Causes of death and post-mortem testing for SARS-CoV-2 in a tertiary hospital during the COVID-19 pandemic in Ghana

Background Causes of death during the coronavirus disease 2019 (COVID-19) pandemic ranhttp://crossmark.crossref.org/dialog/?doi=10.4102/ajlm.v11i1.1766=pdf&date_stamp=2022-11-23ge from direct consequences of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection to deaths unrelated to SARS-CoV-2. Another feature of the pandemic is the post-mortem testing for SARS-CoV-2. Understanding these aspects of COVID-19 are essential in planning and limiting the impact of SARS-CoV-2 virus on healthcare systems. Objective This study investigated the underlying causes of death and the presence of SARS-CoV-2 in bodies received at the 37 Military Hospital, Accra, Ghana, during the COVID-19 pandemic. Methods The study was conducted from 4–27 May 2020. Deceased patients that met the inclusion criteria were prospectively selected during the expanded surveillance period for SARS-CoV-2 testing, autopsy and determination of underlying and immediate cause of death. Results A total of 161 deceased patients were analysed with 53 autopsies. The overall positive test rate for SARS-CoV-2 was 14.9% (24/161 patients), with a positive rate of 5.0% (8/161 patients) for nasopharyngeal samples and 30.2% (16/161 patients) for bronchopulmonary samples. The underlying causes of death were not related to SARS-CoV-2 infection in 85.1% (137/161) of patients, SARS-CoV-2-associated 12.4% (20/161) and SARS-CoV-2-induced in 2.5% (4/161). Cardiovascular complications formed the most common cause of death in patients with or without SARS-CoV-2. Conclusion There was a high positive rate of SARS-CoV-2 in post-mortem cases. However, most deaths were not caused by SARS-CoV-2 but by cardiovascular complications. The high rate of bronchopulmonary positive results for SARS-CoV-2 requires that autopsies be done in suspicious cases with negative nasopharyngeal sampling.


Introduction
In 2019, the world was struck by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) which caused the coronavirus disease 2019 . 1,2 The resulting COVID-19 pandemic continues to pose challenges to healthcare systems. Compared to previous pandemics, the COVID-19 pandemic has exhibited a cyclical nature and unique manifestation in different geographical areas. 2 Various aspects of the disease have been reported from epidemiology to mortality. 3,4,5,6,7 Studies on mortality do not report on other causes of death within the pandemic, 1,6,7 although these are equally important in the fight against the SARS-CoV-2. Autopsies are indispensable in clarifying the cause of death, as well as for quality control and audit in medical practice. It is an established fact that a clinically determined cause of death is not as reliable as an autopsy-determined cause of death. 8,9 as the hospital-based mortality surveillance and routine surveillance. 11 As part of its expanded surveillance protocol, the 37 Military Hospital in Accra, Ghana, screened all deceased patients received at the Department of Anatomical Pathology morgue for SARS-CoV-2 and cause of death in May 2020. The aim of the study was to determine the underlying causes of death and the presence of SARS-CoV-2 using rRT-PCR for all cases admitted to the Department of Anatomical Pathology during the study period.

Ethical considerations
Permission was granted by the 37 Military Hospital's institutional review board for the study with ethical approval number 37MH-IRB IPN/MAST/2020. No consent was required of the next of kin for the autopsies. To ensure privacy and confidentiality of the deceased patients, data was de-identified, coded and stored with passwords on electronic media.

Study design
This prospective study was conducted over a three-week period from 4 to 27 May 2020. This was the first institutional surveillance study in Ghana on deceased patients at an Anatomical Pathology department. The study was conducted at the J.M. Wadhwani Department of Anatomical Pathology at the 37 Military Hospital in Accra, Ghana. The department receives about 4000 deceased individuals of all categories of death from the hospital and its environs annually and conducts approximately 1500 autopsies. The department conducts autopsies for patients who have died from unknown cause for medico-legal or coroner's inquiry, homicide, suicide, accidents, misadventure and routine hospital cases for autopsy (where the cause of death is not known by the attending physician). Deceased patients with a known cause of death, where a medical certificate of cause of death has been provided by the attending physician, are stored by the department prior to burial. 12 Two categories of deceased patients, in-hospital deaths (IHDs) and brought-in-dead cases (BIDs), formed the study population. In-hospital deaths are deaths that occurred while patients were being managed in the hospital. Brought-indead deaths are those that occurred outside the hospital or before the individual had been seen by a physician in the hospital. All deceased individuals received 4-27 May 2020 were included in the study, except for individuals less than one year old, unidentified individuals and patients diagnosed as having COVID-19 in the hospital.
Study procedures were conducted in three phases. First, samples were collected for SARS-CoV-2 testing. Second, an autopsy was completed for a subset of patients when requested by the coroner or physician. Finally, the underlying and immediate cause of death for all individuals was determined.

