key: cord-0987034-o88jywnm
authors: Varisco, Tyler J.; Johnson, Michael L.; Thornton, Douglas
title: Comment on Arshad et al: Treatment with Hydroxychloroquine, Azithromycin, and Combination in Patients Hospitalized with COVID-19
date: 2020-08-06
journal: Int J Infect Dis
DOI: 10.1016/j.ijid.2020.07.071
sha: ff143f2f1cdbde969db7e63b94e3700b84bc9617
doc_id: 987034
cord_uid: o88jywnm

nan

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

We read, with great interest, the results published by Arshad and colleagues concerning the potential association between treatment with hydroxychloroquine with or without azithromycin and in-hospital mortality in patients with COVID-19. 1 The reported treatment benefit contradicts that reported elsewhere, including a recent study at US Veterans Affairs Hospitals that showed an almost perfectly inverse risk of mortality (HR=2.61; 95% CI, 1.10-6.10). 2 Although Lee and colleagues provided an excellent commentary on the potential for immortal time bias and selection bias in this work 3 , the following design and analysis flaws further threaten the validity of the reported findings.

First, corticosteroid use was common in patients who received hydroxychloroquine with or without azithromycin, 79% and 74% respectively. It is unclear when in the course of care adjunctive therapy including steroids and tocilizumab were initiated. Without disclosing the treatment protocol at each of the six centers, all we can conclude is that steroid use was not consistent at baseline and, thus, should have been treated as a time dependent covariate. This may indicate that an initial clinical decline in the hydroxychloroquine arms was masked by subsequent corticosteroid initiation.

The author's efforts to control confounding create additional confusion. Reportedly, 25% of mSOFA scores were missing and "no imputations…[were] made for missing data…" Yet the Cox model was adjusted for "clinical disease severity (mSOFA, O2 saturation)." Table two states that all 2,541 cases were included yet shows no effect for mSOFA score indicating that either all cases were included or the model was not adjusted for severity as discussed in the text.

Dichotomizing age is also concerning given the established association between COVID-19 mortality and age. Dichotomizing age robs information from the model and is a significant threat to validity.

Finally, in the propensity scored (PS) model, the authors redefined exposure as hydroxychloroquine or nothing and the authors did not include azithromycin as a covariate in the J o u r n a l P r e -p r o o f PS generating model. With so few matched pairs included in the model (n=190), the omission of this potential confounder, especially one the Cox model showed to be associated with the outcome, should not be taken lightly. Furthermore, we are left to question why exposure was redefined given the availability of established methods to compare multiple treatment groups using inverse probability of treatment weighting. 5 In light of these significant threats to validity, we urge clinicians treating patients with COVID-19 to interpret and apply the contradictory findings of this study cautiously.

No funding source was used for this commentary.

No ethical approval was necessary for this commentary.

Treatment with Hydroxychloroquine, Azithromycin, and Combination in Patients Hospitalized with COVID-19

Outcomes of hydroxychloroquine usage in United States veterans hospitalized with Covid-19. medRxiv : the preprint server for health sciences

An Observational Cohort Study of Hydroxychloroquine and Azithromycin for COVID-19: Can't Get No Satisfaction

Multicollinearity in Regression Analyses Conducted in Epidemiologic Studies

Exploring large weight deletion and the ability to balance confounders when using inverse probability of treatment weighting in the presence of rare treatment decisions

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.