Real-world effectiveness of hydroxychloroquine, azithromycin, and ivermectin among hospitalized COVID-19 patients: results of a target trial emulation using observational data from a nationwide healthcare system in Peru
This article has been Reviewed by the following groups
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
- Evaluated articles (ScreenIT)
Abstract
Introduction
Peru is one of the most impacted countries due to COVID-19. Given the authorized use of hydroxychloroquine (HCQ), azithromycin (AZIT), and ivermectin (IVM), we aimed to evaluate their effectiveness alone or combined to reduce mortality among COVID-19 hospitalized patients without life-threatening illness.
Methods
Retrospective cohort emulating a target trial, using nationwide data of mid- and high-level hospitals from the Peruvian Social Health Insurance 01/April/2020–19/July/2020. Patients 18 yo and above with PCR-confirmed SARS-CoV-2, and no life-threatening illness at admission were included. Five treatment groups (HCQ alone, IVM alone, AZIT alone, HCQ+AZIT, and IVM+AZIT within 48 hours of admission) were compared with standard of care alone. Primary outcome was all-cause mortality rate; secondary outcomes were all-cause death and/or ICU transfer, and all-cause death and/or oxygen prescription. Analyses were adjusted using inverse probability of treatment weighting. Propensity scores were estimated using machine learning boosting models. Weighted hazard ratios (wHR) were calculated using Cox regression.
Results
Among 5683 patients, 200 received HCT, 203 IVM, 1600 AZIT, 692 HCQ+AZIT, 358 IVM+AZIT, and 2630 standard of care. HCQ+AZIT was associated with 84% higher all-cause death hazard compared to standard care (wHR=1.84, 95%CI 1.12-3.02). Consistently, HCQ+AZIT was also associated with higher death and/or ICU transfer (wHR=1.49, 95%CI 1.01-2.19), and death and/or oxygen prescription (wHR=1.70, 95%CI 1.07-2.69). HCQ only showed higher death and/or oxygen prescription hazard. No effect was found for AZIT or IVM+AZIT.
Conclusions
Our study reported no beneficial effects of hydroxychloroquine, ivermectin, azithromycin. The HCQ+AZIT treatment seems to increase risk for all-cause death.
Funding
Instituto de Evaluación de Tecnologías en Salud e Investigación – IETSI, EsSalud
Article activity feed
-
SciScore for 10.1101/2020.10.06.20208066: (What is this?)
Please note, not all rigor criteria are appropriate for all manuscripts.
Table 1: Rigor
NIH rigor criteria are not applicable to paper type.Table 2: Resources
No key resources detected.
Results from OddPub: We did not detect open data. We also did not detect open code. Researchers are encouraged to share open data when possible (see Nature blog).
Results from LimitationRecognizer: We detected the following sentences addressing limitations in the study:Limitations and Strengths: Despite being a trial emulation, this study still is an observational retrospective cohort. Without a random assignment, residual confounding is a possibility. To control this, we used a robust approach based on incident users, defining a significant time zero to prevent immortal time-bias, allowing a grace …
SciScore for 10.1101/2020.10.06.20208066: (What is this?)
Please note, not all rigor criteria are appropriate for all manuscripts.
Table 1: Rigor
NIH rigor criteria are not applicable to paper type.Table 2: Resources
No key resources detected.
Results from OddPub: We did not detect open data. We also did not detect open code. Researchers are encouraged to share open data when possible (see Nature blog).
Results from LimitationRecognizer: We detected the following sentences addressing limitations in the study:Limitations and Strengths: Despite being a trial emulation, this study still is an observational retrospective cohort. Without a random assignment, residual confounding is a possibility. To control this, we used a robust approach based on incident users, defining a significant time zero to prevent immortal time-bias, allowing a grace period inclusion, and using modern data science techniques to emulate random assignment. Especially, the machine learning approach (generalized boosting model) allow us to include many more potential confounders (∼30 covariates) in the weighting model, than a conventional logistic regression would allow. This is reflected in the appropriate balance and overlapping achieved between treatment and control groups once the weighting scores were applied. Moreover, a sensitivity analysis was performed using doubly robust adjustment in the weighted Cox regression models, obtaining consistently estimates of causal effects. Despite all these robust strategies, we cannot guarantee that there is some degree of residual confusion in our study. Specifically, our finding that IVM could be associated with an increased risk of one of the secondary outcomes, but not with the rest, could be due to residual confusion not properly controlled even after double adjustment. Other possible limitation is the occurrence of non-registered variations of the treatment, as we relied on electronic records of drug dispensing. This process is strictly monitored and even audited; ...
Results from TrialIdentifier: No clinical trial numbers were referenced.
Results from Barzooka: We did not find any issues relating to the usage of bar graphs.
Results from JetFighter: We did not find any issues relating to colormaps.
Results from rtransparent:- Thank you for including a conflict of interest statement. Authors are encouraged to include this statement when submitting to a journal.
- Thank you for including a funding statement. Authors are encouraged to include this statement when submitting to a journal.
- No protocol registration statement was detected.
-
SciScore for 10.1101/2020.10.06.20208066: (What is this?)
Please note, not all rigor criteria are appropriate for all manuscripts.
Table 1: Rigor
Institutional Review Board Statement This target trial protocol was approved by EsSalud’s Institutional Review Board of COVID studies (91-SGRyGIS-DIS-IETSI-ESSALUD-2020) and was also registered in the Peruvian Health Research Projects repository Randomization Because these patients could have potentially been assigned to any group, they were randomly distributed between the control and treatment groups to avoid timedependent bias due to inappropriate exclusion or treatment assignment (12). Blinding During balance optimization, we remained blinded to the outcome results of the study. Power Analysis not detected. Sex as a biological variable Age ranged between 18 and 104 years old with a mean of 59.4 years … SciScore for 10.1101/2020.10.06.20208066: (What is this?)
Please note, not all rigor criteria are appropriate for all manuscripts.
Table 1: Rigor
Institutional Review Board Statement This target trial protocol was approved by EsSalud’s Institutional Review Board of COVID studies (91-SGRyGIS-DIS-IETSI-ESSALUD-2020) and was also registered in the Peruvian Health Research Projects repository Randomization Because these patients could have potentially been assigned to any group, they were randomly distributed between the control and treatment groups to avoid timedependent bias due to inappropriate exclusion or treatment assignment (12). Blinding During balance optimization, we remained blinded to the outcome results of the study. Power Analysis not detected. Sex as a biological variable Age ranged between 18 and 104 years old with a mean of 59.4 years old (SD = 16.3 years old) and 36.8% (n = 2091) of the participants were women. Table 2: Resources
No key resources detected.
Results from OddPub: We did not detect open data. We also did not detect open code. Researchers are encouraged to share open data when possible (see Nature blog).
Results from LimitationRecognizer: An explicit section about the limitations of the techniques employed in this study was not found. We encourage authors to address study limitations.
Results from TrialIdentifier: No clinical trial numbers were referenced.
Results from Barzooka: We did not find any issues relating to the usage of bar graphs.
Results from JetFighter: We did not find any issues relating to colormaps.
About SciScore
SciScore is an automated tool that is designed to assist expert reviewers by finding and presenting formulaic information scattered throughout a paper in a standard, easy to digest format. SciScore checks for the presence and correctness of RRIDs (research resource identifiers), and for rigor criteria such as sex and investigator blinding. For details on the theoretical underpinning of rigor criteria and the tools shown here, including references cited, please follow this link.
-
