Discussion
The pooled analyses in this updated systematic review and meta-analysis indicated no evidence of any significant association between ACEIs/ARBs and any COVID-19 related clinical outcomes; however, the sub-group analyses revealed evidence of a negative impact of ACEIs/ARBs use and some COVID-19 related clinical outcomes such as higher odds of hospitalisation, ICU admission and ventilator use. Contrastingly, a positive impact in terms of lower odds of death/ICU admission, as a composite outcome, and a higher rate of hospital discharge. Furthermore, our study findings, for the first time, showed inter-class variations between ACEIs and ARBs effects on COVID-19 clinical outcomes with low quality evidence indicating lower risk of acquiring COVID-19, less severe COVID-19 infection, higher rate of ICU admission and ventilator use with ARBs but not ACEIs.
Our study findings also showed no significant association between ACEIs/ARBs and mortality, severe COVID-19 infection, or positive tests for COVID-19, in agreement with two previously published systematic reviews (29, 30). This was despite the inclusion of more recently published studies (18, 27, 37, 38, 46, 47, 50), which implies consistency of evidence. This is encouraging given the controversies surrounding hydroxychloroquine. Furthermore, these non-significant associations were also observed for additional COVID-19 related outcomes including ICU admission, hospitalisation, and hospital discharge. However, unlike the previous two systematic reviews (29, 30), our study found evidence of associations between ACEI/ARB use and certain COVID-19 clinical outcomes. Whilst the pooled estimate of the sub-group analyses indicated a higher odds of ICU admission with ACEIs/ARBs among studies conducted in the USA (23, 40, 41) and peer-reviewed studies (23, 25, 41), all these studies were of poor quality and none performed adjusted analyses to account for potential confounders. Confounding by indication is of particular concern with comorbidities such as CVD and diabetes associated with more severe COVID-19 morbidity and mortality (4-6). Similarly, the observed significant associations between ACEIs/ARBs use and high odds of ventilator use and hospital discharge rates were from Benelli et al (38) with crude analysis and non-peer-reviewed and Ip et al (27) and Zeng et al (26) which were both non-peer reviewed, of poor quality and used crude analyses. Similarly, the studies in the pooled analyses that showed significant association of ARBs use and ICU admission (38, 39), lower risk of acquiring COVID-19 infection (45), and severe infection (18, 19) were of poor quality, used unadjusted/crude analyses, and/or non-peer reviewed. In terms of duration of hospital stay, Yang et al (25) and Zeng et al(26) both reported a reduction in hospital stay with ACEIs/ARBs; however, it was not possible to combine them in the meta-analysis as they used different measure of effects with the former reporting the outcome as a mean difference while the latter as a median.
On the other hand, our study findings showed high quality evidence on the association of ACEIs/ARBs and higher odds of hospitalisation but lower odds of death/ICU admission (as a composite endpoint). The higher odd of hospitalisation was observed in the sub-group analyses of studies conducted in the USA (40, 41), used adjusted analyses (44), peer-reviewed (41) and of good quality (44); whereas the studies for lower death/ICU admission were from Europe (37, 42), used adjusted analyses and of good quality (37), although all of them were non-peer reviewed.
Several hypotheses have been suggested to explain the negative and positive effects of ACEIs/ARBs use on COVID-19 clinical outcomes. The former is thought to be related to ACEIs/ARBs potential ability to up-regulate ACE2, the cell entry point for COVID-19; hence facilitate COVID-19 cell entry and its subsequent infectivity/pathogenicity (52); however, the evidence to date demonstrates ACE2’s up-regulation consistently in cardiac and renal tissues in response to ARBs therapy but not ACEIs (4, 53); this observed difference between ARBs and ACEIs has been suggested to be due to the increased level of angiotensin-II, which occurs following ARBs treatment but not ACEIs, which in turn imposes an increased substrate load on ACE2 enzyme requiring its upregulation (54). Importantly, it should be emphasised that evidence of ACEIs/ARBs induced ACE2 upregulation in the respiratory tracts, which is the key entry system for COVID-19, is lacking (53). Furthermore, it should be noticed that alteration in angiotensin-II level, which is only one substrate of ACE2’s multiple substrates, is unlikely to result in any meaningful differences in ACE2 substrate load, hence its upregulation (53); additionally, the fact that people from various sexes, ages, and races are all susceptible to COVID-19 infection suggests that physiological expression of ACE2 might already be sufficient for COVID-19 infection; thus any further ACE2 upregulation might not have effects on the risk/severity of COVID-19 infection (25). Together, these evidence indicate that the concerns around ACEIs/ARBs use in COVID-19 patients might be unjustifiable. On the other hand, the protective effect hypothesises on ACEIs/ARBs protecting against lung injury, through blockage of the harmful angiotensin II- AT1R axis, which gets activated by impairment of ACE2 activity as a result of ACE2’s downregulation results from ACE2’s binding with COVID-19 virus; additionally, the corresponding increase in angiotensin II and angiotensin I, due to ACEIs/ARBs use, would activate the protective axis and hence reducing COVID-19 viral pathogenicity (4). Genetic ACE2 polymorphism among some individuals has been also suggested as potential factor explaining, at least partially, the harmful effects on ACEIs/ARBs among COVId-19 patients (55); but this needs further investigation.