Coronavirus disease 2019 (COVID-19) is an acute respiratory illness with systemic manifestations caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and is associated with considerable morbidity and mortality.1,2 Main symptoms are fever, fatigue, dry cough, and myalgia, which can progress to dyspnea or, in more severe cases, acute respiratory distress syndrome (ARDS). While most adults with COVID-19 recover, a minority develop ARDS requiring hospitalization.3,4 Approximately 80% of patients have mild illness, 15% have moderate to severe disease and 5% have critical illness.3,5

Hydroxychloroquine has generated interest and great debate as a potential treatment for COVID-19 due to widespread availability, low cost, antiviral and immunomodulatory activity, and established safety profile from historical use for other indications such as malaria and autoimmune diseases.6,7 in vitro, hydroxychloroquine limits entry of SARS-CoV-2 into cells and late stage viral replication.6–11 Furthermore, hydroxychloroquine reduces production of several pro-inflammatory cytokines potentially involved in the development of ARDS among infected patients.12–14 Associated with azithromycin, hydroxychloroquine was suggested to decrease SARS-CoV-2 viral load in non-randomized studies.11,15 Based on these mechanisms of action and clinical experience early in the pandemic, hydroxychloroquine was used as a treatment for COVID-19 in some settings and discussion regarding its efficacy is still in course.16

Treatment of COVID-19 with hydroxychloroquine has been recommended in many treatment guidelines, including in China, France, Italy, Netherlands, South Korea, United States and in Brazil, where national regulatory agencies have authorized the use of hydroxychloroquine in hospitalized patients.17–19 Our primary aim was to assess the efficacy of hydroxychloroquine in combination with azithromycin in terms of clinical and biochemical outcomes in adult patients with COVID-19 hospitalized for mild to moderate ARDS. Secondary aim was to describe the occurrence of adverse effects potentially associated with the therapeutic regimen.

This was a retrospective study with analysis of electronic medical records of patients admitted for mild and moderate COVID-19 related ARDS in a single quaternary private hospital in Rio de Janeiro, Brazil, from March through June 2020. Consecutive patients aged 18 years and older with COVID-19 confirmed by a positive reverse transcription polymerase chain reaction (RT-PCR) testing (GeneXpertR Xpress – Cepheid, CA, USA) from nasopharyngeal sample were included. Patients who were initially admitted to the intensive care unit (ICU) or with no confirmation by RT-PCR for SARS-CoV-2 infection were excluded from this study.

The severity of patients was defined as mild to moderate if, at baseline, quick sequential organ failure assessment (qSOFA) score was 0 or 1, and as severe if patients required invasive ventilatory support in the first 24 h of hospital admission.

In our hospital, from March through May 2020, hydroxychloroquine was prescribed as initial therapy associated with azithromycin. Patients received a total dose of 2400 mg of oral hydroxychloroquine prescribed for five days (400 mg b.i.d. in the first day and 400 mg once a day thereafter), associated with azithromycin 500 mg a day by oral or venous route also for five days. Although this combination was prescribed according to our hospital’s initial protocol, some patients refused to take hydroxychloroquine or had contraindication [arrythmia, corrected QT interval (QTc) >470 ms in men or 480 ms in women, renal or hepatic dysfunction] and they constituted the control group.

Data regarding age, sex, severity, body mass index (BMI), comorbidities, previous medication use, and patients’ initial extent of pulmonary involvement on chest computed tomography (CT) scans were collected retrospectively. The following clinical outcomes were assessed: discharge, transfer to ICU for invasive oxygen use, death, duration of hospitalization, duration of non-invasive oxygen use, occurrence of gastrointestinal symptoms (nausea, vomiting, abdominal pain or diarrhea) and QTc abnormalities. The electrocardiogram (ECG) was recorded on a Pagewriter TC30 device (Koninklijke Philips N.V., Netherlands) in 12 leads, with manual, automatic and rhythm measurements, maximized by the Glasgow algorithm. Biochemical assessment included serum global leucometry count and lymphocyte count (Sysmex – XE/XN; Sysmex America Inc./USA), aspartate aminotransferase (AST), alanine aminotransferase (ALT) (Alinity – CI – series; Abbott Diagnostics/USA), troponin, D-dimer and fibrinogen (ACL TOP; Werfen/Spain), and C-reactive protein (CRP) levels (Immage 800 – Beckman Coulper Inc./USA), analyzed at baseline and during hospitalization.

