Key points

  • The role of drugs that are used in the management of hypertension and cardiovascular diseases in the deaths of COVID-19 patients is unclear so far

  • The results of the current analysis have found the non-significant role of these drugs in the deaths of COVID-19 patients

  • Based on current evidence, we suggest to continue the use of drugs particularly ACEi/ARB in the management of hypertension and cardiovascular diseases in COVID-19 patients

Background

Recent, highly contagious novel coronavirus (2019-nCoV) caused by SARS-CoV-2 emerged from Wuhan, China, and rapidly spread over more than 100 countries has caused unprecedented health concerns all over the globe. The first case was recorded in November 2019, and the World Health Organization (WHO) declared a pandemic and a global public health emergency on March 11, 2020 [1]. The virus spread continuously despite many drastic containment measures (complete lockdown, curfews, etc.). On March 13, 2022, more than 456,797,217 cases of COVID-19 were reported across the globe, resulting in approximately 6,043,094 deaths [2]. Health authorities all over the world are struggling to develop possible prevention and therapeutic measures [2, 3]. Fortunately, vaccines are developed and used as a preventive measure across the globe. However, still there is no specific drug available for the treatment of infected patients with SARS-CoV-2. There are a number of research questions that are unanswered so far related to this infection. It has been observed that COVID-19 patients with co-morbid conditions such as diabetes (DM), hypertension (HT), or cardiovascular disease (CVD) are more prone to death [3, 4]. There are a number of explanations for this. It has also been hypothesized that the use of medicines in the management of co-morbid conditions also could be one of the reasons. Hypertension and cardiovascular diseases are the most common co-morbid conditions, and the most commonly used drugs in the management of these conditions are acting on the renin–angiotensin–aldosterone system (RAAS) such as ACEi/ARBs. SARS-CoV-2 enters the cell through the host's angiotensin-converting enzyme (ACE) [4], and drugs acting on the RAAS (ACEi and ARB) may boost ACE2 expression, resulting in greater SARS-CoV-2 binding [5]. The enhanced binding of SARS-CoV-2 to the host could result in severe symptoms or even deaths of COVID-19 patients. The pieces of evidence have been primarily contentious up to this point. Some studies have also shown the protective effect of these medicines in COVID-19 patients [4, 6], whereas some studies have concluded a higher mortality rate [7,8,9,10,11].

To the best of our knowledge, few meta-analyses have also been conducted to find out the association of ACEi/ARBs in the deaths of COVID-19 patients. However, a number of included studies are too less to draw any valid conclusion. Further, some meta-analyses have also included different designs of studies. Therefore, we have performed a systematic review and meta-analysis of observational studies to find out the exact association of ACEi/ARB in the deaths of COVID-19 patients.

Methods

Search strategy

PubMed, Google Scholar, and MedRxiv preprint server were used to identify relevant studies from December 2019 to January 2022 with proper MeSH terms. The MeSH phrases or keywords with Boolean operators were used as followings “(COVID19)” OR “(COVID-19)” OR “(COVID-19 VIRUS INFECTION)” OR “(COVID19 INFECTION)” OR “(SARS COVID 19 INFECTION)” OR “(2019 NOVEL CORONAVIRUS INFECTION)” OR “(SARS COVID DISEASE)” OR “(COVID-19 DISEASE)” AND “(ACE)” OR “(ARB)” OR “(ANGIOTENSIN CONVERTING ENZYME)” OR “(ANGIOTENSIN RECEPTOR BLOCKERS)” which were used are presented in Additional file 1: Table S1. This study was carried out according to the PRISMA [12] and STROBE guidelines [13].

Eligibility criteria

The studies were included or excluded as per the defined inclusion and exclusion criteria. The inclusion criteria include COVID-19 patients, age above 18 years, use of ACEi/ARB classes of drugs alone or in combinations, one of the outcomes was death. The studies were excluded if published other than in the English language, review articles, meta-analyses, case reports, letters, comments or opinions, animal studies, death was not reported as one of the outcomes and editorials.

Screening

The screening of relevant studies as per inclusion and exclusion criteria was done independently by two authors (RS and AK). The PRISMA guideline was followed, and a selection of studies based on titles, abstracts, and full texts was presented in the PRISMA flow chart. The conflict among the authors was resolved after discussion with third (JM), fourth (AKT), and fifth authors (GA).

