Discontinuation of Antihypertensive Medications on the Outcome of Hospitalized Patients With Severe Acute Respiratory Syndrome-Coronavirus 2.

[Figure: see text].

The current severe acute respiratory syndrome-coronavirus 2 (SARS-CoV-2) pandemic has so far led to 2.96 million deaths worldwide (https://ourworldindata.org/covid-deaths, April 13, 2021). Epidemiological data suggests that people with underlying medical conditions like diabetes, hypertension, kidney, or coronary artery disease as well as obesity are more vulnerable to severe disease,[1,2] but this seems to be attenuated after adjusting for age.[3,4] Since the virus uses the membrane-bound ACE2 (angiotensin-converting enzyme 2), an enzyme involved in the renin-angiotensin-aldosterone system (RAAS), to enter and infect the host cells, many have commented on ACE inhibitors or angiotensin receptor blockers (ARB) to result in either a worse or a better outcome.[5,6] On the contrary, it has been suggested, that RAAS blocking agents might increase ACE2 levels and hereby increase the risk of a SARS-CoV-2 infection.[1,2] On the contrary, this has been refuted by others.[7] Interestingly, ARBs are suggested to be beneficial, since they block the deleterious proinflammatory effects of increased angiotensin II levels, which emerge after viral infection.[8,9]

To be able to shed some light on these contradictory issues, many have investigated the effects of RAAS blocking agents on the severity of hospitalized patients with SARS-CoV-2. A most recently updated meta-analysis, that comprised 86 nonrandomized observational studies, showed that ACE inhibitor or ARBs was associated with a small but significant decrease in mortality, among hypertensive patients hospitalized for SARS-CoV-2.[10] Analysis on hospitalized patients is likely to be confounded, since in seriously ill patients antihypertensive and, therefore, RAAS blocking agents are often discontinued.[11,12] If this is the case, it will create a bias in which potentially beneficial medication is withdrawn, leading to a more unfavorable outcome and consequently a neutral net effect on at-home medication use in a SARS-CoV-2 infection. Indeed, recently, others showed that in-hospital withdrawal of ARBs or ACE inhibitor resulted in a higher risk of severe disease or death in patients with a SARS-CoV-2 infection.[11,13,14] Whether this was due to the fact that these patients already had severe disease at admission and whether this was controlled for in the analysis, was not reported. Furthermore, none of the mentioned studies addresses the issue of higher versus lower affinity ARBs. In the previously mentioned meta-analysis, 62% of the data came from Europe and the United States and by checking the prescription level data for individual ARB medications, we found that these countries mainly prescribe low-affinity ARBs.[10] This is another potential reason for the neutral association of ARBs with SARS-CoV-2 related mortality.

Besides, previously, we analyzed the prescription data of different types of antihypertensive medications among 30 countries worldwide and its relationship to the first 3-week mortality due to a SARS-CoV-2 infection. With this analysis, we were able to show that countries prescribing higher amounts of ARBs and especially higher amounts of high-affinity ARBs had lower mortality rates in the first 3 weeks.[15]

Therefore, we hypothesized that discontinuation of RAAS blocking agents in hospitalized patients with SARS-CoV-2, could have an association with worse outcome irrespective of disease severity at admission.

Methodology

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Study Design and Data Collection

In this retrospective cohort study, data on SARS-CoV-2 related hospitalization was collected from 10 participating hospitals in the Netherlands. Baseline characteristics, medication use, comorbidities, and outcomes were extracted from the Electronic Health Records, by a Natural language processing and text mining-based tool validated in previous study.[16] For details see methods in the Data Supplement.

Study Population

Patients admitted to the hospital were included in the study if they were 18 years or older and were diagnosed with SARS-CoV-2. Data on comorbidities, such as hypertension, diabetes, coronary artery disease, and heart failure in the general population, were obtained from the Dutch bureau of statistics (StatLine database Dutch Bureau of Statistics 2018). To compare the use of antihypertensive medication between patients hospitalized with a SARS-CoV-2 infection and the general population personalized prescription data from IQVIA Netherlands was used (methods in the Data Supplement).

