No evidence of clinical benefits of early treatment of COVID-19 patients with hydroxychloroquine and azithromycin

Dear Editor,

We read with interest the retrospective study by Million et al. describing the outcome of 1061 COVID-19 patients receiving ≥3 days of hydroxychloroquine and azithromycin [1]. Based on an “unrestricted massive screening” of cases with possible COVID-19, they identified outpatients with a positive SARS-CoV-2 PCR from nasopharyngeal sample. Their therapeutic strategy relied on the early (median time from onset of symptoms: 6.4 days) prescription of antiviral agents in day-care hospital in order to reduce to risk of clinical worsening, hospitalization and, theoretically, death [1]. We agree with Million et al. that early antiviral mass treatment should rely on cheap, available and well-tolerated drugs with in vitro activity against SARS-CoV-2. The combination of hydroxychloroquine and azithromycin is a seducing candidate, considering their in vitro synergism and preliminary clinical data [1,2]. In their current report on 1061 patients, the authors describe a case fatality rate of 0.9% and a hospitalization frequency of 4.3% [1]. The authors claim that the combination of hydroxychloroquine and azithromycin is associated with a low proportion of clinical worsening and a low case fatality rate. However, without any control group, it seems difficult to attribute this outcome to their therapeutic intervention.

Indeed, data from different countries demonstrate that the case fatality rate of COVID-19 has been overestimated in early reports, probably due to restricted sampling policy, focusing on the most severe cases. In countries where widespread PCR-based testing were performed, case fatality rate was low, such as South Korea (0.7%) [3]. Million and coworkers attribute this low case fatality rate to a frequent prescription of hydroxycholoroquine but in fact Korean guidelines advice against the use of any antiviral drugs for mild disease in young and healthy patients [4]. In France, a recent epidemiological study based on hospital and laboratory data estimated that the case fatality rate of COVID-19 was 0.7% (and less than 0.06% below 50 years-old) and frequency of hospitalization was 3.6% [5]. These data are cornerstone for the interpretation of uncontrolled therapeutic studies.

Furthermore, the authors do not discuss the selection bias that their diagnostic approach induced. By offering PCR sampling for outpatients who can reach a diagnostic facility, they mostly selected young adults (mean age = 43.6 years), with a low frequency of comorbidities (hypertension: 14%, obesity: 5.8%, coronary artery disease: 4.3%) [1]. Older age has repeatedly been associated with a higher likelihood of death, especially beyond 50 yo [5]. Obesity and cardiovascular diseases are also strongly associated with worst outcome [6]. As a consequence, the low case fatality rate observed in the study by Million et al. is close to what is observed in the untreated general population and is influenced by the selection of young, healthy and mildly infected patients.

While demonstrating that the outcome of patients receiving the combination of hydroxychloroquine and azithromycin is similar to that of untreated young adult patients in the general population, the authors leave us with the same unanswered question: should we prescribe early antiviral treatment to our patients with mild-to-moderate COVID-19?

Funding

None.

Declaration of competing interests

The authors declare no competing interests.