Characteristics of sudden death by clinical criteria

Abstract

Sudden death is a leading cause of deaths nationally. Definitions of sudden death vary greatly, resulting in imprecise estimates of its frequency and incomplete knowledge of its risk factors. The degree to which time-based and coronary artery disease (CAD) criteria impacts estimates of sudden death frequency and risk factors is unknown. Here, we apply these criteria to a registry of all-cause sudden death to assess its impact on sudden death frequency and risk factors. The sudden unexpected death in North Carolina (SUDDEN) project is a registry of out of-hospital, adjudicated, sudden unexpected deaths attended by Emergency Medical Services. Deaths were not excluded by time since last seen or alive or by prior symptoms or diagnosis of CAD. Common criteria for sudden death based on time since last seen alive (both 24 hours and 1 hour) and prior diagnosis of CAD were applied to the SUDDEN case registry. The proportion of cases satisfying each of the 4 criteria was calculated. Characteristics of victims within each restrictive set of criteria were measured and compared to the SUDDEN registry. There were 296 qualifying sudden deaths. Application of 24 hour and 1 hour timing criteria compared to no timing criteria reduced cases by 25.0% and 69.6%, respectively. Addition of CAD criteria to each timing criterion further reduced qualifying cases, for a total reduction of 81.8% and 90.5%, respectively. However, characteristics among victims meeting restrictive criteria remained similar to the unrestricted population. Timing and CAD criteria dramatically reduces estimates of the number of sudden deaths without significantly impacting victim characteristics.

1. Introduction

Sudden death is a leading cause of mortality in the United States with an incidence estimated between 150,000 and 450,000^[1–5] deaths annually. The variability among criteria used to define sudden death may contribute to imprecision in the number of lives affected and limit our understanding of the social and clinical risk factors for sudden death. Importantly, there is no single set of criteria to establish sudden death. A commonly cited criteria is the World Health Organization definition, which describes such deaths as sudden, unexpected, natural deaths either witnessed, and within 1 hour of symptom onset or, if unwitnessed, within 24 hours of having been last seen alive and symptom-free.^[6] Others define sudden death as being seen <1 hour prior plus evidence of coronary artery disease (CAD),^[7] and others require attempted resuscitation of victims.^[4] These diverse criteria contribute to confusion about what constitutes a sudden death. Specifically, it is not known if restrictive criteria for sudden death underestimates the number of victims or biases estimates of the social and clinical risk factors for sudden death.

To address these questions, we assessed changes in the number of sudden death victims and their associated social and clinical characteristics by commonly accepted criteria within a population sample of all cause sudden death victims.

2. Methods

The sudden unexpected death in North Carolina project is a registry of 399 out of hospital sudden death cases among adults aged 18 to 64 within Wake County, North Carolina, attended by emergency medical services (EMS) from 2013 to 2015 and adjudicated by cardiologists. The study was designed to make no assumptions of the cause of death, cardiac or otherwise. Detailed screening and adjudication methods have been described previously.^[8] Briefly, possible cases were identified through Wake County EMS documentation. Medical records were obtained for cases of concern for cardiac arrest based on provider impression or treatment with external defibrillation, cardiopulmonary resuscitation, or manual defibrillation. Potential cases were excluded if they were over 64 or <18 years of age, survived transportation to the hospital, experienced an expected death from clinical disease, were not independently living, died of obvious noncardiac etiology such as trauma or overdose, or were not a resident of Wake County. No criteria on causality or time since last free of symptoms were imposed as inclusion requirements for the registry. Potential cases were reviewed by a panel of 10 cardiologists, and if a consensus was reached by 3 panel members, the deaths were deemed expected. We excluded cases with unavailable medical records (n = 28), and cases with a history of chronic kidney disease or heart failure (n = 75) leaving 296 sudden deaths for analysis. Additionally, there were 25 deaths with unknown time of last known well. These were excluded from risk factor analysis. However, these cases were included when considering the proportion of sudden deaths meeting timing and CAD criteria and were considered not to meet time-based criteria since their time of last known well was unknown (Fig. 1). Cutoffs for the time last known alive and free of symptoms and presence of CAD were derived from published criteria for sudden death in the World health organization and from the Atherosclerosis Risk in Communities study, estimated by the time last known well of <24 hours and the presence of CAD^[6] and as time last known well of <1 hour and the presence of CAD,^[7] respectively. Timing criteria alone (last known well of <1 hour; last known well of <24 hours) without requiring CAD are also presented.

