Cardiac Arrest Registries

Cardiac arrest registries can provide valuable information on the demographics of those whosuffer a cardiac emergency, on the circumstances of the collapse, on the emergency responseand on outcomes. Registry data can estimate the likely incidence of cardiac arrests requiringresponse by the EMS. Registries however underestimate the total number of SCDs as they donot include arrests that were not witnessed. Neither do they include events for which nocardiac arrest is called, for example in sudden collapse of someone with other seriousmorbidity.

Death certificate data and data compiled by the EMS were analysed for the State ofWashington in the year 2000.17 There were 3,577 deaths from cardiac disease in thispopulation of nearly 2 million. The EMS responded to 39% of all heart disease deaths,representing 57% of out-of-hospital events (including 128 people who were successfullyresuscitated and discharged from hospital).

Using multiple sources of information, SCD incidence was 53 per 100 000 (median [central]age 69 years, 57% male) in 2002 in Multnomah County in Oregon, representing 6% ofannual mortality.18 Resuscitation was attempted in two-thirds of cases and 8% survived tohospital discharge. One in five cases was identified using sources other than first responders.

A retrospective analysis of death certificates for 2002 using ICD-10 codes and location ofdeath gave an estimated SCD rate of 153 per 100,000 (median age 81, 51% male) whichwould have greatly overestimated the required EMS response.

The Helsinki Cardiac Arrest Registry reported an incidence of out-of-hospital cardiac arrest of 80 per 100,000 inhabitants per year between 1994 and 1999.19 A similar rate (9.2 / 10,000)was reported for 1997 – 2000 by the Maastricht Circulatory Arrest Registry.20 Sudden deathrepresented 19% of all deaths.

If, as in Washington State, the EMS in Ireland respond to 39% of all cardiac deaths, thiswould have corresponded to 3,169 such deaths in 2002, or 81 per 100,000 population, similarto rates reported by other North European countries.

There have been several studies of the circumstances of sudden collapse. The Belfast study ofout-of-hospital deaths in 2003 – 2004 used multiple sources of information to identifySCDs.

There were 297 such deaths in 12 months, 66% of which were in men. The meanage was 68 years (65 in men and 72 in women). The majority, 78%, occurred at home, 16%were in a public place and the remainder in a nursing home; 93% were attended by theEMS. The mean ‘call to response’ interval was 8 minutes and 27% were in ventricularfibrillation (VF) when the EMS arrived. This was higher (53%) in witnessed arrests. In thoseattended by the EMS, nearly one in ten were resuscitated to reach hospital alive and 7%survived to hospital discharge.

Several studies have examined the factors associated with successful resuscitation. Whenresuscitation was attempted in the Helsinki Registry, fifty-seven patients (17%) survived todischarge.20 Nearly one-third survived when collapse was bystander witnessed and thediagnosis was cardiac arrest with VF as the initial rhythm.

The Swedish Cardiac Arrest Registry found the following factors influenced the likelihood ofsurvival at 1 month: an initial rhythm of VF, shorter time to arrival by the EMS (below themedian i.e. better than the response time in the middle when ordered sequentially), arrestoccurring outside the home, arrest which was witnessed, one where bystander CPR wasinitiated and in younger (below the median) people.

In Seattle, Washington, the annual incidence of cardiac arrest with VF as the first identified rhythm decreased from 0.85 per 1,000 population in 1980 to 0.38 per 1,000 in the year2000.The Swedish Cardiac Arrest Registry reported that between 1992 and 2003 the meanage of out-of-hospital cardiac arrests increased from 68 to 70 years and percentage offemales increased from 29% to 32%.25 There was a decrease in cases of cardiac origin from75% to 61% and a decrease from 36% to 25% of arrests with VF as the initial rhythm. Thesechanges reflect the changing patterns of mortality in many developed countries, withdecreasing death rates and a smaller proportion of deaths being attributed to cardiovasculardisease.

