Ongoing Need for Clinical Trials and Contemporary End Points for Outpatient COVID-19

Abstract

The American College of Physicians presents recommendations for the outpatient treatment of COVID-19 based on Sommer and colleagues' systematic review. The editorialists commend the authors of the recommendations and review for trying to summarize the rapidly evolving literature into clear practice points and discuss the challenges of continually updating reviews and associated recommendations as new evidence emerges and relevant outcomes evolve.

In their article, Qaseem and colleagues (1) present version 1 of the American College of Physicians' living, rapid practice points for the outpatient treatment of confirmed COVID-19. These practice points are based on a review by Sommer and colleagues (2) that included all trials on the Epistemonikos COVID-19 L·OVE Platform up to 4 April 2022. Although it is beyond the scope of this editorial to comment on the accuracy of that platform, we commend the authors for trying to summarize the rapidly evolving literature into clear practice points. They have also created a hierarchy of outcomes of interest, including death, recovery, hospitalization, and serious and nonserious adverse events. However, it is a testament to how rapidly things are changing that this living, rapid review is already more than 6 months out of date for many reasons.

With respect to molnupiravir (Practice Point 1), the recommendation for use is based on 2 randomized controlled trials involving 1637 participants. However, there are many more trials involving molnupiravir outside the MOVe-OUT program (3), and up to a dozen trials remain unpublished as of November 2022. Ignoring the major issue of publication bias, the real challenge with this recommendation comes from PANORAMIC (Platform Adaptive trial of NOvel antiviRals for eArly treatMent of covid-19 In the Community [4]) from the United Kingdom. PANORAMIC recruited 25 783 participants between 8 December 2021 and 27 April 2022—more than 10 times the number included in the Annals systematic review. It showed that molnupiravir was not effective at reducing hospitalization or death due to COVID-19 in high-risk outpatients, of whom 98% were vaccinated and who primarily had Omicron variant infections. Although PANORAMIC showed a time-to-recovery benefit, this open-label study is subject to bias for assessing time to recovery. Little evidence of time-to-recovery benefit was present in the original, double-blind, placebo-controlled MOVe-OUT trial (3). It is unlikely there will ever be a larger outpatient randomized COVID-19 trial.

With respect to Practice Points 2 (nirmatrelvir–ritonavir) and 3 (remdesivir), the major issue surrounds the populations that have been studied. The only published randomized controlled trials—EPIC-HR (Evaluation of Protease Inhibition for Covid-19 in High-Risk Patients [5]) (nirmatrelvir–ritonavir) and PINETREE (6) (remdesivir)—involved high-risk outpatients who were unvaccinated and were having their first COVID-19 illness, and these trials were conducted before Omicron. The standard-risk EPIC-SR trial (nirmatrelvir–ritonavir) was designed to involve low-risk unvaccinated and high-risk vaccinated patients, but EPIC-SR failed to meet the primary outcome (time to symptom improvement), and no statistical reduction in hospitalization occurred (7). To our knowledge, remdesivir has never been studied in randomized trials involving vaccinated patients, nor have monoclonal antibodies. Thus, a recommendation to use these products represents a substantial stretch from the actual evidence. In the Omicron era (and later), with natural and vaccine- or booster-derived immunity in most of the population, the effectiveness of these medicines remains unclear. The PANORAMIC platform is recruiting patients to nirmatrelvir–ritonavir at present and may provide the most insights (8).

Repurposed medicines, such as fluvoxamine (Practice Point 14), are recommended against (9), and metformin (10) is not addressed. Of concern is the 2-tiered definition of hospitalization being subtly used, where more than 24 hours of acute care is an acceptable definition of “hospitalization” for the EPIC-HR and MOVe-OUT trials (3, 5) but unallowable for repurposed agents. When fluvoxamine, 100 mg twice daily, is compared using the same definition of 24 hours of acute care or hospitalization, a modest benefit exists (9). Repurposed therapies remain highly relevant for low- and middle-income countries worldwide where expensive therapies are unavailable.

This editorial may feel nihilistic; however, the fact that hospitalization has become extremely uncommon compared with before vaccine availability is a testament to the successes of the public health campaigns that have fueled SARS-CoV-2 vaccination and boosting. Clinicians are still seeing COVID-19 and need guidance. Yet, the creation of guidelines may further erode the equipoise needed to perform the definitive trials that we truly require. Further, the widespread use of these agents in the absence of trials disincentivizes the manufacturers from conducting or allowing such trials. How do we move forward with generating the necessary evidence for a rational COVID-19 outpatient strategy, and how do we keep our recommendations up to date with a rapidly changing evidence base?

First, we really do need evidence generated for therapeutics in persons who have been multiply vaccinated or have recovered from COVID-19. Where the idea of receiving a placebo therapy may be unpalatable to some clinicians and patients, we can use multiple active agents with placebo controls. The ACTIV-6 (Accelerating COVID-19 Therapeutic Interventions and Vaccines) platform is one example of a rigorous active and placebo-controlled randomized trial that could potentially help answer these questions; however, the ACTIV-6 trial is scheduled to end enrollment in early 2023. Second, we will clearly need to move beyond dichotomized thinking surrounding end points. With hospitalization and death now rare in outpatients, trials powered on these outcomes become completely infeasible. What kind of end points should we consider for outpatient COVID-19 trials? The same type of end points that should inform clinical decision making and health care economics moving forward. With emergency departments and urgent care centers worldwide already under tremendous pressure in fall 2022, and with influenza and other respiratory viruses further driving health care use, a good argument could be made to include these types of care in any outcome. Preventing disease progression, resulting in reductions in emergency department visits or hospitalizations, is a clinical benefit to patients and a benefit to society in 2022 and beyond. People who are sick also want to feel better faster and avoid long-term sequelae of infection. Thus, time to recovery and the prevalence of persistent symptoms or “long COVID” become key components of outcome assessments for trials. Because care-seeking behavior and perception of symptoms are subject to bias due to knowledge of treatment assignment, blinded placebo or active control is essential. With these outcomes being collected, we can avoid dichotomized thinking and arrive at a bigger picture of what a medication's effects are in the current COVID-19 era.

Finally, we need a way to have evidence continually updated. The concept of living, rapid reviews is fantastically bold and innovative—but they challenge the traditional peer review and publishing model. Novel strategies to move evidence synthesis as close to publication as possible will be required.