SARS-CoV-2 emerged in 2020, and after rapid global spread, virus variants emerged that had adaptation or immune evasion mutations and were classified as variants of concern (VOCs). Although the first VOCs, Alpha (B.1.1.7) and Delta (B.1.617.2), showed enhanced transmissibility (1), Omicron (B.1.1.529) lineages carry mutations that provide strong immune evasion after infection with previous lineages or mRNA vaccination (2). Because autopsy data are lacking, differences in SARS-CoV-2–related pulmonary disease and tropisms have not been well studied. In this study, we performed full autopsies of persons who died from COVID-19 and conducted comprehensive analyses to characterize COVID-19–related cases macroscopically and microscopically.

Figure

Figure. Prevalence of SARS-CoV-2 variants, pulmonary inflammation, and diffuse alveolar damage in corpses autopsied during 2020 and 2022 at the Institute of Legal Medicine and crematoria in Hamburg, Germany. A) Overall...

All corpses admitted to the Institute of Legal Medicine (n = 8,578) and crematoria (n = 1,705) in Hamburg, Germany, during March 3, 2020–March 31, 2022, were screened for SARS-CoV-2 mRNA by quantitative reverse transcription PCR (3). We found a total of 808 SARS-CoV-2 RNA-positive corpses; median monthly prevalence was 6.54% in 2020, 5.28% in 2021, and 12.50% in 2022 (Figure, panel A). In comparison, the median monthly prevalence of SARS-CoV-2 in Hamburg’s general population was 0.10% in 2020, 0.37% in 2021, and ≈5.20% in 2022 (https://de.statista.com/statistik/daten/studie/1106006/umfrage/entwicklung-der-fallzahl-des-coronavirus-in-hamburg). A considerably higher prevalence of virus in deceased persons than in the general population can be explained by an older age in postmortem cohorts, targeted transport of clinically suspected or confirmed COVID-19 deaths to the Institute of Legal Medicine in 2020, and underreporting of overall SARS-CoV-2 prevalence because of limited availability of molecular testing.

We further characterized SARS-CoV-2 RNA-positive samples by using multiplexed typing quantitative reverse transcription PCR (4,5). The occurrence of VOCs among corpses (B.1.1.7 for December 2020–March 2022, B.1.617.2 beginning in May 2021, and B.1.1.529 beginning in November 2021) reflected their occurrence within the general population but had a 2–4 week delay (https://www.leibniz-liv.de/de/aktuelles/covid-19/daten-der-hamburg-surveillance-plattform).

Of the 808 COVID-19–associated deaths (determined by postmortem SARS-CoV-2 RNA detection), we autopsied 49 corpses and collected multiorgan tissue samples for more detailed analyses. We included 23/49 consecutive deceased persons infected with non-VOC lineages and 26/49 consecutive deceased persons infected with VOCs (B.1.1.7, n = 7; B.1.617.2, n = 4; and B.1.1.529, n = 15) (Table). We processed formalin-fixed paraffin-embedded tissue for histologic and immunohistochemical staining and cryopreserved tissue for molecular analysis as previously described (3,6).

The median nasopharyngeal SARS-CoV-2 RNA load was 5.82 (interquartile range [IQR] 4.08–7.31) log₁₀ copies/mL (Figure, panel B). Nasopharyngeal and organ viral loads for non-VOC and B.1.1.7 were published in part previously (7,8). We observed strong evidence for a difference in mean nasopharyngeal viral loads by virus variant (p = 0.01); by using multiple comparisons, we observed a difference in means between B.1.1.529 and non-VOC lineages (p = 0.002; Figure, panel B). This result supports increased infectivity of B.1.1.529 compared with non-VOC lineages (9). An association between nasopharyngeal virus load and virus variant at the 5% significance level did not persist in a multivariable model adjusted for the deceased’s vaccination status, which might be because of the small sample size (Appendix Tables 1, 2). Of note, the pulmonary virus load was strongly associated with viremia (odds ratio 2.21, 95% CI 1.34–3.63; p = 0.002) and mRNA detection in peripheral organs (odds ratio 1.54, 95% CI 1.10–2.16; p = 0.01) in univariable logistic regression models. However, normalized SARS-CoV-2 RNA loads in peripheral organs did not differ between virus variants (Figure, panel C).

Experienced pathologists performed single-blind histologic assessments. We detected SARS-CoV-2 nucleocapsid protein in the lungs of 25/41 (61%) cases. Using a Likert-like scale, we found the median abundance of SARS-CoV-2 nucleocapsid protein (0, not detected; 1, low abundance; 2, intermediate abundance; 3, high abundance) was 1 (IQR 0–2) for non-VOC lineage, 1.5 (IQR 1–2) for B.1.1.7, 0.5 (IQR 0–1) for B.1.617.2, and 0 (IQR 0–1) for B.1.1.529 cases (p = 0.03) (Appendix Figure).

We detected mild to strong inflammatory changes in the lungs of 36/40 (90%) cases and microscopic signs of diffuse alveolar damage (DAD), indicating acute respiratory distress syndrome, in 17/40 (43%) cases (Figure, panels D, E). As in recent animal experiments (10), pulmonary changes, such as inflammation and DAD, were associated with virus variant at the 5% level (Appendix Tables 3, 4) but not with nasopharyngeal or pulmonary viral load (inflammatory changes, p>0.05; DAD, p>0.05). An association between virus variants and inflammatory changes or DAD at the 5% level did not persist in a multivariable model adjusted for the deceased’s vaccination status (Appendix Tables 5–7).

In conclusion, our data confirm higher SARS-CoV-2 mRNA loads in nasopharynges of deceased persons who were infected with the B.1.1.529 VOC lineage, but we observed no differences in pulmonary or tertiary organ viral loads. However, pulmonary inflammation appeared higher and DAD more frequent in non-VOC than VOC lineages until adjustment for vaccination status. Our results suggest that prior vaccination, rather than reduced virulence of virus variants, might convey protection against pulmonary inflammation and acute respiratory distress syndrome during SARS-CoV-2 infections.

Top