Effect of Ivermectin on Time to Resolution of Symptoms Among Adults With Mild COVID-19

What is the effect of ivermectin on the duration of symptoms in adults with mild COVID-19?

Findings In this randomized clinical trial that included 476 patients, the duration of symptoms was not significantly different for patients who received a 5-day course of ivermectin compared with placebo (median time to resolution of symptoms, 10 vs 12 days; hazard ratio for resolution of symptoms, 1.07).

This meaning The findings do not support the use of ivermectin for the treatment of mild COVID-19, although larger trials may be needed to understand the effects on other clinically relevant outcomes

Abstract

Importance Ivermectin is widely prescribed as a potential treatment for COVID-19 despite uncertainty about its clinical benefit.

Objective To determine whether ivermectin is an efficacious treatment for mild COVID-19.

Design, Setting, and Participants Double-blind, randomized trial was conducted at a single site in Cali, Colombia. Potential study participants were identified by simple random sampling from the state’s health department electronic database of patients with symptomatic, laboratory-confirmed COVID-19 during the study period. A total of 476 adult patients with mild disease and symptoms for 7 days or fewer (at home or hospitalized) were enrolled between July 15 and November 30, 2020, and followed up through December 21, 2020.

Intervention Patients were randomized to receive ivermectin, 300 μg/kg of body weight per day for 5 days (n = 200) or a placebo (n = 200).

Main Outcomes and Measures Primary outcome was time to resolution of symptoms within a 21-day follow-up period. Solicited adverse events and serious adverse events were also collected.

Results Among 400 patients who were randomized in the primary analysis population (median age, 37 years [interquartile range {IQR}, 29-48]; 231 women [58%]), 398 (99.5%) completed the trial. The median time to resolution of symptoms was 10 days (IQR, 9-13) in the ivermectin group compared with 12 days (IQR, 9-13) in the placebo group (hazard ratio for resolution of symptoms, 1.07 [95% CI, 0.87 to 1.32]; P = .53 by log-rank test). By day 21, 82% in the ivermectin group and 79% in the placebo group had resolved symptoms. The most common solicited adverse event was a headache, reported by 104 patients (52%) given ivermectin and 111 (56%) who received a placebo. The most common serious adverse event was a multiorgan failure, occurring in 4 patients (2 in each group).

Conclusion and Relevance Among adults with mild COVID-19, a 5-day course of ivermectin, compared with placebo, did not significantly improve the time to resolution of symptoms. The findings do not support the use of ivermectin for the treatment of mild COVID-19, although larger trials may be needed to understand the effects of ivermectin on other clinically relevant outcomes.

Introduction

Therapeutic approaches are needed to improve outcomes in patients with COVID-19. Ivermectin, a widely used drug with a favorable safety profile, is thought to act at different protein-binding sites to reduce viral replication. Because of evidence of activity against SARS-CoV-2 in vitro and in animal models, ivermectin has attracted interest in the global scientific community and among policymakers. Several countries have included ivermectin in their treatment guidelines, leading to a surge in the demand for the medication by the general population and even alleged distribution of veterinary formulations. However, clinical trials are needed to determine the effects of ivermectin on COVID-19 in the clinical setting.

Viral replication may be particularly active early in the course of COVID-19 and experimental studies have shown antiviral activity of ivermectin in the early stages of other infections. The hypothesis of this randomized trial (EPIC trial [Estudio Para Evaluar la Ivermectina en COVID-19]) was that ivermectin would accelerate recovery in patients with COVID-19 when administered during the first days of infection.

Methods

Study Design and Patients

This study was approved by the Colombian Regulatory Agency (INVIMA No. PI-CEP-1390), the independent ethics committees of Corporación Científica Pediátrica, and collaborating hospitals in Cali, Colombia, and conducted in accordance with Declaration of Helsinki and Good Clinical Practice guidelines. Written informed consent was obtained from all patients. Full details of the trial can be found in the protocol.

This double-blind, randomized trial of ivermectin vs placebo was conducted from July 15 to December 21, 2020, by Centro de Estudios en Infectología Pediátrica in Cali. Study candidates were identified from the state’s health department electronic database of all patients with a positive result from a SARS-CoV-2 reverse transcriptase-polymerase chain reaction or antigen test performed in any of the Colombian National Institute of Health–authorized laboratories in the city of Cali.

