In this analysis of 286 patients hospitalized with COVID-19, early in early phase of the pandemic, we found that patients taking combination RAASi and diuretic for the indication of Hypertension were at higher risk of in-hospital mortality than patients taking RAASi alone. We also found that continuing RAASi therapy was highly beneficial, in terms of in-hospital mortality, for patients who were concomitantly treated with a diuretic. We did not find an association between RAASi (dis)continuation and in-hospital mortality for patients who were not treated with a diuretic. These trends remained true after accounting for age, which was also associated with increased risk of mortality, as well as coronary artery disease and congestive heart failure. We found that RAASi continuation was most important for patients with COVID-19 treated with both an ARB and a diuretic. These findings suggest that RAASi combined with diuretic use may have been a proxy for left ventricular diastolic or systolic dysfunction; thus, these drugs exerted a beneficial effect throughout the COVID-19 hospitalization course, which is invariably influenced by degrees of respiratory failure.
Others have suggested that RAASi treatment may be protective in patients with COVID-19. [10,12,13] Most studies focused on these drugs in the context of hypertension and not in other CV settings such as IHD or CHF. Among those with HF, the compiled results of many studies suggest that continuation of RAASi and other components of goal-directed medical therapy is consistently associated with reduced mortality and hospitalization. [20] This principle appears to extend to patients hospitalized with COVID-19 at high risk for acute cardiorenal syndromes.[21,22] Recently, the BRACE CORONA trial showed no association between RAASi continuation and the primary outcome of death following hospital discharge for patients with COVID-19. [17] Further, a separate study demonstrated that discontinuation of RAASi therapy for patients hospitalized with COVID-19 was associated with increased risk of mortality following discharge. [18] Because patients with severe hypertension and those with other CVD are often treated with a combination of RAASi and diuretics, we investigated the differential effect of (dis)continuing this combination therapy on in-hospital mortality. Our work had two primary findings 1) patients who were treated with diuretics had a higher rate of in-hospital mortality compared to those who were not treated with a diuretic (16.3% vs. 11.1%); 2) we confirmed an overall neutral effect on in-hospital mortality for patients continuing versus discontinuing RAASi, except for patients who were also treated with a diuretic. The risk of in-hospital mortality was significantly increased for patients treated with a diuretic who discontinued RAASi therapy compared to those who continued RAASi therapy (25.5% vs. 7.8%); this finding was largely driven by ARB-users. These observations may support the hypothesis that in COVID-19 an enhancement of RAAS activity is harmful and correlates with the degradation of respiratory function as well as the arterial vascular tone] In this view, the use of diuretics amplifying the activity of RAAS can lead worse outcome.
The roles of RAASi observed in various settings may be explained by the different baseline conditions in terms of age, sex, ethnicity, extra-cardiac comorbidities and frailty. Indeed, the treatment may configure a negative effect in older patients with more advanced systemic diseases and vulnerable general conditions. Therefore, ACE expression varies in relation to race and sex, explaining some discrepancies observed in young people and in black patients, in whom prior ARB use was associated with augmented viral susceptibility and infection severity.[24,25]
There are clinical scenarios in which patients may need to discontinue therapy. Patients with more hemodynamic impairment, lower blood pressure, and dehydration may benefit from discontinuing RAASi. [26] Notably, A double-edged mechanism related to clinical presentation and underlying CVD may be posed. In this context, it is important to know the clinical status related to the underlying CVD and baseline condition prior to infection. [9,27] In the Swedish meta-analysis including 138,700 patients, the protective effect of RAASi was observed for all types of CVD studied. Unfortunately, that study investigated only antecedent use of RAASi; the effect of continuation was not evaluated.[28] In another study comparing CV therapies head-to-head, Reynolds and colleagues did not find relevant differences between any of the five major drug classes. [29] Behind these data, some observations about the role of the ACE system in viral spread may be warranted. The virus could enter directly inside the epithelial cell of the respiratory system via the ACE receptor, inducing an inflammatory cascade via bradykinin escape. [30,31] The subsequent increase in prostaglandins and cyclooxygenases leads to interleukins overproduction. ACE catalyzes the conversion between angiotensin I to angiotensin II, resulting in a potent vasoconstrictive effect, whereas ACE2 converts angiotensin II to angiotensin 1-7, causing significant vasodilatory, anti-inflammatory, and anti-fibrotic effects. [32,33] The consequent ACE2 upregulation mediated by ACE-inhibitors leads to beneficial pulmonary and vascular effects, improving endothelial and epithelial cell metabolism, and antagonizing thrombotic cascade. [16,34] Current mechanisms are likely less pronounced by ARB, in which ACE blockade occurs only at the angiotensin tissue (AT2) receptor level with incomplete protection against ACE system modification in COVID-19. [35] Overall, our findings confirm a relevant role of RAASi in the context of COVID-19 and show the importance of considering the basal clinical condition, severity of infection, and concomitant treatments that may influence drug discontinuation and outcomes. [36]
Limitations
This a retrospective observational analysis with relatively small number of patients meeting inclusion criteria; thus, the findings are limited by the small sample size. Further, due to the small event count, we required specialized statistical methods to produce tractable estimates and confidence intervals for the multivariable adjusted model. Approximately 98% of the patients in this study had hypertension, so the findings may not be widely generalizable. Additionally, we did not record information regarding prior hospitalizations and/or severity of extracardiac diseases affecting these patients. Similarly, we lacked information on RAASi dose, duration, and tolerability with respect blood pressure and kidney function. The observational period was strictly limited to the hospital setting; no data were available post-discharge. Our findings cannot be extended to ambulatory patients with less severe symptoms. Finally, these data reflect patients hospitalized during the first wave of the pandemic in Italy and may not reflect characteristics of patients hospitalized more recently since evidence of viral mutations.