (EMS) to reach any col-apsed subject within the shortest possible time.ontinuous monitoring of EMS is mandatory to

mplement additional strategies to improve theffectiveness of the system, and to check the vari-bles predicting outcome.

Out of hospital, the interval from collapse toefibrillation remains the critical factor to improveurvival. One reasonable approach is early defibril-ation by lay volunteers other than the EMS. Studiesn crowded areas4—6 demonstrated a 49—74% sur-ival rate when immediate cardiopulmonary resus-itation (CPR) is provided by bystanders.6 Locationsther than office towers, sporting venues, massransit facilities, and casinos were also proposedor automated external defibrillator (AED) pro-rammes, but differing results were published onheir additional value in general health districts.7,8

dditional data are needed in rural areas with aifferent population mix and in areas where an effi-ient traditional EMS is operative.

The objective of this study was to monitor theffectiveness of the EMS in subjects resuscitatedrom out-of-hospital cardiac arrest in a general Ital-an health district, with a mixed urban/rural andedium-density population, to verify the variablesredicting a favourable outcome, and to test thencremental benefit of an AED programme, progres-ively implemented in the context of the existingMS.

aterials and methods

tudy design and setting

he EMS is the sole provider of emergency pre-ospital ambulance care for the 171,000 inhabi-ants of the Health District of Forlı with 58% of

rTgp

All rights reserved.

he population living in an urban environment. EMSesponds to all requests for emergency assistancen the area (2380 km2), with an annual volume of5,000 calls. It has been operative for over 25 years,s well standardised and continuously monitored.ll ambulances are equipped with an automatedxternal defibrillator and nurses trained in basic lifeupport-defibrillation (BLS-D). Patients are trans-orted to the Emergency Department (ED) of Forlı,450 acute bed general hospital, providing gen-

ral clinical services (except cardiac surgery andeurosurgery). The ED treats about 50,000 sub-ects per year, the hospital over 21,000 subjects,ith 900 admissions to intensive care. The hospi-

al database is directly connected with the Generalegistry Office of the District.

The EMS of Forlı responds to all emergencies with2-tier emergency system comprising 11 ambu-

ances operating round-the-clock and one heli-opter during daylight. Ambulance configurationncludes the first responder BLS vehicle, equippedith one BLS-D nurse certified according to Amer-

can Heart Association, and one attendant, and andvanced life support (ALS) vehicle, with one BLS-

certified nurse and one ALS emergency physi-ian (certified according to American Heart Asso-iation).

Within this system, the EMS alerted to a possi-le cardiac arrest by an emergency call initiates aoordinated response. The dispatcher asks a spe-ific question ‘‘Is the patient unconscious?’’ If thenswer is positive, a high priority procedure is acti-ated, resulting in (a) dispatch of ambulance by theMS and (b) either a telephone call to the nearestobile AED or a telephone call to the volunteers in

he nearest fixed AED location. Option (b) has beendded in 2002 (see below). The system is aimedo guarantee defibrillation within 8 min in 90% ofases.

The information on ambulance dispatch, timerom call to arrival on the scene and presence ofystanders witnessing the event is registered in a

Out-of-hospital cardiac arrest in a mixed urban and rural setting 181

Conclusions: Even in a mixed urban/rural setting, survival from out-of-hospital car-diac arrest is dependent on well-known predictors. In our setting, the number of cases

esponders equipped with AEDs has been added.wenty-five fixed-location AEDs were placed pro-ressively in crowded areas, 7 mobile AEDs werelaced in police and fire vehicles and 7 in the vehi-

182

Table 1 Location and use of the fixed and mobileAEDs in the general health district of Forlı from 2002to 2004

DeployedN = 39

Used N = 13

Fixed AEDsAthletic centers 9 1Swimming pools 1 —Post office 1 —Airport 1 —Residential facilities andretirement homes

8 —

Public offices 4 —Prison 1 —

Mobile AEDPolice and fire vehicles 7 1Public assistance 7 11

tuCsM

DCcc

D

Fstu

aoSswc

S

CfiobaEaoro

osto

D

O

OaTj

vehicles

cles of the Public Assistance, a non-profit organi-zation of volunteers participating in the assistanceand transportation of patients (Table 1). Thus, atotal of 39 AEDs were deployed to cover a pop-ulation of 171,000 inhabitants (1 AED per 4400inhabitants). During the study period, 900 lay volun-teers were trained and organised to cooperate withthe EMS. They were recruited among volunteers ofnon-profit organisations and among public officersworking where AED were deployed. Considering theprogressive implementation of the programme, theaverage availability of AED instruments was 711days.

