Right ventricular dysfunction in acute heart failure from emergency department to discharge: Predictors and clinical implications

a b s t r a c t

Background: Among Acute heart failure inpatients, Right ventricular dysfunction predicts clinical outcomes independent of left ventricular (LV) dysfunction. Prior studies have not accounted for congestion se- verity, show conflicting findings on echocardiography (echo) timing, and excluded emergency department (ED) patients. We describe for the first time the epidemiology, predictors, and outcomes of RVD in AHF starting with earliest ED treatment.

Methods: Point-of-care echo and 10-point Lung ultrasound were obtained in 84 prospectively enrolled AHF patients at two EDs, <=1 h after first intravenous diuresis, vasodilator, and/or positive pressure ventilation (PPV). Echo and LUS were repeated at 24, 72, and 168 h, unless discharged sooner (n = 197 exams). RVD was defined as

<17 mm tricuspid annulus plane systolic excursion (TAPSE), our primary measure. To identify correlates of RVD, a multivariable linear mixed model (LMM) of TAPSE through time was fit. Possible predictors were specified a priori and/or with p <= 0.1 difference between patients with/without RVD. Data were standardized and centered to facilitate comparison of relative strength of association between predictors of TAPSE. survival curves for a 30- day death or AHF readmission primary outcome were assessed for RVD, LUS severity, and LVEF. A multivariable generalized linear mixed model (GLMM) for the outcome was used to adjust RVD for LVEF and LUS. Results: 46% (n = 39) of patients at ED arrival showed RVD by TAPSE (median 18 mm, interquartile range 13-23). 18 variables with p <= 0.1 unadjusted difference with/without RVD, and 12 a priori predictors of RVD were in- cluded in the multivariable LMM model of TAPSE through time (R2 = 0.76). Missed Antihypertensive medication (within 7 days), ED PPV, chronic obstructive pulmonary disease history, LVEF, LUS congestion severity, and right ventricular systolic pressure (RVSP) were the strongest multivariable predictors of RVD, respectively, and the only to reach statistical significance (p < 0.05). 30-day death or AHF readmission was associated with RVD at ED arrival (hazard ratio {HR} 3.31 {95%CI: 1.28-8.53}, p = 0.009), ED to discharge decrease in LUS (HR 0.11

{0.01-0.85}, p < 0.0001 for top quartile ?), but not LVEF (quartile 2 vs. 1 HR 0.78 {0.22-2.68}, 3 vs. 1 HR 0.55

{0.16-1.92}, 4 vs. 1 HR 0.32 {0.09-1.22}, p = 0.30). The area under the receiver operating curve on GLMM for the primary outcome by TAPSE (p = 0.0012), ?LUS (p = 0.0005), and LVEF (p = 0.8347) was 0.807.

Conclusion: In this observational study, RVD was common in AHF, and predicted by congestion on LUS, LVEF, RVSP, and comorbidities from ED arrival through discharge. 30-day death or AHF-rehospitalization was associ- ated with RVD at ED arrival and ?LUS severity, but not LVEF.

(C) 2021

Abbreviations: AHF, acute heart failure; HF, heart failure; ED, emergency department; RV, right ventricle; RVD, right ventricular dysfunction; LV, Left ventricle; PPV, Positive Pressure Ventilation; POC, Point-of-care; TAPSE, tricuspid annulus plane systolic excursion; EF, ejection fraction; US, ultrasound; LUS, lung ultrasound; echo, echocardiography; ACC, American College of Cardiology; ASE, American Society of Echocardiography; BMI, body mass index; COPD, chronic obstructive pulmonary disease; OSA, obstructive Sleep apnea; CPAP, continuous positive airway pressure; PHTN, pulmonary hypertension; LOS, length of stay; GLMM, generalized linear mixed model; LMM, linear mixed model; VIF, variance inflation factor; AUROC, Area Under the Receiver Operating Characteristic Curve; 95%CI, 95% Confidence Interval; RVOT, Right ventricular outflow tract; VTI, velocity time integral.

* Corresponding author at: 720 Eskenazi Avenue, Fifth Third Bank Building 3rd Floor, Indianapolis, IN 46202, USA.

E-mail address: [email protected] (N.E. Harrison).

