key: cord-1050746-ws3rs4bb
authors: Gautret, Philippe; Ly, Tran Duc Anh; Raoult, Didier
title: Response to the submission IJAA-D-21-00195 Hydroxychloroquine + Azithromycin treatment in elderly patients
date: 2021-03-11
journal: Int J Antimicrob Agents
DOI: 10.1016/j.ijantimicag.2021.106313
sha: a2ba8dd51f4094bac5d1cdaa95bb033b8b567410
doc_id: 1050746
cord_uid: ws3rs4bb

nan

E-mail address: philippe.gautret@club-internet.fr

We thank XX et al [1] , for their letter about our work on the effect of hydroxychloroquine (HCQ) and azithromycin (AZI) combination against COVID-19 in elderly nursing home residents [2] . They discuss reasons for the contradictory results obtained by Magagnoli et al. in US veterans [3] , suggesting that the explanation may be that these authors administered combined treatment at an advanced stage of the disease when patients were hospitalized, while in our study, patients were treated earlier, mostly following mass screening for SARS-CoV-2 at nursing homes. Indeed, the mean time between virological diagnosis and treatment was 2.5 ± 3.0 days in our study. In addition, in the US study, the proportion of patients with low lymphocytes counts (a well-known severity marker in COVID-19 patients) was significantly higher in those allocated HCQ and AZI combined treatment (30%) than in controls (19%), which suggests that treated patients were more severe at enrolment.

Recently, another study conducted on COVID-19 patients at a nursing home in the Principality of Andorra was published [4] . These authors reported that combined treatment with HCQ and AZI independently associated with a significant reduction of mortality rate as compared to beta-lactam, quinolone or no treatment (OR: 0.044, p=0.004), which corroborates our results. Interestingly, 56% patients received the combined treatment within 24 hours of symptom onset. Details of the three studies are presented in Table 1 .

In both the French and Andorran studies there was relative homogeneity of care practice, while in the US study, conducted in medical centers across the country, greater heterogeneity was very likely. In addition, the former studies were based on medical chart retrospective review, whereas the US study was conducted using codes to identify clinical characteristics and study outcomes, which is much less accurate [5]. 3 In conclusion, interpretation of retrospective data addressing the effect of treatments for COVID-19 should be interpreted with caution, notably when a treatment may be more effective at a given stage of the disease. Multicentric retrospective studies, where disease severity is assessed on coded data with little or no information about the timing of therapy, may lead to incorrect interpretation.

Funding: No funding 5 

Pattern of SARS-CoV-2 infection among dependant elderly residents living in long-term care facilities in Marseille

Outcomes of Hydroxychloroquine Usage in United States Veterans Hospitalized with COVID-19. Med (N Y)

COVID-19 mortality risk factors in older people in a long-term care center

Why most published research findings are false

Ethical Approval: Not required