Lay volunteers operate under the medicalresponsibility of the head of the EMS, who certifiesthe qualifying examination, according to interna-tional guidelines General Health District of Forlı,Italy.1 Training courses for lay volunteers include5 h of theoretical and practical lessons, a finalexamination9 and retraining courses (1-h) every 12months. Volunteers, trained in groups of 6, areinstructed to recognise the absence of conscious-ness, the absence of breathing, and to check forsigns of circulation. When these are absent, theyare instructed to turn on the AED and to follow itsvoice instructions. The time from activation of thedevice to initial AED analysis is calculated duringtraining and on the field as a measure of skillful-ness.

The volunteers follow the default defibrillationprotocol of the AEDs, delivering fixed energy at

150 J. After three unsuccessful shocks, the AED ini-tiates a 1-min pause. The responsibility for resus-citation is transferred to the EMS on their arrival,whereupon the standard American Heart Associa-

1gia

A. Fabbri et al.

ion guidelines are followed. The EMS ambulancesse the biphasic Zoll M Series (Medical Corporation,helmsford, MA) and AED programme the bipha-ic Forerunner Heartstart FR2 AED device (Laerdaledical AS, Stavanger, Norway).The protocol is in keeping with the Helsinki

eclaration and is approved by the local Ethicalommittee. A project coordinator (S.M.) regularlyhecks the adherence to the protocol to ensure theompleteness of information.

ata collection and processing

or the purpose of the present report, we used thetudy variables defined from the very beginning ofhe programme and a standard Case Report Form inse with each cardiac arrest case.

In our setting, response times are recordedccording to Utstein definition; the internal clocksf the AEDs are synchronised with the EMS Centraltation and the interval from emergency call to firsthock are determined by synchronising the AEDsith the dispatch-time recorded in the communi-ation center. Intervals are computed in minutes.

election of cases

ase definitions follow the Utstein style guidelinesor reporting cardiac arrest data.10 The flow charts reported in Figure 1. All confirmed cases of out-f-hospital cardiac arrest were eligible, but onlyystander-witnessed patients treated for cardiacrrest of presumed cardiac origin were considered.xclusion criteria were age <18 years and cardiacrrest caused by trauma, drug overdose, drowningr burns. Local protocols enable EMS to withholdesuscitation in the presence of unequivocal signsf death.

In the socio-cultural setting of the health districtf Forlı, it is exceedingly uncommon that uncon-cious subjects are resuscitated by bystanders. Forhis reason, we did not consider this possibility inur analysis.

The study was carried out from July 1, 1994 toecember 31, 2004.

utcome measures

utcome measures were (1) favourable outcomet discharge, (2) favourable outcome at 1 year.he outcome was considered favourable when sub-ects survived with an overall capability category

—2, classified by the Overall Performance Cate-ories (OPC),11 with level 1 indicating good capabil-ty, and 2 moderate capability (patient is consciousnd performs independent activities of daily life,

Out-of-hospital cardiac arrest in a mixed urban and ru

Fga

bsuwon4

camntrcuHw

The median age was 73 years (IQR, 65—81 years),66.2% were males. The events mainly occurred athome; the initial rhythm was shockable (ventricu-lar fibrillation or ventricular tachycardia) in 50% ofcases.

