Monoclonals for patients hospitalised with COVID-19.

Monoclonal antibodies that neutralise SARS-CoV-2 have consistently reduced hospitalisation or death in outpatients with mild to moderate COVID-19.1, 2, 3 Conversely, results of randomised trials in patients who are hospitalised are mixed.4, 5, 6, 7, 8 In The Lancet Respiratory Medicine, Thomas L Holland and colleagues present results of the ACTIV-3 trial comparing intravenous tixagevimab−cilgavimab with placebo for patients hospitalised with COVID-19. Although tixagevimab−cilgavimab did not improve the primary outcome of time to sustained recovery (rate ratio [RR] 1·08 [95% CI 0·97–1·20]; p=0·21), it was associated with improved 28-day (6% vs 9%; p=0·02) and 90-day (9% vs 12%; p=0·03) mortality.

This study represents the third trial in which intravenous monoclonal antibody treatment was associated with decreased mortality in some patients who are hospitalised. The RECOVERY trial compared casirivimab–imdevimab with standard care in 9785 patients hospitalised with COVID-19. Although treatment was not associated with a 28-day mortality benefit in the overall cohort (RR 0·94 [95% CI 0·86–1·03]), mortality was lower in patients who were seronegative at the time of enrolment (RR 0·80 [0·70–0·91]). In the company-sponsored trial of 1223 patients who were hospitalised, casirivimab–imdevimab treatment was associated with a significant reduction in 28-day mortality (relative risk reduction [RRR] 35·9% [95% CI 7·3–55·7]), most predominately observed in the seronegative subgroup (RRR 55·6% (24·2–74]).

Conversely, three monoclonal antibodies (bamlanivimab, sotrovimab, BRII-196/BRII-198) previously evaluated in the ACTIV-3 platform failed the early futility analysis, which assessed pulmonary function at day 5 by means of a seven-category ordinal scale.6, 7 Enrolment was subsequently terminated, limiting the number of participants receiving each intervention to less than 200. It is noteworthy that neither the primary outcome of time to sustained recovery nor pulmonary function at day 5 (odds ratio 1·08 (95% CI 0·89–1·30) was improved with tixagevimab–cilgavimab. However, tixagevimab–cilgavimab passed the early futility analysis and was permitted to continue enrolment. Therefore, it is unknown whether these failed monoclonal antibodies would have shown a mortality benefit in a larger trial despite no effect on the ordinal outcome scales, as was the case with tixagevimab–cilgavimab.

The effect of various therapies evaluated for COVID-19 on ordinal outcome scales has been inconsistent, and these scales have plagued findings of pandemic trials for several reasons. First, each step on the scale is not necessarily of equivalent clinical significance. Second, multiple non-clinical and non-COVID-19-related factors can influence recovery, depending on how recovery is defined. Finally, an intervention might halt progression of the disease course to more severe illness (a clinically important endpoint) yet fail to hasten symptom resolution or return to baseline functional status. Therefore, when evaluating COVID-19 therapeutics in patients hospitalised with severe disease, it might be more prudent to power studies to assess the objective and more important endpoint of mortality. ACTIV-3 illustrates this point. This trial was powered on the basis of a failed primary outcome of improvement in time to sustained recovery, and it was only by coincidence that the mortality rate in patients receiving placebo was high enough to show a benefit with tixagevimab–cilgavimab. If the Data Safety Monitoring Board had not made the decision midway through the study to allow enrolment of patients on high-flow nasal cannula or had the more severe delta variant not emerged—both of which significantly increased mortality rates beyond what was anticipated at the onset of the study—it is likely that this study would have concluded no benefit to tixagevimab–cilgavimab.

The question now facing clinicians is whether the results of this trial should lead to the recommendation of intravenous tixagevimab–cilgavimab therapy for patients who are hospitalised. This is challenging to answer for several reasons, relating to the dynamic nature of both the virus and the host. To first consider the virus, the in vitro activity of tixagevimab–cilgavimab, like other monoclonal antibodies, varies with emerging SARS-CoV-2 variants. It is unclear whether changes in in vitro potency are clinically meaningful and how those changes affect efficacy. Ideally, each monoclonal antibody would be studied clinically against each SARS-CoV-2 variant. However, this is not feasible, which makes application of these results to present and future variants difficult. Furthermore, even if the relative efficacy shown in these trials is consistent across variants, the absolute benefit of therapy will change as mortality associated with SARS-CoV-2 changes. It is not appropriate to extrapolate the magnitude of a mortality benefit in a predominantly delta variant landscape to the current disease course of an omicron-subvariant-infected patient, especially when the former variant was associated with significantly worse outcomes.

To consider the host, 73% of patients in this study were unvaccinated, and the mortality benefit observed with monoclonal antibodies in patients who are hospitalised appears limited to patients who are seronegative (table ). Indeed, in ACTIV-3, 28-day mortality was identical (13%) in the tixagevimab–cilgavimab and placebo groups in the subset of patients who were vaccinated. With up to five vaccine doses now recommended, and the decreased risk of death in those vaccinated, the efficacy of tixagevimab–cilgavimab in these patients is unclear. Furthermore, nearly all unvaccinated patients have now been previously infected with SARS-CoV-2. The presence of infection-derived immunity further limits the seronegative population in which monoclonal antibodies have shown benefit. These factors collectively underscore the challenges of developing and evaluating monoclonal antibody therapies in the face of a rapidly mutating virus and evolving host population.

Table

28-day mortality rates by serostatus

	Seronegative		p value	Seropositive		p value
	Monoclonal	Placebo		Monoclonal	Placebo
Casirivimab–imdevimab	396/1633 (24%)	451/1520 (30%)	0·0010	411/2636 (16%)	383/2636 (15%)	0·30*
Casirivimab–imdevimab	24/360 (7%)	24/160 (15%)	0·0047*	26/369 (7%)	18/201 (9%)	0·42*
Tixagevimab–cilgavimab	15/307 (5%)	30/337 (9%)	0·059	22/380 (6%)	27/339 (8%)	0·29

Data are n/N (%); n is event rate; N is population in the study.

Calculated by authors of this commentary; not provided in manuscript.

Should we use intravenous tixagevimab–cilgavimab in patients who are hospitalised? It might be reasonable to consider therapy in patients either known to be seronegative or severely immunocompromised and unlikely to respond robustly to vaccination. Although it might be enticing to expand tixagevimab–cilgavimab use beyond these populations, it is not supported by the evidence and future studies are needed to ascertain the role of monoclonal antibodies in a population that is no longer immune-naive.

Importantly, these same limitations and questions hold true for other antiviral agents used in both the inpatient and outpatient setting for COVID-19. The authors and study teams of ACTIV-3 and all previous trials should be applauded for their impressive work to establish the benefit of various therapies during a rapidly evolving pandemic landscape. Their work has saved countless lives to date. However, new trials adequately powered to current event rates within the emerging variant and immunity landscape are needed to establish whether there is a benefit to any antiviral or immunomodulatory therapy for patients with COVID-19.

JMP discloses serving as a consultant to Merck, Shionogi, and Roche. EKM discloses serving as a consultant to Shionogi and is the Director of infectious diseases improvement and clinical research innovation at UPMC. In this role, EKM oversaw clinical trials involving use of sotrovimab (donated by GSK) and evaluation of tixagevimab–cilgavimab in outpatients (funded by AstraZeneca). She does not receive salary support or funding for her involvement in these trials.