These findings offer a natural history of COVID-19 that will inform discussions and research concerning antiviral therapy, lead author Jonathan Z. Li, MD, associate professor of infectious disease at Brigham and Women’s Hospital and Harvard Medical School, both in Boston, and colleagues reported in Annals of Internal Medicine.
Dr. Jonathan Z. Li
“There are increasing reports that high-risk patients are avoiding nirmatrelvir-ritonavir due to concerns about post-Paxlovid rebound, but there remains a gap in our knowledge of the frequency of symptom and viral relapse during untreated natural infection,” Dr. Li said in a written comment.
To address this gap, Dr. Li and colleagues analyzed data from 563 participants from the placebo group of the Adaptive Platform Treatment Trial for Outpatients with COVID-19 (ACTIV-2/A5401).
From days 0-28, patients recorded severity of 13 symptoms, with scores ranging from absent to severe (absent = 0, mild = 1, moderate = 2, severe = 3). RNA testing was performed on samples from nasal swabs on days 0–14, 21, and 28.
“The symptom rebound definition was determined by consensus of the study team, which comprises more than 10 infectious disease, pulmonary, and critical care physicians, as likely representing a clinically meaningful change in symptoms,” Dr. Li said.
Symptom scores needed to increase by at least 4 points to reach the threshold. For instance, a patient would qualify for relapse if they had worsening of four symptoms from mild to moderate, emergence of two new moderate symptoms, or emergence of one new moderate and two new mild symptoms.
The threshold for viral relapse was defined by an increase of at least 0.5 log10 RNA copies/mL from one nasal swab to the next, while high-level viral relapse was defined by an increase of at least 5.0 log10 RNA copies/mL. The former threshold was chosen based on previous analysis of viral rebound after nirmatrelvir treatment in the EPIC-HR phase 3 trial, whereas the high-level relapse point was based on Dr. Li and colleagues’ previous work linking this cutoff with the presence of infectious virus.
Their present analysis revealed that 26% of patients had symptom relapse at a median of 11 days after first symptom onset. Viral relapse occurred in 31% of patients, while high-level viral relapse occurred in 13% of participants. In about 9 out 10 cases, these relapses were detected at only one time point, suggesting they were transient. Of note, symptom relapse and high-level viral relapse occurred simultaneously in only 3% of patients.
This lack of correlation was “surprising” and “highlights that recovery from any infection is not always a linear process,” Dr. Li said.
This finding also suggests that untreated patients with recurring symptoms probably pose a low risk of contagion, according to David Wohl, MD, coauthor of the paper and professor of medicine in the division of infectious diseases at the University of North Carolina at Chapel Hill.
Paxlovid may not be to blame for COVID-19 rebound
“These results provide important context for the reports of Paxlovid rebound and show that baseline rates of symptom and viral relapse should be accounted for when studying the risk of rebound after antiviral therapy,” Dr. Li said.
Dr. Wohl suggested that these data can also play a role in conversations with patients who experience rebound after taking antiviral therapy.
“Many who have a return of their symptoms after taking Paxlovid blame the drug, and that may be justified, but this study suggests it happens in untreated people too,” Dr. Wohl said in a written comment.
Longer antiviral therapy deserves investigation
This is a “very important study” because it offers a baseline for comparing the natural history of COVID-19 with clinical course after antiviral therapy, said Timothy Henrich, MD, associate professor in the division of experimental medicine at University of California, San Francisco.
“Unlike this natural history, where it’s kind of sputtering up and down as it goes down, [after antiviral therapy,] it goes away for several days, and then it comes back up; and when it comes up, people have symptoms again,” Dr. Henrich said in an interview.
This suggests that each type of rebound is a unique phenomenon and, from a clinical perspective, that antiviral therapy may need to be extended.
“We treat for too short a period of time,” Dr. Henrich said. “We’re able to suppress [SARS-CoV-2] to the point where we’re not detecting it in the nasal pharynx, but it’s clearly still there. And it’s clearly still in a place that can replicate without the drug.”
That said, treating for longer may not be a sure-fire solution, especially if antiviral therapy is started early in the clinical course, as this could delay SARS-CoV-2-specific immune responses that are necessary for resolution, Dr. Henrich added,
“We need further study of longer-term therapies,” he said.
Dr. Aditya Shah
An array of research questions need to be addressed, according to Aditya Shah, MBBS, an infectious disease specialist at Mayo Clinic, Rochester, Minn. In a written comment, he probed the significance of rebound in various clinical scenarios.
“What [type of] rebound matters and what doesn’t?” Dr. Shah asked. “Does symptom rebound matter? How many untreated and treated ‘symptom rebounders’ need additional treatment or health care? If rebound does not really matter, but if Paxlovid helps in certain unvaccinated and high-risk patients, then does rebound matter? Future research should also focus on Paxlovid utility in vaccinated but high-risk patients. Is it as beneficial in them as it is in unvaccinated high-risk patients?”
While potentially regimen-altering questions like these remain unanswered, Dr. Henrich advised providers to keep patients focused on what we do know about the benefits of antiviral therapy given the current 5-day course, which is that it reduces the risk of severe disease and hospitalization.
The investigators disclosed relationships with Merck, Gilead, ViiV, and others. Dr. Henrich disclosed grant support from Merck and a consulting role with Roche. Dr. Shah disclosed no conflicts of interest.