Individual variation in viral dynamics could underlie inconsistencies in COVID-19 antiviral treatments

By James DuckerReviewed by Emily Henderson, B.Sc.Jul 5 2021

Due to variation in viral dynamics between people, improving the overarching experimental design could help limit the observed differences in trial outcomes for developing effective antiviral treatments for coronavirus disease 2019 (COVID-19).

Research. Image Credit: Shoya Iwanami, CC0

Understanding and limiting variation to design effective antiviral drugs

Clinical trials for antiviral COVID-19 treatments have been a focal objective for research efforts since the onset of the pandemic, with results showing varying effectiveness in success.

The inconsistency in findings is a limiting factor when developing successful drugs, and a new modeling study suggests that underlying variation may originate from person-to-person differences in viral dynamics.

Understanding and addressing sources of variation is a key process to develop widely applicable treatments, and intraspecific variation has often been a challenging and contested focus in developing effective viral treatments.

According to Shoya Iwanami and Shingo Iwami of Nagoya University in Aichi, Japan, Keisuke Ejima of Indiana University, Indiana, USA, and colleagues, who present the new findings in the journal PLOS Medicine, adapting experimental designs is key to addressing challenges in individual variation.

The major finding of the modeling study suggests that recruiting trial participants shortly after symptoms begin could reduce the number of participants required to detect significant antiviral drug effects, and improve the overall outlook of the treatment in development.

Developing a successful antiviral drug for COVID-19 has global health benefits, but clinical trials to date that test candidate drugs have produced inconsistent results, perhaps due to flaws in the way the trials are designed and conducted.

To address this potential shortcoming, Iwanami and colleagues used a model of the dynamics of SARS-CoV-2 once it has infected a person, which was then combined with clinical data to examine how viral load changes over time.

The models showed significant variation in the rate of decline in viral load between patients. The authors then hypothesized such differences between patients may underlie the inconsistent results reported in non-randomized clinical trials so far.

Improving designs for clinical trials developing COVID-19 antiviral treatments

From the first findings, the researchers then used it to simulate a series of hypothesized findings from randomized clinical trials for COVID-19 antiviral drugs that successfully interrupt virus replication.

This added computational simulation found that even if a potential drug decreased reduced viral replication by 95%, the associated randomized clinical trial would require over 13,000 people to receive the drug tested, plus the same number of people to receive a placebo for comparison, to detect statistically significant differences in viral load.

Such a large sample size would be unreasonable and difficult to test in most treatments.

However, when the models factored in participants that were treated within 24 hours of the onset of their symptoms, they found that up to 600 participants were needed for each treatment group.

The reduction in sample size indicates that randomized clinical trials for COVID-19 drugs could be improved by enrolling participants as soon as possible after symptoms appear, or by setting enrolment criteria based on the time that has passed since symptom onset.

Future studies could therefore employ more detailed models of SARS-CoV-2 dynamics to produce more reliable calculations of the numbers of participants needed for randomized clinical trials to produce consistent results. Improving model accuracy and data quality by incorporating elements such as regional variation, age group, or the presence of variants, could further benefit the development of antiviral treatments.

We found that if patients are recruited to clinical trials regardless of the time since symptom onset, the number should be over 10,000, which is unreasonably large," he then added. "This is because many patients are recruited too late to see antiviral treatment effect. Thus, we suggest recruiting only those who are still 'new' since symptom onset.

If we recruit only patients in 2 days since symptom onset, only 500 patients need to be recruited. The approach we developed can be applied to other types of drugs and different infectious diseases. We are hoping to develop an online platform which supports designing clinical trials.”

Dr. Iwami