Cardiovascular Risk Factors and Clinical Outcomes among Patients Hospitalized with COVID-19: Findings from the World Heart Federation COVID-19 Study

Background and aims: Limited data exist on the cardiovascular manifestations and risk factors in people hospitalized with COVID-19 from low- and middle-income countries. This study aims to describe cardiovascular risk factors, clinical manifestations, and outcomes among patients hospitalized with COVID-19 in low, lower-middle, upper-middle- and high-income countries (LIC, LMIC, UMIC, HIC). Methods: Through a prospective cohort study, data on demographics and pre-existing conditions at hospital admission, clinical outcomes at hospital discharge (death, major adverse cardiovascular events (MACE), renal failure, neurological events, and pulmonary outcomes), 30-day vital status, and re-hospitalization were collected. Descriptive analyses and multivariable log-binomial regression models, adjusted for age, sex, ethnicity/income groups, and clinical characteristics, were performed. Results: Forty hospitals from 23 countries recruited 5,313 patients with COVID-19 (LIC = 7.1%, LMIC = 47.5%, UMIC = 19.6%, HIC = 25.7%). Mean age was 57.0 (±16.1) years, male 59.4%, pre-existing conditions included: hypertension 47.3%, diabetes 32.0%, coronary heart disease 10.9%, and heart failure 5.5%. The most frequently reported cardiovascular discharge diagnoses were cardiac arrest (5.5%), acute heart failure (3.8%), and myocardial infarction (1.6%). The rate of in-hospital deaths was 12.9% (N = 683), and post-discharge 30 days deaths was 2.6% (N = 118) (overall death rate 15.1%). The most common causes of death were respiratory failure (39.3%) and sudden cardiac death (20.0%). The predictors of overall mortality included older age (≥60 years), male sex, pre-existing coronary heart disease, renal disease, diabetes, ICU admission, oxygen therapy, and higher respiratory rates (p < 0.001 for each). Compared to Caucasians, Asians, Blacks, and Hispanics had almost 2–4 times higher risk of death. Further, patients from LIC, LMIC, UMIC versus. HIC had 2–3 times increased risk of death. Conclusions: The LIC, LMIC, and UMIC’s have sparse data on COVID-19. We provide robust evidence on COVID-19 outcomes in these countries. This study can help guide future health care planning for the pandemic globally.


INTRODUCTION
The coronavirus disease 2019 (COVID-19) pandemic has presented an unprecedented global challenge to health care communities. The pandemic continues to affect the lives of millions worldwide, with substantially growing numbers of infections and deaths in all countries [1][2][3]. Although a large proportion of patients (~80%) with COVID-19 have mild to moderate symptoms, the case-fatality rate is highly variable [4][5][6]. The mortality rate among adults with COVID-19 ranges between 2-7% overall, and up to 20% among the elderly [7,8]. Prior studies suggested that people with established cardiovascular disease (CVD), or those at high CVD risk, develop a more severe course of COVID-19 needing admission to an intensive care unit (ICU) [9] and experience higher mortality [10,11]. Furthermore, recent reports suggest an excess of cardiovascular complications among COVID-19 patients, including acute coronary events, myocardial injury, and heart failure [12].
However, most published reports (~95%) on the epidemiology and management of COVID-19 patients are from high-income countries (HICs) [13]. There is limited documentation from low-, lower-middle-, and upper-middle-income countries (LIC, LMIC, and UMIC) where >90% of the poorest billion in their most productive age-group live [14,15]. Also, it is unclear whether the commonly prevalent CVDs in LICs such as rheumatic heart disease, congenital heart disease, and peripartum and other cardiomyopathies, increase the risk of severe course of COVID-19, leading to poor outcomes. Many LMICs have a high burden of CVD and its risk factors that are associated with greater morbidity and mortality following a COVID-19. Managing seriously ill patients with COVID-19 requires vast resources (medical supplies, ventilators, etc.), emerging as a daunting challenge even in HICs. To plan credible policies to combat the COVID-19 pandemic, a better understanding of how resource-constrained countries are dealing with the COVID-19 pandemic is imperative. This World Heart Federation (WHF) global study aimed to describe cardiovascular risk factors and clinical outcomes among patients hospitalized with COVID-19 across diverse populations to inform clinical and policy practices.