Definitions and Study Outcome

Patients were classified according to the severity of disease in 3 groups: mild disease, severe disease, or death. Patients with normal arterial oxygen partial pressure (pO2 ≥80 mm Hg) at admission and no admission to the intensive care unit or death during hospitalization were classified as having mild disease. Severe disease was defined as either a low pO2 (pO2 <80 mm Hg) at admission or intensive care unit admission during hospitalization. The final group consists of the patients who died of SARS-CoV-2. Patients using any of the 5 antihypertensive medication groups, irrespective of the underlying disease it was prescribed for, were classified as antihypertensive users. Data of the five most important antihypertensive medications, that is, ACE inhibitor, ARBs, β-blockers (BB), calcium channel blockers (CCB), and diuretics, either used at home or during hospitalization, were gathered. Information on when and for what reason medication was discontinued during hospitalization was not available. We also defined which ARB had a higher or lower receptor affinity, taking into account the differences in terms of receptor binding kinetics, pharmacodynamics as well as pharmacokinetic properties[17-19] (methods in the Data Supplement).

Statistical Analysis

All analyses were performed using SPSS version 26.0 (SPSS Inc, Chicago, IL). Results were reported as mean±SD for continuous variables and number and percentage (%) for categorical data, except when indicated otherwise. Baseline characteristics were compared using Fisher exact test. We conducted 3 analyses: first, we analyzed the proportions of antihypertensive medication use in all hospitalized patients on the severity of SARS-CoV-2 infection as compared with the general population. Second, we analyzed the impact of discontinuation of antihypertensive medication during hospitalization on the severity of SARS-CoV-2 infection. Finally, we analyzed the effect of discontinuing higher or lower affinity ARBs on severity of SARS-CoV-2 disease. The final model was corrected for sex, age, and modified early warning score. The model with outcome death was additionally adjusted for pO2 at admission (methods in the Data Supplement).

Results

Table 1 shows the baseline characteristics of admitted patients with SARS-CoV-2 as compared with the general population. A total of 2289 patients admitted for a SARS-CoV-2 infection were included in the database, for 1584 (69.20%) data was available on at-home and in-hospital antihypertensive medication use as well as all the desired outcome parameters. Of these 1584 patients, 914 (57.70%) used any of the antihypertensive medications at home. As expected, patients with a SARS-CoV-2 infection were older (68.77±13.52 versus 50.25±18.60; P<0.05), more often of male sex (59.2% versus 49.18%; P<0.05), and had significantly more often hypertension (39.58% versus 20.4%), diabetes (25.06% versus 8.0%), coronary artery disease (13.82% versus 9.5%), and heart failure (9.97% versus 1.6%) as compared with the general population (all P<0.05). Besides, patients admitted for a SARS-CoV-2 infection more often used β-blockers (57.33% versus 50.69%; P<0.05) or diuretics (47.81% versus 38.23%; P<0.05).

Table 1.

Baseline Characteristics of All Patients Admitted With SARS-CoV-2 Infection as Compared With the General Population

Table 2 shows the baseline characteristics of the study population according to disease severity. Of the 1584 patients admitted to the hospital, 418 (26.39%) had mild disease, 791 (49.94%) had severe disease, and 375 (23.67%) patients died. The patients who died were older (77.10±8.61 versus 65.66±12.77 versus 67.19±15.34), more often of male sex (67.43% versus 59.67% versus 51.19%), and had more comorbidities such as hypertension (49.33% versus 33.50% versus 42.58%), diabetes (33.60% versus 21.11% versus 24.64%), coronary artery disease (19.73% versus 10.74% versus 14.35%), and heart failure (13.86% versus 7.45% versus 11.24%) compared with the patients with severe or mild disease, respectively. Besides, patients who died used significantly more often ACE inhibitor (26.13% versus 15.80% versus 20.57; P<0.05) and CCBs (28.27% versus 16.02% versus 17.22%; P<0.05) compared with patients with severe disease and mild disease, respectively. Furthermore, patients who died used significantly more often β-blockers (40% versus 28.70%; P<0.05) and diuretics (33.07% versus 28.91%; P<0.05) compared with patients with severe disease. Patients who developed severe disease were more often men (59.67% versus 51.19%; P<0.05) and significantly less often used ARBs (13.78% versus 18.18%; P<0.05), ACE inhibitor (15.80% versus 20.57%; P<0.05), and β-blockers (28.70% versus 35.17%; P<0.05) compared with those who developed mild disease.

Table 2.

Baseline Characteristics of All Patients Admitted With SARS-CoV-2 Infection With Mild or Severe Disease or Whom Died

At-Home Antihypertensive and In-Hospital Prognosis

When analyzing only individuals with at-home antihypertensive medication, patients who died were older (77.76±8.02 versus 69.77±10.85 versus 72.80±11.77), more often men (65.12% versus 57.35% versus 52.02%), and had more comorbidities such as hypertension (57.75% versus 46.32% versus 55.24%), diabetes (38.37% versus 24.51% versus 26.61%), coronary artery disease (21.70% versus 13.72% versus 19.76%) compared with the patients with severe disease or mild disease, respectively (Table 3).