In patients who were receiving hydroxychloroquine, QTc interval was analyzed daily. Hydroxycholoquine was withdrawn if QTc interval increased 60 ms from baseline or was superior to 470 ms in men or 480 ms in women, and an ECG was repeated to monitor cQT interval.

Chest CT scans were obtained in helical tomography with 64 channels (Brilliance 40, Philips, Medical Systems, OH, USA). In none of the tests performed, iodinated contrast medium was used. The extent of pulmonary involvement was defined as mild (when viral pneumonia affected less than 25% of the lungs), moderate (26%–50%) and severe (>50%). To make the assessment less subjective, a scoring scheme was used, with one point for each affected lung segment. The score varied from 0 to 40 and was graded as follows: from 1 to 10 points: mild (up to 25%); 11–20 points: moderate (26–50%), and >20 points: severe (>50%).

All data were collected in spreadsheets in Microsoft Excel version 365 (Microsoft Corp., Redmon – WA – USA) and the statistical analysis was performed using SAS Guide 4.3 (SAS Institute Inc., Cary – NV – USA).

Descriptive analyses were performed for all variables. Kolmogorov-Smirnov test was used to check if the continuous variables had normal distribution. As the continuous variables did not present a normal distribution, non-parametric tests were used for all variables. For continuous and discrete countable variables, the Wilcoxon Mann–Whitney test and the Kruskal–Wallis test were used to compare two or more categories, respectively. For dichotomous variables, the chi-square test was performed. For all tests, the p-value was calculated within 95% of confidence. Statistical significance was considered when p < 0.05.

The study is in accordance with ethical standards and was approved by local Ethics Committee (National Commission of Ethics in Research/CONEPCAAE 31634420.0.0000.5282/Opinion n. 4.028.151) prior to data collection.

Four hundred and nineteen adult patients were initially screened. From these, 36 (8.6%) had negative RT-PCR for SARS-CoV-2 infection and were excluded from this study. Out of 383 patients, 190 (49.6%) were initially admitted to the intensive care unit (ICU) for severe COVID-19 related ARDS and were also excluded. One hundred and ninety-three patients were consecutively admitted for mild or moderate COVID-19 related ARDS and were included in this study. Of those, 101 (52.3%) used hydroxychloroquine associated with azithromycin (active group) and 92 (47.7%) did not, being considered as the control group. Reasons for not taking hydroxychloroquine combined with azithromycin were: 44 (47.8%) patients had contraindications and 32 (34.8%) refused to take this combination. In 16 (17.4%) patients, this reason was unknown.

Demographic and clinical characteristics are provided in Table 1. The majority of patients were male (117/193; 61%). Median age was 68 years (IQR 47−72) in the total cohort, with 59 years (IQR 47–70) in the active group and 65 years (IQR 47−77) in the control group (p < 0.05). The median BMI was 28.3 kg/m2 in the total cohort and both groups had similar median BMI [28 (IQR 25−31) in the active group and 29 (IQR 26−33) in the control group (p > 0.05)]. Hypertension was the most common chronic disease reported, affecting 51.8% (100/193) of the studied population, followed by 29.5% (57) with diabetes mellitus, 15.0% (30) heart disease, 6.7%13 chronic pulmonary disease, and 5.2%10 cancer. The frequency of previous comorbidities was not different between groups.

Regarding previous use of medication, angiotensin-converting enzyme inhibitor (ACEI) and angiotensin receptor blocker (ARB) use was reported by 41.5% (80/193) of the patients,