By Salim S. Abdool Karim, M.B., Ch.B., Ph.D., and Nikita Devnarain, Ph.D.
In practicing evidence-based medicine, physicians use the best evidence currently available on safety and efficacy in making decisions on treatment choices for their patients. During the Covid-19 pandemic, some of the early treatment trials were rushed, leading to studies that were badly conducted1 or had too few patients.2 As a result, initial evidence of the efficacy of some Covid-19 treatments could not be replicated,3,4 but these drugs were already in widespread use by then, and some clinicians have been reluctant to change to proven efficacious alternatives. Ivermectin and fluvoxamine, in particular, are still widely prescribed, even though evidence has been steadily accumulating to indicate that both treatments at acceptable doses are not effective for Covid-19.3-5
In this issue of the Journal, Bramante et al.6 report the results of the COVID-OUT randomized, controlled trial of oral metformin, ivermectin, and fluvoxamine for the early treatment of SARS-CoV-2 infection in 1323 outpatients. The investigators found no reductions in hypoxemia, emergency department visits, hospitalization, or death associated with any of the three drugs. A strength of the trial is the selection of adults between the ages of 30 and 85 years who were at high risk for severe Covid-19 because of overweight or obesity. However, as a result, the trial may not be readily generalizable to patients at lower risk for severe disease. One secondary analysis, which should be interpreted with caution, suggested that metformin may reduce a composite of emergency department visit, hospitalization, or death in this population with overweight or obesity, a finding that indicates no more than the need for further investigation at this time.
When this trial was initiated in 2020, evidence on the three treatments was either unavailable or equivocal. Since then, data have been accumulating from several clinical trials, including meta-analyses of metformin, ivermectin, and fluvoxamine. In a combined analysis of antidiabetic agents involving more than 3 million patients with diabetes and Covid-19 in 24 observational studies and 110 patients in one clinical trial,7 the investigators found that the use of metformin before hospital admission, but not in-hospital use, correlated with reduced mortality. In a meta-analysis of fluvoxamine involving 2208 outpatients with nonsevere cases of Covid-19 in three trials,8 investigators found that those who received fluvoxamine did not have a lower incidence of hospitalization, mechanical ventilation, or death than those in the control groups. For ivermectin, a meta-analysis of 16 trials8 involving 2407 patients with both severe and nonsevere illness showed no reliable evidence of reductions in mechanical ventilation, hospital admission, duration of hospitalization, clinical severity, or mortality; in addition, the investigators found no effect related to the dose of ivermectin. In light of this available evidence of nonefficacy for ivermectin and fluvoxamine, how much evidence of nonefficacy is enough?