STUDY DESIGN AND SETTING
The full details of the study design, patient recruitment, eligibility criteria, and assessments were reported previously [13]. Briefly, we conducted a prospective cohort study of adults with COVID-19 from hospitals in LIC, LMIC, UMIC, and HIC [16], with a 30-day post-admission followup. We invited all WHF members from 100+ countries to identify three recruiting centres in their respective countries. Each centre was expected to recruit between 50 and 500 consecutive COVID-19 patients, depending on the size of the country. Forty hospitals from 23 countries participated in this study. Figure 1 shows the location of the participating countries.

OUTCOMES
Study outcomes at hospital discharge included the need for ICUs admission, need for a ventilator, death (with cause), major adverse cardiovascular events (MACE; myocarditis, arrhythmia, heart failure including left ventricular ejection fraction, acute coronary event including cardiac arrest, and acute heart failure), neurological outcomes (ischemic stroke, transient ischemic attack, haemorrhagic stroke), acute renal failure and pulmonary outcomes such as pneumonia, and acute respiratory distress syndrome. In addition, post-discharge death and rehospitalization up to 30-days were collected.

SAMPLE SIZE AND ANALYSIS
With a sample size of ≥5200 eligible COVID-19 hospitalized patients, and assuming an incidence of 8% of an adverse outcome such as all-cause mortality and cardiovascular events, this study had more than 90% power to detect adverse outcomes (mortality, MACE) with 95% confidence and a margin of error of 2% (confidence interval). Further, with a sample of ≥5000 COVID-19 patients, and assuming a proportion of potential cardiovascular risk factors such as hypertension or diabetes at 20%, the study has more than 80% power to detect a relative risk of 2.0 for poor outcomes (death and MACE).