Table 3.

Baseline Characteristics of Patients on Antihypertensive Medication Admitted With SARS-CoV-2 Infection With Either Mild or Severe Disease or Whom Died

The number of patients in whom at least one antihypertensive medication was discontinued during hospitalization was 76 for ARBs (28.14%), 100 for ACE inhibitor (37.04%), 33 for BB (12.22%), 66 for CCB (24.44%), and 99 for diuretics (36.66%), as depicted in Table 1. Among the patients who died, discontinuation of ACE inhibitors (46.93% versus 24.0% versus 27.91%; P<0.05) and β-blockers (11.33% versus 4.85% versus 3.40%; P<0.05) was reported significantly more often compared with patients who developed severe disease and mild disease, respectively (Table 3).

When analyzing the risk of dying or developing severe disease in patients with a SARS-CoV-2 infection, we only observed that individuals using CCB had a higher risk of dying, independent of any of the adjustments including modified early warning score score (Table 4).

Table 4.

Association of At-Home Antihypertensive Medications With Outcome in Patients on Antihypertensive Medication Hospitalized for SARS-CoV-2

Discontinuation of Antihypertensive Medications During Hospitalization

When analyzing the effect of in-hospital discontinuation of antihypertensive medication on the severity of SARS-CoV-2 infection, we observed that discontinuation of ARBs, ACE inhibitor, and BB was associated with an increased risk of dying from a SARS-CoV-2 infection, even after adjusting for confounders such as age, sex, and modified early warning score at admission (β, ARB, 2.65 [95% CI, 1.17–6.04]); ACE inhibitor (β, 2.28 [95% CI, 1.15–4.54]); BB (β, 3.60 [95% CI, 1.10–10.27]; Table 5), where as discontinuation of CCBs or diuretics was not. Finally, to further substantiate our findings, we also analyzed the effect of discontinuing antihypertensive medication association with severity of a SARS-CoV-2 infection in a selected population of hypertensive patients only and found similar results (Table S1 in the Data Supplement).

Table 5.

Association of Discontinuing At-Home Antihypertensive Medications With Outcome in Hospitalized Patients With SARS-CoV-2

High- Versus Low-Affinity ARBs

To further substantiate our observations, we also analyzed the distribution of higher or lower affinity ARBs on disease severity. We observed that with increasing severity of disease, a smaller number of individuals used high-affinity ARBs, which might suggest that the affinity of ARBs is relevant in preventing disease severity. These findings were not statistically significant (Figure; Table S2).

Distribution of high- versus low-affinity angiotensin receptor blockers (ARB) in individuals with mild or severe disease or whom died.

Discussion

In the present study, we show that discontinuation of ARBs, ACE inhibitor, and β-blockers during hospitalization for a SARS-CoV-2 infection is associated with a 2 to 4× increased risk of death, independent of the severity of disease at admission.

In animal models, it has been suggested, that ARBs might alleviate the deleterious effects of ever-increasing Ang II (angiotensin II) levels, due to the SARS-CoV-2 infection.[20,21] Since ACE2 is internalized when the SARS-CoV-2 virus binds to it, the enzyme no longer converts proinflammatory Ang II into anti-inflammatory Ang 1 to 7, leading to overstimulation of the Ang II type 1 receptor, inducing lung damage and respiratory distress syndrome. Therefore, since ARBs and ACE inhibitor block this pathway and maintain an equilibrium, discontinuation of these medications in patients with SARS-CoV-2 infection could lead to worsening of disease.[22] Besides, sympathetic overdrive due to an excessive increase in the level of Ang II could override the β1 adrenergic receptor blocked by daily use of β-blockers. The above mechanisms as well as the effect increasing Ang II has on superoxide production was previously described by us.[23]

During the initial phase of the pandemic, others have suggested that the use of ARBs and ACE inhibitor could lead to an increased susceptibility of a SARS-CoV-2 infection as ARBs and ACE inhibitor could upregulate ACE2 and facilitate the entry of the virus. If this would be the case, patients infected with the SARS-CoV-2 virus would more often use ARBs and ACE inhibitor compared with the general population, which has been shown not to be the case.[24,25] In addition, if the use of ARBs and ACE inhibitor would be associated with a more severe SARS-CoV-2 infection, represented by hospitalization, patients using ARBs and ACE inhibitor would be overrepresented among hospitalized SARS-CoV-2 patients as compared with the general population, which was not the case in the present study. Furthermore, previous studies were not designed to answer the question whether ARBs and ACE inhibitor lead to more severe SARS-CoV-2 disease in hospitalized patients, since they did not take into account the discontinuation of at-home medication during admission.[10]

