Epidemiological profiles and causes of sudden deaths of various ages in Ethiopia: an autopsy-based study

Introduction

Sudden death is an important global public health issue, especially when it occurs in apparently healthy individuals due to the bearing these losses have on individuals, families, communities and wider society. The definition of sudden death varies according to authorities and conventions.¹ According to the World Health Organization (WHO), death is considered a sudden death (SD) when a nonviolent and unexpected death occurs within 24 hours after the commencement of a fatal illness.² The frequency and pattern of sudden death in different parts of the world vary greatly due to the diversity of prevalent diseases in various nations and due to diverse genetic and environmental factors.^3,4

Sudden unexpected natural death is often the initial clinical appearance of an underlying condition in people who had previously been asymptomatic and seemingly healthy.^3,5 Furthermore, since many of these deaths are unwitnessed and unattended by clinicians, the extent of the problem is difficult to determine. In this circumstance, an autopsy is an initial and sole opportunity to determine and document the precise cause of death.^1,4,6 In Ethiopia, SD is regarded as a type of medicolegal death for which an autopsy is required to determine the cause.

Public health authorities, therefore, depend heavily on vital statistics systems with a focus on the cause of death (COD) for formulating their programs, both curative and preventive specialist services. Mortality statistics are an essential component of the nation's vital statistics system. They are the ingredients to measure population growth and provide a demographic perspective for the planning of development in social services. Cause-specific mortality rates are crucial measures of population health trends. There are several challenges in interpreting epidemiological information in resource-limited settings, including a lack of uniformity and low quality of death certificates as well as the utilization of verbal autopsies.^6–8 Thus, vital health information may not reach the national registry, which could prevent the implementation of crucial interventions and prevention measures. Accurate evidence on causes of death is necessary to recognize the general epidemiological profile of diseases in Ethiopia and to support decision-makers in prioritizing the public health agenda. An autopsy is an important source of epidemiological data, as the considerable causes of SD remain hermetic until postmortem examination.¹ Therefore, epidemiological autopsy-based data on sudden death causes are vital to understand the characteristics of the affected population and customize public health programs. To our knowledge, this is the first study paper on SD in Ethiopia and is devoted to documenting the sociodemographic, behavioral, clinical and pathological characteristics of sudden death victims of various ages.

Methods

This is an observational, prospective, descriptive study that included all sudden unexpected natural cases observed from July 1, 2019, to June 30, 2020, in the Department of Forensic Medicine and Toxicology of St. Paul’s Hospital and Millennium Medical College (SPHMMC), Addis Ababa, Ethiopia. SPHMMC is a tertiary hospital and medical college governed by a board under the Federal Ministry of Health. SPHMMC in Addis Ababa and Aider Hospital in the Tigray region were the only two centers in the country providing forensic medicine and toxicology services during the study period. The Department of Forensic Medicine and Toxicology, SPHMMC, offers postmortem services to almost all regions of Ethiopia, except cases from the Tigray region served by Aider Hospital.

This study included all SD cases (according to the WHO definition) brought for autopsy from July 1, 2019, to June 30, 2020. The WHO defines SD as death that occurs within 24 hours after the disease’s commencement in a person not known to have been diagnosed with a serious disease, accident, or poisoning.² In consideration of this definition, two factors were considered: 1) the time of onset of terminal signs or symptoms until death and (2) the expectation of death at the time of occurrence. The definition of the onset of terminal signs or symptoms used in this study was the time when an individual had to change his activity because of the illness. The time of death was defined as the time an individual was pronounced dead. For unwitnessed deaths in which an individual was known to have been alive within 24 hours before the pronounced time of death, they were considered sudden death. The second factor in deciding whether or not the death was sudden was the unexpected nature of the occurrence. The degree of disability reported before death was used as a measure of the expectation of death at the time of the event. Individuals who died but were not confined to their homes, health facilities, or other related institutions due to illness for more than 24 hours before death were not considered unexpected deaths and were not included.

Principal investigators and five well-trained medical doctors collected the data. Information on the biodata of the victims, clinical data and circumstances of the death was collected from all potential sources, such as police files, medical records, and direct interviews with the eyewitnesses, relatives, and friends of the victims. The duration of hospital stay before death was extracted from the hospital record, autopsy referral papers, and police requests.