STATISTICAL ANALYSES
The hospital-level resources and facilities available are reported as numbers (percentages), and by WHO region and World Bank income groups. We report overall demographics, clinical characteristics, medication use (pre-admission), laboratory parameters of the study participants (and by status as survivors), in-hospital deaths, and post-discharge 30-day deaths. Data are reported as a number (percentage) for categorical variables, mean (SD) for normally distributed continuous variables, and median (IQR) for skewed distributions. The overall p-value for differences between survivors,   Table 2b describes the COVID-19 symptoms, and co-morbidities among study participants overall and according to vital status at discharge and 30 days. Overall, almost all hospitalized patients (95%) were diagnosed for COVID-19 by RT-PCR. The median time from symptom onset to admission was 5 (IQR: 3 to 8) days. The most common presenting symptom was cough (68.2%), followed by fever or chills (66.4%), and dyspnoea (62.3%). A history of cardiovascular disease was reported in 32.2% -the most common of which were coronary artery disease (10.9%), heart failure (5.5%), stroke (3.7%) and arrhythmia (3.0%). In terms of cardiovascular risk factors,     47.3% had hypertension and 32.0% had diabetes mellitus. Non-survivors more often presented with significantly higher heart rate, lower diastolic blood pressures, shortness of breath and more frequently had hypertension, diabetes, coronary heart disease, atrial fibrillation, rheumatic heart disease, Chagas disease, valvular disease, and chronic kidney disease (Table 2b).  and Asians (13%) had significantly higher proportions of in-hospital deaths. Furthermore, Middle Eastern (4%), and Asians (3%) had relatively higher post-discharge 30-day deaths than other race/ethnicities (p < 0.001). Figure 2b shows the in-hospital deaths and post-discharge 30-days death by World Bank income groups. Compared to HIC (4%), COVID-19 patients recruited from LIC (10%), LMIC (15%), and UMIC (19%) had significantly higher in-hospital deaths (p < 0.001). eTables 1a and 1b shows the demographics and clinical characteristics of study participants, and COVID-19 symptoms/vital signs by World Bank income groups, respectively. Furthermore, a greater number of patients from LMIC (38.9%) and LIC's (31.7%) reported pre-existing diabetes versus UMIC (20.5%) or HIC (22.0%). eTable 2 shows pre-admission medications. Angiotensin converting enzyme (ACE) inhibitors or angiotensin II receptor blockers were used by 27.1%, beta blockers by 16.9%, diuretics by 11.8%, aspirin by 18.3%, and NSAIDs by 3.5%. Anti-coagulants/ antiplatelets, nitrates, diuretics, aldosterone antagonists, oral hypoglycaemic, and insulin were more commonly used by non-survivors than by survivors (p < 0.05 for each). However, NSAIDs and antidepressants were more commonly used by survivors than non-survivors (P < 0.05). Supportive care and medications prescribed during hospitalization are shown in eTable 3. Overall, two-thirds of participants (63.7%) required oxygen therapy, 12.6% required non-invasive ventilation, and 7.4% required invasive ventilation. More than half of the participants were given intravenous fluids (55.8%), corticosteroids (69.5%) and antibiotics (68.1%). Less than a quarter of the participants (23.1%) were prescribed RAAS inhibitors during hospitalization.
The incidence of individual cardiovascular events, and other clinical outcomes stratified by survival status at discharge and 30-day follow-up, is presented in Table 4. Overall, the median    (IQR) length of hospital stay was 9 (6 to 14) days. About one-third of participants (31.4%) required ICU admission, and the median (IQR) number of days in ICU was 7 (3 to13) days. The most frequently reported clinical outcomes at discharge were pneumonia (37.5%), acute respiratory distress syndrome (13.6%), acute renal injury (8.2%), cardiac arrest (5.5%), shock (5.4%) and acute heart failure (3.8%). Non-survivors more frequently had ICU admission (p < 0.001) and major adverse cardiovascular events, and other adverse clinical outcomes (p < 0.001). The overall vital status at discharge and 30-day follow-up is presented in eTable 4. The most commonly reported causes of death included respiratory failure (39.3%), sudden cardiac death (20.0%) and other cardiovascular deaths (17.4%). A total of 565 participants (16.6%) were not fully recovered from the COVID-19 at 30-day follow-up, and a relatively small proportion of participants (3.8%) responded that their ability to self-care at discharge was worse than before COVID-19. Table 5a shows the factors associated with overall mortality from any cause using the log binomial regression models. As per the final regression Model 3 adjusted for demographics, clinical characteristics and smoking, the elderly (≥60 years) was at three times higher risk of death than younger (<45 years) patients. Male vs. female sex increased the risk of death by at least 16%. Compared to Caucasians, Asians, Blacks, and Hispanics had almost 2-4 times higher risk of death. Likewise, patients recruited from LIC, LMIC, UMIC vs. HIC had 2-3 times increased risk of death. Lastly, pre-existing coronary heart disease, renal disease, diabetes, oxygen therapy use, ICU admission, and higher respiratory rate were all significantly associated with the risk of overall death (p < 0.05 each).   eTable 5 shows the factors associated with in-hospital death and post-discharge 30-day death from any cause, using the multinominal logistic regression models. In addition to older age, male sex, ethnicity (Asian/Middle Eastern) and pre-existing CHD, renal disease, and diabetes, HIV status emerged as a significant factor associated with post-discharge 30-day death but not for in-hospital death.

