We designed a prospective registry that included consecutive laboratory-confirmed COVID-19 patients admitted at our institution. The period of inclusion ranged from 27 February to 7 April 2020. Clinical outcomes were monitored up to 2 May 2020. Demographic characteristics (age and sex), clinical data (both CV and other significant comorbidities, laboratory findings, treatments, complications, and in-hospital outcomes) and results of cardiac examinations (electrocardiography and echocardiography) were collected from medical records and independently entered into a dedicated electronic database. A predefined laboratory test (including a complete blood count, coagulation tests, liver and renal function, electrolytes, C-reactive protein, procalcitonin, lactate dehydrogenase, creatine kinase, d-dimer, NT-proBNP and high-sensitivity cardiac-specific troponin-T (hs-cTnT) was performed on admission. Levels of NT-proBNP and hs-cTnT were measured thereafter if clinically indicated.

This study was performed according the Declaration of Helsinki, ISO 14155, and clinical practice guidelines. The study protocol was approved by the Institutional Ethics Committee and the hospitals’ research commissions of our institution. Written informed consent was waived in light of the urgent need to collect data.

Only laboratory-confirmed cases were included. A laboratory-confirmed COVID-19 was defined by a positive result on a reverse-transcriptase–polymerase-chain-reaction (RT-PCR) assay of a specimen collected on a nasopharyngeal swab. RT-PCR assays were performed according to World Health Organization interim guidance document.4

The electrochemiluminescence-based immunoanalytical system, Elecsys 2010 (Roche Diagnostics Ltd., Germany) was used to determine plasma levels of NT-proBNP and hs-cTnT. Hs-cTnT had a limit of detection of 5ng/L and a 99 percentile in the healthy population of 14.0ng/L. The manufacturer recommends a cut-off value for a positive test (for myocardial injury) of 14.0ng/L.5 For NT-proBNP, a level ≥ 300pg/mL on the acute setting was considered positive, following European Society of Cardiology recommendations.6

The presence of CV risk factors was defined as the existence of at least one item among hypertension, hyperlipidemia, diabetes, history of smoking or obesity. CVD was defined as the presence of coronary artery disease, chronic congestive heart failure (CHF), valvular heart disease, atrial fibrillation/atrial flutter, stroke or peripheral artery disease. Cardiac injury was diagnosed if serum levels of hs-cTnT were above the 99 percentile upper reference limit, regardless of new electrocardiographic or echocardiographic abnormalities, according to the Fourth Universal Definition of Myocardial Infarction.7 Acute respiratory distress syndrome (ARDS) was diagnosed according to the Berlin criteria as acute-onset hypoxemia (ratio of arterial oxygen partial pressure to fractional inspired oxygen expressed as a fraction [PaO2/FiO2]<300) associated with bilateral pulmonary opacities on chest imaging that were not fully explained by CHF or other forms of volume overload.8

Quantitative variables are presented as mean±standard deviation or median [interquartile range]. Categorical variables are presented as number (percentage). The Student t test or the Mann–Whitney U were used to compare continuous variables. Pearson's chi-square test was used for categorical variables.

We designed a multivariable model to find independent predictors of death during hospitalization. To build this model, we analyzed a total of 19 independent variables using a bivariate logistic regression analysis. We included age > 65 years old, sex, CV risk factor (hypertension, hyperlipidemia, diabetes mellitus, smoking habit, obesity), CVD and other significant comorbidities (chronic kidney disease (CKD) – defined as a glomerular filtration rate < 60mL/min per 1.73m2, chronic obstructive pulmonary disease (COPD), bronchial asthma and history of previous cancer). We included other 2 variables, active treatment with angiotensin-converting enzyme inhibitors (ACEIs) or angiotensin receptor blockers (ARBs), based on previous studies describing their potential role on SARS-CoV-2 mechanisms of infection. For the selection of variables, we used a generous threshold of P<.1 to determine statistical significance to ensure that we did not drop any potentially relevant variable from the final model. Overfitting of the multivariable model was ruled-out on the basis of McFadden's pseudo R-squared values and cross-validation method. For the rest of the test, a value of P<.05 was considered statistically significant. All tests were performed with STATA 12 (StataCorp LLC, United States).

