Increased Mortality in Patients With Acutely Decompensated Heart Failure During the COVID-19 Pandemic in Toronto, Canada

Background Coronavirus disease 2019 (COVID-19) has resulted in a reduction in patients seeking timely consultation for illnesses that are not related to COVID-19. Previously, we reported a decline in the number of emergency department (ED) visits and hospitalizations for acute decompensated heart failure (ADHF) during the 2020 COVID-19 pandemic vs that in 2019. We aimed to determine the consequences of these early trends on ADHF-patient morbidity and mortality. Methods We compared consecutive patients presenting with ADHF to 3 academic medical centres in Toronto, Canada from March 1-September 28, 2020, vs those from the same time period in 2019. We used multivariate logistic regression models to evaluate whether the odds of hospitalization after presenting to the ED, recurrent ED visits or readmission within 30 days, and in-hospital all-cause mortality differed by timeframe. Results We observed that, during the COVID-19 pandemic, a lower total number of patients presented to the hospital with ADHF, vs that in 2019. Despite this difference, the probability of being admitted to the hospital did not differ for patients seen in 2020 vs 2019. Among ADHF patients admitted to the hospital, however, we observed a significantly higher proportion being admitted to the intensive care unit, and a relative 66% increase in in-hospital mortality during the 2020 COVID-19 era, compared to that in 2019. Conclusions Our findings suggest that improved messaging may be needed for patients living with chronic health conditions, including HF, during the pandemic, to educate and encourage them to present to hospital services when in need.

, the world continues to battle the pandemic and a myriad of other collateral economic, societal, and health crises. 1 Public health stay-at-home orders, physical distancing measures, growing amounts of cases, hospitalizations, ventilator-dependence, and deaths, as well as the subsequent reconfiguration of healthcare delivery have all had unintended consequences, including amplified patient fear and anxiety. 2 This situation has led to a reduction in patients seeking timely consultations for illnesses that are not related to  In 2020, during the early stages of the pandemic, we described the volume and characteristics of patients with acute decompensated heart failure (ADHF) presenting to the University Health Network (UHN), in Toronto, Canada, compared with those of a time-matched 2019 cohort. 4 In this early study, we reported a decline in ADHF-related emergency department (ED) visits and hospitalizations. Unclear at the time was whether this decline reflected a tendency of patients to simply avoid the hospital setting, or whether changes in patient behavioural patterns combined with effective medical management via other platforms had circumvented the need for presentation at the hospital or for admission. Understanding the impact of the observed changes is critical, as it would highlight the need for either improved public health messaging, to reassure and educate patients on hospital safety measures and the importance of seeking timely medical attention, or adoption of newer patient-management strategies within routine practice. Therefore, to determine the consequences of the observed early trends on ADHF-patient morbidity and mortality, we sought to compare ADHF-related admission rates, the number of recurrent ED visits and readmissions within 30 days, and mortality, in consecutive patients presenting to 3 academic medical centres in Toronto, Canada from March 1 to September 28, 2020, vs those in the same time period in 2019.

Study design
In this multicentre cohort observational study, we compared consecutive patients presenting with ADHF to the UHN (Toronto General Hospital and Toronto Western Hospital), Mount Sinai Hospital (MSH), and Sunnybrook Health Sciences Centre (SHSC), in Toronto, Canada, from March 1 to September 28, 2020, and from March 1 to September 28, 2019 (time-matched control cohort). Public health messaging, including stay-at-home messaging, as well as adjustments made at cardiac clinics were initiated as of March 1, 2020, and this timeframe also contains the sentinel public lockdown date in mid-March within Ontario, Canada of these early trends on ADHF-patient morbidity and mortality. Methods: We compared consecutive patients presenting with ADHF to 3 academic medical centres in Toronto, Canada from March 1-September 28, 2020, vs those from the same time period in 2019. We used multivariate logistic regression models to evaluate whether the odds of hospitalization after presenting to the ED, recurrent ED visits or readmission within 30 days, and in-hospital all-cause mortality differed by timeframe. Results: We observed that, during the COVID-19 pandemic, a lower total number of patients presented to the hospital with ADHF, vs that in 2019. Despite this difference, the probability of being admitted to the hospital did not differ for patients seen in 2020 vs 2019. Among ADHF patients admitted to the hospital, however, we observed a significantly higher proportion being admitted to the intensive care unit, and a relative 66% increase in in-hospital mortality during the 2020 COVID-19 era, compared to that in 2019. Conclusions: Our findings suggest that improved messaging may be needed for patients living with chronic health conditions, including HF, during the pandemic, to educate and encourage them to present to hospital services when in need.

