Cardiac Dysfunction in Critically Ill patients with COVID-19 – A Multicentre Observational Study

Introduction: The importance of cardiac dysfunction in critically ill patients with COVID-19 is not well studied. The aim of the study was to assess the incidence, clinical risk factors, and prognosis of cardiac dysfunction in critical illness caused by COVID-19, and to evaluate if cardiac biomarkers can detect this condition. Methods: This was a multicentre observational study performed in ve intensive care units (ICUs) in Sweden. Patients admitted to participating ICU with COVID-19 were examined with echocardiography within 72 hours from admission to the ICU and again after four to seven days. Cardiac biomarkers and clinical data were collected at the time of echocardiography. Cardiac dysfunction was dened as either left ventricular (LV) dysfunction (having an ejection fraction < 50% and/or regional hypokinesia) or right ventricular (RV) dysfunction (having a tricuspid annular plane systolic excursion (TAPSE) < 17mm or a moderate/severe RV dysfunction assessed visually). Results: We included 132 patients of whom 94 (71%) were included prospectively. The vast majority were intubated (n=127). At the time of admission to ICU, 35 (27%) patients had cardiac dysfunction and 7 patients (5%) had cardiac dysfunction detected later in the ICU-period. LV dysfunction was found in 18 patients and RV dysfunction in 17 patients, 7 patients had both RV and LV dysfunction. Noradrenaline > 0.20µg/kg/min was the only clinical variable associated with a higher risk of cardiac dysfunction. RV dysfunction was associated with an increased risk of death in a risk-adjusted model (OR 3.98, p = 0.013). Troponin and N-terminal pro b-type natriuretic peptide (NTproBNP) had moderate values in detecting cardiac dysfunction (AUC 0.729 and AUC 0.744, respectively). A combination of troponin < 1.44 times the upper reference limit and NTproBNP < 857ng/L had 85% probability of excluding cardiac dysfunction. Conclusions: Cardiac dysfunction is common in critically ill patients with COVID-19. Although not easily detected with clinical variables, cardiac biomarkers might be helpful.

Results: We included 132 patients of whom 94 (71%) were included prospectively. The vast majority were intubated (n=127). At the time of admission to ICU, 35 (27%) patients had cardiac dysfunction and 7 patients (5%) had cardiac dysfunction detected later in the ICU-period. LV dysfunction was found in 18 patients and RV dysfunction in 17 patients, 7 patients had both RV and LV dysfunction. Noradrenaline > 0.20µg/kg/min was the only clinical variable associated with a higher risk of cardiac dysfunction. RV dysfunction was associated with an increased risk of death in a risk-adjusted model (OR 3.98, p = 0.013). Troponin and N-terminal pro b-type natriuretic peptide (NTproBNP) had moderate values in detecting cardiac dysfunction (AUC 0.729 and AUC 0.744, respectively). A combination of troponin < 1.44 times the upper reference limit and NTproBNP < 857ng/L had 85% probability of excluding cardiac dysfunction.
Conclusions: Cardiac dysfunction is common in critically ill patients with COVID-19. Although not easily detected with clinical variables, cardiac biomarkers might be helpful. RV dysfunction is associated with an increased risk of death, these patients might bene t from further investigation or treatments.