0735-6757/(C) 2021

  1. Introduction

acute heart failure is a common presenting problem to emer- gency departments (EDs) and is associated with high morbidity and mortality [1]. Half of chronic HF patients have right ventricular dysfunc- tion (RVD) [2-8], which predicts hospitalization, mortality, and other adverse events independent of left ventricular (LV) function, thus RVD may be an important contributing factor to development of AHF. In AHF patients undergoing invasive hemodynamic monitoring, improve- ment in RV function through the course of treatment, but not LV func- tional improvement, independently predicts better outcomes [9]. Inva- sive monitoring to assess RV function in response to treatment is rare, but non-invasive assessment of RV function through ultrasound is feasi- ble for emergency physicians [2,10,11] and other clinicians with cardiac ultrasound experience.

Though a useful tool for assessing RVD, it has been recently sug- gested [12] that the correlation of clinical outcomes with a particular echo measure of RVD may differ early versus late in hospitalization. Prior investigations have mainly included studies of inpatients enrolled no earlier than 12 h after hospital admission [12], long after multiple in- terventions such as diuresis, vasodilators, and positive pressure ventila- tion (PPV) may have been administered in the ED. Consequently, the epidemiologic, physiologic, and clinical implications of RVD at the time of ED treatment for AHF are not understood.

Moreover, data on repeated measures through the treatment time course are lacking. Since prior research has focused on inpatient care, most data have included a technologist-performed echocardiogram without repeat or trend. As such, the hypothesis that RV function and other echo assessments are dynamic through hospitalization is largely based on comparison of separate studies performed at different timepoints, rather than observing changes over time directly in a single population [12]. Point-of-care (POC) echocardiography has been sel- dom utilized in the study of RVD in AHF, but offers the potential for as- sessment both at the earliest stages of ED treatment and repeated mea- surement through downstream care.

Our group previously performed a small retrospective pilot study

[13] which showed a relationship between AHF readmission and RVD defined as Tricuspid annular plane systolic excursion < 17 mm in ED patients. However, this pilot was underpowered to detect more than simple association. Important questions remain addressed, such as how common RVD is in ED patients, whether it is found outside of known prior pulmonary hypertension (PHTN), how RV function changes over time from ED to hospital discharge, which clinical features are shared by patients with vs. without RVD, and whether RVD relates to clinical outcomes independent of LV ejection fraction (EF) and conges- tion severity.

  1. Methods

Right-ventricular Evaluation in Emergency Department AHF (REED- AHF) is a prospective study of point-of-care echo prognostication starting at the earliest phases of ED treatment and continuing through ED or hospital discharge. Patients were enrolled from two academic EDs with 80-100 thousand visit/year volumes, from September 2019- February 2020 and November 2020-March 2021 (interrupted by the COVID-19 Pandemic). The study is approved by the Wayne State Uni- versity (WSU) Institutional Review Board (IRB).

    1. Aims of the current report

Interim results of the a priori primary aims in REED-AHF were pre- sented previously [2,14] with final analysis ongoing. The primary and secondary aims of REED-AHF are described in the Supplement. The present report is a preplanned secondary analysis with three descriptive goals: 1. Describe RVD prevalence at ED arrival in the REED-AHF cohort

2. Identify and describe clinical variables associated with the severity

and trend of RVD through time 3. Describe the association between RVD and clinical outcomes observed in our pilot study [13], adjusting for LV function (by LVEF) and congestion severity (by lung-ultrasound


    1. Screening, inclusion/exclusion, and diagnostic adjudication

Patients age >= 18 were eligible if the treating ED clinician suspected AHF and at least one of three AHF treatments was intended: IV loop di- uretics, intravenous (IV) or sub-lingual (SL) nitroglycerin, and/or PPV. An echocardiogram and 10-point LUS were obtained at the point-of- care on a portable US platform (Vivid Q, GE) by one of four emergency physicians (EPs) or two research echocardiography technologists em- ployed by the WSU Department of Emergency Medicine Division of Re- search. The initial echo and LUS were required <=1 h after the administra- tion of any of the three qualifying AHF treatments. Research techs screened patients 7 days per week between 7 am and 4 am, but enroll- ment was convenience-limited to study sonographer availability to per- form the ED exam within the time limit. 25% were enrolled between 5 pm-8 am and 18% on a weekend. Patients were excluded if the treating ED physician diagnosis changed (e.g. after further testing), or if they had cardiogenic shock, ST-elevation myocardial infarction, or tachyarrhythmia requiring IV Rate control. Written consent was ob- tained from the patient or next-of-kin within 24 h of ED scanning, or else the patient was excluded and data deleted per IRB regulation. A car- diologist expert (AA) reviewed all cases to adjudicate ED AHF diagnosis, with disagreements resolved by a third expert (MF).

