1.1 Study design
1.1.1 Study type
This study was a phase 3 randomized, double-blind clinical trial evaluating the efficacy and safety of CVO+C for the treatment of COVID-19.
1.1.2 Study site
The study was performed in the Analamanga region of Madagascar. Subjects were recruited from Voara village in Andohatapenaka, Antananarivo, Madagascar, and from three secondary healthcare centers in Antananarivo.
Follow-up was conducted at the health center of the CNARP (Centre National d’Application de Recherches Pharmaceutiques)
1.2 Patients
Subjects with positive RT-PCR test results for SARS-CoV-2 were asked to participate in the study. They were included if they met the inclusion criteria for the study: adult male or non-pregnant (checked with a pregnancy test) female patient capable of understanding the information given, providing written informed consent (including consent for the collection of oropharyngeal swabs and venous blood in accordance with the protocol), with laboratory-confirmed (RT-PCR on nasopharyngeal/throat swabs) SARS-CoV-2 infection. The illness could be of any duration, but the patients had to display infiltrates on radiological imaging (chest X ray or CT-scan) or clinical signs on examination (rales/crackles). Patients had to have a creatinine concentration ≤ 110 µmol/L, a creatinine clearance rate (EGFR) ≥ 60 mL/min /1.73 m2, ASAT and ALAT levels no higher than five times the upper limit of the normal range (ULN), and TBIL levels no higher than twice the ULN. In addition, patients had to have a normal baseline ECG for inclusion in this study, and ECG results had to remain normal throughout the study.
Patients were excluded if their ALAT/ASAT levels were more than five times the ULN, if they were pregnant or breast-feeding, had severe (stage 4) chronic kidney disease or required dialysis (eGFR<30), were expected to be transferred to another non-participating hospital site within 72 hours, were allergic to any of the study drugs, had shortness of breath, known prolonged QT syndrome or were using other drugs known to prolong the QT/QTc interval. We also excluded patients with other types of viral pneumonia
1.3 Randomization and blinding
The patients were randomly assigned to double-blind treatment or placebo, in a 1:1 ratio. One of the investigators provided code following simple randomization procedures using a computer-generated randomization. The code was kept by the study pharmacist, who provided participants medication for 15 days. Patients and physicians kept blinded to the allocation.
1.4 Study medication
The patients were randomized to two arms. The patients in the treatment arm received CVO+C (Pharmalagasy, batch no. 1923-001) for oral administration, and were required to take three capsules per day for 15 days. The patients in the control arm received a placebo (Pharmalagasy, beta cyclodextrin and magnesium stearate, batch no. 1919-001) under the same conditions (3 capsules per day for 15 days) as the study treatment. The study treatment and the placebo were identical in physical appearance and taste.
1.5 Follow-up
The patients enrolled in the study were required to attend follow-up visits on days 7, 14, 21 and 28. At each visit, any adverse events or serious adverse events were recorded, viral load was determined (CT method for the ORF1ab and N genes of thevirus), and clinical, hematological and biological parameters (liver and kidney function) were assessed.
1.6 Outcomes
The primary outcome was the proportion of patients with complete clearance of SARS-CoV-2 virus from oropharyngeal samples on day 28 and an absence of severe and serious adverse events. The secondary outcomes were recovery time, hematological parameters (leukocytes, lymphocytes and ESR), biomarkers of liver function (ASAT and ALAT), renal function (creatinine levels) and blood sugar levels on days 7, 14, 21 and 28. Safety outcomes included adverse events occurring during treatment, serious adverse events, and premature discontinuation of treatment. Treatment was stopped early if a severe or serious adverse event occurred. The participant received the official standard of care.
1.7 Statistical analysis
1.7.1 Required sample size
Assuming a treatment efficacy of 40% for the placebo arm and 60% for the CVO+C arm, with equal numbers of patients assigned to each arm, a margin of error of 0.03, a power of 80% and an alpha risk of 5%, we would need to include 306 patients, with 153 patients per arm, to detect a significant difference between the two arms. Assuming an attrition rate of 10% due to deaths and withdrawals from the study, a total sample size of 338 patients would be required.
1.7.2 Statistical analyses
A per-protocol analysis was performed, by the endpoint analysis method. For quantitative endpoints and dichotomous qualitative endpoints measured at regular intervals, the last known values were used if the subjects were lost to follow-up. The criteria for treatment efficacy were a negative RT-PCR test for SARS-CoV-2 at the end of the study, with no adverse events or with only mild (grade 1) or moderate (grade 2) adverse events.
Regardless of the RT-PCR result at the end of the study (positive or negative), treatment was considered to have failed if the patient died, displayed serious adverse events (grade 3 or 4), was hospitalized with or without oxygenotherapy (via a mask, nasal prongs, high-flow oxygen treatment, intubation and mechanical ventilation with or without additional organ support, such as renal replacement therapy, or extracorporeal membrane oxygenation) or had been released from hospital but was unable to perform everyday activities normally.
Adverse events were considered mild (grade 1) if they were transient, required little or no treatment and did not interfere with everyday activities. They were considered moderate (grade 2) if they interfered with everyday activities, causing discomfort but no significant or permanent risk of harm to the subject, and were alleviated by additional treatment.
Two grades of serious adverse events were considered. Grade 3 severe adverse events were defined as events disrupting everyday activities or significantly affecting clinical status, generally incapacitating the subject and possibly requiring intensive treatment interventions. Grade 4 serious adverse events were defined as potentially life - threatening.
Statistical analyses were performed with STATA® 13.1 (Copyright 1985-2013 StataCorp LP, Statistics/Data Analysis, StataCorp, 4905 Lakeway Drive, College Station, Texas 77845 USA)
Qualitative variables are expressed as percentages or composition ratios for descriptive statistical analysis. They were analyzed with c² tests, Fisher’s exact tests and Wilcoxon Mann-Whitney tests. Quantitative variables are expressed as medians and interquartile range. Normally distributed quantitative variables were analyzed with t-tests, whereas Wilcoxon rank-sum tests were used to compare the two groups for non-normally distributed variables. All tests were one-tailed, and a p-value < 0.05 was considered significant. The odds ratio (OR) corresponding 95% confidence intervals (CI) were calculated in patients treated with CVO+C compared with placebo. The overall probability of conversion to negativity for RT-PCR tests was estimated by determining the time taken for viral RNA levels to become undetectable. This analysis was performed by the Kaplan-Meier method, with log-rank tests used for comparisons.