Background
Coronavirus disease 2019 (COVID-19) is caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). As of 21 December 2020, more than 70 million cases have been diagnosed worldwide and 1.6 million people have died from the disease [1]. The clinical characteristics of COVID-19 range from asymptomatic to critical infection (respiratory failure, shock, or multiorgan dysfunction), affecting around 5% of cases [2,3]. Acute respiratory distress syndrome (ARDS) is the main complication. Other major complications in patients requiring intensive care and mechanical ventilation are cardiac injury and arrythmias [4].
Myocardial injury with elevated troponin or creatine kinase is reported in 12-38% of hospitalized patients with COVID-19 [5,6] and in 59% of patients treated in ICUs [7]. Cardiac complications, on a case report basis, include myocarditis [8][9][10], COVID-19 associated Takotsubo cardiomyopathy [11,12], and myocardial infarction [12]. Several studies have reported abnormalities in left and right ventricular function in a large proportion of patients with COVID-19 undergoing echocardiography [12][13][14][15]. Studies have also shown that left and right ventricular dysfunction are independently correlated to increased mortality in patients with COVID-19 [14,16]. Few prospective studies of critically ill patients with COVID- 19 undergoing echocardiography exist and are limited in scope [13,17,18]. Thus, the clinical importance of cardiac dysfunction in critical illness caused by COVID- 19 is not yet established. In addition, studies linking echocardiographic ndings with clinical data, cardiac biomarkers, and mortality are lacking.
The aim of the present study was to describe the incidence, risk factors, and mortality related to cardiac dysfunction in critical illness caused by COVID-19, and to assess the value of biomarkers for the detection of this condition.

Methods
This was a multicentre observational study. It was approved by the Swedish Ethics Review Authority Sweden (approval number 036 − 18, 2020 − 01684 and 2020-03815) and registered in the international database at Clinicaltrials.gov (registration number NCT04524234).

Study design and inclusion
The study was performed at ve ICUs in Sweden from 20 April to 30 August 2020 (details of each participating ICU, the ICU settings, and principles of patient treatment are described under Supplemental data). Most patients were included prospectively (n = 94, 71%), however 30 patients were included retrospectively from one of the ve study sites. In addition, nine patients who had followed the study protocol before the ethics review was approved were included retrospectively from the other four centres.
All patients ≥ 18 years of age who were admitted to a participating ICU with veri ed COVID-19 were eligible for inclusion. Entering patients had an echocardiogram performed within 72 hours from admission to the ICU. Echocardiography was repeated in four to seven days as study resources became available or if there was evidence of clinical deterioration. Cardiac biomarkers, routine laboratory tests, and clinical data were recorded at each exam. Consent to be included in the study was obtained from the patient or patient's next of kin whenever possible. Retrospective inclusion of patients who had followed the study protocol was approved by the ethics review authority, and for these patients consent was waived.
Echocardiography, de nitions, recordings, and measurements Echocardiography was performed according to standard protocol (see Supplemental data) for assessment of left ventricular (LV) or right ventricular (RV) dysfunction and estimation of pulmonary artery pressure (PAP). Tissue Doppler was not available in all centres and this modality was not further investigated. Cardiac dysfunction was de ned as having either LV or RV dysfunction. LV dysfunction was de ned as ejection fraction < 50% with or without regional hypokinesia assessed with eyeballing. RV dysfunction was de ned as having a tricuspid annular plane systolic excursion (TAPSE) < 17 mm or a moderately or severely depressed RV dysfunction qualitatively judged visually. Elevated PAP was de ned as jet velocity of a tricuspid regurgitation > 2.9 m/s (TR Vmax), or, in patients with no tricuspid regurgitation present, having indirect signs of elevated PAP such as RV dilation and a pulmonary acceleration time < 100 ms [19,20]. Velocity Time Integral (VTI), area of LV out ow tract (LVOT) with calculation of stroke volumes, and cardiac output were assessed, wherever image quality was acceptable. Examinations were recorded and, reviewed by a physician certi ed in echocardiography, if performed by a non-certi ed intensivist.
At the time of echocardiography, blood samples for measurement of the cardiac biomarkers troponin and N-terminal pro b-type natriuretic peptide (NTproBNP) were obtained and clinical data was recorded. Four sites used highly sensitive troponin T and one site used troponin I. The levels are presented as times above upper limit of normal. The following clinical data was recorded on admission: age, sex, medical history, and Simpli ed Acute Physiology Score 3 (SAPS 3) [21]. Along with each echocardiography we also recorded systolic blood pressure, mean arterial pressure, diastolic blood pressure, noradrenaline dosage, lactate levels, respiratory settings, and PaO2/FiO2-ratio. Routine laboratory tests, including Creactive protein (CRP), leukocyte count, creatinine levels, and d-dimer, were obtained on the same day as echocardiography. Mortality at 30 days and death occurring within 90 days of admission were registered.