    1. Lung ultrasound and echocardiography protocol

10-point LUS was performed with the phased array probe optimized for B-line detection. This included the same 8-point protocol for image acquisition and B-line scoring (0-20 per zone) previously described for the BLUSHED study, an ED-based pilot randomized trial of LUS- guided decongestion in AHF [15,16]. Two additional views of the costophrenic angles to detect Pleural effusions were obtained, adding 10 points per side if present. Total 10-point LUS scores therefore ranged 0-180. The pleural effusion views and scoring were originally a compo- nent of an earlier protocol version in BLUSHED.

ECG-gated point-of-care (POC) echocardiograms included 2D, dopp- ler, M-mode, and speckle-tracking assessments in the parasternal and apical windows of the LV, RV, RV outflow tract, and mitral and Tricuspid valves. LVEF was determined by Simpson’s biplane. Tricuspid annulus plane systolic excursion (TAPSE) was chosen as the measure of interest for RVD based on our prior investigations in ED patients [2,13,14] and other groups’ inpatient investigations [7,12,17].

Ideal cutoff values for TAPSE to define RVD among AHF patients in the ED have not been defined. We chose to validate TAPSE<17 mm, being the RVD cutoff associated with AHF-rehospitalization in our pilot study [13], and the suggested threshold for RVD by American Col- lege of Cardiology(ACC)/American Society of Echocardiography(ASE) guidelines [18].

The LUS and echo protocol was repeated a maximum of three times, at 24, 72, and 168 h from the initial scan, ending when the patient was discharged from the ED or hospital.

    1. Ultrasound quality assessment and interpretation

One investigator made measurements on the POC platform, while a second repeated all measurements in offline analysis software (GE echoPAC). Assessors were blinded to one another and to outcomes at the time of measurement. Interclass correlation (ICC) was calculated. Measures with ICC < 0.7 were overread by an expert testamur and Fel- low of the ASE. For analysis, the original measurements on the POC plat- form were used.

    1. Data collection, clinical outcomes and follow-up

Data were prospectively obtained by interviewing the patient, the treating clinician, and review of the medical record, and stored in RED- Cap (V10.3.3, Vanderbilt University). Post-hospital follow-up by phone was completed 30-45 days after discharge by research assistants blinded to ultrasounds. Two blinded investigators, with a third for dis- agreements, reviewed outcomes through 90 days in the EMR, including a healthcare information exchange compiling information across 4 large healthcare networks in Southeast Michigan. The primary clinical out- come for this investigation was death or AHF-readmission at 30 days. AHF as the cause of readmission was adjudicated by two blinded inves- tigators with a third to break ties.

    1. Variables of interest

We suspected a priori that RV function and ultrasound Image quality would be associated with the following: age, sex, body mass index (BMI), LVEF, RV diameter, E/a ratio, e’ velocity, PPV during scanning, LUS B-line score, and history of chronic obstructive pulmonary disease (COPD), obstructive sleep apnea (OSA), Continuous positive airway pressure , and/or pulmonary hypertension (PHTN). Descriptive

statistics were calculated for the remaining observed variables (listed in Tables 1-3) in patients with/without RVD at each timepoint. Variables with p > 0.1 difference were added to the list of potential correlates of RVD and assessed in multivariable analysis. Multiple imputation (see Supplemental Methods) was used for variables missing at random. Missingness was 0% for TAPSE and clinical outcomes, and < 5% for LUS and LVEF.

    1. Multivariable modeling

We hypothesized patients with worse baseline RV function would have longer length of stay (LOS), confounding the odds of RVD through time. For instance, RVD could appear artificially more common if pa- tients with worse RVD stayed longer and those without RVD were discharged sooner. To assess the patient-specific trend in RVD likelihood through time, a binary generalized linear mixed model (GLMM) was fit for the odds of RVD by timepoint and LOS), with a random intercept for patient and random slope for LOS.

A multivariable linear mixed model (LMM) was fit for the prediction of TAPSE by the suspected predictors as defined above, in addition to time. Patient was modeled as a random intercept, and time as a random slope. Variables were standardized and centered by the method of

Table 1

Past medical history and demographics right ventricular dysfunction

defined as TAPSE <17 mm


No RVD in ED


*p < 0.05

**p < 0.01

{n = 84}

{n = 45}

{n = 39}

***p < 0.001


Age (years)

62 (54-70)

61 (55-70)

64 (54-70)

p = 0.986

Male Sex

60% (50)

62% (28)

56% (22)

p = 0.750

African American Race Body Mass Index

89% (75)

28.2 (23.7-36.4)

89% (40)

31.1 (24.4-36.8)

90% (35)

27.1 (22.1-32.6)

p = 0.880

p = 0.080

Past Medical History Pacemaker or AICD

42% (35)