Sample collection for SARS-CoV-2
Nasopharyngeal samples were collected from all deceased patients who met the inclusion criteria by the anatomical pathology team using sterile swab sticks. Additional bronchopulmonary samples were collected from patients who underwent autopsy as part of their post-mortem examinations. For the bronchopulmonary specimens, one sample was taken from swabs of secretions in the tertiary bronchi and another from fluid squeezed from the lungs, after sectioning, into an ordinary sterile specimen container. These samples were sent for rRT-PCR test at a national reference laboratory, Noguchi Memorial Institute for Medical Research in Accra, Ghana.

Autopsy
A complete autopsy was done by certified department pathologists on individuals only when directed by the coroner or referred by the attending physician as per department policy. All available case notes and records were reviewed by the pathologists. Selected organs, including the lungs, heart, kidney, liver, spleen, intestines and brain, were sampled for histopathological examination and some photographs taken.

Determination of underlying cause of death
Cause of death was considered to be the disease or injury which initiated the train of events leading directly to death or the circumstances of an accident or violence which produced the fatal injury. At least two certified pathologists determined the underlying cause of death for all individuals studied and reported on the histological findings. The pathologists analysed the clinical notes, as well as the cause of death provided by the attending physician, to determine the underlying cause of death for cases that did not undergo autopsy. In cases that underwent autopsy, the underlying cause of death was determined by evidence after histopathological examination and using the clinical notes, if any.

Data analysis
The data collection form captured the biodata and the clinical cause of death for all patients who did not undergo autopsy. The cause of death for those who underwent autopsy was also recorded. The rRT-PCR test results were received in batches from the Noguchi Memorial Institute for Medical Research and matched to the individual patient to complete the data collection process. Underlying causes of death were categorised into three groups. The first group comprised deaths not related to SARS-CoV-2 infection. This group included patients who tested negative for SARS-CoV-2 and who had various underlying causes of death (e.g. massive cerebral haemorrhage due to hypertension); these are referred to as 'non-SARS-CoV-2-related deaths'. The second group comprised deaths associated with, but not caused directly by, SARS-CoV-2 infection. This group included patients with a positive SARS-CoV-2 test but whose underlying cause of death was not directly attributable to SARS-CoV-2 (e.g., perforated duodenal ulcer); these are referred to as 'SARS-CoV-2-associated deaths'. The last group comprised of deaths directly induced by SARS-CoV-2 infection. This group included patients with a positive SARS-CoV-2 test and whose underlying cause of death was attributable to SARS-CoV-2 (e.g., acute respiratory distress syndrome with or without pneumonia); these are referred to as 'SARS-CoV-2induced deaths'. Collected data were managed in Microsoft Excel (2016; Norristown, Pennsylvania, United States), which was used to calculate frequencies and proportions.

Study population
A total of 161 deceased patients, including 96 IHDs and 65 BIDs, were examined over the study period ( Table 1). The mean age of the study population was 53.3 years (Range: 18 months -90 years) with a male-to-female ratio of 1:1. Among the age groups above 40 years, there was an increasing ratio of SARS-CoV-2 infection compared to non-SARS-CoV-2 infection. A total of 53 autopsies were performed. All deceased patients had nasopharyngeal samples collected for testing, and those that underwent full autopsy (n = 53) also had a bronchopulmonary sample collected. There was a total of 24 rRT-PCR tests positive for SARS-CoV-2 from both nasopharyngeal and bronchopulmonary samples. The overall positive test rate from the study population was 14.9% (24/161), with a positive rate of 5.0% (8/161) for nasopharyngeal samples and 30.2% (16/53) bronchopulmonary samples. Of patients who had no autopsy, eight had SARS-CoV-2-associated causes of death and a positive nasopharyngeal test and 100 had a negative nasopharyngeal test and a cause of death that was not associated with SARS-CoV-2. Eight patients with a negative nasopharyngeal test had a bronchopulmonary specimen that tested positive for SARS-CoV-2. Of the 161 patients included in the study, four patients' (2.5%) deaths were directly caused by SARS-CoV-2, 20 patients (12.4%) tested positive for SARS-CoV-2 but had another cause of death; for the majority of patients, the cause of death was not related to SARS-CoV-2 infection (n = 137, 85.1%).