DISCUSSION
This study represents the first comprehensive global data on mortality, cardiovascular outcomes, and cardiovascular risk factors among hospitalized COVID-19 patients recruited from diverse global populations. We found a high prevalence of hypertension, diabetes, and cardiovascular disease among patients admitted with COVID-19. The main predictors of mortality or MACE in this WHF study cohort included older age, male sex, pre-existing coronary heart disease, diabetes, renal disease, severe COVID-19 infection with higher respiratory rates and requiring ICU admission and oxygen therapy. We did not find an association of ACE-inhibitors or ARBs with either mortality or cardiovascular events. In the unadjusted regression model, hypertension was significantly associated with mortality, but it did not reach statistical significance in the final model adjusted for both demographic (age, sex, ethnicity, smoking status) and clinical characteristics (diabetes, COPD, asthma). However, hypertension significantly increased the risk of MACE in our cohort. Interestingly, HIV infection was significantly associated with postdischarge 30-day death but not in-hospital death.
Our study is particularly notable for its prospective recruitment of consecutive COVID-19 hospitalized patients across multiple hospital sites spread across LMICs, MICs, and HICs. Further, this is the first global study to explore the relationship of country income status with the clinical outcomes. We found COVID-19 patients recruited from LIC, LMIC and UMIC's were at significantly greater risk of mortality than HICs. Likewise, patients from LMIC and UMICs vs. HICs had almost 2-3 times increased risk of MACE, but patients from LIC's had lower risk of MACE, which can be partly explained due to the variation in the demographic characteristics, and pre-existing chronic conditions. Our analysis demonstrated a greater rate of in-hospital deaths, post discharge 30-day deaths and MACE among Hispanics, and Asian populations compared to Caucasians. Higher prevalence of comorbidities such as hypertension, diabetes, renal disease and obesity among Asians, Hispanics, and other populations (such as Blacks and Middle Eastern populations) may play a role in the increased mortality and MACE in our cohort of COVID-19 patients. The association of ethnicity with in-hospital deaths in patients with COVID-19 has been reported in previous studies [23,24]. Multiple studies from the United States report that Hispanic, Asian or Pacific Islander, and African American patients had higher risk of COVID-19 associated hospitalization and higher mortality compared with White patients after adjusting for sociodemographic factors, and comorbidities [25]. However, disparities in access to health care or potential biases associated with hospital admissions were not considered in many studies, which may influence the rate of COVID-19-related complications and mortality.
Respiratory infections, including previous epidemics of SARS, are known to increase the risk of MACE and mortality [26]. This is particularly well-established for seasonal influenza infections, where vaccination appears to reduce cardiovascular morbidity and mortality by 15% to 20% among high-risk individuals [27][28][29]. Both age and male sex predict adverse outcomes among patients with influenza, and this association remains significant among patients with COVID-19 [28]. However, the exact mechanism for male predominance in the context of COVID-19 infections remains ambiguous. It may be partially explained by sex-related differences in innate and adaptive immunity-links with estrogen receptor signalling [9][10][11]. Further, expression of angiotensin-converting enzyme 2 receptor of SARS-CoV-2 is high in the testes and testosterone, which can cause increased rates of COVID-19 infection among men leading to even higher mortality in critical cases [30,31].
Our findings are also consistent with prior studies that suggested pre-existing CVD was a strong negative prognostic factor [32][33][34][35][36]. Recent studies also showed an excess of cardiovascular complications among COVID-19 patients, such as acute coronary events, myocardial injury, and heart failure [12]. Several systematic reviews and meta-analyses have confirmed that older age, pre-existing CVD, higher troponin T concentration, higher C-reactive protein, and lower albumin are associated with increased mortality among patients hospitalized with COVID-19 [37][38][39][40][41][42]. Indeed, the increased physiological demands imposed by severe infection with COVID-19 may affect people with established CVD more seriously than those without CVD. Poor cardiovascular reserve also unfavourably affects the immune system, potentially predisposing to infection. In our study, although hypertension was significantly associated with mortality in bivariable analysis, it did not emerge as an important risk marker in the adjusted multivariable regression analysis. However, hypertension did significantly increase the risk of MACE in our cohort. A 2020 meta-analysis of 11 studies (all from China) involving 2552 COVID-19 patients reported an OR, 95% CI: of 2.49 (1.98-3.12) for severe disease, and mortality (2.42; 1.51-3.90) in the presence of hypertension [43]. A retrospective cohort study of 1883 COVID-19 patients found two-fold increased risk of COVID-19 mortality in the hypertension vs. non-hypertensive group [44]. An important limitation of the previous literature regarding hypertension and severity of COVID-19 and mortality is the lack of age-adjusted data, and further meta-regression analysis have shown that COVID-19 severity and mortality is seen only in older patients (≥60 years), in whom prevalence of hypertension increases with age.
We found no harmful effects of RAAS inhibitors or NSAIDs. Both classes of drugs received immense attention at the beginning of the COVID-19 pandemic due to their implied role in upregulating the expression of ACE-2, the receptor used by SARS-CoV2 for endocytic internalization [45,46]. Our results are supported by other observational studies of these drug classes and endorse the position statement of the European Society of Cardiology that treatment with ACE inhibitors should not be discontinued in patients with COVID-19. Our study found that HIV status was positively associated with 30-day mortality but not with in-hospital deaths, which suggests a differential immune response like lower likelihood of cytokine storm during acute illness but other susceptibilities coupled with health system factors affecting increased 30-day deaths. Therefore, providing structured and long-term follow-up care for patients with HIV remains an important policy and clinical practice recommendation.
A very low percentage of the patients had received vaccinations at the time of the WHF study being conducted. We are planning to further extend patient recruitment and propose conducting long-term follow-up up to 12 months to study the impact of COVID-19 vaccination, persistence of immunity, and the potential implication of anti-microbial resistance with the severity of COVID-19 illness and adverse outcomes. Further, the long-term follow-up study will address 'long COVID' (i.e., long-term downstream clinical consequences and symptoms of COVID-19) as well as aims to elucidate the potential mechanisms/factors contributing to a sizeable proportion of the sudden cardiac deaths observed in our cohort.

