With a verifiable total of 43,829 patients in the analysis, including 11,166 exposed to ACEI/ARB, our meta-analysis consistently revealed the absence of an association between ACEI/ARB use and mortality, disease severity, and hospitalization risk in COVID-19, a finding to be validated as further evidence accumulates. The country cluster sensitivity analysis explained several of the differences seen in the mortality outcome. The UK/US cluster revealed an increased risk of mortality with two of the three studies [13,14] having OR adjusted for comorbidities. The study by Richardson et al. was a large case series in New York conducted between March 1 and April 4, 2020, dates that coincide with that state’s first wave of the pandemic prior to extreme measures such as stay-at-home orders being taken [15,16]. A similar pattern emerges with the UK study by Bean et al. in which the country experienced their first wave prior to lockdown measures [1, 14, 17, 16]. Studies in France, Italy, China, Denmark, Spain and Korea were also conducted in their first wave; however, in contrast, during that time they had already implemented multiple NPI strategies (including a national lockdown) to mitigate the impact of the pandemic [1, 18, 19]. Based on our analysis, two of these countries (France and China) saw a decrease in mortality among ACEI/ARB users (shown in Figure 2). It would be difficult to determine whether this effect was due to an underlying mechanism of ACEI/ARB protection or if it was the result of very strict NPIs having been executed. With cautious interpretation, among many other factors, this may suggest that countries with much stricter NPI policies may have been successful in reducing mortality among ACEI/ARB users suffering from COVID-19 in the early stages of the pandemic.
The meta-regression sensitivity analyses by country clusters did not explain the heterogeneity in the disease-severity endpoint very well. Indeed, harmonization of disease-severity definitions may strengthen future studies and meta-analyses; however, more importantly, some researchers suggest the existence of ACE-2 overexpression polymorphism as a potential explanation for the severity of the COVID-19 presentation [20]. The country-cluster analysis is not well-suited to answer this question, and more studies among different ethnic groups are needed. For example, in a large cohort in the UK, it was found that Blacks being treated with ACEI/ARB were more susceptible to COVID-19 compared to Whites [21]. Whether this ethnic difference is real is yet to be determined, and the discordant country-specific results are points for future attention as more data start to accumulate. Lastly, the sensitivity analysis by country revealed more hospitalizations in the US among ACEI/ARB users. We believe that the hospitalization findings should be considered informative, at best, as these are influenced by clinical decisions and hospital-bed availability and not based on objective criteria.
An advantage of this current analysis is that we captured published studies from early stages of the pandemic in each country. Therefore, we think that we were able to estimate the risk of mortality during the implementation of NPI measures (especially national lockdowns) in a number of countries. The South Korean national response was a highly successful model for handling the pandemic. With aggressive measures that included contact tracing to prevent community transmissibility, South Korea reported the largest numbers of cases in the first two months of the pandemic. Among other factors, these large numbers of cases did not translate into increased mortality among ACEI/ARB users [19].