Referring to the study: Hydroxychloroquine and azithromycin as a treatment of COVID-19: results of an open label non-randomized clinical trial

The novel coronavirus (SARS-CoV-2) first described in China has been spreading rapidly throughout the world. Thousands of deaths mainly in China and Italy aroused the attention of the entire medical and scientific community as well as the general population. For this reason, emergency treatment could save tens of thousands of lives affected by the SARS-CoV-2. An off-label treatment with the use of hydroxychloroquine with or without azithromycin association emerged in a time of emergency [1]. However, some serious scientific limitations should be pointed out in this recent study.

Firstly, this is not a randomized, double-blind, placebo-controlled trial. Although the authors included a control group, it is poorly described throughout the study and we do not know how these control patients were treated. Moreover, the safety and possible adverse effects such as cardiac abnormalities [2] were not mentioned.

Secondly, the authors calculated 48 patients assuming a 50% efficacy of hydroxychloroquine in reducing the viral load at day 7. However, the study included only 36 patients and 6 of them dropped out of the study without any report of the reason for exclusion. This is a concern because the study does not seem to have enough sample power to complete the primary outcome. The study allocated only 30 patients in three subgroups: Hydroxychloroquine (n=14), Hydroxychloroquine + Azithromycin (n=6) and control (n=16). Again, it is not clear from which groups those 6 patients were excluded. Finally, there are not any clinical characteristics of the patients such as blood pressure, oxygen saturation, medications, kidney or hepatic markers, and comorbidities.

Thirdly, means of quantitative data were compared using Student's t-test. This is a repeated measure study, which should have included an analysis of variance (ANOVA) for repeated measures. Student's t-test, in this case, increases the type I error and false-positive results. Moreover, neither standard deviations nor confidence intervals are demonstrated. In addition, the difference in age between groups is huge, which clearly shows that the study was not balanced between groups.

Finally, the presence of SARS-CoV-2 was evaluated in the liquid collected from the nasopharyngeal swab, and real-time reverse transcription-PCR was not performed in 49 repeated samples throughout D0 and D6, mainly in the control group. Why were these PCRs not evaluated? How did the authors deal with the missing data? There is no data regarding the presence of SARS-CoV-2 and the viral load in the plasma and serum of patients. Therefore, we do not know whether the viral load could be present systemically.