Severe COVID-19 Disrupts the Human Virome: Persistent Anellovirus Activity Linked to Long COVID

Viruses are not present in the human body only during an infection. Many can remain latent or maintain extremely low levels of activity over long periods, usually causing almost no symptoms. A severe illness, however, may disrupt this balance. A new study indicates that severe COVID-19 not only causes harm through SARS-CoV-2 itself but is also accompanied by the reactivation of several existing viruses in the body, with some signals persisting into the long COVID phase.

The research team analyzed 1,154 hospitalized, unvaccinated COVID-19 patients from the prospective IMPACC cohort. Participants were drawn from 20 hospitals at 15 academic institutions in the United States. Researchers collected nasal, blood, and, from some patients, endotracheal samples during hospitalization and conducted follow-ups at 3, 6, 9, and 12 months after discharge. By integrating viral RNA sequencing with data on immune cells, cytokines, proteomics, and metabolomics, they mapped changes in the human virome before and after infection.

During the first 40 days after hospitalization, signs of reactivation appeared among viruses in the herpesvirus, enterovirus, and anellovirus families. Transcriptional signals from cytomegalovirus, Epstein–Barr virus, herpes simplex virus type 1, and anelloviruses were generally associated with more severe acute disease. Among the most severely ill patients, cytomegalovirus signals were also associated with a higher risk of death within one year, while the detection of some viruses coincided with complications including shock, stroke, or secondary infections.

Different viruses left distinct immune footprints. Multi-omics analysis showed that viral reactivation was accompanied by elevated inflammatory signals including IL-6, CXCL10, and TNF, increased numbers of activated CD4 and CD8 T cells, and changes in the expression of genes related to cell replication. These associations remained observable after accounting for COVID-19 severity and SARS-CoV-2 viral load, suggesting that they were not entirely incidental consequences of critical illness.

Anelloviruses were the most notable finding. These viruses are widespread in the human body and generally have no clear pathogenicity; their abundance is sometimes regarded as an indirect indicator of immune surveillance capacity. The study found that patients who continued to show anellovirus transcriptional signals at least 3 months after discharge more often exhibited a long COVID phenotype characterized primarily by fatigue and poorer physical function. If replicated in other populations, this signal could potentially be developed into a biomarker for risk stratification and provide an entry point for investigating immune dysregulation in long COVID.

However, the simultaneous presence of viral activity and symptoms does not mean that the former causes the latter. The study detected viral transcriptional signals, which cannot uniformly be equated with intact, infectious viruses. Anelloviruses may also be merely a readout of immune dysfunction rather than a driver of disease. The sample was also concentrated among hospitalized, unvaccinated patients and may not apply to people with mild infections after vaccines became widespread. The next step is to validate the findings across populations from different periods of the pandemic and with varying disease severity, and to use mechanistic studies to determine whether reactivated viruses participate in the course of long COVID or instead leave measurable traces of deeper immune disruption.