Risk factors associated with severe outcomes of COVID-19: A systematic rapid review to inform national guidance on vaccine prioritization in Canada

Abstract

Background: To inform vaccine prioritization guidance in Canada, we systematically reviewed evidence on the magnitude of association between risk factors and severe outcomes of COVID-19. The urgent nature of this review necessitated an adapted methodology, which may serve as an exemplar for reviews undertaken under strict timelines. Methods: We updated our existing review by searching online databases and websites for cohort studies providing multivariate adjusted associations. After piloting, one author screened studies and extracted data. Two authors estimated the magnitude of association between exposures and outcomes as little-to-no (odds, risk, or hazard ratio <2.0, or >0.50 for reduction), large (2.0-3.9, or 0.50-0.26 for reduction), or very large (≥4.0, or ≤0.25 for reduction), and rated the evidence certainty using GRADE. Results: Of 11,734 unique records we included 134 reports. There is probably (moderate certainty) at least a large increase in mortality from COVID-19 among people aged 60-69 vs. <60 years (11 studies, n=517,217), with ≥2 vs. no comorbidities (4 studies, n=189,608), and for people with (vs. without): Down syndrome (1 study, n>8 million), type 1 and 2 diabetes (1 study, n>8 million), end-stage kidney disease (1 study, n>8 million), motor neuron disease, multiple sclerosis, myasthenia gravis, or Huntington’s disease (as a grouping; 1 study, n>8 million). The magnitude of association with mortality is probably very large for Down syndrome and may (low certainty) be very large for age 60-69 years, and diabetes. There is probably little-to-no increase in severe outcomes with several cardiovascular and respiratory conditions, and for adult males vs. females. Conclusion: There is strong evidence to support at least a large increase in mortality from COVID-19 among older adults aged 60 to 69 years versus <60 years; people having two or more versus no comorbidities; and for people affected by several pre-existing conditions. The methodology employed in this review may provide an important exemplar for future syntheses undertaken under urgent timelines. Systematic review registration: PROSPERO #CRD42021230185.