Table 2 Number of events and clinical character-istics of bystander-witnessed out-of-hospital cardiacarrests of cardiac origin in the general health districtof Forlı from 1994 to 2004

N = 479

Events per 100000/year 27 [18—38]a

Male sex 317 (66.2)Age, years 74 [66—84]b

Urban setting 343 (71.6)Home location 357 (74.5)Week-end 122 (25.5)Night time 199 (41.5)Shockable rhythms 241 (50.3)Discharged alive 73 (15.2)Favourable outcome at discharge 55 (11.5)Favourable outcome at 1 year 38 (7.9)

igure 1 Report of out-of-hospital cardiac arrest in theeneral health district of Forlı between 1994 and 2004,ccording to simplified Utstein criteria.

ut has moderate cerebral or non-cerebral organ-ystem dysfunction). The outcome was definednfavourable for subjects who died or survivedith OPC score 3 (severe disability: dependent onthers for daily support and severe cerebral andon-cerebral organ-system dysfunction) and score(coma state).Categories were assigned according to physi-

ians’ descriptions of patients’ neurological statust discharge. At 1 year, the outcome was deter-ined by a physician unaware of the initial diag-

osis on the basis of the response to a structuredelephone call. The last patient who survived wasesuscitated on April 18, 2004, thus permitting a

omplete analysis of the results. Four cases withncertain neurological status were handled by theead of the Department (VA). Three cases (0.6%)ere lost to follow up.

ral setting 183

Primary data analysis

Subjects with favourable outcome were comparedwith subjects with unfavourable outcome (sub-jects dying or surviving with OPC 3—4), usingMann—Whitney test, chi-square test or logisticregression analysis.12 Medians and interquartileranges (IQR) and frequencies were used to describedata distribution. The following variables weretested in multivariate logistic regression analysis:sex, age, co-morbidity (history of diabetes, hyper-tension, myocardial infarction, congestive heartfailure), seasonality, day-week, day-times, urbansetting, home location, response times, initialrhythm. All analyses were performed with SPSS/PC+statistical package on a personal computer. Two-tailed P-values less than 0.05 were considered sta-tistically significant.

Results

Of 998 consecutive subjects with sudden cardiacarrest in whom resuscitation was attempted, 843(84.5%) events were of presumed cardiac aetiol-ogy (Figure 1), and 479 of these cases (56.8%)were bystander-witnessed. The rate of events ofwitnessed cardiac origin totaled 27 cases (95% CI,18—38) per 100,000 population per year (Table 2).

Data are expressed as number of cases (%) unless otherwisespecified.

a Mean [95% confidence interval].b Median [interquartile range].

184 A. Fabbri et al.

Table 3 Clinical characteristics of subjects and main specificities of the events in relation to favourable outcomeat discharge after out-of-hospital cardiac arrest in the general health district of Forlı from 1994 to 2004

Characteristics of events Favourable outcomeN = 55 (11.5)

Unfavourable outcomeN = 424 (88.5)

Odds ratio (95%confidence intervals)

Males (%) 44 (80.0) 273 (64.4) 2.21 (1.11—4.41)Age <74 years 38 (69.1) 197 (46.5) 2.58 (1.41—4.71)History of hypertension 7 (12.7) 118 (27.8) 0.38 (0.17—0.86)History of diabetes mellitus 7 (12.7) 122 (28.8) 0.36 (0.16—0.82)History of congestive heart failure 1 (1.8) 130 (30.7) 0.04 (0.03—0.31)History of myocardial infarction 5 (9.1) 112 (26.4) 0.28 (0.11—0.72)Winter season 14 (25.5) 124 (29.2) 0.83 (0.43—1.57)Week-end 15 (27.3) 107 (25.2) 1.11 (0.59—2.09)Night time 22 (40.0) 177 (41.7) 0.93 (0.52—1.65)Urban setting 45 (81.8) 298 (70.3) 1.90 (0.93—3.89)Home location 35 (63.6) 322 (75.9) 0.55 (0.31—1.00)

n ana

tntbtwcw(

aa9

bw1sco

Shockable rhythms 50 (90.9)

Data are tested for significance by univariate logistic regressio

After resuscitation and hospital admission, 73(15.2%) subjects were discharged alive, and 55(11.5%) fit the criteria for favourable outcome (OPC1 or 2) at discharge. Subjects with favourableoutcome were more likely to be males, younger,and had a higher prevalence of shockable rhythmswhen rescued (50 (90.9%) versus 191 (45.0%); OR95% CI 8.07 (4.03—16.17); P versus unfavourableoutcome <0.001). They were also less likelyto have a history of hypertension, diabetes,myocardial infarction, and congestive heart failure(Table 3).