SARS-CoV-2-induced deaths
Of the four SARS-CoV-2-induced deaths, three were IHDs and one was BID ( Table 2). All cases had severe pulmonary oedema and diffuse alveolar damage (DAD) with prominent hyaline membrane formation ( Figure 1). There was severe congestion with intra-alveolar deposition of fibrinous tissue in Case 1 ( Figure 1a). Case 2 also showed large pneumocytes with enlarged nuclei and granular amphophilic cytoplasm and microthrombi in the smaller pulmonary arteries (Figures 1a and Figure 1c) as well as edema, congestion, proliferation of fibroblasts and  lymphocytic infiltrates in the interstitium (Figure 1d). Case 3 showed very prominent DAD and hyaline membrane formation, interstitial fibrosis, infiltration of the pulmonary interstitium and alveolar spaces by macrophages and scattered multinucleated giant cells. Microthrombi were also noted in smaller pulmonary arteries (Figure 1c). Case 4 revealed thrombi in branches of the coronary and renal arteries (Figure 1e and Figure 1f).

SARS-CoV-2-associated deaths
Of the 20 patients with a positive SARS-CoV-2 test and an unrelated cause of death, 13 were IHD and seven were BID ( Table 3). The mean age was 54.7 with a range of 24-90 years. The male:female ratio was 1:1. Cardiovascular disorders constituted almost half of the underlying causes of death in this category (n = 9). Hypertension was the most common cardiovascular disorder among both IHDs and BIDs. Other causes of death included tumours (n = 1), accidents and trauma (n = 4), infections (n = 1), and others (n = 5).

Non-SARS-CoV-2-related deaths
Of the 137 non-SARS-CoV-2-related deaths, 80 were IHD and 57 were BID. The age range was 2-88 years with a male: female ratio of 1:1 For patients who had an autopsy, the clinical diagnoses were confirmed, except for one case of primary liver cell carcinoma, which had been diagnosed clinically as cirrhosis of the liver. Broad categories of causes of death identified were: cardiovascular disorders (n = 49), tumours (n = 23), accident and trauma (n = 13), diabetes mellitus-related complications (n = 11), infections (n = 18), liver disorders (n = 6), renal disorders (n = 1), and others (n = 16). Hypertension and its related complications formed the most common cause of death in patients without SARS-CoV-2 (Table 4). Two patients who tested negative for SARS-CoV-2 on rRT-PCR for both the nasopharyngeal and bronchoalveolar samples had lung tissue histopathology consistent with acute respiratory distress syndrome or DAD. There were five IHD and two BID with 'miscellaneous' causes of death. The BID deaths were from viral hepatitis (n = 1) and severe asthma (n = 1), whereas the IHD deaths were from intravascular haemolysis (n = 1), viral meningitis (n = 1) and malaria (n = 3).