Some studies of cardiac arrest use data from cardiac arrest registries; others add supplementary data from other sources. The case definition for registration may vary, toinclude all cases of sudden collapse or only those cases where resuscitation was attempted.The International Resuscitation Network registry aims to study emergency response andoutcomes.
Audit of emergency response in Ireland will be facilitated by the establishment of a cardiac arrest registry, with data collection to international standards. Chapter 6 of this report considers information systems and surveillance, and makes recommendations toimprove the collection of data relevant to SCD in Ireland.

Post-mortem studies

Several retrospective autopsy (post-mortem) series have been reported. These do not include SCDs where a general practitioner (GP) certified (without a post-mortem) a patient with known CHD who died suddenly as having died, for example, of a ‘myocardial infarct’. This is usually done with the knowledge of the coroner.

The Report of the Task Force on Sudden Cardiac Death Chapter 2

In one series in the UK, 692 SCD autopsy results were analysed for white Caucasians between 16 and 64 years of age with no history of cardiac disease. Eighty-two per cent of deaths were due to CHD and 12% had other cardiac pathology. Of the remainder, 2% had a history of some other condition but in 4% no cause of death could be identified. The annual rate of SCD (including unidentified causes) in this population was 11 per 100,000.

A study in Belfast identified SCDs from EMS and autopsy reports. There were autopsy reports for 37% of cases.8 In the autopsy group, 60% had no known history of cardiac disease but severe coronary artery disease was found in 94%.

Nearly 2,000 reports were analysed of post-mortems at Connolly Hospital, Dublin between January 1987 and December 2001.9 Of these, 777 (69% male, 31% female) were performed because of sudden, unexpected, out-of-hospital death occurring within 1 hour of symptom onset. The underlying cause of death was cardiac in 608 subjects (78%). In these, a possible mechanical cause of death was found in 28%, therefore an arrhythmia was the presumed mechanism of death in 72%.


There has been particular interest in the findings at post-mortem of SCD in young people. In a retrospective series looking at 72 SCDs under 35 years of age in Dublin between 1993 and 2002, death was most frequently attributed to CHD, followed by hypertrophic cardiomyopathy (HCM) (24% of deaths). Patients with congenital heart disease and those younger than 10 years of age were excluded from this study.

Between 1992 and 1999 the Swedish national database of forensic medicine recorded 181 cases of SCD in the 15 to 35 year age group, 73% male and 27% female.11 The incidence rate of 0.93 / 100,000 was stable over time. A cardiomyopathy was found in 23%, CHD in 18% and myocarditis in 11%. No structural cardiac abnormality was found in 21%.

An autopsy series of 193 cases of SCD aged 35 or younger in Sydney, Australia between 1994 and 2002 found CHD in 24%, cardiomyopathy in 15% and myocarditis in 12%. No structural abnormality was found in 31% and death was presumed to be due to a disorder leading to a fatal cardiac arrhythmia. Approximately 5% of SCDs are unexplained, and no structural abnormality is found at postmortem.

The proportion rises to 20 – 30% in those under the age of 35 years. The percentage in whom no structural abnormality is found is higher in young women than in young men, 50% compared to 24% in one series of cases in the 35 to 44 year age group in the US.14 In the 28% of cases in whom no structural abnormality was found in an Italian series of cases aged 35 or younger, two-thirds were men and one-third women. A genetic factor can be identified in only a small proportion of cases at present but it is expected that this will increase with the identification of further gene defects.

The Royal College of Pathologists has published guidelines on autopsy practice in sudden
death with likely cardiac pathology.16 The contribution of specialist cardiac pathologists to
the surveillance of SCD in Ireland is discussed in Section 6.2 of this report.

Death certificate data

In the United States an analysis of death certificates estimated that the overall SCD rate in 1999 was 3 per 100,000 from birth to age 34 (3,976 deaths), 75 per 100,000 (78,456 deaths) in the 35 to 64 year age group and 1,100 per 100,000 (379,869 deaths) in those aged 65 years or older.4 Of all SCDs in 1998 in those aged 35 or over, CHD was the underlying cause on 62% of death certificates.