Response times were shorter in subjects withfavourable outcome (Table 4). The first respond-ing unit of EMS arrived at the scene within 8 minin only 287 cases (59.9% of total events). Response-times were lower for AED-equipped volunteers andlay persons (6 min, IQR 5—7; 13 cases) in com-parison to EMS units (8 min, IQR 6—10; 466 cases;P < 0.013). Time-to-treatment (the time from emer-gency call to either BLS-D or ALS procedure) of

the first responding unit was 8 min [IQR 6—10], andshorter in events with favourable outcome (6 min(4—7) versus 8 min (6—10) in unfavourable outcome;P < 0.001).

Pw(A

Table 4 Response times for bystander-witnessed subjectsdiac origin in relation to favourable outcome at discharge in

Response times Faou

Emergency call to EMS activation 2 (Emergency call to EMS arrival 5 (Time to BLS-D 6 (Time to ALS 9 (Time to shock (in subjects with shockable rhythms) 7 (

Data (min) were analyzed by Mann—Whitney test.

191 (45.0) 8.07 (4.03—16.17)

lysis.

The absence of a history of myocardial infarc-ion or congestive heart failure, defibrillation, a lowumber of defibrillation shocks, and lower responseimes were associated with favourable outcomeoth at discharge and at 1 year (Table 5). In addi-ion, age <74 years, and no history of hypertensionere associated with favourable outcome at dis-harge, whereas absence of diabetes, and male sexere associated with a favourable 1-year outcome

Table 5).With reference to response times, the chance of

favourable outcome decreased by 28% for everydditional minute from event to response (OR, 0.72;5% CI, 0.62—0.83).

The effects of AED implementation were testedy a separate analysis of the years 2002—2004,hen AEDs became operative. During this period,3 events were treated by AEDs, the rhythm provedhockable in 5, and 1 subject had a favourable out-ome. There was no systematic increase in the ratef AED use with years (chi-square test for trend,

= 0.668). The AED programme was not associatedith a favourable outcome at univariate analysis

OR, 1.18; 95% CI, 0.13—10.67). The fixed-locationEDs were used only in one case.

resuscitated from out-of-hospital cardiac arrest of car-the general health district of Forlı from 1994 to 2004

vourabletcome (N = 55)

Unfavourableoutcome (N = 424)

P-value

1—2) 2 (1—2) 0.2734—7) 7 (6—8) <0.0014—7) 8 (6—10) <0.0016—11) 10 (8—15) <0.0016—9) 10 (7—12) <0.001

Out-of-hospital cardiac arrest in a mixed urban and rural setting 185

Table 5 Results of logistic regression analysis of factors contributing to favourable outcome at discharge and1-year survival after out-of-hospital cardiac arrest

Estimatea Odds ratio 95% CI

Favourable outcome at dischargeDefibrillated 2.12 8.39 3.10—22.73Number of defibrillations <2 1.42 4.12 2.02—8.41Age <74 years 0.88 2.42 1.15—5.08Time to treatment −0.28 0.75 0.65—0.86History of myocardial infarction −0.91 0.40 0.16—1.00History of congestive heart failure −1.01 0.37 0.14—0.99History of hypertension −1.07 0.34 0.14—0.83

Favourable outcome at 1 yearDefibrillated 3.65 38.32 4.98—295.0Number of defibrillations <2 1.11 3.04 1.26—7.33Male sex 1.25 3.50 1.18—10.36Time to treatment −0.35 0.70 0.58—0.85History of diabetes mellitus −1.19 0.30 0.10—0.91History of congestive heart failure −2.79 0.06 0.01—0.50History of myocardial infarction −3.04 0.05 0.01—0.49

s dichotomous variables; 95% CI: 95% confidence intervals.s (b)

D

Iwicads

apis3hgros(cOIi

dsosro

ftrta1slrarp

j6fldprmrgvdas

All variables (excluding time to treatment) were considered aa The parameter estimate reports the estimated coefficient

iscussion

n the setting of a general Italian health districtith a mixed urban/rural population, cardiac arrest

ntervention produces a favourable outcome at dis-harge in 11.5% of resuscitated subjects and in 7.9%fter 1 year. These data fit with previous studies inifferent settings, and identify predicting variablesimilar to those observed in urban areas.