Full postmortem examinations were carried out in each instance using the Letulle evisceration procedure, systematically inspecting all cavities, including the cranial, cervical, thoracic, and abdominal cavities. Organs were dissected, examined and checked for signs of gross pathological changes and violence. Autopsied specimens (whole organs or organ pieces) were fixed in 10% formalin for histopathological examination. Multiple sections with thicknesses of 4-5 mm were taken. The tissues were processed, subjected to paraffin sectioning at a thickness of 4 micrometers and then stained using hematoxylin and eosin staining.

A structured data collection form was used to obtain all relevant clinical, epidemiological and pathological data. The predesigned data collection form was pretested in 50 cases to maintain data quality, and the necessary amendment was made to the form before the actual data collection. Data entry was performed using Microsoft Excel 2016 and exported to the Statistical Package for Social Science (SPSS window version 26) for analysis. Descriptive summary measures were used to characterize sociodemographic, behavioral, clinical characteristics and circumstances of death. A chi-square test was used to compare categorical variables, and the level of significance was set at a p-value <0.05.

Ethics approval was obtained from the SPHMMC Institutional Review Board (Ethical clearance reference no: PM 23/188). The study was conducted as per the Declaration of Helsinki. All information was treated anonymously and confidentially using unique identification codes rather than individual names and identifications.

Results

Epidemiological characteristics

A total of 4,942 medicolegal autopsies were performed during the study period, of which 568 cases were due to SD, accounting for 11.5% of the total autopsied cases. The youngest case was a 1-day-old male newborn, and the oldest was a 98-year-old man, with a mean age of 44.81±17.349. Males (n = 482/586) were predominant over females (n = 86/586) at a ratio of 5.6:1.²⁷

The maximum number of sudden deaths (24.3%) was in the age group of 31-40 years, followed by 41-50 years, which represents 19.6% of all SD cases. SD was less prevalent in extreme age groups, less than 10 years (1.8%) and over 70 years (6.7%). Males outnumbered females in all age groups. The fourth decade was the most predominant age group seen in males and females, accounting for 23.9% and 26.7% of total sex-specific cases, respectively. The number of cases by age and sex is summarized in Figure 1.

Figure 1. Age and sex distribution of sudden death cases (n=568).

More than two-thirds of the victims of SD (70%) belonged to urban areas, whereas 30% of victims were from rural areas. There was a statistically significant difference between the residence of the victims and sex (p<0.001). Most of the victims were from Addis Ababa city, the capital city of Ethiopia, accounting for 59.2% (n=336) of cases, followed by the Oromia region (23.6%), and the remaining 17.2% were from six other regions of the country. Table 1 summarizes the sociodemographic characteristics of the victims of sudden death.

Table 1. Summary of sociodemographic characteristics of sudden death victims.

Sociodemographic characteristics	Category	Frequency	Percentage
Sex	Male	482	84.9
	Female	86	15.1
Residence	Urban	397	70.0
	Rural	171	30.0
Region/Administrative city	Addis Ababa	336	59.2
	Oromia	134	23.6
	SNNPRS*	54	9.5
	Amhara	28	4.9
	Diredawa	7	1.2
	Harar	4	0.7
	Afar	3	0.5
	Gambela	2	0.4
Occupation	Not applicable (<18 years/student)	37	6.5
	Employed	325	57.2
	Unemployed	49	8.6
	Retired	5	0.9
	Unknown	152	26.8
Education	Preschool	10	1.8
	Illiterate	68	12.0
	Primary school	22	3.9
	Middle school	69	12.1
	High school	165	29.0
	Graduate/more	68	12.0
	Unknown	166	29.2
Marital Status	Unmarried	103	18.1
	Married	237	41.7
	Divorced	15	2.6
	Widowed/r	57	10.1
	Unknown	156	27.5
Religion	Orthodox	181	31.9
	Protestant	141	24.8
	Muslim	85	15.0
	Catholic	11	1.9
	Unknown	150	26.4

* SNNRS-Southern Nations Nationalities and Peoples’ Regional States.

