Safety and efficacy of hydroxychloroquine for treatment of non-severe COVID-19 among adults in Uganda: a randomized open label phase II clinical trial

Diagnosis of COVID-19 was performed using RT-PCR at the government approved laboratories. Patients diagnosed with COVID-19 were included in the study if they were 18 years and above and had a diagnosis of COVID-19 made in the last 3 days. Patients were excluded if they had known allergies to HCQ or chloroquine, were on medications that have clinically significant interactions with HCQ, had a positive rapid test for malaria, were diagnosed with severe/critically ill COVID-19 (WHO Ordinal Scale of ≥ 5), had QTc prolongation of > 450 ms for males and > 470 ms for females, were pregnant or breastfeeding or were on chronic HCQ use. Participants found to have hypo- or hyperkalemia at baseline were withdrawn from the study.

Randomization was performed by an independent statistician using a computer generated randomization code with block randomization with varied sizes. Patients were allocated in a ratio of 1:1; to receive either HCQ 400 mg twice a day for the first day followed by 200 mg twice daily for the next 4 days plus SOC treatment or SOC treatment alone. The SOC treatment at the time included vitamin C and zinc supplementation. Symptomatic patients also received azithromycin and analgesics if necessary. Computer-generated randomization codes were enclosed in sequentially numbered opaque sealed envelopes containing treatment allocation. After meeting study eligibility criteria, the study nurse assigned the next envelope to the participant, opened the envelope and assigned treatment allocation. Treatment was immediately initiated. Participants and the trial team were not blinded. Participants who progressed to WHO ordinal scale ≥ 5 (severe or critical disease) during the study were managed according to the national clinical guidelines for COVID-19 which includes intravenous antibiotics and anticoagulation with low molecular weight heparin.

Participants were evaluated daily for new clinical symptoms, worsening or improvement of existing symptoms and adverse events during admission. An ECG was obtained at baseline, day 2 and 4. Where the QTc interval on ECG exceeded 500 ms or increased by > 60 ms above the baseline, the ECG was repeated. If the repeat QTc interval remained above these values, HCQ was discontinued. Serum ALT, visual tests using Snellen’s and Ishihara charts were measured at baseline and day 4 while serum potassium was measured only at baseline. Participants who developed grade 3 or 4 clinical or laboratory based adverse events were discontinued from medication. Participants were discharged after 2 consecutive negative SARS COV-2 PCR tests, generally after 10–14 days.

SARS COV-2 viral load (CT values) from RT-PCR testing of samples collected using nasal/orapharyngeal swabs was performed at baseline, day 2, 4, 6, 8 and 10. Following nasal/oropharyngeal swabbing, samples were stored in sterile saline solution and transported in a cooler box to the Infectious Diseases Institute Core Laboratory, which is a government approved laboratory for SARS COV2 PCR testing.

Ethics approval was obtained from the Makerere University School of Medicine Research and Ethics Committee (#REC REF 2020-137), the Uganda National Council for Science and Technology (RESCLEAR/05) and the National Drug Authority (CTA 0143). Written informed consent was obtained from all study participants and the trial was conducted according to Good Clinical Practice Guidelines in accordance with the Declaration of Helsinki. The trial was registered with ClinicalTrials.gov on 26/04/2021 (registration number NCT04860284).

An independent data and safety monitoring board (DSMB) reviewed the study protocol and oversaw the progress of the trial. Progressive data review for safety and efficacy was planned after 25% of the participants had completed 10 days of follow-up, and another as deemed necessary by the DSMB. Stopping guidelines were provided to the DSMB with the use of a Lan–DeMets spending function for the primary outcome. The first interim analysis was performed and presented to the DSMB committee on February 17, 2021, for their recommendation on whether to stop the trial for safety concern or futility or any other reason given by the committee. The trial stopped because of the national roll-out of HCQ as standard of care by the Uganda Ministry of Health.

Primary and secondary outcomes were analyzed on intention-to-treat population, and other outcomes and safety data were analyzed on complete cases. The primary outcome was median time from randomization to SARS COV-2 viral clearance by day 6. Viral load clearance was defined as a negative SARS COV-2 PCR test with no subsequent positives. Analysis of time to viral clearance was performed using Kaplan–Meier methods, and compared across the two treatment arms using log-rank test. We used Cox regression model to compare the secondary outcome of rates of viral load clearance in the two arms. Proportional hazard assumption was checked using schonefeld residuals. For other outcomes: the proportion of PCR negative conversion by day 6 and day 10, and proportion of participants with 25% reduction of SARS COV-2 viral load (CT-values) from baseline at day 6 were compared between treatment arms using Chi square test; change in SARS COV-2 viral load (CT-values) over time in the two arms compared using student’s T-test; Time to symptom clearance by day 10 was summarized using median and inter-quartile range and compared using Wilcoxon rank-sum test. Safety outcomes like incident elevated ALT (> 40 IU), incident elevated QTc interval (QTc > 450 ms in males and QTc > 470 ms in females), incident color vision loss/deficiency and adverse events were summarized using frequencies and percentage, across the treatment arms. We conducted all analyses with STATA software, version 15.1 (Texas, USA), according to the intention-to-treat principle, with two-sided type I error of 5%. A participant who did not achieve viral clearance or was lost to follow-up, withdrawn or died before analysis time (day 6 or day 10) or had missing SARS COV-2 viral load and PCR test results, was assigned non-viral load clearance at their time of censoring. Sensitivity analyses of viral clearance were performed using an adjusted Cox regression model, specified a priori included baseline age and sex. Baseline CT values and body mass index (BMI) were not included in the adjusted regression model due to very high missing values. Baseline SARS COV-2 CT-values were measured at patient enrolment. However, majority of the patients were enrolled at day 4 after first positive PCR tests. There was no correction for multiplicity on tests for secondary/other outcomes so results are reported as point estimates and 95% confidence intervals. Confidence interval widths are not adjusted for multiplicity, so intervals should be interpreted with caution.

We assumed HCQ would lower median time to viral clearance from 7 days (as per standard of care) to 4 days, with power of 80% and 5% 2-sided significance level, we calculated a sample size of 284 patients (142 per group) after accounting for 25% loss to follow-up or missing data.

There were no significant differences in adverse events such as elevated alkaline phosphatase and prolonged QTc interval among patients in the intervention and control group. Details of clinically significant laboratory abnormalities are as shown in Table 6.