Recently published studies that investigated the association between discontinuation of RAAS blocking agents and outcome in hospitalized SARS-CoV-2 patients, found similar results, namely that discontinuation of these medications during the course of hospitalization was associated with an increased risk of dying from SARS-CoV-2 infection,[11-13] whereas the continuation of these medications was associated with lower mortality.[14,26] On the contrary, a recent large randomized controlled trial on the discontinuation of RAAS inhibitors found no difference in the duration of hospital or death due to a SARS-CoV-2 infection but had a serious flaw in the study design. Namely, randomization was broken due to the fact that around 80 individuals were left out of the analysis, due to misconduct of one of the study sites. Interestingly, additional on treatment analysis showed a similar 2× higher risk of dying in the group randomized to discontinuation of ARB medication during hospitalization.[27]

Another important confounder to consider, is that a favorable outcome of ARBs, in case of a SARS-CoV-2 infection, can only be anticipated if the Ang II type 1 receptor is sufficiently blocked. The currently most prescribed ARBs, losartan and valsartan, have a low Ang II type 1 receptor affinity, a short half-life and are prescribed in a once daily dose. Therefore, they only modestly block the Ang II effects. Meaning that we cannot expect sufficient Ang II type 1 receptor blockade to affect the outcome of a SARS-CoV-2 infection. Since most of the published studies mainly prescribe low-affinity ARBs[10] and since we already showed that worldwide higher affinity ARBs are associated with lower first 3-week mortality,[15] we suggest that this should be better looked into. The current analysis only shows a slight decrease in the number of individuals on a high-affinity ARB, in the group that died, but the data was underpowered to show significance.

Strength and Limitations

One could speculate that discontinuation of ACE inhibitor, ARBs, and BB among patients who died of SARS-CoV-2, could be due to the fact that these patients were already on the verge of dying and for that reason antihypertensive medication was not deemed necessary and therefore discontinued. On the contrary, if this was the case, then this would also hold true for CCB and diuretics, which was not the case. Besides, information on when or for what reason medication was discontinued during hospitalization was lacking. Although medications will often be discontinued because of more severe disease, we tried to control for that by correcting our data for the modified early warning score score, which did not change the results. Therefore, we speculate that higher risk of mortality is related to the discontinuation of RAAS and β-blockers themselves, although this analysis cannot confirm this because of its cross-sectional nature.

Another limitation might be that the analysis could be influenced by the fact that antihypertensive drugs are prescribed for different underlying conditions. We therefore, we adjusted the logistic regression analysis for comorbidities, which did not influence the results. In addition, we also analyzed hypertensive patients only and found similar results. Nevertheless, it is important to stress, that one should consider all antihypertensive medication users as a whole, irrespective of the underlying disease it is prescribed for, since its association with a SARS-CoV-2 infection will not differ. Whether the underlying disease and its reaction to a SARS-CoV-2 infection will differ, is still to be debated. Most of the underlying diseases, such as hypertension, diabetes, obesity, and cardiovascular disease overlaps within patients and probably have a common denominator, such as insulin resistance. However, we besides considering underlying comorbidities we also explored the association with the number of prescribed antihypertensives.

Another issue is the retrospective observational nature of the study, which means that the data is acquired in the past and one is dependent on the quality of this data. Therefore, we could not use the appropriate standard definition of SARS-CoV-2 severity but instead used the most crude and conservative definitions related to pO2 at admission. In addition, retrospective studies would not allow you to assess causality, therefore, we tried to substantiate our findings by assessing a kind of dose-response relationship by analyzing the effects of lower or higher affinity ARBs.

Conclusions

In conclusion, discontinuation of at-home ACE inhibitor, ARBs, or BB in patients hospitalized for a SARS-CoV-2 infection was associated with an increased risk of dying, whereas discontinuation of CCBs and diuretics was not.

Perspectives

In our multicentre study from the Netherlands, discontinuation of ACE inhibitor, ARBs, and BB in hospitalized patients with SARS-CoV-2 infection was associated with an increased risk of dying, irrespective of the severity of disease at admission. The observational study design, prevents us from concluding causality. However, this study provide association in the protective roles of ARBs, ACE inhibitor and β-blockers in hospitalized patients with SARS-CoV-2.

Sources of Funding

None.

Disclosures

None.