Quality assessment

The Newcastle–Ottawa (Questionnaire) Scale (NOS) was used to determine the quality of the studies and measuring the risk of bias in cohort and case–control studies [14]. Three reviewers (RS, AK, and JM) have conducted quality assessments of included studies, and disagreements were resolved after discussion with the fourth (AKT) and fifth (GA) authors. The following are the major components used for the quality assessment: comparability, selection of non-exposed cohort, representativeness of the exposed cohort, ascertainment of exposure, outcome assessment, demonstration that the outcome of interest was not present at the start of the study, adequacy of cohort follow-up and follow-up time. The quality rating scale runs from 0 to 10, with a score of > 7 stars indicating high-quality content.

Data extraction

The data were extracted from 68 studies and cross-checked by both authors (RS and AK). The data were extracted in an M.S Excel sheet which contains the columns like authors’ first names, type of study, location of study, total sample size, number of males/females, age groups, the total number of patients in the ACEi/ARBs and Non-ACEi/ARB’s groups, number of subjects died in the ACEi/ARBs and non-ACEi/ARB’s groups.

Data analysis

The overall estimate was calculated in terms of odds ratios with a 95% confidence interval. The random-effect model was preferred over the fixed-effect model due to variations among included studies. The Chi-square statistic and the I2 z test were used to measure heterogeneity. The funnel plot was used to determine whether there was any publishing bias. The sensitivity analysis was performed to check the effects of outliers on the outcome. For the data analysis, RevMan 5 was employed.

Results

Search results and study characteristics

The initial search identified 94,184 studies. A total of 224 duplicates were found and the remaining 93,940 studies were further screened based on the titles. A total of 2907 studies were found relevant which were further screened based on abstracts. Further, a full text of 200 studies was downloaded, and finally, 68 studies were found relevant for quantitative analysis as per the aim and objective of the current study. Out of these 68 studies, 61 studies were published in peer-reviewed journals, whereas the remaining 7 studies were preprints. The step-by-step screening and selection of studies are presented in Fig. 1 as per the PRISMA flow chart. Out of 68 selected studies [7,8,9,10,11, 15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77], 56 were cohort and the remaining 12 were case–control studies. A total of 128,078 patients were found. The study characteristics of included studies are compiled in Table 1.

Fig. 1
figure 1

Selection of studies as per the PRISMA Checklist

Full size image
Table 1 Characteristics of included Studies
Full size table

Quality evaluation

The Newcastle–Ottawa Scale was used to determine the quality of the studies. The study's cohort and case–control classes were assessed on a scale of 0 to 10, with low risk of bias (8–10), moderate risk (5–7), and high risk (0–4) assigned to each. A total of 52 studies were found to be of excellent quality, and 16 studies were of fair quality as compiled in Table 2. In the case of cohort studies, 49 studies were found to be excellent and 7 were of fair quality, whereas in the case of case–control studies, out of 12 studies, 3 studies were found to be excellent and the remaining 9 were of fair quality.

Table 2 Quality assessment using the Newcastle–Ottawa scale
Full size table

ACEi/ARB and deaths of COVID-19 patients

A total of 68 studies were included, with a total of 128,078 COVID-19 cases. A total of 31,625 patients were on ACEi/ARB, whereas the remaining 96,453 were in non-ACEi/ARB group. The pooled odds ratio was found to be 1.14 [0.95, 1.36] which indicates a non-signification association of ACEi/ARB in deaths of COVID-19 patients as compared to the non-ACEi/ARB group (Fig. 2). However, the heterogeneity among studies was found to be 92% which is quite high as indicated by I2 statistics and the Chi-square test (p < 0.00001). Therefore, further, sub-group analysis was also done to find out the reasons for heterogeneity.

Fig. 2
figure 2

Pooled analysis results using a random effect model ACEi/ARB (forest plot)

Full size image

Publication bias

The funnel plot was used to analyze publication bias qualitatively. The shape of the plot revealed some degree of asymmetry (Fig. 3) which indicates the involvement of publication bias.

Fig. 3
figure 3

Funnel plot for the assessment of publication bias (ACEi/ARB)

Full size image

Sub-group analysis

The subgroup analysis was done to check the effect of ACEi/ARB individually on the outcome.