Study setting
At each of these 3 centres, rapid virtualization of cardiac care began in March 2020 (March 9, March 16, and March 27 at UHN, MSH, and SHSC, respectively), at which point in-person clinic appointments were rescheduled or replaced by videoconferencing visits or telephone calls. These changes in structure of the heart function clinics resulted in an initial drop in the total number of appointments, followed by a subsequent return to the usual visit volume to approximately 250 visits per month at UHN (97% virtual), 235 visits per month at MSH (35% virtual), and 115 visits per month at SHSC (34% virtual) by September 2020. Additionally, at UHN, patients were invited to enroll in the existing Medly Program, a mobile phone-based telemonitoring program designed to provide remote clinical support for patients with heart failure (HF). 5 After 2 months of rapid patient onboarding in March and April 2020 (63 and 59 patients, respectively), the Medly Program had an average net growth of 30 patients per month from May to September 2020 (compared to 12 patients per month in the 12 months leading up to the COVID-19 pandemic). Telemonitoring was not available at MSH or SHSC.

Study population
We screened adult patients (aged ! 18 years) who had presented to the ED with a triage diagnosis of "shortness of breath" and/or "leg swelling/edema," and patients who were directly admitted to the hospital from an ambulatory clinic with an admitting diagnosis of ADHF, and including, of these patients, those who had ADHF. To qualify as having ADHF, all patients were required to have clinical records recording a diagnosis of ADHF based on clinical symptoms consistent with ADHF and either of the following: (i) a brain natriuretic peptide (BNP) level of ! 100 pg/mL or an N-terminal -proBNP (NT-proBNP) level of ! 300 pg/mL at the time of presentation; or (ii) a left ventricular ejection fraction (LVEF), as measured by echocardiography within the last year or during the index hospitalization, of 50%. Although the LVEF threshold of 50% was used as one inclusion criterion for identifying patients with ADHF, we did include patients with any LVEF if they satisfied one of the other inclusion criteria. Patients were excluded if they had an implantable ventricular assist device or a cardiac transplant, as they represent a unique cohort of patients for whom HF is managed by a specialized multidisciplinary team. Patients presenting with ADHF secondary to an acute coronary syndrome or secondary to severe pulmonary disease were also excluded. Informed consent was obtained for patients who presented after the study initiation (27 patients did not consent to participate and were thus excluded from the study sample); consent was waived for patients who presented to the hospital before study initiation. All study data were collected retrospectively.

Data collection
We extracted clinical and laboratory data from electronic medical records, including the following: demographics (age, sex, residence location [forward sortation area of postal codes]; engagement with an HF program [ie, patient previously followed by an HF clinic]; full code status; comorbidities (ie, atrial fibrillation/flutter, diabetes mellitus, dyslipidemia, hypertension, smoking, chronic obstructive pulmonary disorder, liver cirrhosis, cancer, use of dialysis, dementia, peripheral vascular disease, cerebrovascular accident); HF characteristics and history (etiology of cardiomyopathy, time of initial HF diagnosis, pre-presentation New York Heart Association functional class, LVEF, HF medications at time of presentation, use of an implantable cardioverter defibrillator and/or cardiac resynchronization therapy); physical examination parameters (body mass index, heart rate, systolic blood pressure, respiratory rate); and laboratory data (hemoglobin, creatinine, sodium, and BNP/NT-proBNP levels). COVID-19 status was determined in the 2020 patient cohort according to  institutional guidelines, based on nasopharyngeal swab testing and application of COVID-19 real-time reverse transcriptasepolymerase chain reaction.

