Abstract

METHODS

The study population included all decedents undergoing forensic postmortem examination (including toxicology for drugs and alcohol) by the OCME from May 24 to June 30, 2020 from whom 5 mL whole blood could be collected. The Johns Hopkins Bloomberg School of Public Health Internal Review Board (IRB) determined that the study is not human subjects research, and the Maryland Health Department IRB provided “exemption” status.

In April 2020, the Maryland OCME implemented a protocol that included questioning living relatives or acquaintances about symptoms in the decedent or recent contact with possible COVID-19-infected person before death. Reverse-transcription polymerase chain reaction (RT-PCR) SARS-CoV-2 testing is performed by the medical examiner if there is suspicion of infection in the decedent.

Severe acute respiratory syndrome coronavirus-2 immunoglobulin (Ig)M and IgG antibodies were detected using the CoronaChek serologic lateral flow assay (purchased from CLIAwaived and made by Hangzhou Biotest Bioctech Co., Ltd.). Studies of CoronaChek on positive and negative control specimens from Maryland showed good sensitivity (95%; 95% confidence interval [CI], 83%–99% in convalescent plasma donors >28 days after symptom onset; 100% [95% CI, 89%–100%] in COVID-19-confirmed hospitalized individuals 15 days after symptom onset) and specificity (100% [95% CI, 94%–100%] in patients infected with rhinoviruses and other coronaviruses before the COVID-19 pandemic) [7].

Age at death, sex, race, Hispanic ethnicity, homelessness, and cause and manner of death were abstracted from the OCME electronic database. Covariates were compared by SARS-CoV-2 antibody status using Student’s t tests and Fisher’s exact χ² tests for differences in means and proportions, respectively. The prevalence of SARS-CoV-2 antibodies was estimated for each week during the study period. The 7-day average number of new COVID-19 cases and the positivity rate in Maryland were plotted over the same period [2, 8]. Log-binomial regression models estimated adjusted prevalence ratios (aPRs) and 95% CIs; adjusted models include age (<60 and ≥60 years, sex (male and female), race (white, black, and Hispanic), homeless status, and manner and cause of death (natural, homicide, suicide, accidental illicit drug intoxication, accidental motor vehicle crash, other accident, undetermined, and manner is still pending).

RESULTS

Of the 720 decedents examined, 500 had a viable whole blood specimen for testing. There were 2 Asian decedents, 5 of other race, and 1 with unknown race, none of whom were seropositive. Among the 500 decedents, 19 were SARS-CoV-2 IgM or IgG antibody indeterminate due to hemolysis of the sample and were classified as negative. The seroprevalence of SARS-CoV-2 IgG or IgM antibodies was 10% (50 of 500; 95% CI, 9%–11%); 15 were IgG positive only, 10 were IgM positive only, and 25 were both IgG and IgM positive.

Comparing antibody positive to negative decedents, there was no difference in age (mean age 45.2 vs 45.2 years, P = .99) or homeless status (14% vs 11%, P = .64), but there was a greater proportion of antibody-positive decedents among Hispanic men (20% vs 5%, respectively, P < .01) and among those who died in a motor vehicle crash (18% vs 7%, respectively, P < .01) and a lower proportion among decedents of an accidental illicit drug intoxication (8% vs 15%, P = .28), and decedents with an undetermined manner and cause of death (10% vs 23%, P < .03). Twelve of the 500 decedents were RT-PCR positive for SARS-CoV-2 via the OCME protocol (10 of whom were antibody positive); all 12 were classified as natural deaths.

The weekly seroprevalence of SARS-CoV-2 antibodies ranged from 5% (95% CI, 3%–7%) to 21% (95% CI, 16%–26%) per week (Figure 1). During the same period, the 7-day average of the number of confirmed cases (SAR-CoV-2 RT-PCR positive) in the State of Maryland decreased from 1008 on May 30, 2020 to 364 cases on June 30, 2020; the 7-day average testing positivity rate also decreased from 10.9% to 4.7%.

Open in a separate window

Figure 1.

