LAUSR

The prevention and treatment algorithm for COVID-19 begins with vaccination. Although vaccines have proved exceptional, the need for effective therapeutics for SARS-CoV-2 infection remains: Vaccine and booster uptake is heterogeneous, infection can occur in vaccinated persons, protection can wane with time, additional immune-evasive subvariants continue to emerge, some people are at higher risk for severe disease, and immunocompromised persons may mount a suboptimal protective response even if they are vaccinated or have previously been infected. Based on evidence and regulatory approval, the ambulatory COVID-19 treatment algorithm has included the antivirals nirmatrelvir plus ritonavir (a boosted SARS-CoV-2 protease inhibitor), molnupiravir, and remdesivir and an ever-evolving (and shrinking) list of monoclonal antibodies. In their article, Dryden-Peterson and colleagues report the results of a retrospective cohort study to assess the real-world efficacy of nirmatrelvir plus ritonavir for early outpatient COVID-19 in an era of prevalent SARS-CoV-2 immunity and immune-evasive Omicron subvariants (1). This study included more than 44000 nonhospitalized adults aged 50 years or older with COVID-19. The primary exposure was a prescription for nirmatrelvir plus ritonavir within 2 days of diagnosis, mimicking an intention-to-treat analysis, and the primary outcome was hospitalization for any cause within 14 days of diagnosis or death from any cause within 28 days. The study was designed to emulate a clinical trial and used inverse probability weighting, a propensity score–matching approach, to account for potential bias in real-world prescribing decisions. The primary finding was an adjusted risk ratio for hospitalization or death of 0.56 (a 44% reduction and a number needed to treat of 239) favoring nirmatrelvir plus ritonavir. This study has several strengths. It seeks to answer an ever-evolving question: What is the real-world efficacy of nirmatrelvir plus ritonavir, the leading component of our outpatient COVID-19 armamentarium, in the setting of a changing population immunity and viral lineage landscape? The study was large, rigorously analyzed, and well presented. The retrospective approach is both an asset—in measuring real-world efficacy, which is perhaps what matters most from a population health standpoint—and a drawback, with the normal limitations of residual bias. There are 2 important retorts. First, the authors used a thoughtful, though not infallible, approach to risk adjustment, thus reducing confounding by indication. Second, immortal time bias was mitigated by the exclusion of patients from both groups who were hospitalized within 2 days of diagnosis. Additional limitations are a lack of safety outcomes and the exclusion of patients with relative contraindications to nirmatrelvir plus ritonavir, which excluded many immunosuppressed patients. As in all COVID-19 literature, the viral subvariants under study are, to some extent, out of date by the time the article goes to press. This study covered the Omicron subvariants BA.1, BA.2, and BA.2.12.1 as well as the start of the BA.5 era, but it did not include the rapidly emerging next era of increasingly immune-evasive subvariants, such as BA.4.6, BQ.1, BQ.1.1, and BA.2.75.2. The only published randomized controlled trial to evaluate nirmatrelvir plus ritonavir was the manufacturersponsored EPIC-HR (Evaluation of Protease Inhibition for COVID-19 in High-Risk Patients) study (2). This trial enrolled high-risk persons in the Delta era without previous self-reported COVID-19 or, importantly, vaccination. The group receiving nirmatrelvir plus ritonavir had a lower rate of all-cause 28-day hospitalization or mortality than the placebo group (0.72% vs. 6.45%; P< 0.001). As the pandemic continued to evolve, data reflecting broader comorbidities, vaccination status, and the Omicron era were needed. To expand the eligibility pool, a second manufacturer-sponsored trial, EPIC-SR (Evaluation of Protease Inhibition for COVID-19 in Standard-Risk Patients), was initiated that included unvaccinated persons without risk factors for severe disease and high-risk vaccinated persons. The results, available only via a manufacturer press release and not yet published or peer-reviewed, show that the trial was halted early for low rates of hospitalization and death, with no difference in the primary outcome of sustained alleviation of symptoms (3). Secondary analyses showed a nonsignificant reduction in hospitalizations and death compared with placebo for high-risk vaccinated patients. In one Omicron-era retrospective study of adults aged 18 years or older, nirmatrelvir plus ritonavir was associated with a 45% relative risk reduction in all-cause 30-day emergency department visits, hospitalization, and death (4). In a subsequent analysis among adults aged 40 years or older, the benefit of nirmatrelvir plus ritonavir in preventing hospitalization and death was confirmed primarily in those aged 65 years or older and not in those aged 40 to 64 years. Furthermore, the protective effect was amplified among persons aged 65 years or older who did not have prior immunity (5). The real-world event rates and relative risk reductions reported by Dryden-Peterson and colleagues were similar to other previously reported retrospective results in vaccinated persons (6). A smaller subgroup of unvaccinated persons in the present study had an 81% relative risk reduction in all-cause hospitalization and death, which correlates with EPIC-HR results from the Delta era but is a larger point estimate than those in subsequent studies in the Omicron era, all of which have low event rates and high numbers needed to treat. New results in the present study include a greater benefit in partially vaccinated persons, those who were vaccinated more than 20 weeks prior, and those with