COVID-19 Rebound With and Without Paxlovid: What the Evidence Shows

Early treatment with antiviral drugs prevents hospitalization and death among people with mild-to-moderate COVID-19 who are at risk of severe disease. The National Institutes of Health’s COVID-19 Treatment Guidelines recommend first-line nirmatrelvir/ritonavir (Paxlovid) or remdesivir, and second-line molnupiravir. The two oral drugs, Paxlovid and molnupiravir, were widely available but underused — possibly because of reports that people “rebounded” after treatment, especially with Paxlovid.

SARS-CoV-2 rebound usually means symptoms returning, or a new positive test, after someone has recovered from COVID-19. Rebound was described before antivirals existed, and linked to immunity and individual factors. CDC’s May 2022 health advisory on rebound after the 5-day Paxlovid course noted that untreated people rebounded too. The Food and Drug Administration, which approved Paxlovid on May 25, 2023, for adults at high risk after authorizing it for emergency use in December 2021, found no consistent link between treatment and rebound in its randomized EPIC-HR trial.

To sharpen the picture, CDC reviewed rebound studies published from February 1, 2020, to November 29, 2023.

The review

Searches of PubMed, JSTOR and Google Scholar returned 303 publications. After duplicates and irrelevant or excluded studies were removed, 23 met the criteria, and 7 — one randomized trial and six observational studies — compared rebound between treated and untreated patients.

How studies were selected for the review. Image from CDC’s page.

What the studies found

• Four studies, including the randomized trial, found no statistically significant difference in rebound between treated and untreated people.
• Three found more rebound among treated people, but with caveats: one studied people with chronic lymphocytic leukemia; one compared treated people who were older (median age 57 against 39), more vaccinated and far more often immunosuppressed (32% against 9%); and one had only limited follow-up.
• A large observational study found rebound in 6.6% of people treated with Paxlovid, 4.8% with molnupiravir and 4.5% with no treatment — differences that were not significant. People with weakened immune systems were more likely to rebound whether or not they were treated.
• The EPIC-HR trial found low, similar rebound rates in its treated and untreated groups, and rebound was not tied to low drug levels, hospitalization or death, severe relapse, vaccination, or new mutations.
• No hospitalizations or deaths were reported among outpatients who rebounded; their symptoms were mild.

How rebound was defined, and how often samples were taken, varied widely between studies, and so did the rate found.

How rebound behaves

In 22 treated patients from three studies, the median time to a negative test was 6 days after the first positive one; rebound came at a median of 9 days after diagnosis and had resolved by a median of 16 days into the illness. Overall, the evidence suggests rebound appears as a mild illness 3–7 days after the first illness resolves, in treated and untreated people alike, and is not tied specifically to Paxlovid.

One study found people who rebounded shed infectious virus longer (14 days against 3) but found no drug-resistance mutations. Another, of six patients who rebounded after Paxlovid, found a strong immune response during rebound, likely lowering the risk of worsening, and no resistance either.

Why treated people may rebound

People given antivirals are, by design, at high risk of severe disease, and may have traits — such as immunosuppression — that make the virus’s course more variable. Treatment may also hold the virus down early, letting it resume once the course ends if the body clears it slowly; some immunocompromised people may need longer treatment. Two clinical trials were studying longer courses of Paxlovid for immunocompromised people and retreatment after rebound. Rebound probably does not reflect reinfection or resistance, though more study is needed, and completing the full course matters to avoid encouraging resistance.

Limitations

• Studies did not share a standard definition of rebound.
• The search may have missed relevant publications.
• Observational studies could rarely confirm that courses were completed or that vaccination and past infection were recorded accurately.
• Few studies matched symptoms to viral load.
• Treated people are followed closely and more likely to report returning symptoms, which may explain why early reports centered on Paxlovid, the most-used oral antiviral in the United States.

Bottom line

The benefit of antiviral treatment for people at risk of severe COVID-19 outweighs the risk of rebound, which resolves quickly and does not make symptoms more severe. As the NIH guidelines say, the possibility of rebound should not stop clinicians from prescribing these lifesaving drugs when they are indicated.

Sources

Based on Smith DJ, Lambrou A, Patel P, "SARS-CoV-2 Rebound With and Without Use of COVID-19 Oral Antivirals," Morbidity and Mortality Weekly Report, Centers for Disease Control and Prevention; a work of the United States government in the public domain. Where the report’s body calls the four studies that found no difference “retrospective cohort studies,” this page follows its abstract, which counts the randomized trial among them.