Outcomes
We collected information on the number of visits to the ED for patients with ADHF, along with admission rates for ADHF. Among patients admitted for ADHF, we evaluated the need for intensive care throughout the index admission, inhospital all-cause mortality rates, and 30-day readmission or ED visit rate.

Statistical analysis
Continuous variables are summarized by means and standard deviations, or medians and interquartile ranges (IQRs). Dichotomous and categorical variables are summarized as frequencies and proportions. We compared characteristics of patients seen during the 2020 pandemic to those of the 2019 time-matched cohort using Wilcoxon rank-sum tests or t-tests for continuous variables, and c 2 or Fisher's exact tests for categorical variables. We used a generalized linear mixed model with random intercepts corrected for time as a weekly variable to estimate the average percent change in ED visits for ADHF, and hospitalizations for ADHF from the ED and from an HF clinic during the pandemic and previous year.
We used univariate and multivariate logistic regression models to evaluate whether the odds of hospitalization after presenting to the ED, 30-day readmission or ED visits, and in-hospital mortality differed by timeframe. We selected covariates based on their clinical importance and used automated forward (based on a P value < 0.15 on univariable analysis) and backward (based on a P value < 0.15 after entering multivariable analysis) for final variable selection to generate the multivariable regression models. We forced era, age, sex, and engagement with an HF program into the model. For in-hospital mortality, to account for patients with multiple admissions within a short timeframe, admissions that were within 7 days of the previous discharge date were considered to be one continuous admission, in which mortality occurring during the continuous hospitalization was considered to be in-hospital mortality. We also conducted simplified logistic regression models that included era and centre, to evaluate whether the probability of each of these outcomes in 2020 vs 2019 differed by centre. For outcomes in which a significant interaction was observed, we calculated pairwise estimates to determine the difference across eras at each centre for each outcome.
We considered a P value < 0.05 to be statistically significant. All statistical analyses were performed using Stata 16 (StataCorp, College Station, TX).

Results
Overall, we recorded 1884 patients with ADHF who presented to the ED or were admitted directly from a clinic from March 1 to September 28, 2020, and the number of patients from the same timeframe in 2019. Of these patients, 804 presented during the 2020 COVID-19 pandemic, whereas 1080 presented during the 2019 study period (Fig. 1). The mean age of patients was 76 AE 15 years; 872 patients (46%) were female, and 391 patients (21%) were being followed by an HF program at the time of presentation to the hospital. Of the patients presenting in 2020, a total of 11 (1.4%) were identified as being positive for COVID-19. Supplemental Table S1 presents characteristics of these patients, overall and stratified by era. In general, patient characteristics were similar among those who presented to the hospital in 2020 vs in 2019.

Characteristics of patients admitted to the hospital during the 2020 COVID-19 pandemic vs in 2019
Of the 1884 patients included in our study, 1584 patients were admitted to the hospital either from the ED or directly from a clinic, 685 during the 2020 COVID-19 pandemic, and 899 during the same timeframe in 2019. The median length of stay did not differ between patients admitted during the COVID-19 era vs those admitted in the previous year (7 days (IQR 4-13) vs 7 days (IQR 4-12), P ¼ 0.27). Among this group of hospitalized patients (including those admitted from the ED and from a clinic in 2019 and 2020), the mean age was 76 AE 15 years; 748 patients (47%) were female, and 325 patients (21%) were being followed by an HF program at the time of admission. In general, patient characteristics were similar between the 2 eras, apart from the fact that more of the patients admitted in 2019 had a history of atrial fibrillation (Table 1). Of the patients admitted in 2020, a total of 10 (1.5%) were identified as being positive for COVID-19.