Previous observational data described worse outcomes in patients with CVD and COVID-19 infection. Guo et al. described a higher mortality in patients with CVD (including hypertension, coronary artery disease and cardiomyopathy) compared to those without CVD in 187 COVID-19 patients in China.9 Li et al. found a higher prevalence of hypertension, diabetes and CVD among severe COVID-19 patients admitted to the ICU compared with those admitted to non-ICU units.10 In our cohort, those patients with CVD or CV risk factors had a significantly higher risk of death during hospitalization. Furthermore, those patients with CVD or CV risk factors more frequently showed elevated cardiac biomarkers (both hs-cTnT and NT-proBNP) on admission. Recently, Guo et al. described a higher risk of death in patients with CVD and associated myocardial injury (evaluated by elevation of troponin T) among 187 patients with COVID-19.9 Although the mechanisms of cardiac injury of SARS-CoV-2 remain not fully understood, several potential mechanisms have been proposed, including direct myocardial damage by the virus, systemic inflammatory response to infection or indirect damage induced by hypoxia.9 Whether worse outcomes in patients with CVD are mediated by this mostly subclinical cardiac injury remains unknown. On the other hand, most CVD components and CV risk factors failed to show an independent relationship with in-hospital death, with only CHF appearing as an independent predictor. This can be explained by baseline differences between groups of patients with/without CVD and CV risk factors. In our study, patients with CVD and CV risk factors were significantly older. The age difference may explain the increased risk found in these subgroups. With this in mind, we would better consider the presence of CVD and CV risk factors as a risk marker. Furthermore, as Fig. 3 illustrates, it seems that there is an interaction between age and CVD or CV risk factors. The presence of CVD or CV risk factors was not associated with worse outcomes in younger patients. On the contrary, the presence of CVD or CV risk factors in older patients identifies a subgroup with particularly poor prognosis in COVID-19.

Infections are a frequent cause of hospitalization in patients with CHF. Alon et al. described respiratory infections as the most frequent focus of infection in patients with CHF.11 Furthermore, infectious-related admissions, compared with other reasons for admission, seem to be associated with an increased 30-day mortality in patients with CHF.11 Moreover, in our cohort, previous history of CHF was higher (5 vs 1%) than in previous reports regarding other coronaviruses like Middle East respiratory syndrome coronavirus (MERS-CoV).12

The linkage between ACEIs/ARBs and COVID-19 is based on the association between angiotensin-converting enzyme 2 (ACE2) and coronaviruses like SARS-CoV-1 and SARS-CoV-2.13,14 In these viruses, ACE2 acts as a co-receptor for viral entry in the cell. Initial reports of case series from China reported a higher risk of ARDS and death in patients with hypertension.15,16 These early data raised concern about a possible harmfull effect of ACEIs/ARBs in patients with COVID-19. Several scientific societies, like the Council on Hypertension of the European Society of Cardiology, promptly replied arguing that there was no scientific evidence to suggest that treatment with ACEIs or ARBs should be discontinued in patients with COVID-19. In our study, both ACEIs and ARBs failed to show an independent relationship with an increased risk of death in COVID-19 patients, reassuring about the safety of these therapies in the context of SARS-CoV-2 infection.

Initial reports described a higher risk of death during admission by COVID-19 infection. Zhou et al. described 54 deaths among 191 patients with COVID-19 (28%) admitted at 2 hospitals in Wuhan.15 Similarly, Wu et al. described 44 deaths among the 201 COVID-19 patients (22%) admitted at 1 institution in Wuhan.16 In our cohort 103 patients (12%) died during the initial admission. As age has been systematically found as a strong predictor of death in COVID-19, differences in mortality might be explained by differences in age between different cohorts. However, in our cohort, mean age was considerably old (62 years old) and very similar to previous data on retrospective cohorts.

Our work has several limitations that deserve discussion. First of all, due to its observational non-randomized design, we cannot exclude the presence of potential confounders that were not identified. Secondly, despite our protocol included the measurement of both hs-cTnT and NT-proBNP levels at the time of admission, only 75% of patients for hs-cTnT and 58% of patients for NT-proBNP were finally evaluated. Therefore, this may constitute a cause of uncontrolled selection bias. Thirdly, the role of age in the prognosis of COVID-19 may also be a matter of debate. Even if it seems clear that older patients have a worse prognosis in COVID-19, part of this poorer outcomes may be linked to less access of these patients to some therapies that might reduce mortality, as might be the case for mechanical ventilation or new drugs. Although in our cohort, we found no differences between older and younger patients in the indication of mechanical ventilation, we cannot exclude that some older patients with aggressive disease were not selected for mechanical ventilation based on opportunity cost criteria in a situation where healthcare system was on the verge of collapse. Fourthly, most patients in our study required hospital admission. Only a group representing 10% of the total cohort was managed in a hospital-at-home program without hospital admission. Furthermore, milder cases without respiratory compromise and with only minor symptoms may have been managed as outpatients by general practitioners and are not well represented in our cohort.