STRENGTHS AND LIMITATIONS
A major strength of the present study was the global cohort with representation across country income classification (LIC, LMIC, UMIC and HIC). Other important strengths of this study were the geographical coverage (all continents included), consecutive enrolment of hospitalized COVID-19 patients and very high follow-up rate at 30-days (>98%), access to complete medical histories of >95% hospital-screened RT-PCR-diagnosed COVID-19 patients. This allowed us to investigate the importance of pre-existing CVD comorbidities on the absolute risk of severe COVID-19 outcomes in a global cohort of hospitalized COVID-19 patients controlling for age, sex, ethnicity, and other clinical characteristics using multivariate regression analysis.
There are some limitations to our study. As by nature of observational study designs, the results prevented us from making definite inferences regarding causality. It remains uncertain whether the COVID-19 infection is directly involved in the pathogenesis of or acts as a trigger for, cardiovascular events in people with elevated risk. Given the diverse patient population and ethnic groups and multiple comparisons, the univariable and bivariable associations we observed should be interpreted cautiously. Our results apply only to hospitalized patients, and we relied on the routine practice followed at the participating hospital sites for treatment and management of COVID-19, so we did not have comprehensive biomarkers and imaging measurements for all patients recruited in this study.

CONCLUSIONS
This WHF COVID-19 study demonstrated that patients hospitalized with COVID-19 were predominantly middle-aged men with a high prevalence of cardiovascular risk factors such as diabetes and hypertension, and increased mortality at 30 days. The key predictors of mortality were older age (≥ 60 years), male sex, Asian/Hispanic/Black ethnicity, pre-existing coronary heart disease, diabetes, renal disease, severe infection of COVID-19 requiring ICU admission, oxygen therapy and higher respiratory rates, but no significant association was found with hypertension or RAAS inhibitors. Further, the COVID-19 patients from LIC, LMIC, UMIC's vs. HIC also experienced greater in-hospital mortality, but LIC patients had lower post-discharge 30-day deaths. This study uniquely provides robust evidence on COVID-19 outcomes from LIC, LMIC, and UMICs, which have sparse data on COVID-19 and guides future health care planning for the pandemic globally.

ADDITIONAL FILES
The Additional files for this article can be found as follows: