First report from the German COVID-19 autopsy registry

Abstract

Background: Autopsies are an important tool in medicine, dissecting disease pathophysiology and causes of death. In COVID-19, autopsies revealed e.g., the effects on pulmonary (micro)vasculature or the nervous system, systemic viral spread, or the interplay with the immune system. To facilitate multicentre autopsy-based studies and provide a central hub supporting autopsy centres, researchers, and data analyses and reporting, in April 2020 the German COVID-19 Autopsy Registry (DeRegCOVID) was launched.

Methods: The electronic registry uses a web-based electronic case report form. Participation is voluntary and biomaterial remains at the respective site (decentralized biobanking). As of October 2021, the registry included N=1129 autopsy cases, with 69271 single data points including information on 18674 available biospecimens gathered from 29 German sites.

Findings: In the N=1095 eligible records, the male-to-female ratio was 1·8:1, with peaks at 65-69 and 80-84 years in males and >85 years in females. The analysis of the chain of events directly leading to death revealed COVID-19 as the underlying cause of death in 86% of the autopsy cases, whereas in 14% COVID-19 was a concomitant disease. The most common immediate cause of death was diffuse alveolar damage, followed by multi-organ failure. The registry supports several scientific projects, public outreach and provides reports to the federal health authorities, leading to legislative adaptation of the German Infection Protection Act, facilitating the performance of autopsies during pandemics.

Interpretation: A national autopsy registry can provide multicentre quantitative information on COVID-19 deaths on a national level, supporting medical research, political decision-making and public discussion.

Funding: German Federal Ministries of Education and Research and Health.Hintergrund: Obduktionen sind ein wichtiges Instrument in der Medizin, um die Pathophysiologie von Krankheiten und Todesursachen zu untersuchen. Im Rahmen von COVID-19 wurden durch Obduktionen z.B. die Auswirkungen auf die pulmonale Mikrovaskulatur, das Nervensystem, die systemische Virusausbreitung, und das Zusammenspiel mit dem Immunsystem untersucht. Um multizentrische, auf Obduktionen basierende Studien zu erleichtern und eine zentrale Anlaufstelle zu schaffen, die Obduktionszentren, Forscher sowie Datenanalysen und -berichte unterstützt, wurde im April 2020 das deutsche COVID-19-Autopsieregister (DeRegCOVID) ins Leben gerufen.Methoden: Das elektronische Register verwendet ein webbasiertes elektronisches Fallberichtsformular. Die Teilnahme ist freiwillig und das Biomaterial verbleibt am jeweiligen Standort (dezentrales Biobanking). Im Oktober 2021 umfasste das Register N=1129 Obduktionsfälle mit 69271 einzelnen Datenpunkten, die Informationen über 18674 verfügbare Bioproben enthielten, die von 29 deutschen Standorten gesammelt wurden.Ergebnisse: In den N=1095 ausgewerteten Datensätzen betrug das Verhältnis von Männern zu Frauen 1,8:1 mit Spitzenwerten bei 65-69 und 80-84 Jahren bei Männern und >85 Jahren bei Frauen. Die Analyse der Sequenz der unmittelbar zum Tod führenden Ereignisse ergab, dass in 86 % der Obduktionsfälle COVID-19 die zugrunde liegende Todesursache war, während in 14 % der Fälle COVID-19 eine Begleiterkrankung war. Die häufigste unmittelbare Todesursache war der diffuse Alveolarschaden, gefolgt von Multiorganversagen. Das Register unterstützt mehrere wissenschaftliche Projekte, die Öffentlichkeitsarbeit und liefert Berichte an die Bundesgesundheitsbehörden, was zu einer Anpassung des deutschen Infektionsschutzgesetzes führte und die Durchführung von Obduktionen in Pandemien erleichtert.Interpretation: Ein nationales Obduktionsregister kann multizentrische quantitative Informationen über COVID-19-Todesfälle auf nationaler Ebene liefern und damit die medizinische Forschung, die politische Entscheidungsfindung und die öffentliche Diskussion unterstützen.Finanzierung: Bundesministerien für Bildung und Forschung und für Gesundheit.

Keywords: Autopsy; COVID-19; Cause of Death; Registry; SARS-CoV-2.

Figure 1

a) Flow diagram of included and excluded cases. b) COVID-19 autopsies per site. From N=1129 autopsies, contributed by N=29 university and non-university autopsy centers in N=27 cities, N=1095 autopsy cases were eligible for analyses. (Map source: Map Data from OpenStreetMap. This data is available under the Open Database License and under Creative Commons Attribution-Share Alike 2.0 license.) NV = no value

Figure 2

a) COVID-19 autopsies per calendar week (N=1094, 1·18% of all COVID-19 deaths). COVID-19 autopsies (purple line) follow the pandemic peaks of reported COVID-19 deaths (grey area, data: RKI). Note high autopsy rate in the Northern region during the 1^st pandemic wave (blue line, one case excluded due to missing value). b) COVID-19 autopsies by age and sex (N=1094). Age and sex distribution of COVID-19 autopsy cases shows male predominance in patients from 50-80 years, but more female autopsy cases in patients older than 85 years. The age peaks ≥85 years and female sex and 65-69 years and male sex showed a formally significant association, while the age peak 80-84 years and male sex did not. (p < 0·001, Fisher's exact test, two tailed P value). c) Age and sex distribution of COVID-19 autopsy numbers by age and sex stratified by pandemic wave. Comparison of age groups between different waves in female or male sex showed no significant differences. d) The comparison of the percentage of age groups of female or male sex of the total number of female/male COVID-19 deceased persons per pandemic wave showed a significant decrease of female cases in the age group of 85-89 years from the 1^st to the 3^rd and 80-84 years from the 2^nd to 3^rd pandemic wave (p < 0·05). Comparison between male and female age groups during each wave showed no significant results (two-way ANOVA with Bonferroni post-test). d) Disease duration (N=870). The disease duration from first COVID-19 symptoms/Positive SARS-CoV-2 test to death was less than two weeks in 41-52% of the analysed cases, while it was between 2 and 5 weeks in 37-45% and longer than 5 weeks in 11-14% of cases. CW = calendar week; RKI = Robert Koch-Institute

Figure 3

a) Results of SARS-CoV-2 test at autopsy (N=889). When autopsies took place within three weeks from the first COVID-19 symptoms/SARS-CoV-2 test, the test for SARS-CoV-2 at autopsy was positive in more than half of the cases. When the disease duration and post-mortem interval were longer than three weeks, the percentage of positive SARS-CoV-2 tests at autopsy decreased to less than 50%. b) Cause of death at autopsy (COVID-19 autopsies with positive clinical or post-mortem test for SARS-CoV-2, N=986). COVID-19 was the underlying cause of death in 86·2% of COVID-19 autopsies, whereas COVID-19 was a concomitant disease in 13·8% of COVID-19 autopsies. The most common immediate cause of death in COVID-19 deaths was diffuse alveolar damage/acute respiratory distress syndrome (DAD/ARDS), followed by multiorgan failure. c) Comparison of the disease duration of the most common immediate causes of death as shown in b) by pandemic wave (Tukey method for plotting the whiskers and outliers, no significant differences between waves, two way ANOVA with Bonferroni post test).

Figure 4

a) Immediate cause of death at autopsy (weeks after first COVID-19 symptoms/SARS-CoV-2 test, N=771). Radial heat map (scaled to rows) of the immediate cause of death sorted by interval between first COVID-19 symptoms/SARS-CoV-2 test and death shows highest prevalence of DAD/ARDS as immediate cause of death in the first two weeks of the disease, decreasing subsequently, and an inverse development for multiorgan failure as the second most prevalent immediate cause of death (no significant differences between disease duration intervals). b-e) A further stratification of a) by region shows similar regional trends for cause of death per week after first COVID-19 symptoms/SARS-CoV-2 test. The higher prevalence of DAD/ARDS in COVID-19 autopsies after a disease duration ≤2 (p < 0·01) and >2 ≤5 weeks in the South compared to the East (p < 0·05) and ≤2 weeks in the South compared to the West were significant (p < 0·05), as well as the decrease of DAD/ARDS from ≤2 weeks to >5 weeks in the South (p < 0·05) and higher prevalence of multiorgan failure after >5 weeks in the East compared to the North (p < 0·05) and to the West (p < 0·01), two-way ANOVA with Bonferroni post-test). All radial heatmaps are scaled to row, colour codes and labelling identical to a).

Figure 5

a) Immediate cause of death at autopsy (weeks after first COVID-19 symptoms/SARS-CoV-2 test, N=985). Radial heat map (scaled to rows) of the immediate cause of death sorted by pandemic wave shows decreasing incidence of bacterial pulmonary superinfection and pulmonary embolism with each wave (no significant differences between the waves). b-d) A further stratification of a) by pandemic wave and interval from first COVID-19 symptoms/SARS-CoV-2 test shows similar percentages of persons deceased from DAD/ARDS during the three waves, but decreasing bacterial pulmonary superinfection with increasing disease duration. The decrease in DAD/ARDS from ≤2 to <5 weeks (p < 0·001) and the increase in multiorgan failure from ≤2 to subsequent weeks (p < 0·05) in the third wave d) were significant (two-way ANOVA with Bonferroni post test). e) The comparison between sex and cause of death regardless of disease duration showed similar distributions of cause of death diagnoses between COVID-19 autopsies of female and male deceased persons (no significant differences). All radial heat maps are scaled to row, colour codes and labelling identical to a).

Figure 6

a) Schematic overview of registration of decentrally archived COVID-19 autopsy tissue samples. Samples are categorized into central nervous system, thorax, intraperitoneal space and retroperitoneal space. b) Available tissues from registered COVID-19 autopsies for different fixation methods (minimum N=18674 tissues from 1129 COVID-19 autopsies, numbers represent different organs/tissues, not total amount of samples). Decentrally archived samples from COVID-19 autopsy tissue comprise mainly formalin-fixed, paraffin-embedded samples, followed by formalin-fixed, cryopreserved and otherwise preserved biomaterials e.g., in gutaraldehyde for electron microscopy, or in RNA preserving fixation medium for RNA studies). CNS samples appear overrepresented in comparison to non-CNS samples due to numerous sampling localizations, compared to less detailed sampling localizations for non-CNS samples. CN = cranial nerve; CNS = central nervous system; FFPE = formalin-fixed, paraffin embedded; LC = locus coeruleus; MO = medulla oblongata; SN = substantia nigra; VPM = nucleus ventralis posteromedialis