Absolute number of ED visits, hospital admissions, and outcomes during the 2020 COVID-19 pandemic vs in 2019
Between March 1 and September 28, 2020, we recorded a numerically lower total number of ED visits for ADHF, compared with the number of ED visits from the same timeframe in 2019 (729 vs 991, respectively; Table 2). Among patients presenting to the ED, the absolute number of admissions from the ED were lower in 2020 compared to that in 2019 (610 and 810, respectively), as was the absolute number of patients admitted directly from a clinicd75 patients in 2020 vs 89 in 2019. Compared to 2019, the average weekly number of ED visits in 2020 decreased by 26% (95% confidence interval [CI] 19% to 32%), the average weekly number of admissions from an ED visit decreased by 25% (95% CI 16% to 32%), and the average weekly number of all admissions (from an ED and directly from a clinic) decreased by 18% (95% CI 12% to 40%; Fig. 2).

Probability of admission after ED presentation and outcomes during hospitalization in the 2020 COVID-19 pandemic vs in 2019
Following a visit to the ED, the probability of being admitted to the hospital did not differ for patients seen in 2020 vs in 2019. Multivariate analyses indicate that the odds of being admitted to the hospital following presentation to the ED were not significantly increased (odds ratio (OR) 1.15, 95% CI 0.88 to 1.49) in 2020 vs the odds in 2019 after adjusting for age, sex, residence location, engagement with an HF program, LVEF, New York Heart Association class, history of dyslipidemia, history of dementia, history of peripheral vascular disease, systolic blood pressure, heart rate, respiratory rate, and hemoglobin and creatinine levels (Tables 3 and 4). In a simplified model including era and centre evaluating the probability of admission after ED presentation (Supplemental Table S2), we identified a significant interaction effect, suggesting that the probability in 2020 vs 2019 differs between UHN and SHSC (OR 1.50, 95% CI 1.08 to 2.1 at UHN; OR 0.83, 95% CI 0.36 to 1.88 at MSH; OR 0.66, 95% CI 0.41 to 1.06 at SHSC, for interaction effect P < 0.02).
Among ADHF patients admitted to the hospital, we observed a significantly higher proportion of patients being admitted to the intensive care unit (ICU) during the 2020 COVID-19 pandemic vs in 2019 (21.8% vs 17.5%, P ¼ 0.03). In-hospital mortality was significantly higher among the patients admitted during the COVID-19 era vs in 2019 (8.9% vs 5.7%, P ¼ 0.01). Multivariable analysis indicates that the odds of in-hospital mortality was 66% higher (OR 1.66, 95% CI 1.08 to 2.53) among patients who were admitted to the hospital in 2020 vs in 2019 (Table 3;  Supplemental Table S3). No differences in in-hospital mortality by era were found across centres (Supplemental Table S2).
Multivariable analysis indicates that after hospital discharge, the odds of 30-day readmission or recurrent ED visit among patients discharged alive was not significantly increased (OR 1.08, 95% CI 0.73 to 1.60) in 2020 vs 2019 (Table 3; Supplemental Table S4).

Outcomes for patients presenting to the hospital with ADHF and COVID-19-positive status
We identified 11 patients (1.4%) who presented to the hospital with ADHF during the 2020 COVID-19 pandemic who had a positive COVID-19 status. Of these patients, 10 presented to the ED, and 1 patient was admitted directly from a clinic. All patients were admitted to the hospital, except for 1 patient who left against medical advice and was alive at 30 days post-ED visit. Among the patients admitted to the hospital, 1 patient was admitted to the ICU (and discharged alive), and 1 patient died in-hospital from COVID-19-related pneumonia.

Summary of findings
Previously, we reported a decline in ADHF-related ED visits and hospitalizations during the early stages of the 2020 COVID-19 pandemic, vs the number in 2019. 4 In the current study, which includes 2 additional study centres, as well as a longer timeframe, we continued to observe a lower total number of patients presenting with ADHF to the ED at 3 tertiary care centres in Toronto, Canada, throughout the COVID-19 pandemic, as compared to a time-matched cohort in 2019. Despite the lower volume of patients presenting to the ED and the lower absolute number of admissions to the hospital for ADHF, the probability of being admitted to the hospital did not differ for patients seen in 2020 vs in 2019 after adjusting for several important covariates. Among ADHF patients admitted to the hospital, however, we observed a significantly higher proportion of patients being admitted to the ICU, as well as a relative 66% increase and an absolute 3% increase in in-hospital mortality during the 2020 COVID-19 era vs in 2019.

Relation to previous work
Throughout the COVID-19 pandemic, healthcare centres around the world have had to restructure their delivery of HF care in unprecedented ways to effectively manage patients while minimizing the risk of COVID-19 transmission. In efforts to reduce the spread of the novel severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) infection, HF centres have transitioned to virtual care platforms, and in some cases, deferred or delayed patient management. 1 Although the consequences of the SARS-CoV-2 infection are profound, the indirect effects of the COVID-19 pandemic, including delay or foregoing of health care for patients living with other chronic conditions, are believed to carry an equal, if not greater burden on subsequent cardiovascular mortality and morbidity. 1,6 Our study findings demonstrate that ADHF presentations to healthcare facilities have decreased during the COVID-19 pandemic, and as a result, these patients may Table 3. Three logistic regression models evaluating the impact of COVID-19 on the probability of admission to the hospital among all presentations to the emergency department (ED), in-hospital mortality among all admissions to the hospital, and 30-day readmission or ED visit among all admissions to the hospital of patients discharged alive, evaluated by comparing patients presenting in 2020 vs in 2019

Outcome
Unadjusted OR (95% CI) P Adjusted* OR (95% CI) P CI, confidence interval; OR, odds ratio. * Multivariate model was adjusted for multiple baseline characteristics, comorbidities, laboratory values, and heart failure medications as described in Table 4 and  Supplemental Tables S3 and S4. be increasing their risk of further HF progression and mortality. These observations parallel our findings from the early stages of the COVID-19 pandemic 4 and demonstrate that the trend in reductions of both ADHF presentations and admissions to the hospital has continued throughout this 7-month study period. Many other studies have reported reductions in ADHF-related ED visits and hospital admissions since the start of the COVID-19 pandemic, [7][8][9][10] and this phenomenon has been observed across other non-COVID-19-related conditions also, 3,[11][12][13] including acute cardiovascular conditions such as stroke, 14,15 and myocardial infarction. 16,17 Although we are not able to confirm the reasons for the observed behaviour in this study, it is likely a consequence of the collective pandemic response and the fear imposed by public health messaging regarding stay-athome orders. 3,6 Among the patients who did present to the hospital in our study, the probability of admission (ie, receiving in-hospital care for HF management) was no different between 2019 and 2020. Moreover, we did not observe any difference in the risk of readmission or recurrent ED visits within 30 days of discharge during the COVID-19 pandemic vs in 2019. This finding may suggest that the resources available to patients and the quality of HF care have not changed, despite the shift in healthcare practices during the COVID-19 pandemic. This notion should be emphasized in public health messaging, as patients living with chronic health conditions, including HF, should be encouraged to seek timely medical care when needed throughout the pandemic. Given the increased rates of ICU admission and in-hospital mortality observed among patients who were hospitalized for ADHF in 2020, compared to those in 2019, our study demonstrates that patients are likely waiting too long at home before presenting to the hospital. This theory is further supported by the fact that ICU resources were not exhausted, and no additional triage strategies were implemented, across the 3 study institutions during this time. Only 1 death in our study was attributed to COVID-19, which makes us wonder if many of the other deaths we observed could have been avoidable given stronger public health messaging and earlier presentation to the hospital. 13 Other centres around the world have demonstrated similar increases in in-hospital mortality rates since the start of the COVID-19 pandemic. We observed a 66% increase in the odds of in-hospital mortality among patients admitted to the hospital during the COVID-19 era vs in 2019. Similarly, Germany has reported a 27% increase in the relative risk of inhospital mortality for March-April 2020 vs 2 control periods, 6 and another study from the United Kingdom demonstrated a 2-fold increase in in-hospital mortality from January-June 2020, compared to that in the same timeframe in 2019. 8 In Canada, similar trends have been reported by the Canadian Cardiovascular Society COVID-19 Challenge for Canada Initiative (CCS-C3I), which has demonstrated an w30% decrease in HF admissions, and increased in-hospital mortality rates from 9.5% (April 2019) to 10.5% (April 2020), for patients with HF in Canada during this first wave of the COVID-19 pandemic. 18