Potential study participants were identified and selected by simple random sampling from the state’s database. Adult men and non–pregnant or breastfeeding women were eligible if their symptoms began within the previous 7 days and they had mild disease, defined as being at home or hospitalized but not receiving high-flow nasal oxygen or mechanical ventilation (invasive or noninvasive). Patients were excluded if they were asymptomatic, had severe pneumonia, had received ivermectin within the previous 5 days, or had hepatic dysfunction or liver function test results more than 1.5 times the normal level. Details of selection criteria can be found in the protocol. Health disparities by race/ethnicity have been reported in COVID-19 infections. Hence, information on this variable was collected by study personnel based on fixed categories as selected by the study participants.

Randomization

Eligible patients were randomly assigned in a 1:1 ratio to receive either oral ivermectin or placebo in solution for 5 days. Patients were randomized in permuted blocks of 4 in a randomization sequence prepared by the unblinded pharmacist in Microsoft Excel version 19.0 who provided masked ivermectin or placebo to a field nurse for home or hospital patient visits. Allocation assignment was concealed from investigators and patients.

Interventions

Study patients received 300 μg/kg of body weight per day of oral ivermectin in solution or the same volume of placebo for 5 days. Ivermectin was provided by Tecnoquímicas SA in bottles of 0.6% solution for oral administration. Patients were asked to take the investigational product on an empty stomach, except on the first study day, when it was administered after screening and randomization procedures took place.

Up to August 26, 2020, the placebo was a mixture of 5% dextrose in saline and 5% dextrose in distilled water, after which the placebo was a solution with similar organoleptic properties to ivermectin provided by the manufacturer. Because blinding could be jeopardized due to the different taste and smell of ivermectin and the saline/dextrose placebo, only 1 patient per household was included in the study until the manufacturer’s placebo was available. Bottles of ivermectin and placebo were identical throughout the study period to guarantee double-blinding.

Procedures

A study physician contacted potential study participants by telephone to verify selection criteria for eligibility and obtain informed consent. Patients were then visited at home or in hospital by a study nurse who drew blood for liver enzyme evaluations and performed a urine pregnancy test. Eligible patients were revisited by a study nurse for enrollment, documentation of baseline demographic and clinical information, and dispensing of the investigational product. The investigational product was left with the patient for self-administration on days 2 through 5. Subsequently, patients were contacted by telephone by study staff on days 2 through 5, 8, 11, 15, and 21 for a structured interview. A study of physician-reviewed medical records of hospitalized patients to obtain the information required by the protocol. After the study end (day 21), unused or empty investigational product bottles were collected to certify adherence. Data were entered into an electronic database and validated by the site’s quality management department.

Outcome Measures

The primary outcome was the time from randomization to complete resolution of symptoms within the 21-day follow-up period. The 8-category ordinal scale used in this trial has been used in different COVID-19 therapeutic trials and is recommended by the World Health Organization’s R&D Blueprint. It consists of the following categories: 0 = no clinical evidence of infection; 1 = not hospitalized and no limitation of activities; 2 = not hospitalized, with limitation of activities, home oxygen requirement, or both; 3 = hospitalized, not requiring supplemental oxygen; 4 = hospitalized, requiring supplemental oxygen; 5 = hospitalized, requiring nasal high-flow oxygen, noninvasive mechanical ventilation, or both; 6 = hospitalized, requiring extracorporeal membrane oxygenation, invasive mechanical ventilation, or both; and 7 = death. Time to recovery was defined as the first day during the 21 days of follow-up in which the patient reported a score of 0.

Secondary outcomes included the proportion of patients with clinical deterioration, defined as those with worsening by 2 points (from the baseline score on the 8-category ordinal scale) since randomization. Additional secondary outcomes were the clinical conditions as assessed by the 8-category ordinal scale on days 2, 5, 8, 11, 15, and 21; however, data for days 2 and 15 are not reported here. The proportion of patients who developed fever and the duration of fever since randomization and the proportion of patients who died were also reported. Proportions of patients with new-onset hospitalization in the general ward or intensive care unit or new-onset supplementary oxygen requirement for more than 24 hours were combined into a single outcome called escalation of care. The frequency of incident cases of escalation of care, as well as the duration in both treatment groups, was reported. Evaluation of adverse events (AEs) included solicited AEs, AEs leading to treatment discontinuation, and serious AEs. AEs were classified according to the National Cancer Institute’s Common Terminology Criteria for Adverse Events version 5.0.

Post Hoc Analysis

Given that some patients’ need for escalation of care was imminent when randomized, the frequency of incident cases of escalation of care occurring 12 or more hours after randomization and the duration up to day 21 in both treatment groups were reported. A comparison of the proportions of patients who required emergency department (ED) or telemedicine consultation was also performed.

Statistical Analysis

The primary outcome was originally defined as the time from randomization until worsening by 2 points on the 8-category ordinal scale. According to the literature, approximately 18% of patients were expected to have such an outcome. However, before the interim analysis, it became apparent that the pooled event rate of worsening by 2 points was substantially lower than the initial 18% expectation, requiring an unattainable sample size. Therefore, on August 31, 2020, the principal investigator proposed to the data and safety monitoring board to modify the primary endpoint to time from randomization to complete resolution of symptoms within the 21-day follow-up period. This was approved on September 2, 2020. The original sample size of 400 based on the log-rank test for the new primary endpoint was kept, using ivermectin to placebo assignment ratio of 1:1. This would allow the detection of 290 events of interest (symptom resolution), assuming that 75% of patients would have the outcome of interest at 21 days, with a 2% dropout rate. This would provide an 80% power under a 2-sided type I error of 5% if the hazard ratio (HR) comparing ivermectin vs placebo is 1.4, corresponding to a 3-day faster resolution of symptoms in patients receiving ivermectin, assuming that time to resolution of symptoms is 12 days with placebo.24 With an HR of 1.4, 75% and 85% of patients in the placebo and ivermectin groups, respectively, would experience the outcome of interest at 21 days.

On October 20, 2020, the lead pharmacist observed that a labeling error had occurred between September 29 and October 15, 2020, resulting in all patients receiving ivermectin and none receiving placebo during this time frame. The study blind was not unmasked due to this error. The data and safety monitoring board recommended excluding these patients from the primary analysis but retaining them for sensitivity analysis. The protocol was amended to replace these patients to retain the originally calculated study power. The primary analysis population included patients who were analyzed according to their randomization group, but excluded patients recruited between September 29 and October 15, 2020, as well as patients who were randomized but later found to be in violation of selection criteria. Patients were analyzed according to the treatment they received in the as-treated population (sensitivity analysis).

The primary endpoint of time from randomization to complete resolution of symptoms with ivermectin vs placebo was assessed by a Kaplan-Meier plot and compared with a log-rank test. The HRs and 95% CIs for the cumulative incidence of symptom resolution in both treatment groups were estimated using the Cox proportional hazards model. The proportional hazards assumption was tested graphically using a log-log plot and the test of the nonzero slope. There was no evidence to reject the proportionality assumption.

The time to complete resolution of symptoms was assessed after all patients reached day 21. Data for patients who died or lacked symptom resolution before day 21 were right-censored at death or day 21, respectively. Evaluation of the effect of the treatment in each study visit using the 8-point ordinal scale was estimated using the proportional odds ratio (OR) with its respective 95% CI with an ordinal logistic regression. The proportional odds assumption was met according to the Brant test. The 8-point ordinal scale was inverted in its score, where 0 corresponded to death and 7 to a patient without symptoms.

For sensitivity analysis, primary and secondary endpoints were compared in the as-treated population.

Clustered standard errors were estimated to adjust for the correlation between multiple patients from the same household. Statistical significance was set at P < .05, and all tests were 2-tailed. Because of the potential for type I error due to multiple comparisons, findings for analyses of secondary endpoints should be interpreted as exploratory. Statistical analyses were done with Stata version 16.0 (StataCorp). Bootstrapping 95% CIs for differences of medians were calculated with R statistical package version 3.6.3 (The R Foundation).