Figure 1.

Details of case ascertainment and identification of sudden deaths in Wake County, North Carolina. EMS = emergency medical services, SUDDEN = sudden unexpected death in North Carolina.

Demographic risk factors related to sudden death included age, sex, race, marital status, high school graduation status, 2020 census tract rurality, and median income. Populations with < 20% living in an urbanized area were categorized as highly rural, 20% to 60% urban were designated as mixed rural-urban, and > 60% urban were designated as highly urban as described in previous methods.^[9] Clinical risk factors included were CAD, hypertension, diabetes mellitus, hyperlipidemia, chronic respiratory disease, smoking history, mental illness, and alcohol and substance misuse disorders. Presence of CAD was determined from available medical records and by autopsy. Both current and former smokers were considered to have a positive smoking history. Mental illness included a composite of substance misuse disorder, depression, anxiety, schizophrenia, or bipolar disorder. The comparison of risk factors between 2 groups was implemented in R version 4.0.2 (https://www.r-project.org) using a 2-sample t test for continuous variables and Fisher exact test for categorical variables. Each combination of criteria (timing <24 hours, timing <1 hour, both timing <24 hours and CAD, and both timing <1 hour and CAD) were compared to the remaining cases in the overall study population (no timing or CAD criteria). For example, of the 271 cases included in risk factor analysis, 90 met criteria for timing <1 hour. The risk factors among these 90 cases were compared to the remaining 181 cases in the overall study population that did not meet criteria. This was then repeated for each set of criteria.

3. Results

Of 296 study victims, 222 (75%) victims had a time last known well of <24 hours, and 90 (30.4%) of <1 hour. Sixty-eight (23%) individuals had CAD, of which 40 were identified by autopsy and 28 identified by review of medical records. In addition, 54 (18.4%) individuals had CAD and a time last known well <24 hours, and 28 (9.5%) had CAD and a time last known well <1 hour (Fig. 2).

Figure 2.

Reduction in sudden deaths by time-based criteria and prior diagnosis of coronary artery disease. The SUDDEN registry, with no time-based criteria or prior CAD diagnosis, includes 296 sudden deaths. Timing criteria of <24 hours and <1 hour refer to when a victim was last seen alive and well prior to sudden death. Implementation of each criterion results in a percent reduction of sudden deaths compared to no criteria within the SUDDEN registry. CAD = coronary artery disease, SUDDEN = sudden unexpected death in North Carolina.

Time-based criteria for sudden death substantially reduced available cases, by 25.0% for time <24 hours and by 69.6% for time <1 hour. Utilizing criteria for CAD reduced cases by 77.0% for no time criteria and, when added to time criteria, by 81.8% and 90.5% for 24 hours and 1 hour, respectively (Fig. 2). Despite major reductions in available cases, few differences in demographic, social, and clinical characteristics emerged between these restricted samples and the overall study population (Table 1). For example, when comparing the most restrictive criteria group (timing <1 hour and coronary artery disease; n = 28) to remaining cases, the more restrictive group had a younger mean age, higher mean income, and experienced less substance misuse, but only the difference in substance misuse was statistically significant (P = .037).

Table 1.

Demographic and clinical risk factors for sudden death by time and coronary artery disease criteria in Wake County 2013–2015.

