COVID-19 Updates: Remdesivir & Serology
Angela Myers, MD, MPH | Director, Division of Infectious Diseases | Associate Director, Infectious Diseases Fellowship Program; Associate Professor of Pediatrics, UMKC School of Medicine
Mary Anne Jackson, MD |Dean and Professor of Pediatrics - UMKC School of Medicine | Medical Editor, The Link Newsletter
The data safety monitoring board for the large randomized clinical trial of remdesivir, a new SARS-CoV-2 antiviral, met on April 27 to review the results of the primary trial end point: time to recovery defined as ready for hospital discharge or return to normal activity. Of note, the primary endpoint was adjusted on April 16 from mortality to time to recovery. This adjustment was made before any data was released or analyzed. As you may recall, time to recovery is a measure often used in influenza treatment trials (e.g. oseltamivir and baloxivir). The trial enrolled 1,063 patients, and the early data have shown that treatment hastened the time to recovery from SARS-CoV-2 disease by ~1/3 (p<0.001). This amounted to a median time to recovery of 11 days for patients in the treatment arm compared to 15 days for patients in the placebo arm.1 No significant clinical difference was seen between the 5-day treatment arm and the 10-day treatment arm. While of borderline statistical significance, there may be a reduced mortality benefit as well, with 8% mortality in the treatment arm compared to 11.6% in the placebo arm (p=0.059).
The Food and Drug Administration made the announcement on May 1, 2020 that this medication now has emergency use authorization in hospitalized patients with COVID-19. More details from this trial will be published in the coming weeks in the peer-reviewed literature. In addition, there are ongoing clinical trials to continue to determine which patients benefit the most from treatment with remdesivir.
However, another clinical trial of remdesivir failed to show a benefit in a study of adults with SARS-CoV-2 infection at 10 hospitals in China.2 This study enrolled patients <2 weeks from symptom onset with low 02 saturations and radiographically confirmed pneumonia. A total of 237 patients received remdesivir IV or placebo for 10 days in a 2:1 ratio. Patients could also receive other medications intended to treat SARS-CoV-2 infection including corticosteroids, lopinavir-ritonavir, and interferons. The primary endpoint for this study was time to clinical improvement, a different outcome than the U.S. study. In this study, remdesivir was not associated with reduced time to clinical improvement (HR 1.27 [95% CI 0.89-1.80]), mortality (1.1% difference [95%CI -8.1-10.3]), or time to clearance of virus (virus undetectable in 78% of both groups by day 28).2 Adverse events were similar in treatment (66%) and placebo (64%) patients, although more patients in the treatment group stopped therapy early due to gastrointestinal side effects (e.g. nausea, vomiting), elevations in transaminases, and worsening cardiopulmonary status (18% vs. 5%). It should be noted that only a small proportion of patients in this study were severely ill with 0.4% requiring mechanical ventilation, and that study enrollment did not meet the target enrollment due to a rapid decline in new cases due to implementation of public health restrictions. This brought their statistical power down from 80% likelihood of detecting a difference (typical in a clinical trial) to only 58%. So, there is a 42% chance that there was a difference, but it was not detected due to small subject numbers, making it difficult to draw conclusions from the data.
Now three months into the COVID-19 pandemic, attention has turned to defining the seroprevalence from this disease and understanding the nature of and durability of the immunologic response.
Immunity from infection and protection from future disease follows three scenarios:
- Antibody develops after primary infection
- Antibody is passively delivered and lasts ≤3 months through plasma or immunoglobulin infusion
- Active immunity following vaccination
Immunity following natural infection is complicated and is not always enduring. Life-long immunity that protects one from infection after subsequent exposure to viruses necessitates a strong, specific and durable response to the primary infection. In some cases, antibody develops after primary infection and that antibody protects against reinfection. The immunologic response after varicella or measles infection, for instance, protects an individual across their lifetime and this protection is replicated following varicella and MMR vaccine. In other cases, antibody develops but wanes and is not protective against repeat exposure and reinfection in subsequent seasons. For example, while the production of virus-neutralizing antibody follows RSV infection, the antibody response is not sustained and, for children and even in adults, infections occur throughout life.