Pre-de ned outcomes
The pre-de ned primary outcome was 30-day mortality in patients with, versus those without, cardiac dysfunction on admission. Sub-group analysis was performed for LV and RV dysfunction. Pre-de ned secondary outcomes were a) incidence of LV and RV dysfunction within 72 hours of admission and during ICU stay; b) clinical risk factors associated with cardiac dysfunction; and c) levels of the cardiac biomarkers troponin and NTproBNP in patients with, as opposed to without cardiac dysfunction. Other analyses were not pre-de ned at the initiation of the study and are considered exploratory.

Statistics
A statistical analysis plan was written before the analyses were performed. Continuous variables were checked for normality. Normally distributed variables are presented as mean ± standard deviation; nonnormally distributed variables are presented as median and interquartile range (IQR). Student´s T-test was used to compare means of normally distributed variables and the Mann-Whitney U test was used for comparison of distributions of non-normally distributed variables. Fisher's exact test was used for comparison of binary outcomes between two groups. A generalized estimating equation was used for identi cation of clinical variables associated with an increased risk of experiencing cardiac dysfunction at any time in the ICU. Logistic regression was used to evaluate mortality at 30 days between patients with or without cardiac dysfunction in a non-adjusted and a risk-adjusted analysis (primary outcome). Kaplan-Meier methodology with a log rank test were used to compare incidences over time. Receiver operating characteristics (ROC) analyses were used for estimation of sensitivity and speci city of troponin and NTproBNP for detection of cardiac dysfunction. IBM SPSS Statistics Version 26 were used for the statistical analyses.
No power analysis was performed, as little data of the study population was available when the study was initiated but we aimed for an inclusion of at least 100 patients.

Results
Out of 344 patients admitted to the participating ICUs during the study period, 137 were included in the study. Inclusion failure was mainly because no investigator involved in the study was available for an inclusion or echocardiographic examination. In a sensitivity analysis regarding SAPS score and age, the study population was representative of each ICU´s patient population [22]. Two patients were excluded due to their echocardiography being performed > 72 hours from admission. An additional three patients were excluded as a result of poor echocardiographic image quality. Thus, a total of 132 patients included in the nal analysis ( Fig. 1).

Study population
The median age of the cohort was 63 years (IQR 53-70); 34 (26%) were women. Median body mass index (BMI) was 30 (IQR [25][26][27][28][29][30][31][32][33]. Seventeen patients (13%) had a history of cardiac disease and 74 others (56%) had a risk factor for cardiovascular disease. A history of coronary artery disease, valvular disease, or hypertension was more common in patients with cardiac dysfunction. The mean SAPS score was 52 ± 9 points (Table 1). A total of 127 patients (96%) were intubated during their ICU stay. Patients were given low-molecular weight heparine (LMWH) in doses corresponding to 0.77 (IQR 0.61-0.95) times the treatment dose for acute thrombotic disease. Patients with LV dysfunction had a larger LV diameter and lower ejection fraction, stroke volume, and cardiac index compared to patients with normal LV function (  Clinical characteristics and routine laboratory data Clinical characteristics did not differ between those patients with and those without cardiac dysfunction at the time of the rst echocardiography, except that patients with cardiac dysfunction more often had a high dose of noradrenaline (> 0.20 µg/kg/min). In a generalized regression analysis, taking repeated examinations into consideration, a dose of noradrenaline > 0.20 µg/kg/min was the only clinical variable associated with an increased risk of cardiac dysfunction (see Supplemental data). Levels of d-dimer were slightly higher (2.0 mg/L vs 1.6 mg/L, p = 0.006) in patients with cardiac dysfunction, but there were no differences in other routine laboratory data (Table 3). PCV, n (%) 9 (9) 8 (23) APRV, n (%) 9 (9) 4 (11) High ow oxygen, n (%) 9  (Table 4). There was also an increased risk of death during the rst 90 days from admission in patients with RV dysfunction or elevated PAP (Fig. 2). NTproBNP > 857 ng/l. Using a combination of troponin > 1.44 times the upper reference limit and NTproBNP > 857 ng/l had a positive predictive value (PPV) of 81% for detection of cardiac dysfunction. Troponin < 1.44 times the upper reference limit and NTproBNP < 857 ng/l had 85% probability of excluding cardiac dysfunction. (Table 5). Cardiac biomarkers obtained from echocardiography at rst presentation of cardiac dysfunction. In patients with normal echoes, cardiac biomarkers used were also obtained at rst echocardiography. AUC = area under curve, PPV = positive predictive value, NPV = negative predictive value, NTproBNP = N-terminal pro b-type natriuretic peptide