	No timing or CAD criteria	Timing <24 h		Timing <24 h and CAD			Timing <1 h			Timing <1 h and CAD
	n = 271*	n = 222	P value	n = 54	P value	n = 90		P value	n = 28		P value
Age, mean (SD)	52.1 (9.6)	51.6 (9.8)	.054	49.0 (10.4)	.015	52.0 (9.9)		.96	48.8 (9.1)		.051
White, n (%)	185 (68.3)	147 (66.2)	.13	37 (68.5)	1	57 (63.3)		.27	19 (67.9)		1
Male, n (%)	188 (69.4)	152 (68.5)	.61	39 (72.2)	.74	65 (72.2)		.49	22 (78.6)		.39
Married, n (%)	101 (37.3)	82 (36.9)	.87	20 (37.0)	1	32 (35.6)		.69	11 (39.3)		.84
Non High school graduate, n (%)	42 (15.5)	35 (15.8)	.13	8 (14.8)	1	17 (18.9)		.29	5 (17.9)		.78
Census tract†, n (%)
Highly rural (< 20% urban)	7 (2.6)	5 (2.3)	.77	1 (1.9)	1	1 (1.1)		.57	1 (3.6)		.76
Rural-urban mix (20–60% urban)	24 (8.9)	20 (9.0)		5 (9.3)		7 (7.8)			2 (7.1)
Highly urban (> 60% urban)	240 (88.5)	197 (88.7)		48 (88.9)		82 (91.1)			25 (89.3)
Income, mean (SD)	62.9 (24.9)	63.9 (24.8)	.18	64.6 (28.8)	.62	65.9 (26.8)		.18	71.6 (34.1)		.099
Coronary artery disease‡, n (%)	61 (22.5)	54 (24.3)	.25	N/A			28 (31.1)	.021	N/A
Hypertension, n (%)	135 (49.8)	116 (52.3)	.11	29 (53.7)	.55	53 (58.9)		.04	15 (53.6)		.69
Diabetes, n (%)	59 (21.8)	50 (22.5)	.57	12 (22.2)	1	22 (24.4)		.53	5 (17.9)		.81
Hyperlipidemia, n (%)	83 (30.6)	74 (33.3)	.041	20 (37.0)	.25	37 (41.1)		.012	11 (39.3)		.29
Chronic respiratory disease, n (%)	80 (29.5)	66 (29.7)	1	15 (27.8)	.87	28 (31.1)		.78	7 (25.0)		.67
Smoker, n (%)	106 (39.1)	89 (40.1)	.52	27 (50.0)	.086	42 (46.7)		.086	15 (53.6)		.11
Mental illness§, n (%)	153 (56.5)	122 (55.0)	.34	26 (48.1)	.17	45 (50.0)		.15	12 (42.9)		.16
All substance misuse, n (%)	100 (36.9)	75 (33.8)	.033	16 (29.6)	.27	27 (30.0)		.11	5 (17.9)		.037
Alcohol misuse, n (%)	71 (26.2)	48 (21.6)	<.001	9 (16.7)	.085	17 (18.9)		.057	3 (10.7)		.067

P values represent comparison of each criteria subgroup to cohort with no timing or CAD criteria.

CAD = coronary artery disease, SD = standard deviation.

^*Twenty-five cases were excluded from the total 296 sudden death cases due to unknown time of last known well.

^†Urban-rural categories derived from 2010 census describing percent population living in urbanized area. Income estimated using median income data from 2010 census tract and is reported in thousands of US dollars.

^‡CAD determined by autopsy for 40 victims and by medical records for 28 victims. Seven of these victims were excluded from this analysis due to an unknown time of last known well.

^§Medical history of depression, anxiety, bipolar disorder, schizophrenia, or substance misuse were considered to have a positive history of mental illness.

There were no major differences in demographic, social, or clinical characteristics between the included 271 victims with known time last well and the excluded 25 victims with unknown time last known well, beyond more victims with diabetes (21.8% vs 36.0%, P = .12) and hypertension (49.8% vs 76.0%, P = .014) in the excluded victims.

4. Discussion

Stringent timing and CAD criteria dramatically reduces the number of deaths classified as sudden death and did not greatly change the demographic, social, or clinical characteristics of sudden deaths. Our findings are consistent with previous studies showing time-based definitions influencing the epidemiology of sudden death. Specifically, a 24 hour timing cutoff increases the proportion of natural death classified as sudden, but has the potential to exclude natural deaths that should be studied as sudden deaths.^[10] Additionally, previous studies have demonstrated that increasing the time cut off for sudden death from 1 hour to 24 hours increases the percentage of natural deaths attributed to sudden death from 12% to 18.5%.^[10,11] Our study demonstrates that removal of the 24 hour cutoff increases the number of cases of sudden death by approximately 33% without a major change to population characteristics. As such, more inclusive criteria for sudden death would provide a sample size sufficient for further epidemiologic study and intervention.