Large studies have long identified the main vari-bles predicting recovery and survival.13—15 Com-arisons are difficult, because of differences eithern patient selection or in outcome measures, withurvival at hospital discharge ranging from 3 to3%.15 We considered only patients dischargedome or to an acute rehabilitation facility withood recovery defined by OPC,12 having an accu-acy easy to determine. Our population was prettyld, reflecting health district demography, and 39ubjects were classified as OPC 2 before the eventno cases with OPC 3—4 were rescued). After resus-itation, 25 patients were classified as OPC 1, 30 asPC 2, 16 as OPC 3, and 2 were left in a coma state.

f all patients discharged alive from hospital werencluded, our survival rate would rise to 15.2%.

Our series confirm that time-to-treatment,efibrillation and the number of defibrillationhocks are predictors of survival, together with

ther variables previously identified in differenttudies (younger age,16 location of the event,17

esponse times of advanced rescuers18 and absencef comorbidities19).

aeIp

of the fitted logistic regression model.

Response times were a very critical issue, andavourable outcomes were strictly dependent onime-lag from event to shock. In a cohort study,educing the 90th percentile for response timeo 8 min increased the predicted survival to 8%,nd reducing it to 5 min increased survival to0—11%.20 We failed to reach the desired goal of ahort standard time-to-rescue limit in our series—aikely reason for the lower prevalence of shockablehythms compared to the literature,7 but the over-ll favourable outcome was similar to the highesteported standards, which also did not have an AEDrotocol.

Data on the benefits of AED programmes are sub-ect to criticism.8,21 In a widely cited study,22 only

subjects were saved by AEDs deployed on theeet of American Airlines, with 24,000 flight atten-ants trained and over 70 million passengers trans-orted. Two more studies in airports4 and casinos5

eported a survival rate of 30% after AED imple-entation in these very specific settings, but the

esults are overoptimistic in the ‘real world’ of aeneral district, and particularly in a rural setting. Aery recent report of a national English programmeemonstrated that only 172 cases were treated in4-year period using 681 AEDs distributed in 110

ites, and only 39 subjects were discharged alive.7

The additional benefits of AED programmes

re obviously dependent on the efficiency of thexisting EMS and the location of cardiac arrest.23

n communities without EMS early defibrillationrogrammes, survival rates averaged 5%, and spe-

of cases (with four cases dying of non-cardiacevents).

The limited benefit of the AED programme in thispreliminary analysis of our setting (one life saved in

2eEbrao

L

WIsifcctss

oneoae

cctaiaebtott

coti

C

Iihhce

A. Fabbri et al.

years) is a matter of concern. With a number ofvents of 50/year and a success rate of 11.5% withMS, an optimistic 20% increase in the recovery ratey an extended AED programme would theoreticallyeach a statistical significance after over 30 years,time span that is not compatible with the needs

f public health planners.

imitations

e aimed to collect data on resuscitation in antalian health district with a mixed urban/ruraletting. The results may be used as backgroundnformation and not for comparison between dif-erent organisation models. Very large, uniformlyollected databases are needed for drawing firmonclusions on the effects of AEDs, since fac-ors other than medical performance or organi-ation might explain the differences in terms ofurvival.

The cardiac origin of the arrest was derivednly on the judgment of the ambulance person-el, introducing a bias, which cannot be clinicallyliminated. An additional bias is due by the lackf consideration of any resuscitation procedurettempted by bystanders before the arrival of themergency personnel.

Another limitation is that our method of dataollection might have failed to capture all theardiac arrests that occurred during the study. Par-icularly important would be any failure to capturecase with neurologically intact survival. This lim-

tation is largely unlikely, since our district hospitaldmits all cases accepted by the ambulances of themergency service. The estimation of the time-lagetween collapse and the initiation of resusci-ation might be inaccurate, and the unreliabilityf such estimates was the basis for our decisiono use time intervals rather than to use absoluteimes.