Discussion
The main ndings of this study are that (1) cardiac dysfunction commonly occurs in critical illness caused by COVID-19; (2) RV dysfunction, de ned as a TAPSE < 17 mm or moderately to severely depressed function assessed visually, was associated with an increased risk of death; (3) cardiac dysfunction is not easily recognized by clinical variables; and that (4) cardiac biomarkers have a moderate value in detecting cardiac dysfunction in patients critically ill with COVID-19.
We found that 32% of the patients had LV or RV dysfunction at some time during their stay in the ICU. Cardiac dysfunction was more common at the time of admission to the ICU than later in the ICU-period.
The incidence of cardiac dysfunction was in line with other studies of patients hospitalized, whose COVID-19 disease severities ranged from mild to severe, and where LV dysfunction was found in 10 to 42% and RV dysfunction in 14 to 39% [15,23,24]. One reason for the relatively low incidence of RV dysfunction may have been the liberal use of thromboprophylaxis that was introduced early in the study period [25], leading to less pulmonary embolism than in many earlier studies [15,26]. We did not assess patients with tissue doppler or strain analysis, modalities that are more sensitive for the detection of ventricular dysfunction, which could explain a comparative low incidence of cardiac dysfunction in our study.
We found a number of different types of LV dysfunction. Two patients had suspected COVID-19 myocarditis; the diagnosis was veri ed in one patient. Knowledge of COVID-19 myocarditis is still very limited. Patients are reported to present in various ways and they may have either global or regional hypokinesia, with or without preserved ejection fraction, and their recovery time may be brief or prolonged [27,28]. There are no strict criteria for COVID-19 myocarditis. In the present study we can neither con rm nor exclude additional cases within our cohort. Five patients presented with a clinical presentation of the Takotsubo syndrome. Severe respiratory distress is an established trigger of Takotsubo, and the incidence we found is in agreement with other studies of Takotsubo syndrome in critically ill patients [29]. Moreover, Takotsubo has also been reported in several case studies of patients with COVID-19 [11,12]. One patient was diagnosed with PIMS-TS, a condition associated with cardiac dysfunction in COVID-19 [30]. Other plausible causes of LV dysfunction are secondary effects due to hypoxia, hypotension, or a toxic effect due to the in ammatory state [31]. RV dysfunction was seen in 18% of participants, and there was a close correlation between elevated PAP and RV dysfunction. Thus, RV dysfunction in most cases is probably attributable to an increased RV afterload. Elevated PAP was common in our study population, having been seen in nearly one-third of all patients. Both pulmonary embolism and pulmonary microangiopathy with microthrombosis are common in severe cases of COVID-19 and are known causes of elevated PAP [32,33]. Pulmonary hypoxia with hypoxic vasoconstriction was another likely cause of elevated PAP among our study patients [34]. Increased right chamber afterload due to mechanical ventilation is of course a common cause of elevated PAP in ICU patients in general.
Cardiac dysfunction was not associated with an increased risk of death in our study. However, patients with cardiac dysfunction had a more complex course of disease with a greater need for renal replacement therapy and less ICU-free days. It is unclear if this is a causal relationship, or if cardiac dysfunction is a marker for more severe disease. In a sub-group analysis, RV dysfunction and elevated PAP were independently associated with an increased risk of death. This is consistent with ndings from other studies of patients with ARDS and hospitalized patients with COVID-19 [13,14]. Regrettably, the main cause of death was not registered in the present study, and patients were not systematically assessed for pulmonary embolism. It is, therefore, unclear if RV dysfunction and elevated PAP are only markers of more severe pulmonary disease, with a subsequent risk of pulmonary collapse and respiratory death, or if there are other direct causes as well. It seems reasonable to assume that patients with RV dysfunction, elevated PAP, or both will bene t from further investigation by computer tomography for diagnosis of pulmonary embolism, worsening of ARDS, COVID-19 typical in ltrates, or secondary bacterial infection. Furthermore, pulmonary vasodilators could be tried to see if there is improvement in RV function or decreased PAP [35]. In severe cases of COVID-19, RV dysfunction could be supportive in a decision to prepare for, or initiate, extracorporeal membrane oxygenation (ECMO).
Cardiac dysfunction was not easily detected by clinical characteristics. Our hypothesis was that cardiac dysfunction would be associated with a more severe clinical picture, such as higher levels of oxygen, respiratory support, or elevated lactate levels. The only variable associated with an increased risk of having cardiac dysfunction was a high dose of noradrenaline. It is likely because that critical illness from COVID-19 is such a severe respiratory disease, the contribution of cardiac dysfunction is not detectable by clinical variables in such a patient population. However, high doses of noradrenaline should be considered a marker for a more severe cardiovascular deterioration. In patients requiring noradrenaline > 0.20 µg/kg/min, cardiac dysfunction should be suspected and echocardiography ought to be performed.
The cardiac biomarkers troponin and NTproBNP only showed a moderate ability to detect cardiac dysfunction. However, a combination of troponin at less than 1.44 times its upper reference limit and NTproBNP < 857 ng/l had a negative predictive value of 85% of excluding cardiac dysfunction and might be used as a tool to rule out the need of echocardiography. This could be important in ICUs treating patients with COVID-19, where there is a need to conserve personal protective equipment and minimize the number of individuals interacting with contagious patients. On the other hand, levels above these limits had a predictive value of 81% for detection of cardiac dysfunction, suggesting that echocardiography should be performed.
Our study has some limitations. Inclusion rates were relatively low, as only 40% of the potential study population was included, mostly due to a lack of study resources. Moreover, a number of patients were retrospectively included. Nevertheless, a sensitivity analysis showed the study population was a representative sample of each site´s total population. The main strength of the study is the multi-centre design. Although the sample size is relatively low in absolute numbers, it is the largest study to date with echocardiographic evaluation, cardiac biomarkers, and clinical data carried out among COVID-19 patients in an ICU-setting.