Seasonal influenza disease occurs because of antigenic drift and individuals remain at risk for infection in subsequent years. That is why we require annual vaccination. Influenza pandemics occur when antigenically shifted virus emerges and these viruses transmit rapidly and effectively, as the entire population lacks immunity to the newly emergent strain. Even after a large-scale population exposure and influenza pandemic, because antigenic shift continues the following year, there is no guarantee that one will be immune to the next season’s virus and protection requires vaccination.
In the case of COVID-19, we have not identified the emergence of significant genetically mutated virus, so it is possible that antibody which develops after infection may be protective against future exposures for a period of time. Numerous antibody tests have come to the market in recent weeks. Their goal is to help determine who has been infected and to confirm when enough herd immunity is present (estimates from the University of Minnesota report 60-70% of the population need be infected) so that we can try for some form of social normalcy. Uncertainty over the utility of serology center around at least 7 unknowns regarding the clinical value/utility of these tests.
- First, we want a test that detects antibody to the right antigen to predict protection. Right now, serologic tests that evaluate response to different viral antigens or different portions of antigens (e.g. nucleoprotein or spike protein) show varying performance—and we don’t know precisely which antigenic response is protective.
- Second, we want a test that accurately confirms past infection. Many of the tests look only at IgM or IgG, some evaluated combined IgM plus IgG antibodies, and some look at IgM separately from IgG. IgM responses which are generally short lived, may not accurately identify an individual who was infected if obtained later in the course, when the IgM response has waned. Conversely, IgG antibody that requires 2-4 weeks to develop, could result in a negative test in an infected person if obtained too early after infection. When IgM and IgG are combined into one test, it is impossible to accurately identify when the infection occurred. Was it the most recent URI illness, or was it the one they had a month ago?
- Third, we need to understand which antibodies may be protective.3 And if protective, we want to understand the duration of protection. Although measles antibody confers life-long immunity, this is not true for many other respiratory viruses (e.g. RSV some other human coronaviruses) where reinfections occur throughout life.
- Fourth, if protective, at what level is it protective? Most of the antibody tests provide a +/- or qualitative result, not a quantitative result or titer.
- Fifth, a good serologic test must measure specific SARS-CoV-2 antibody, and not reflect past experience with a common coronavirus as we know that at least 85% of adults are likely to be seropositive for at least one of the common 4 coronaviruses.4
- Sixth, we need to understand which antibody response is the critical signal of protection. Coronaviruses are known to elicit several functional categories, e.g. mucosal, neutralizing, or binding versions. Current commercial tests do not differentiate these categories in the detected antibody. Mucosal – not serum -antibody and perhaps just secretory IgA to SARS-CoV-2 may be what is protective. Binding antibodies may not be protective as they could exacerbate disease by facilitating virus entry into host cells as opposed to a subset of circulating neutralizing antibodies that may be the only protective category. So, serum antibody may need to be considered as a surrogate for protective antibody – and only through much research can surrogate protective concentrations be determined.
- Seventh, disease prevalence is important in interpreting the positive predictive value of any test regardless of whether it is a molecular or antibody test. When disease prevalence is ≤5%, then the likelihood of a false positive is greater than a true positive.4 The same is true for RSV and influenza testing outside of the respiratory season.
As with all things COVID-19 related, we will continue to amass more data to help guide diagnostic and treatment decisions over the coming weeks and months. It is likely we will find the right test and that seroprevalence data will be available in the near future.
- Accessed May 1, 2020. https://www.niaid.nih.gov/news-events/nih-clinical-trial-shows-remdesivir-accelerates-recovery-advanced-covid-19
- Wang Y, Zhang D, Du G, et al. Remdesivir in adults with severe COVID-19: a randomized, double-blind, placebo controlled multicenter trial. The Lancet. April 29, 2020.
- Torres R, Rinder HM. Double-edged Spike-Are SARS-CoV-2-Serologic Tests Safe Right Now? Laboratory Medicine. 2020. https://academic.oup.com/labmed/advance-article/doi/10.1093/labmed/lmaa025/5823966
- Accessed May 1, 2020. https://www.idsociety.org/globalassets/idsa/public-health/covid-19/idsa-covid-19-antibody-testing-primer.pdf