The traditional dominant risk factor for sudden death is coronary artery disease. However, we have found, as have others,^[12,13] that coronary artery disease is less common in the modern era than anticipated.^[14] Other risk factors are common and include hypertension,^[15,16] left ventricular hypertrophy,^[16,17] diabetes,^[18] metabolic syndrome,^[19] chronic respiratory disease,^[20,21] mental illness,^[22,23] and substance misuse.^[22] These risk factors are consistent with pro-arrhythmic effects leading to ventricular fibrillation.^[24,25] For example, alcohol misuse may lead to arrhythmia and sudden death.^[26,27] Our results support that among high confidence sudden death victims included by restrictive timing criteria of <1 hour (n = 90), requiring coronary artery disease for classification of sudden death fails to capture approximately 70% of victims. Additionally, obtaining medical records to prove a history of CAD to satisfy criteria necessary for inclusion as sudden death data can be difficult, and may unnecessarily exclude cases from study. In the present study, we excluded 28 cases, or roughly 10% of the final number of cases, due to missing records. Therefore, broad criteria for sudden death, not restricted by the assumption that coronary artery disease is its main cause, expands the number of vulnerable adults, men, African Americans, and individuals living in poverty and apparent social isolation,^[9,28] allowing a broader study of risk factors and true incidence.

Furthermore, these results have important potential implications for clinicians. Our results suggest sudden death may be more prevalent than currently thought, and clinicians should be aware it is common and has the potential to impact their patients. This awareness is particularly important because sudden death victims access healthcare with increasing frequency in the 6 months prior to their death, often in the outpatient setting and to address common chronic conditions that are risk factors for sudden death.^[29] As our results show, these risk factors extend beyond traditional cardiac risk factors such as CAD. The absence of CAD does not exclude the possibility of sudden death in a patient. Clinicians should be aware that other medical and social characteristics confer increased risk for sudden death, particularly those that can be modified such as alcohol misuse and social isolation. Further study is needed to understand how targeting these factors may prevent sudden death.

Our study has limitations. Information regarding timing of last known well was abstracted from death certificates and EMS reports and is subject to error. EMS reports may have inaccurate times since last known well due to emergent clinical care needs taking precedence over detailed history-taking, or inaccuracies in history from distraught bystanders/family. Additionally, while Wake County is a racially and socioeconomically diverse county, our findings were drawn from a single county over a 2 year period, and may not be generalizable to other times and populations. Our small study sample and missing medical records may limit the detection of differences between groups. Additionally, our study is not a comprehensive exploration of every risk factor for sudden death, and there may be other important risk factors not explored here. One such example is obesity.

The criteria for sudden death are confusing and variable. Such variability in criteria complicates efforts to study sudden death by limiting sample size and impeding comparisons across studies. Timing of last seen alive and free of symptoms and the presence of coronary artery disease appear to be unnecessary for the definition of sudden death. These restrictive criteria may result in artificially low estimates of the incidence of sudden death, minimizing its importance as a public health issue and inhibiting the study of important risk factors outside of traditional cardiac factors. Avoiding these unnecessary criteria simplifies classification, allows a more accurate estimate of sudden death incidence and risk factors, and may facilitate the development of interventions to prevent sudden death, particularly among vulnerable populations.

Acknowledgements

The SUDDEN project is supported in part by funding from the NC TraCS grant (UL1TR002489). The SUDDEN project has been approved by the UNC Institutionl Review Board and is reviewed yearly. The Wake County EMS Data System supports, maintains, and monitors EMS service delivery, patient care, and disaster preparedness for the Wake County, NC community at large. This manuscript has been reviewed by Wake County EMS Data System investigators for scientific content and consistency of data interpretation with previous Wake County EMS Data System publications.

Author contributions

Conceptualization: Christopher Sefton, Susan Keen, Caroline Tybout, Golsa Joodi, Jefferson G Williams, Ross J Simpson Jr.

Data curation: Feng-Chang Lin, Huijun Jiang.

Formal analysis: Christopher Sefton, Susan Keen, Feng-Chang Lin, Huijun Jiang.

Investigation: Christopher Sefton, Caroline Tybout, Ross J Simpson Jr.

Methodology: Christopher Sefton, Susan Keen, Feng-Chang Lin, Ross J Simpson Jr.

Resources: Ross J Simpson Jr.

Supervision: Susan Keen, Ross J Simpson Jr.

Validation: Christopher Sefton, Susan Keen, Ross J Simpson Jr.

Writing – original draft: Christopher Sefton, Ross J Simpson Jr.

Writing – review & editing: Christopher Sefton, Susan Keen, Caroline Tybout, Feng-Chang Lin, Golsa Joodi, Jefferson G Williams, Ross J Simpson Jr.