Dormant Viruses Track Severe and Long COVID
A 1,154-patient longitudinal study found viral reactivation alongside severe and long COVID, but it does not show that reactivation causes either outcome.
The 60-second version
Chronic-virus transcripts were common in a 1,154-person hospitalized COVID cohort and tracked severe disease, inflammation and some long-COVID outcomes.
Key points
- The prospective cohort sampled blood and airways during hospitalization and up to 12 months.
- EBV appeared early, while HSV1 and CMV tended to peak later; 47.9% had at least one non-SARS-CoV-2 virus detected in acute data.
- Anelloviridae correlated with impaired physical function during recovery, but causation and treatment benefit were not tested.
Verdict. A strong longitudinal association study that motivates causal trials, not evidence for routine antivirals in long COVID.
Study designA large early-pandemic hospital cohort
IMPACC prospectively followed 1,154 adults hospitalized with COVID-19 at 20 US hospitals between May 2020 and March 2021. All were unvaccinated at enrolment. Samples were collected during admission and scheduled through 12 months.
MeasurementRNA revealed different viral timelines
Deep RNA sequencing of nasal swabs, peripheral blood cells and ventilated patients' airway aspirates detected transcripts from EBV, CMV, HSV and Anelloviridae. EBV appeared early; HSV1 and CMV tended to appear later. Immune cells, cytokines, proteins and metabolites were analysed alongside the viral signals.
| EBV | Detected near admission in 260 of 1,080 participants, or 24%, then declined. |
|---|---|
| Any detected chronic or other non-SARS-CoV-2 virus | Found in 550 of 1,148 participants with adequate acute data, or 47.9%. |
| HSV1 and CMV | Later peaks were estimated from smaller specimen-specific denominators, especially airway aspirates. |
AssociationSignals tracked severity and inflammation
Viral detection was associated with worse COVID trajectories, inflammatory cytokines, activated immune cells and clinical outcomes. Those patterns remained informative after several adjustments, but severe illness, treatment and immune disruption can all confound the direction of effect.
A virus can be a driver, a passenger or a marker. This study cannot choose among them.
Long COVIDAnelloviruses produced a candidate marker
During recovery, Anelloviridae transcripts in blood cells were more prevalent in a patient-reported group with impaired physical function after adjustment for age, sex, immunosuppressive medication and acute severity. The pathogenic role of this virus family remains uncertain, so the result is a biomarker hypothesis rather than a treatment target.
LimitsThe cohort is not today's average infection
- 1. Participants were hospitalized, unvaccinated and mainly exposed to ancestral SARS-CoV-2.
- 2. Follow-up dropout reduced power for long-COVID analyses.
- 3. RNA transcripts are less conventional than targeted RT-qPCR for viral load, and only three sample compartments were sequenced.
- 4. Observational adjustment cannot establish cause or prove that antivirals would help.
Next evidenceWhat would change clinical practice
Prospective targeted testing must show when reactivation occurs, and controlled trials must show that treating a specific virus improves outcomes. Until then, the paper does not justify routine antiviral treatment or broad screening for long COVID.