ACEi

Out of 68 studies, only 10 studies mentioned specifically about ACEi and contain relevant data (Additional file 1: Table S2). The pooled odds ratio was found to be 1.43 [0.83, 2.47] which indicates the non-signification association of ACEi in deaths of COVID-19 patients as compared to the non-ACEi group (Fig. 4a).

Fig. 4
figure 4

Forest plot showing overall estimate in terms of odds ratio using random effect model a ACEi b Ramipril

Full size image

Further, we have also tried to check the effect of individual ACEi (captopril, enalapril, lisinopril, perindopril, ramipril, and zofenopril) on the outcome, however, we have got relevant information related to ramipril only (Additional file 1: Table S3). The pooled odds ratio was found to be 1.02 [0.44, 2.36] which indicates the non-significant association of ramipril in deaths of COVID-19 patients as compared to the non-ramipril group (Fig. 4b). The funnel plot also indicated the involvement of publication bias as shown in Fig. 5.

Fig. 5
figure 5

Funnel plot for the assessment of publication bias (ACEi)

Full size image

ARB

A total of 10 studies mentioned specifically ARB and contain relevant data (Additional file 1: Table S4). The pooled odds ratio was found to be 1.37 [0.68, 2.77] which indicates the non-significant association of ARB in deaths of COVID-19 patients as compared to the non-ARB group (Fig. 6a). However, the heterogeneity among studies was found to be 92% which is quite high as indicated by I2 statistics and the Chi-square test (p < 0.00001).

Fig. 6
figure 6

Forest plot showing overall estimate in terms of odds ratio using random effect model a ARBs b Losartan c Valsartan

Full size image

Further, the effects of individual ARB (candesartan, irbesartan, valsartan, losartan, telmisartan, eprosartan, fimasartan, azilsartan, and olmesartan) on the outcome have also been tried. However, we have found relevant data on losartan (Additional file 1: Table S5) and valsartan only (Additional file 1: Table S6). The pooled odds ratio was found to be 1.20 [0.11, 13.39] and 2.78 [0.45, 17.12] for losartan and valsartan, respectively, which also indicates the non-signification association of losartan and valsartan in the deaths of COVID-19 patients as compared to non-losartan and valsartan group (Fig. 6b, c). The funnel plot indicated involvement of publication bias as shown in Fig. 7.

Fig. 7
figure 7

Funnel plot for the assessment of publication bias (ARB)

Full size image

Sensitivity analysis

The sensitivity analysis was performed to check the effects of outliers on the outcome.

ACEi/ARB

We have identified 4 studies with a high sample size [48, 60, 62, 63] and one study with a low sample size [20]. The analysis was done again after the exclusion of these studies and pooled odds ratios were found to be 1.08 [0.91, 1.29] which also shows non-significant reductions in deaths of COVID-19 patients in the ACEi/ARB group as compared to non-ACEi/ARB group (Fig. 8a).

Fig. 8
figure 8figure 8

Sensitivity analysis a forest plot of ACEi/ARB after exclusion of high sample size studies (Rosenthal 2020 [63], Rodilla 2020 [62], Rezel-Potts 2021 [60], Lee 2020 [48]) and low sample size study (Banerjee 2020 [20]). b Forest plot of ACEi after exclusion of high sample size studies (Fosbol 2020, Lee 2020) [8, 48] and low sample size study (Banerjee 2020) [20]. c Forest plot of ramipril after exclusion of high sample sizes (Braude 2020) [23] and low sample size (Banerjee 2020) [20]. d Forest plot of ARBs after exclusion of high sample size (Fosbol 2020, Lee 2020) [8, 48]

Full size image

ACEi

The analysis was also done after the exclusion of Fosbol and Lee [8, 48] (high sample size) and Banerjee [20] (low sample size) studies. The pooled odds ratio was found to be 0.92 [0.45, 1.88] which also indicates non-significant reductions in deaths of COVID-19 patients in the ACEi group as compared to the non-ACEi group (Fig. 8b). The overall effect of ramipril was also calculated after the exclusion of Braude [23] (high sample size) and Banerjee [20] (low sample size) studies. The overall odds ratio after exclusion was found to be 0.59 [0.30, 1.14] which also indicates a non-significant reduction in deaths of COVID-19 patients in the ramipril group as compared to the non-rampiril group (Fig. 8c).