The analyses in the US found that 63% of all cardiac deaths were SCDs, occurring out-ofhospital (47%), or in emergency departments or ‘dead on arrival’ (16%), with almost all of the remaining cardiac deaths occurring as non-sudden deaths in hospital.4 SCD accounted for 74% of all cardiac deaths in the 35 to 44 year age group, 73% in those 45 to 54, 66% in those 55 to 64, 58% in those 65 to 74 and 69% in those aged 85 or over.5 Between 1989 and 1998, rates of SCD in those aged over 35 years decreased by 12% in men and by 6% in women.

According to the CSO, 11,652 people died of diseases of the circulatory system in 2002 in Ireland (10,608 in 2004), including 8,126 deaths attributed to cardiac causes.* Assuming that, as in the United States, 63% of cardiac deaths are SCDs, this would translate into 5,119 such deaths in Ireland in 2002. Apart from this method of estimation, it is not possible at present to obtain the number of SCDs from Irish vital statistics reports.

In Ireland data on all deaths is collected by the Central Statistics Office (CSO), on the basis of medical certificates and death registration information collected by local Registrars’ Offices or on the basis of Coroners’ report forms. Death rates for a particular condition are estimated from the information recorded on death certificates and subsequently coded by the CSO. This information is subject to inaccuracies for a variety of reasons. For many causes of death the data are sufficiently accurate for the purposes of identifying high risk groups, tracking trends and planning health services. The nature of SCD makes it more subject to misclassification than other causes of death.

Currently the CSO is using the Ninth Revision of the International Classification of Diseases (ICD). Code 410 (acute myocardial infarction) is the most frequently used code when SCD occurs. The code for ‘cardiac arrest’ or ‘cardiorespiratory arrest’, code 427 is not used in this country, though 463 deaths in 2002 were coded as 427, ‘cardiac dysrhythmias’. Code 798, sudden death, cause unknown, was used for 48 deaths in that year.
The CSO has informed the Task Force that it plans to introduce ICD Version 10 for coding cause of death. This version has a specific code for Sudden Cardiac Death (146.1). Relevant ICD-9 and ICD-10 codes are listed in Appendix 3.

Attempting to quantify the incidence of SCD without a specific code is unsatisfactory. Instructions to those completing a death certificate do not allow use of the term SCD. Instructions on death certificates in Ireland forbid recording the mode of death, such as cardiac arrest. Instead, the certifier must state the underlying cause e.g. acute myocardial infarction (AMI).

For the purposes of completing death certificates from which national statistics are compiled, SCD is therefore coded according to its likely underlying cause, including AMI, coronary thrombosis, angina, heart failure, pulmonary oedema (fluid in the lungs), ruptured aortic aneurysm (a split in the wall of the main blood vessel from the heart), or myocarditis (infection of the muscle of the heart). A post-mortem is not carried out in all cases and so a ‘best guess’ may be made as to the underlying cause of death. Even where a post-mortem is carried out, there may be difficulty in identifying the cause of death.

With the introduction of ICD-10 it will be possible to estimate the total number of SCDs from death certificate data. In calculating the number of SCD cases it will be important to exclude victims who were not actively resuscitated because of end stage disease. These cases are not relevant to planning emergency medical services in the community.

A number of countries such as the USA, France and Italy, now include a ‘tick box’ to record cardiac arrest on the death certificate. Similar information for this country would contribute to planning and evaluation of pre-hospital and hospital care of cardiac arrest, as well as the prevention of SCD. Death certificate data will not include those who survive a cardiac arrest. Although this number is currently small, it is likely to grow over the coming years. Such data could be collected in a cardiac arrest registry (see Sections 2.2.3 and 6.4).

Epidemiology of SCD

SCD is defined as death due to natural causes within an hour of the onset of symptoms, in
the absence of any other cause, and assumed or proven to have a cardiac cause.2,3 Preexisting
heart disease may have been known to be present but the time and mode of death
are unexpected. The death may be witnessed or not witnessed.

There are three sources of information on the incidence of SCD and its underlying causes :

• death certificate and death register data
• post-mortem studies, and
• cardiac arrest registries.

Death certificate data is collected routinely and has the potential to provide information on
trends over time and to facilitate international comparisons. Analysis of post-mortem results
can provide more detailed information on the underlying pathology. Cardiac arrest registries
provide information on the circumstances of the collapse and the underlying heart rhythm,
on the emergency response and on outcomes.