Clinical implications
Altogether, these findings have important implications for the clinical care of patients living with HF and other chronic conditions. Given the significantly worse patient outcomes observed during the COVID-19 era compared to those in previous time-matched control periods, critical to the approach taken in the event of other emergent situations, be they COVID-19-related or not, is the development and adoption of stronger patient-management strategies. In Ontario, approximately 21,000 adults are admitted to the hospital with HF each year, 19 and of these patients, approximately 11% will die in-hospital (2300 patients per year). 20 Applying the findings on increased in-hospital mortality risk from these 3 institutions to other institutions across Ontario, the results from our study suggest that the changes in patient care seen as a result of the public measures and the subsequent fear developed during this pandemic may have led to approximately 1500 additional HF-related deaths in Ontario in 2020. This simple mathematical impact analysis does not consider the reduction in the number of hospital admissions and the possible unmeasurable increase in HF-related deaths occurring outside the hospital setting during the COVID-19 era (ie, HF patients dying at home). Although these numbers are specific to the HF population, similar suboptimal care of patients living with other diseases also has been reported during the COVID-19 pandemic. 14-17

Limitations
Our study has limitations. The COVID-19 pandemic has affected individual centres differently. This heterogeneity, as demonstrated by the interaction effect we observed regarding the probability of admission after ED presentation across the 3 study centres, may increase the chances of our study results being representative of other academic centres in Ontario. Our results, however, may not be generalizable to nonacademic medical centres. The number of patients presenting to the ED could have been underestimated, as patients who presented with triage diagnosis codes other than "shortness of breath" or "leg swelling/edema" were not screened for inclusion in this study. However, such underestimation would have affected both the 2020 and 2019 study cohorts, thereby imposing minimal bias on the study findings. An underestimation of the number of events reported is also possible, namely readmissions and deaths, as some patients might have presented to a different hospital for their follow-up care, or died at home. Moreover, patients who did not present to the hospital at all could not be captured in this study, possibly leading to an underestimation of ADHF-related mortality during the COVID-19 pandemic if these patients died outside of the hospital setting. Another possibility, which may even be likely, is that patients who did not seek medical attention for their ADHF were at an increased risk of mortality, compared with those who presented to the hospital during the COVID-19 pandemic. We also did not have access to disease-specific outcome data.
The findings presented in this article also may be influenced by the virtual care programs established at each of these 3 institutions. Although we were not able to explore this possibility in the current study, the observed reductions in ED visits and admissions during the COVID-19 pandemic could have been influenced by an increase in virtual appointments since the start of the pandemic. Also possible is that the virtual care system did not allow for adequate recognition of patients at higher risk, thereby contributing to the increased in-hospital mortality rates observed in this study. Perhaps HF patients with a higher comorbidity burden and more pronounced cardiac dysfunction should, at a minimum, be required to have a scheduled face-to-face interaction periodically, as opposed to treatment provided primarily via a virtual care system. A recent study from the US, however, has demonstrated that the telehealth model for managing outpatients with HF during the COVID-19 pandemic was safe and effective and did not result in increases in acute HF care or mortality. 21

Conclusions
During the COVID-19 pandemic, we observed a lower volume of patients presenting to the ED with ADHF, and a lower absolute number of ADHF admissions, as compared to those in a time-matched cohort in 2019. Despite these trends, the probability of being admitted to the hospital did not differ, but the risk of in-hospital mortality was significantly higher during the 2020 COVID-19 era vs in 2019. Our findings suggest that improved messaging may be needed for patients living with chronic health conditions, including HF, during the pandemic, to educate and encourage them to present to hospital services when in need.

Funding Sources
The authors have no funding sources to declare.

Disclosures
Dr Alba was awarded the Clinician-Scientist Award offered by the Heart and Stoke Foundation of Canada. The authors have no conflicts of interest to disclose.