Bias as a source of inconsistency in ivermectin trials for COVID-19: A systematic review. Ivermectin’s suggested benefits are mainly based on potentially biased results

Link to article at PubMed

J Clin Epidemiol. 2021 Dec 18:S0895-4356(21)00422-4. doi: 10.1016/j.jclinepi.2021.12.018. Online ahead of print.


OBJECTIVE: The objective of this systematic review is to summarize the effects of ivermectin for the prevention and treatment of patients with COVID-19 and to assess inconsistencies in results from individual studies with focus on risk of bias due to methodological limitations.

STUDY DESIGN AND SETTING: We searched the L.OVE platform through July 6, 2021 and included randomized trials (RCTs) comparing ivermectin to standard or other active treatments. We conducted random-effects pairwise meta-analysis, assessed the certainty of evidence using the GRADE approach and performed sensitivity analysis excluding trials with risk of bias.

RESULTS: We included 29 RCTs which enrolled 5592 cases. Overall, the certainty of the evidence was very low to low suggesting that ivermectin may result in important benefits. However, after excluding trials classified as "high risk" or "some concerns" in the risk of bias assessment, most estimates of effect changed substantially: Compared to standard of care, low certainty evidence suggests that ivermectin may not reduce mortality (RD 7 fewer per 1000) nor mechanical ventilation (RD 6 more per 1000), and moderate certainty evidence shows that it probably does not increase symptom resolution or improvement (RD 14 more per 1000) nor viral clearance (RD 12 fewer per 1000).

CONCLUSION: Ivermectin may not improve clinically important outcomes in patients with COVID-19 and its effects as a prophylactic intervention in exposed individuals are uncertain. Previous reports concluding important benefits associated with ivermectin are based on potentially biased results reported by studies with substantial methodological limitations. Further research is needed.

PMID:34933115 | PMC:PMC8684188 | DOI:10.1016/j.jclinepi.2021.12.018

Leave a Reply

Your email address will not be published.