SARS-CoV-2 variants with shortened incubation periods necessitate new definitions for nosocomial acquisition.

To the Editors,

In the last issue, Lumley et al. provide compelling evidence that indeterminate cases of nosocomial SARS-COV-2 infection, who develop symptoms 3–7 days after hospital admission, are indeed genuine nosocomial acquisitions. By applying genomic data, they found 26/33 (79%) of sequenced indeterminate nosocomial cases share genomic similarity with their putative nosocomial transmission clusters formed through epidemiological linkage.

Interestingly, the definitions of nosocomial acquisition are based on the incubation period of ancestral strains of SARS-CoV-2. There is increasing evidence that SARS-CoV-2 variants of concern may have increased viral fitness, such as higher viral loads, longer viral shedding, and also shorter incubation periods. Also reported in this journal, the median incubation period for the alpha variant has been estimated at around 3 days, compared to around 5 days for ancestral strains. , In another study, delta variant was similarly shown to have a shorter incubation period compared to ancestral strains (4 days vs 6 days). A shorter median incubation period increases the likelihood that indeterminate cases as defined by current definitions are indeed nosocomial acquisitions. Notably, Lumley et al. study captures the second wave of the pandemic in the UK which was dominated by cases of the alpha variant.

As viral variants with increased fitness replace ancestral strains of SARS-CoV-2, definitions of nosocomial acquisition may therefore need to be altered to reflect the shorter incubation time. Moreover, a narrower distribution of incubation periods may make judging the probability of nosocomial acquisition vs community-onset acquisition more difficult, especially when using epidemiology alone. This further suggests genomic linkage may be required for accurate resolution of putative nosocomial transmission clusters.

In addition, in the context of shortened incubation periods and continued high prevalence of infection, methodology for using genomic-linkage to resolve transmission clusters can be strengthened. Genomic analysis of nosocomial transmission could include not just similarity of genomes to other epidemiologically-linked cases, but also a measure of how common this haplotype (or variant) is in the community. Confidence of transmission between epidemiologically-linked cases can be increased not only if the cases share haplotypes, but also if this haplotype is rare compared to those circulating in the community. This measure has been introduced previously, perhaps most robustly by Stirrup et al. who provide a statistical tool for this purpose.

Declaration of Competing Interest

No competing interests to declare. No funding applicable to this correspondence.