Conclusions
Cardiac dysfunction is common in critical illness caused by COVID-19, but it is not easily detected on the basis of clinical characteristics. Cardiac biomarkers here prove be helpful. 2020-01684 and 2020-03815). Consent to be included in the study was obtained from the patient or patient's next of kin whenever possible. For patients retrospectively included, patients consent was waived.

Consent for publication
No individual data was reported in the study.

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.
The study was funded by the Swedish Heart-Lung foundation (number 20200253 and 2017063).
Authors' contributions JH initation of the study, conceptualization and study design, drafting and revising the manuscript, echocardiographic examinations, data acquisition and interpretation. JBF study design, inclusion of patients, data acquisition, compilation and interpretation, drafting and revising the manuscript. CJ study design, echocardiographic examinations, data acquisition. KD, NN echocardiographic examinations, data acquisition. SM, JC data acquisition. ERW data acquisition and compilation. OC reviewing echocardiographic examinations, data acquisition. AY, study design, substantively revised the manuscript.
MC data acquisition, substantively revised the manuscript. BR study design and conceptualization, statistical analyses, funding. JO initiation of the study, conceptualization and study design, echocardiographic examinations, data acquisition and interpretation, statistical analyses, funding, drafting and revising the manuscript. All authors read and revised the paper. All authors have agreed to be accountable for all aspects of the work. All authors approved the nal version of the manuscript.