RE: “EARLY OUTPATIENT TREATMENT OF SYMPTOMATIC, HIGH-RISK COVID-19 PATIENTS THAT SHOULD BE RAMPED UP IMMEDIATELY AS KEY TO THE PANDEMIC CRISIS”

In his recently accepted manuscript on treatment of coronavirus disease 2019 (COVID-19), Dr. Harvey Risch criticizes the National Institutes of Health and the Food and Drug Administration because their reviews “have omitted the 2 critical aspects of reasoning about these drugs: use of HCQ [hydroxychloroquine] combined with AZ [azithromycin] or with DOX [doxycycline] and use in the outpatient setting” (1, p. 000).

We would like to highlight the uncritical appraisal he made on the available evidence and our surprise that such a viewpoint could be accepted in this journal.

Gautret et al. (2) recently led an open-label single-arm clinical trial to prove the efficacy of hydroxychloroquine (HCQ) alone in hospitalized patients (European Union Drug Regulating Authorities Clinical Trials Database (EudraCT) number 2020-000890-25). In the paper reporting on the trial, there was an ad-hoc untreated comparison group comprising 16 patients who either refused to take part in the trial, had noninclusion criteria, or were hospitalized at other clinical centers (2). Among the 26 patients treated with HCQ, 6 were excluded from the analysis without justification. We do not know whether they also received azithromycin (AZ), but given the severity of the disease among the 4 of them who progressed (3 to the intensive care unit, 1 to death), it is quite likely that several of them received the combination of HCQ and AZ. Among the 20 remaining patients, 6 were treated with the combination; thus, the claim of the efficacy of the combination is based on analysis of these 6 participants. Risch reports the P value from the test conducted by Gautret et al. to compare the 3 groups (untreated, HCQ, and HCQ + AZ). However, this test is not appropriate for claiming that the combination is more effective than HCQ alone. The appropriate test would compare the results for the 14 participants who received HCQ alone with results for the 6 who received the combination, and the P value from this test is not significant (P = 0.39).

Risch states that the only valid criticism against this trial is the lack of randomization (1). He rejects criticism about the small study size, saying that size matters only when no evidence is found. But no evidence is found in this study, as shown above. Besides, a larger sample size would allow a more precise and robust estimate of the effect size. And in fact, there are many other valid criticisms. There were more missing values in the untreated group than in the treated group, and the missing-data replacement technique used favored the treated group. The virological assessment was not the same for the treated group and participants in the untreated group followed in other clinical centers, raising an issue about the validity of the measurement of the primary endpoint and precluding assessment and comparison of the initial values in each group. Moreover, an appropriate primary endpoint would be a clinical one—hospitalization for an outpatient and progression to the intensive care unit or death for a hospitalized patient—not a virological outcome. We also find it worrying that Gautret et al. (2) describe daily follow-up during a period of 14 days and report having received ethical approval on March 6, 2020; but the paper was published on March 20, 2020, which is impossible. Therefore, it is absolutely impossible to use this trial to support the use of the combination in an outpatient setting, because it did not include outpatients, its primary endpoint was not appropriate, and it did not show any benefit of the combination on the primary endpoint used.

The only other published study discussed by Risch was carried out in both outpatients and hospitalized patients, with a low proportion of at-risk participants (3). Risch and the authors conclude that the treatment was efficacious; however, a study with no control group for a disease for which natural history is not completely understood cannot produce a conclusion of efficacy. All of the other studies reviewed were unpublished, poorly designed studies whose quality was even lower than that of the studies discussed above, but the limited space available prevents us from providing a detailed critical analysis.

Finally, it is relevant to mention a study performed in a macaque model of the disease (4) that found no efficacy of HCQ alone or in combination with AZ in preventing or curing COVID-19 when administered early. These results are in line with a recent report of the absence of efficacy of HCQ in preventing the disease in exposed individuals (5).

We strongly think that there is no convincing evidence to support the claim made by Risch on the emergency need to treat symptomatic high-risk COVID-19 patients with a combination of HCQ and AZ.

Acknowledgments

Conflict of interest: none declared.