Discussion

This is the first study article on sudden deaths in Ethiopia and is devoted to evaluating the sociodemographic, behavioral, clinical and pathological characteristics of 568 sudden deaths of various ages in Ethiopia. The frequency and pattern of sudden death in different parts of the world vary greatly due to the diversity of prevalent diseases in various nations and due to various genetic and environmental factors.^3,4 The present study reveals that sudden unexpected death with a known natural cause of death constitutes approximately 11.5% of all autopsy cases. This finding matches those of other developing countries.³ This finding is inconsistent with that reported in developed countries.^5,9 It is challenging to compare the magnitude of SD in different parts of the globe because it varies mainly as a function of the diversity of prevalent diseases in various nations and due to diverse genetic and environmental factors.^3,4 Furthermore, various definitions of SD, inclusion criteria and age groups that were evaluated in the study all contribute to the variation in SD incidence described in different studies.

The findings revealed a mean age of 44.8117.349 years. This is consistent with the studies conducted in Nigeria and Libya.^3,4,10 The maximum number of sudden death cases (24.3%) was in the age group of 31-40 years, followed by 41-50 years, representing 19.6% of all SD cases. This finding matches various studies.^11–13 All of these studies noted that middle age groups of 30 to 50 years are at high risk for SD. This finding appears to reflect behavioral and environmental factors that impact the health of society resulting from urbanization, the shift to a Western lifestyle, and the rapid nutritional transition and sedentary lifestyle with increased substance abuse habits.

Of 224 sudden death cases, men (n = 482/586) were predominant over women (n = 86/586), with a ratio of 5.6:1. This finding is consistent with many similar studies.^{1,3,4,11,14,15} In our study, the majority (70%) of SD victims were from urban areas, which is consistent with other similar studies.^4,11 This is due to the sedentary lifestyle and westernization with the increased smoking and alcohol consumption habits adopted by people in urban areas. In addition, the stress levels of urban and rural life are well known to differ.

According to the current study, more than three-quarters of sudden deaths were unwitnessed. This could be because the shorter the survival period, the more likely it would go unnoticed, meaning that the likelihood of death being unwitnessed or unattended is directly correlated with the duration of survival.

Our study found that chest pain, dyspnea, and syncope were the most common terminal symptoms reported in 68.6% of SD victims. These are the principal symptoms of cardiovascular and respiratory diseases, which were the two leading causes of SD. The finding of chest pain as the most common prodromal symptom coincided with a study in South Africa.¹⁴ In light of this, we advise emergency medical professionals to pay more attention to people who exhibit these prodromal symptoms.

The fact that 86.6% of all sudden deaths occurred outside of a hospital setting shows that the majority of these victims may not have known about their underlying medical problems or may not have been able to pay for the necessary medical care. This shows that most cases of SD are silent without preexisting symptoms. Therefore, it is vital to develop public health measures that will help educate the community about the importance of recognizing the manifestation of various clinical conditions and the need to seek immediate clinical help. Furthermore, efforts should be made to make healthcare facilities accessible and affordable with adequate diagnostic and management capacity. Two additional findings from this study further support this view. The first is that 34% of the SD victims had visits to medical facilities in the weeks before death, followed by deaths at home, which could be an indication that the diagnosis was incorrect, that therapy had failed, or that the victim lacked sufficient financial power to complete further investigation or management. The other reason is the high frequency of SD cases from chronic illnesses diagnosed for the first time at autopsy.

The occurrence of SD in males outnumbers females in all systems except GUS. Our findings coincided with those of the studies conducted in Tukey.^4,15 In the current study, men were predominantly affected by cardiovascular causes of sudden natural death, and a significant association was observed between CVS disease and sex. Our findings coincided with the results of the United Kingdom.^3,10,16 The fact that men are more frequently exposed to CVS pathologies can be attributed to the fact that estrogen hormone in women acts as a protective factor against most cardiovascular events, explaining the male preponderance that coexists with the high prevalence of CVS causes of sudden death.^3,10 Furthermore, it could also be the result of men's higher rates of substance misuse compared to women, as well as higher levels of financial stress.

Cardiovascular and respiratory diseases are the most common causes of SD. These results are consistent with those of various studies conducted in different parts of the world.^{1,3,4,14,15,17} The fact that CVS pathologies are becoming the leading cause of SD could probably be the shift to the Western type of lifestyle that our societies are acquiring.