Because the assignment of cerebral performanceategory scores by the study coordinator was basedn descriptions of patients’ neurological status inheir medical records, there may have been somenaccuracies.

onclusions

n a mixed urban-rural setting, typical of most Ital-an areas, where 74.5% of cardiac arrests occur at

186

cific programmes increased survival to 15—40%.12

In an Italian health district similar to our setting,survival rate to hospital discharge tripled from4.3% with EMS intervention to 15.5% with an AEDprogramme, with a full-neurology recovery rateincreasing from 2.4 to 8.4%,24 a figure similar tothat of our study. Therefore, the advantage of AEDcomes from districts with a very low survival ratewith EMS intervention only. In large cities, survivalranged from less than 4.0 to 32.0% in relation tothe population density.25 In our study a favourableoutcome was obtained in 63/241 subjects withventricular fibrillation (26%), the second best resultof cited studies.25 This may also explain the lackof any additional benefit of AED programme thatwe found.

The impact of AED programmes is based onthe assumptions that an extensive network pro-vides defibrillator access to all cases occurringin public places, that arrests are witnessed andrecognised, and that bystanders are willing andable to act immediately and appropriately. Only21% of cardiac arrests occurred in a public place inthe Goteborg study,23 with 17% of cases potentiallyamenable to be treated by AED-equipped units.In London, Davies et al. reported that the num-ber of arrests treated for deployed defibrillatorper year is on average 0.10, similar to our rateof 0.17. However, the percentage of shockablerhythms was much lower in our setting (38.5%versus 78.0%), probably as a consequence ofthe lag time, with an unfavourable effect onsurvival.

The geography and the demographic compositionof the community are an essential component of anAED success. Survival is influenced by ambulanceresponse time and distance, with 12 miles (19 km)being a cut off point.26 Subjects living at greaterdistance from ambulance dispatch points (e.g., ina rural setting) have a significantly lower survivalrate.27 In our series, 136 (28.4%) cases of cardiacarrest occurred in rural setting, without differ-ences of response times compared to the urbansetting.

Age is another important factor. The median ageof patients included in previous series ranged from5328 to 72 years.24 In older people, the AED pro-gramme might prolong life only temporarily, withpatients rapidly developing untreatable end-stageheart disease.29 The median age of our subjectswas as high as 74 years, but the recovery ratewas nonetheless favourable at 1 year in two-thirds

ome, in older persons used to an efficient EMS,ealth care planners must carefully consider theost of an AED implementation programme and itsxpected benefit in terms of life saved per year.

d ru

Gfi

N

C

N

A

Wp

R

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

2

Out-of-hospital cardiac arrest in a mixed urban an

rants or other financial support,nancial interest

one declared.

onflict of Interest

one declared.

cknowledgment

e are grateful to the Emergency Medical Systemersonnel for helpful support.

eferences

1. Guidelines for cardiopulmonary resuscitation and emergencycardiovascular care: an international consensus of science.Circulation 2000;200(Suppl. 8);102.

2. Weston CFM, Wilson RJ, Jones SD. Predicting survival fromout-of-hospital cardiac arrest: a multivariate analysis. Resus-citation 1997;34:27—34.

3. De Maio VJ, Stiell IG, Wells GA, Spaite DW. Optimal defibrilla-tion response intervals for maximum out-of-hospital cardiacarrest survival rates. Ann Emerg Med 2004;42:242—50.

4. Caffrey SL, Willoughby PJ, Pepe PE, Becker LB. Publicuse of automated external defibrillators. N Engl J Med2002;347:1242—7.

5. Valenzuela TD, Roe DJ, Nichol G, Clark LL, Spaite DW,Hardman RG. Outcomes of rapid defibrillation by secu-rity officers after cardiac arrest in casinos. N Engl J Med2000;343:1206—9.

6. Myerburg RJ, Fenster J, Velez M, et al. Impact of community-wide police car deployment of automated external defibril-lators on survival from out-of-hospital cardiac arrest. Circu-lation 2002;106:1058—64.

7. Davies CS, Colquhoun MC, Boyle R, Chamberlain DA. Anational programme for on-site defibrillation by lay peo-ple in selected high risk areas: initial results. Heart2005;91:1299—302.

8. Eisenberg M. Over-the-counter automated external defibril-lators? Show me the data! Ann Em Med 2005;45:96—7.

9. Cummins RO, Hazinski MF, Kerber RE, et al. Low-energybiphasic waveform defibrillation: evidence-based reviewapplied to emergency cardiovascular care guidelines: astatement for healthcare professionals from the AmericanHeart Association Committee on Emergency CardiovascularCare and the Subcommittees on Basic Life Support, AdvancedCardiac Life Support, and Pediatric Resuscitation. Circula-

tion 1998;97:1654—67.