Weekly seroprevalence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) antibodies among decedents examined by the Maryland Office of the Chief Medical Examiner, and the 7-day average number of cases and testing positivity in Maryland, May 24–June 30, 2020. The 7-day average (average of the specified day and the previous 6 days) number of confirmed cases in Maryland was estimated by The New York Times coronavirus database, available at https://www.nytimes.com/interactive/2020/us/maryland-coronavirus-cases.html. The 7-day average testing positivity in Maryland was from the Maryland Department of Health, available at https://coronavirus.maryland.gov/.

Correlates of antibodies included an approximately 4-fold increase in prevalence in Hispanic compared with non-Hispanic white decedents (aPR = 3.88; 95% CI, 1.65–9.11); there was no difference in the prevalence in black decedents (aPR = 0.93; 95% CI, 0.51–1.69). The 21% increase in prevalence in homeless individuals (vs housed individuals) was not statistically significant (aPR = 1.21; 95% CI, 0.52–2.81). Motor vehicle crash decedents had a similar prevalence of SARS-CoV-2 antibodies compared with those who died of natural causes (aPR = 1.15; 95% CI, 0.44–3.04). Accidental illicit drug intoxication decedents had a 71% decrease in the prevalence of antibodies compared with those who died of natural causes (aPR = 0.29; 95% CI, 0.10–0.86); a decrease was also found in homicide decedents (aPR = 0.37; 95% CI, 0.12–1.19), those who died from suicide (aPR = 0.60; 95% CI, 0.18–2.02), those who died from accidents unrelated to illicit drug intoxication or motor vehicle crashes (aPR = 0.40; 95% CI, 0.8–2.13), and those who have an undetermined manner of death (aPR = 0.51; 95% CI, 0.22–1.20), although these differences were not statistically significant.

DISCUSSION

Although the population undergoing forensic postmortem examination by the OCME may differ in lifestyle from those in general population seroprevalence studies, SARS-CoV-2 seroprevalence in this population reflected ethnicity patterns of COVID-19 burden reported in healthcare settings and surveys of testing in Maryland [3, 4].

Our reported seroprevalence of 10% is higher than that reported in seroprevalence studies before June 1 in the United States outside of New York City [5, 9–11]. There may be a higher seroprevalence in decedents undergoing forensic examination due to shared risk factors for nonnatural causes of death and COVID-19, such as socioeconomic disadvantage at the individual- and neighborhood-level [12, 13].

Decedents undergoing forensic investigation provide insight into populations that may be underrepresented in general population seroprevalence studies. We found a lower prevalence of SARS-CoV-2 antibodies in accidental illicit drug intoxication decedents compared with those who died of natural causes. Network studies among drug users have demonstrated that smaller and stable personal networks (compared with larger networks) are protective against other infectious diseases, such as human immunodeficiency virus [14]. Because overdose deaths may be increasing during the US COVID-19 epidemic, data describing SARS-CoV-2 infection in drug users and the epidemic’s impact on their networks are needed to understand syndemic potential [15]. Conversely, we found 23% (9 of 39) seroprevalence of SARS-CoV-2 antibodies in those who died from motor vehicle crashes; this seroprevalence was similar to those with a natural death (28% or 14 of 69), which included the 12 decedents who were found to be SARS-CoV-2 infected at the time of death via the OCME protocol. Given their high prevalence, motor vehicle crashes decedents may be a unique sentinel population for SARS-CoV-2 infection.

Limitations to our study include 96 decedents whose cause and manner of death are still pending. Of those with an undetermined manner of death, approximately half are illicit drug-related deaths; the lower prevalence of SARS-CoV-2 antibodies among accidental illicit drug intoxication decedents is reflected in the undetermined group. Finally, other antibody tests (eg, enzyme-linked immunosorbent assay [ELISA]) may have produced different results. The rapid flow assay utilized has validated high sensitivity and specificity and is relatively cheap and easy to perform and has been compared with numerous ELISAs for comparison of our results to forthcoming seroprevalence studies that utilize ELISAs [7].

CONCLUSIONS

Our study demonstrates 10% SARS-CoV-2 seroprevalence in decedents undergoing forensic postmortem examination and reflects ethnicity patterns of acute disease burden. These decedents provided timely information and allowed for inclusion of those who may be underrepresented in general population seroprevalence studies. Our approach could be replicated among decedents in other locations.

Acknowledgments

References