The most common underlying cause of SD is ischemic heart disease. This finding is consistent with those of various studies conducted in different parts of the world.^{1,4,5,14,15,17} Cardiomyopathy was the second most common COD in the cardiovascular system, accounting for 5.3% of all sudden deaths. This agrees with the result achieved in Libya.⁴

In many series, sudden cardiac death (SCD) is the initial manifestation of the condition in 20%–40% of cases.^16,18 In the current study, 29.8% of victims of cardiovascular death had no antecedent symptoms, and SD was the first manifestation of sudden cardiac death in 29.8% of victims, which is consistent with many similar studies.^16,18

Because SD is frequently the first sign of the disease, it is impossible to detect high-risk individuals, making the prevention of SCD considerably more challenging. Primary prevention is highly challenging because early identification of subjects in the community who are at a high risk of SCD is impossible owing to the significant percentage of SCD occurring in individuals with no previously known disease. However, sudden cardiac deaths have a relatively high prevalence of CVRFs. The risk of SCD in asymptomatic subjects with CVRFs is higher than that in the general population but lower than that of symptomatic patients with a diagnosed condition.¹⁶

As a result, early identification and management of modifiable CVRFs is one potential strategy to reduce the burden of sudden cardiac deaths in the community. In this study, significantly higher CVRF frequencies were observed in SCD victims, including diabetes mellitus (31.7%), hypertension (17.6%), and substance addiction (22%). Therefore, community education on preventive strategies, early detection, and control of CVRFs together with preventive strategies for substance abuse would be effective for the prevention of SD.

Respiratory causes of SD accounted for 32.6% of cases in the current study. Pneumonia was the leading cause of death in the respiratory system, representing 18.7% of sudden death cases, making it the second most common cause of SD. This is closely followed by tuberculosis, comprising 6.3% of all sudden death cases. This is consistent with the results achieved in Libya and South Africa.^4,14

Gastrointestinal pathologies were the third leading cause of SD, occurring in 111 (19.5%) cases. Liver and pancreatic pathologies were the two most common causes of sudden death among GIS pathology, accounting for 53.2% (n=59 cases) and 37.8% (n=42 cases), respectively. Hepatic and pancreatic pathologies together contribute to 90.1% of GIS causes of sudden death and are generally the third and fourth underlying causes of death, respectively. This finding is inconsistent with various similar studies showing that CNS pathologies are the third most common cause of SD.^3,4,14,15,17 A possible explanation could be the high frequency of substance abuse reported among victims of GIS. A history of substance abuse was reported in 46.8% (n= 52) of 111 victims of sudden death from GIS pathology. Chronic alcohol use was the most common substance used and was reported in 50 (45%) of 111 victims, followed by khat and cigarettes in 19.8% and 4.5% of victims, respectively. Alcohol has a causative association with both liver and pancreatic diseases.^19–22 Additionally, there is a growing body of evidence linking khat to the emergence of both acute and chronic liver disease.^23–26 Khat (Catha edulis), a plant that is chewed for its psychedelic effects, is widely used in the eastern and southern parts of Ethiopia, but less frequently in the northern region.²³

Central nervous system pathologies accounted for 10.6% of SD cases in this study. Cerebrovascular accidents were the most common causes (81.7% of all CNS causes), followed by infectious causes (15% of all CNS causes). This is in agreement with a study from Turkey.¹⁵ The number of male versus female cases in all organ systems was higher except for GUS. Our findings coincided with various studies.^4,14,15 The reason is that in our study, a high frequency of maternal deaths was recorded.

Conclusions

Sudden death is an important global public health issue. Cardiovascular, respiratory and gastrointestinal system pathologies were the most common causes. The two main underlying causes of sudden death were ischemic heart disease and pneumonia. Although the causes of SD observed in the current study were comparable to those previously reported elsewhere, the rate of occurrence of certain conditions was different, particularly the higher frequency of liver and pancreatic pathologies. The association of these diseases with chronic alcohol and khat (Catha Edulis) abuse was also documented. Most of the causes of SD in Ethiopia can be prevented and treated. The majority of sudden deaths are silent without preexisting symptoms and occur outside a hospital setting. Therefore, it is vital to develop public health measures that will help educate the community about the importance of recognizing the manifestation of various clinical conditions and the need to seek immediate clinical help. Furthermore, efforts should be made to make healthcare facilities accessible and affordable with adequate diagnostic and management capacity. Documentation of autopsy-based data could provide important epidemiological information to guide medical services, prevention efforts, and control measures.

Conflicts of interest

The authors declare that they have no conflicts of interest.