0. Cummins RO, Chamberlain DA, Abramson NS, et al. Rec-ommended guidelines for uniform reporting of data fromout-of-hospital cardiac arrest: the Utstein style. Circulation1991;84:960—75.

2

ral setting 187

1. The Brain Resuscitation Clinical Trial II Study Group. Arandomized clinical trial of calcium entry blocker admin-istration to comatose survivors of cardiac arrest: design,methods, and patient characteristics. Control Clin Trials1991;12:525—45.

2. Bunch TJ, White RD, Gersh BJ, et al. Long-term outcomes ofout-of-hospital cardiac arrest after successful early defibril-lation. N Engl J Med 2003;348:2626—33.

3. Eisenberg MS, Hallstrom AP, Bergner L. The ACLS Score: pre-dicting survival from out-of-hospital cardiac arrest. JAMA1981;246:50—2.

4. Eisenberg MS, Horwood BT, Cummins RO, Reynolds-HaertleR, Hearne TR. Cardiac arrest and resuscitation: a tale of 29cities. Ann Emerg Med 1990;19:179—86.

5. Eisenberg MS, Mengert TJ. Cardiac resuscitation. N Engl JMed 2001;344:1304—13.

6. Tresch DD, Thakur RK. Cardiopulmonary resuscitation in theelderly: beneficial or an exercise in futility? Emerg Med ClinNorth Am 1998;16:649—63.

7. Litwin PE, Eisenberg MS, Hallstrom AP, Cummins RO. Thelocation of collapse and its effect on survival from cardiacarrest. Ann Emerg Med 1987;16:787—91.

8. Stiell IG, Wells GA, Field BJ, et al. Improved out of hospitalcardiac arrest surviving through the inexpensive optimiza-tion of an existing defibrillation program. OPALS Study PhaseII. JAMA 1999;281:1175—81.

9. Hallstrom AP, Cobb LA, Yu BH. Influence of comorbidity onthe outcome of patients treated for out-of-hospital ventric-ular fibrillation. Circulation 1996;93:2019—22.

0. Pell JP, Sirel JM, Marsden AK, Ford I, Walker NL, Cobbe SM.Potential impact of public access defibrillators on survivalafter out of hospital cardiopulmonary arrest: retrospectivecohort study. BMJ 2002;325:515.

1. Pell JP. The debate on public place defibrillators: chargedbut shockingly ill informed. Heart 2005;89:1375—6.

2. Page RL, Joglar JA, Kowal RC, et al. Use of automatedexternal defibrillators by a US airline. N Engl J Med2000;343:1210—6.

3. Engdahl J, Herlitz J. Localization of out-of-hospital cardiacarrest in Goteborg 1994—2002 and implications for publicaccess defibrillation. Resuscitation 2005;64:171—5.

4. Capucci A, Aschieri D, Piepoli MF, Bardy GH, IconomuE, Arvedi M. Tripling survival from sudden cardiacarrest via early defibrillation without traditional educa-tion in cardiopulmonary resuscitation. Circulation 2002;106:1065—70.

5. Eckstein M, Stratton SJ, Chan LS. Cardiac arrest resuscita-tion evaluation in Los Angeles: CARE-LA. Ann Emerg Med2005;45:504—9.

6. Lyon RM, Cobb SM, Bradley JM, Grubb NR. Surviving out ofhospital cardiac arrest at home: a postcode lottery? EmergMed J 2004;21:619—24.

7. Pell JP, Sirel JM, Marsden AK, Ford I, Cobbe SM. Effect ofreducing ambulance response times on deaths from out ofhospital cardiac arrest: cohort study. BMJ 2001;322:1385—8.

8. Hallstrom AP, Ornato JP, Weisfeldt M, et al. Public access

defibrillation trial investigators. Public-access defibrillationand survival after out-of-hospital cardiac arrest. N Engl JMed 2004;351:637—46.

9. Engdahl J. Outcome after cardiac arrest outside hospital.BMJ 2002;325:503—4.