Discussion
The study identified three categories of deaths that occurred during the pandemic. There were a significant number of undiagnosed SARS-CoV-2 infections among deceased individuals received by the Anatomical Pathology Department of the 37 Military Hospital with unique histopathological features of SARS-CoV-2 induced deaths. The potential for community spread of the virus from handling of unsuspected SARS-CoV-2 deceased individuals without personal protective equipment, as well as the SARS-CoV-2 cases found during autopsy, are a missing link in efforts to control the pandemic. There are other documented missing links affecting overall SARS-CoV-2 control such as non-uniform data collection, vaccine hesitancy and non-reliable energy infrastructure in low-resource environments. 13,14,15 In an Italian multicentre study conducted in 2020, the presence of SARS-CoV-2 in bronchoalveolar samples was 32.7% of suspected cases with 76% having a double negative nasopharyngeal test. 16 Our study also found SARS-CoV-2 in the bronchopulmonary specimens of cases with negative nasopharyngeal tests and in cases who died from other conditions. These findings need to be factored into the current health policy on SARS-CoV-2.
The underlying cause of death is critical in accurate reporting and documentation for better epidemic surveillance and effective planning of control measures. 17 This is the first reported post-mortem study data for Ghana that has categorised underlying causes of death into three groups: non-SARS-CoV-2 related (85.1%), SARS-CoV-2 associated (12.4%) and SARS-CoV-2 induced (2.5%). The case fatality rate in Ghana for SARS-CoV-2 infection was reported to be an average of 0.5% in July 2020. 18 This did not include deaths due to SARS-CoV-2 as identified by post-mortem examinations. The presence of post-mortem-diagnosed cases of SARS-CoV-2-induced deaths in this study implies that the overall case count is higher than reported for Ghana. The need to include cases identified by pathology departments during post-mortem examinations in defining all case mortality has been emphasised. 19 This highlights the importance of post-mortem examination in understanding the natural history of diseases and defining a true mortality rate.
Accurate reporting is a requirement in the documentation of cause of death. In its guidelines for certification and classification of COVID-19 as a cause of death, the WHO