ARB

The ARB has also shown a non-significant effect after the exclusion of high sample size studies (Fosbol 2020, Lee 2020) [8, 48] (Fig. 8d).

Discussion

It has been observed that COVID-19 patients with co-morbid conditions such as diabetes (DM), hypertension (HT), or cardiovascular disease (CVD) are more prone to death. There is a number of reported explanations in the literature. The use of medicines could also be one of the reasons. ACEi and ARB are commonly used in hypertensive or cardiovascular disease patients. Both groups of medicines work by inhibiting the RAAS. Angiotensin-converting enzyme inhibitors prevent angiotensin-I from converting to angiotensin-II, whereas angiotensin receptor blockers prevent angiotensin II from acting, resulting in vasodilation and decreased aldosterone output. Angiotensin-converting enzyme-2 (ACE2) is found in a variety of organs, including the alveoli of the lungs, and is related to angiotensin-converting enzyme 1 (ACE1), which plays a role in RAAS [21, 55]. The SARS-CoV-2 uses the angiotensin-converting enzyme (ACE) of the host to enter inside the cell, and some of the classes of drugs (ACEI and ARB) could increase ACE2 expression which can result in increased binding of the SARS-CoV-2. The increased binding of SARS-CoV-2 with the host might result in severe conditions for the patients. Ferrario et al. (2005) have reported safety concerns regarding the use of RAAS inhibitors in COVID-19 patients due to increased ACE2 expression [78]. The COVID-19 hypertensive patients using ACEi/ARB have more tendencies to develop severe pneumonia compared to those not using ACEi/ARB [64]. The literature has shown conflicting findings regarding the use of ACEIs and ARB in COVID-19 patients.

To the best of our knowledge, very few meta-analyses have also been conducted to find out the association of ACEi/ARB in the deaths of COVID-19 patients. Recently, the meta-analysis results of Dai et al. (2021) have reported a non-significant association of ACEIs/ARBs in the deaths of COVID-19 patients. However, data included in this analysis was up to June 20, 2020 [79]. The meta-analysis conducted by Singh et al. (2022) has also shown similar results, however, studies were included up to January 18th, 2021 [80]. Hasan et al. (2020) have conducted a meta-analysis of 24 studies and also reported a non-significant association of ACEIs/ARBs in the deaths of COVID-19 patients [81]. The meta-analysis results of Wang et al. (2021) have concluded that ACEi/ARB treatment was significantly associated with a lower risk of mortality in hypertensive COVID-19 patients. The studies were included up to October 12, 2020 [82]. Recently, the meta-analysis conducted by Azad and Kumar (2022) has shown no significant association of ACEi/ARB in the deaths of COVID-19 patients [83]. We have included 68 observational studies, and the results of our meta-analysis have also shown a non-significant association of ACEi/ARBs in the mortality of COVID-19 patients. We have also tried to find out the effect of individual ACEi/ARBs and also found a non-significant association. Further, the sensitivity analysis results have also shown the non-significant impact of outliers on the outcome.

The failure to publish the results of specific studies due to the direction, nature, or strength of the study findings is known as publication bias. Outcome-reporting bias, time-lag bias, gray-literature bias, full-publication bias, language bias, citation bias, and media-attention bias are all examples of publishing bias in academic articles [84, 85]. The funnel plots of the current investigation have indicated the involvement of publication bias.

Heterogeneity refers to the differences in research outcomes between studies. Heterogeneity is not something to be terrified of; it simply implies that your data are variable. When multiple research projects are brought together for a meta-analysis, it is apparent that differences will be discovered [86]. The current analysis results have also shown heterogeneity among included studies as indicated by I2 statistics.

Limitations

We have included seven studies from the medRxiv.org databases that had not yet been peer-reviewed. We saw this as a drawback because peer-reviewers would be able to see more flaws in reporting techniques and other details. The majority of this research, however, was expected to be peer-reviewed. We didn’t find sufficient data to check the effect of all individual ACEi/ARBs on the outcome. The studies published in the English language are only considered. The search is limited to selected databases only. The funnel plots have also indicated the involvement of publication bias.

Conclusions

In conclusion, to date, the use of ACEi and ARB classes of drugs for the management of co-morbid conditions of COVID-19 patients has not been linked with increased deaths. However, more evidence is required.