A new study examining COVID-19 outcomes over time found evidence supporting the use of dexamethasone and raised concerns about using nonconcurrent controls in large platform trials.
As the standard of care for COVID-19 evolved and patient composition changed, outcomes improved, according to a new study of recovery and mortality during the pandemic.
The study, published in the Annals of Internal Medicine, examined data from the Adaptive COVID-19 Treatment Trial, a series of phase 3 trials of COVID-19 therapeutics at 94 hospitals in 10 countries from February 2020 through May 2021.
“We found that COVID-19 outcomes for hospitalized adults in the US improved over time,” lead author Gail Potter, PhD, of the National Institute of Allergy and Infectious Diseases told Contagion. “Improvements between the period Feb-May 2020 to May-July 2020 were explained by changing patient composition. The improvements seen from the period May-July 2020 through Aug-Dec 2020 were not explained by changes in patient composition. When we examined possible explanatory factors, the main difference we saw between these two periods was an increase in the use of dexamethasone (11% to 77%). Since this is an observational study, there could also be unobserved factors that contribute to these improvements."
The study included four sequential, double-blind, randomized, placebo-controlled trials of investigational therapeutics and measured 28-day recovery and mortality between comparable cohorts. The ACTT-1 trial spanned from February to May 2020; ACTT-2 was from May to July 2020; ACTT-3 was from August to December 2020; and ACTT-4 was from December 2020 to May 2021.
Changes in the standard of care from ACTT-1 to ACTT-2 included decreases in use of hydroxychloroquine and empirical antibiotics. Investigators found a 25% drop in the odds of baseline intubation. Recovery and mortality improved numerically, but adjusted hazard ratios were near 1, (1.04 and 0.90 respectively) suggesting that impvovements in outcomes were due to differences in patient composition rather than standard of care.
Between ACTT-2 and ACTT-3 changes to standard of care included an increased use of dexamethasone to 77% of remdesivir recipients up from 11%. Antibiotic use continued to decrease. Improvements in outcomes were greater, with a hazard ratio for mortality of 0.45 (confidence interval, 0.21 to 0.97).
Between ACTT-3 and ACTT-4, outcomes remained similar.
“We found improvements in COVID-19 outcomes over fairly short time periods because of changing patient composition and evolving standard of care,” Potter said. “This means that in this setting, nonconcurrent controls in large platform trials are not comparable to enrolled control participants, so including them can cause biased treatment effects unless these differences are accounted for statistically. In addition, this study provides circumstantial evidence in support of dexamethasone for treatment of COVID-19 in hospitalized adults.”
The study highlighted issues related to changes over time that could influence study results. For example, patients may be more hesitant to go to the hospital early in the pandemic and more hestitant to enroll in trials if their disease was less severe. Care and outcome definition also evolved over time, with intubation practices becoming more conserative, which could affect analysis of treatment effect.
Limitations of the study include that not all confounders were measured, mortality comparisons were limited to a 28-day interval.
“It is not clear why we found lower mortality rates than other, much larger studies, that we cited in our paper,” Potter said. “This may have to do with the group of clinics that participated in our trials. It would be interesting to investigate whether the lower mortality related to differences in patient composition, treatments given, hospital surges, resources available, or other factors.”