Rebound of COVID-19 With Nirmatrelvir–Ritonavir Antiviral Therapy

Abstract

In their article, Edelstein and colleagues provide the results of an observational study of virologic response in patients who received treatment with nirmatrelvir–ritonavir (N-R) versus those who received no COVID-19 therapy. The editorialists discuss the findings and emphasize the need for further consideration of the dosage, timing, and duration of treatment to inform optimal use of N-R.

The COVID-19 pandemic sparked an urgent search for treatments to prevent progression of the disease. Several effective monoclonal antibodies were quickly developed for use in outpatients (1) and were granted emergency use authorization by the U.S. Food and Drug Administration (FDA). Unfortunately, the emergence of SARS-CoV-2–resistant variants eliminated their use.

In May 2023, the protease inhibitor nirmatrelvir–ritonavir (N-R) became the first oral COVID-19 treatment to be fully licensed by the FDA. A large phase 3 randomized clinical trial (2) and observational studies (3) demonstrated the clinical and virologic benefits of N-R, with minimal adverse events. However, with broader use, many reports have suggested that N-R leads to rebound of infection even more often than was initially reported (4).

In their article, Edelstein and colleagues (5) provide the results of an observational study of 127 participants, 72 of whom received a 5-day course of N-R. The study included polymerase chain reaction testing and viral culture of self-collected nasal samples. Fifteen (20.8%) of those who received N-R had viral rebound, all but 2 of whom were symptomatic. In contrast, rebound was noted in only 1 (1.8%) of the untreated participants. Rebound was most common among older and immunosuppressed patients and, paradoxically, among those who had received more vaccine doses (that is, vaccination per se did not reliably prevent rebound). An intriguing finding—which is discussed later—is that rebound was more frequent with earlier treatment than with delayed treatment.

The primary limitations of the study, as acknowledged by the authors, are the small sample size and the unavoidable bias ascribed to the decision to offer N-R. The untreated group (who did not require intervention) had less risk for progression of infection and possibly less inherent risk for rebound as well. Patients who received N-R in this study likely had clinical benefit (2, 3), and it is important to note that rebound was not observed in most treated participants. On the other hand, N-R did seem to place patients at greater risk for rebound, as characterized by a longer duration of shedding of replication-competent SARS-CoV-2, recurrent symptoms, and a longer duration of contagion. These concerns suggest that the decision to use N-R is not straightforward and demands exploration of improved COVID-19 treatment management strategies and/or alternative oral therapies.

Understanding the basis of COVID-19 rebound with N-R is important. Edelstein and colleagues’ study and many others have found no evidence that COVID-19 rebound represents evolution of viral resistance during treatment. Alternatively, it has been suggested that early N-R treatment might limit the immune response, leaving the patient vulnerable to recurrent illness. However, a prior small but detailed study of patients treated with N-R who had symptom rebound (6) found high levels of SARS-CoV-2 antispike IgG and increased levels of antinucleocapsid IgG and Omicron-specific neutralizing antibodies in these patients. SARS-CoV-2–specific T-cell responses were greater in patients with COVID-19 rebound than in untreated control participants with early COVID-19.

The most compelling explanation for the relationship between N-R treatment and COVID-19 rebound is that 5 days of treatment at the current dosage is inadequate (4, 7). The selection of the duration of N-R treatment seems, at least in part, to have been based on a dose-escalation pharmacology study, concomitant modeling estimates, and consideration of results of earlier studies with monoclonal antibodies (8). However, results from monoclonal antibody studies seem unlikely to properly inform an estimation of N-R treatment duration.

Perelson and colleagues (9) developed a mathematical model that suggests that early treatment with N-R leaves too many target cells susceptible to infection; incomplete suppression of replication-competent SARS-CoV-2 by N-R would be expected to lead to rebound. The report from Edelstein and colleagues noted that earlier treatment was associated with greater risk for rebound (5).

Another oral protease inhibitor, ensitrelvir (ClinicalTrials.gov: NCT05305547), which is under “Fast Track” development at the FDA, appears to have a long half-life (48 hours). Tsuge and colleagues (10) recently reported no increase in viral or clinical rebound after treatment with ensitrelvir compared with placebo.

Nirmatrelvir–ritonavir and other oral agents are critical treatment options for mild and moderate COVID-19, so strategies to eliminate rebound are important for both personal and public health. Perelson and colleagues suggested that delaying (but not prolonging) N-R treatment would reduce rebound (9). Pfizer is studying the benefits of a second course of N-R treatment for patients with rebound (ClinicalTrials.gov: NCT05567952). The French research agency ANRS | Emerging Infectious Diseases is comparing 5 versus 10 days of N-R in immunocompromised patients (ClinicalTrials.gov: NCT05587894). Until recently, the U.S. government provided N-R free of charge. However, with full FDA approval of the drug, a price of about $1390 for a 5-day course has been established; therefore, added days of treatment would be very expensive in the absence of a change in the standard treatment plan.

In summary, the detailed report from Edelstein and colleagues (5) and earlier results (7) seem to demonstrate that COVID-19 rebound is observed at an increased but unpredictable rate in patients receiving N-R. Further consideration of the dosage, timing, and duration of treatment with N-R is essential to inform optimal use of this drug.