Liver disorders
All liver-disorder-related causes -3 6 9 Liver cirrhosis Raptured hepatoma 1 2 3 Liver failure 1 4 5 Raptured oesophageal varices 1 -1 Table 4 continues →  20 This was to encourage uniformity of data collection to guide health policy, but differences in classification are emerging. 20,21 The evolution of the pandemic and the understanding of the disease requires a simple categorisation of underlying of causes of death. The categorisation adopted in this study is useful for determining accurate case counts during clinical audits of deaths.
Identifying subgroups of COVID-19 case mortality is useful in managing the pandemic. 22 Of particular interest were the 20 SARS-CoV-2-associated deaths in our study. These cases were categorised as having died from their underlying medical conditions, rather than from SARS-CoV-2. The prioritisation of care for SARS-CoV-2 infections by the health system, rather than the underlying condition, might have contributed to their mortality. Most deaths during the pandemic have been reported as being related to underlying conditions. 8,23 This raises the question as to whether SARS-CoV-2 has the potential to delay care among patients with underlying medical conditions. The presence of SARS-CoV-2 among the BIDs patients in our study suggests a high community transmission of the disease. Such cases may be asymptomatic and likely to result in an increased disease burden in the absence of strict protocols.
Interestingly, our study found a high proportion (85.1%) of deaths not attributable to SARS-CoV-2. The impact of the pandemic on health systems and the lack of provision of optimum care are attributable causes to the phenomenon of high proportions of non-SARS-CoV-2-related deaths. 24 Among the BID cases in our study not attributable to SARS-CoV-2, hypertension, infection, liver cirrhosis, diabetes mellitus and pulmonary embolism-related complications were the leading causes of death. Causes of death not attributable to SARS-CoV-2 for cases from the hospitals (in descending order) included hypertension, tumours, accident and trauma, diabetes mellitus and infections. In Ghana, the prevalence of major non-communicable diseases, especially hypertension, and their risk factors has increased and contributes significantly to the disease burden. 25 Hypertension and its related complications were the leading causes of death in both groups of patients. Non-communicable metabolic diseases, tumours and traumarelated deaths rank high in the study. This emphasises the need for aggressive clinical care for non-communicable diseases even in the face of a positive SARS-CoV-2 test.
Our study found an overall SARS-CoV-2 case positivity rate of 14.90% among deceased patients received by the Anatomic Pathology Department; these are not included in Ghana's national data on SARS-CoV-2. The study found a mean age of 55 years for male patients and 53 years female patients with an equal gender ratio. This is a different demographic pattern from the reported increased susceptibility to COVID-19 among the elderly, those with severe comorbidities and among male patients. 26 While our study found a high proportion of comorbidities among the elderly, the deaths were not directly from SARS-CoV-2, but from the associated cormorbidities. 23 It can be extrapolated that, in the context of an underlying medical condition, the primary disease possess more risk than the SARS-CoV-2 infection.
Severe acute respiratory syndrome corona virus 2 has been known to survive longer in cold temperatures and on fomites, with guidelines emerging for the handling of an autopsy in suspected cases. 27,28 In our study, the pathology department discovered 24 SARS-CoV-2-infected deceased patients, of which 20 were previously undiagnosed. The absence of appropriate personal protective equipment and standard protocols predisposed staff to a high risk of infection. It is recommended that there should be a national protocol for the safe performance of autopsies during the pandemic.
The site of sample collection has a direct correlation with the sensitivity of rRT-PCR tests, and bronchopulmonary samples give the highest viral yield. 29,30,31 In our study, bronchopulmonary specimens gave a higher positivity rate than nasopharyngeal specimens, implying that our bronchopulmonary specimens also had a higher viral load.
The higher level of virus in bronchial secretions should be a recognised risk for medical staff during oropharyngeal endoscopic procedures and anaesthesia. The necessary awareness and preventive practices to avoid contracting the virus should be in place.
The analysis of the histology of specimens from COVID-19 cases provides useful information about the disease. Our study found inflammatory features in the lungs and thrombus formation in the arterial systems of major organs. This combination of pulmonary congestion with peri-vascular lymphocytic infiltrates, interstitial/intra-alveolar fibrosis, type 2 pneumocyte hyperplasia, DAD with hyaline membrane formation, as well as arterial thrombosis, are likely to result in the rapid and severe hypoxemia noticed among COVID-19 patients. An ensuing cellular response in the presence of an underlying medical disease could contribute to the high mortality in these patients. A higher proportion of such patients in our study had cardiovascular disease. The role of the thrombus formation and subsequent hypoxia on vascular endothelium need to be investigated.
In our study, there were two cases with the clinical and histopathological features of acute respiratory distress syndrome/DAD that had negative SARS-CoV-2 rRT-PCR test results. There were no definitive diagnoses at the time of autopsy. Even though both patients had comorbid conditions, evidence at post-mortem did not attribute their deaths to the comorbid conditions. A recent case series report ascribed such situations to a 'probable' case of COVID-19 as the cause of death. 30 Such cases, who test negative for SARS-CoV-2 but have clinical features suggestive of COVID-19, should be considered as a subgroup of COVID-19 to help in case management and prevention protocols, considering the high mortality associated with the disease. The negative PCR results of these two cases are most likely to be false-negative results. Such false-negative errors may be the result of pre-analytical issues, 32 such as poorly collected samples, mislabelling, improper storage or transport, or interference from medications. That notwithstanding, there are other causes of acute respiratory distress syndrome/DAD that should be considered, and an autopsy should therefore be encouraged to distinguish between SARS-CoV-2-induced deaths and SARS-CoV-2-associated deaths.
The study showed that a significant number of clinically diagnosed cases of COVID-19 may be false. Deaths could have been wrongly labelled as due to COVID-19, because of a positive SARS-CoV-2 test by the nasopharyngeal route and presence of respiratory difficulty. Such patients would have been denied dignified burials by their relatives, instead of the practice of mass burial as mandated by national policy at the beginning of the pandemic. These findings are a useful addition to Ghana's national data to determine which deceased patients can be safely released to families and which require supervised burials. The importance of routine screening in high-risk clinical practice among asymptomatic patients and guidelines to protect staff must be emphasised.

Limitations
It would have been ideal to conduct autopsies for all 161 patients in this study, in order to determine the pathological changes present, presence of SARS-CoV-2 in bronchopulmonary specimens, as well as the autopsy cause of death. Unfortunately, this was not possible, because in some cases a medical certificate of cause of death had been issued by the attending physician so no autopsy was requested. This study was also limited to one institution and may not reflect the nationwide pattern.

Conclusion
There was a high positive rate of SARS-CoV-2 infection among the patients in this study who underwent postmortem examination. However, the majority of those deaths were not caused by SARS-CoV-2 but by cardiovascular complications. Thus, there should be aggressive clinical care for non-communicable diseases. The high SARS-CoV-2 infection rate among the deceased may reflect a high rate in the population. The high rate of bronchopulmonary positive results for SARS-CoV-2 requires that autopsies be conducted in suspicious cases with negative nasopharyngeal sample test results. Falsenegative rRT-PCR test results for nasopharyngeal samples were common enough that there is a need to intensify preventive protocols.