38% (17)

46% (20)

p = 0.579

Cardiac Ablation

7% (6)

4% (2)

10% (4)

p = 0.544

Atrial Fibrillation or Flutter

32% (27)

27% (12)

39% (15)

p = 0.357

Home Supplemental Oxygen

16% (13)

11% (5)

21% (8)

p = 0.376

Chronic Kidney Disease

51% (33)

47% (21)

56% (22)

p = 0.501


5% (4)

4% (2)

5% (2)

p = 1

Chronic Heart Failure

87% (73)

87% (39)

87% (34)

p = 1

Coronary Artery Disease

57% (48)

47% (21)

69% (27)

p = 0.062


39% (33)

33% (15)

46% (18)

p = 0.329

Myocardial Infarction

49% (41)

44% (20)

54% (21)

p = 0.522

Obstructive Sleep Apnea

29% (24)

27% (12)

31% (12)

p = 0.863

Chronic Obstructive Pulmonary Disease

57% (48)

51% (23)

64% (25)

p = 0.328

Obesity Hypoventilation Syndrome

6% (5)

6% (3)

5% (2)

p = 1

Pulmonary Hypertension

32% (27)

29% (13)

36% (14)

p = 0.651

Diabetes Mellitus

46% (39)

49% (22)

44% (17)

p = 0.790


95% (80)

93% (42)

97% (38)

p = 0.714

Peripheral Vascular Disease

25% (21)

20% (9)

31% (12)

p = 0.377


71% (60)

73% (33)

69% (27)

p = 0.863


35% (29)

33% (15)

36% (14)

p = 0.987

Hepatic Disease

8% (7)

7% (3)

10% (4)

p = 0.843

Tobacco Smoking

79% (66)

78% (35)

80% (31)

p = 1

Alcohol Use

45% (38)

47% (21)

44% (17)

p = 0.950

Other Substance Use

24% (20)

18% (8)

31% (12)

p = 0.280

Active Cancer

1% (1)

2% (1)

0% (0)

p = 1

Pulmonary Embolus or Deep Venous Thrombosis

24% (20)

24% (11)

23% (9)

p = 1

Medications Prescribed

ACE inhibitor or Angiotensin Receptor Blocker

38% (32)

40% (18)

36% (14)

p = 0.872

Angiotensin Neprilysin Inhibitor

16% (13)

13% (6)

18% (7)

p = 0.779

Other Antihypertensive Agent

55% (46)

60% (27)

49% (19)

p = 0.414


1% (1)

0% (0)

3% (1)

p = 0.943

Loop diuretic

63% (53)

62% (28)

64% (25)

p = 1

Beta Blocker

71% (60)

67% (30)

77% (30)

p = 0.426

Antiplatelet Agent

61% (51)

51% (23)

72% (28)

p = 0.087


23% (19)

20% (9)

26% (10)

p = 0.723

% (count) or median (IQR) p-value

Presented as median (interquartile range {IQR}) for continuous variables and %(n) for categorical; comparisons made with the Wilcoxon rank sum and chi-square tests, respectively. AICD = automated Implantable cardioverter-defibrillator, PCI = percutaneous coronary intervention, CABG = coronary artery bypass graft, ACE = angiotensin converting enzyme.

Clinical characteristics

right ventricular dysfunction defined as TAPSE <17 mm % (count) or median (IQR) p-value


No RVD in ED


*p < 0.05

**p < 0.01

{n = 84}

{n = 45}

{n = 39}

***p < 0.001

Emergency Department (ED) Electrocardiogram (ECG) and Vital Signs

Wide QRS on ECG in ED

21% (17)

18% (8)

23% (9)

p = 0.780

ED ECG Rhythm = Atrial Fibrillation or Flutter Diastolic Blood Pressure (mmHg) – Max Systolic Blood Pressure (mmHg) – Max Temperature (C?) – Max

Heart Rate – Max Respiratory Rate – Max SpO2 (%) – Arrival

11% (9)

100 (90-114)

156 (143-179)

36.7 (36.5-36.9)

98 (88-111)

20 (20-26)

97 (95-98)

11% (5)

100 (90-115)

164 (146-190)

36.7 (36.5-36.9)

91 (84-111)

20 (20-26)

96 (95-98)

10% (4)

99 (90-110)

151 (137-161)

36.7 (36.4-36.9)

98 (94-111)

20 (20-26)

98 (96-99)

p = 1

p = 0.654

p = 0.005**

p = 0.457

p = 0.192

p = 0.801

p = 0.014*

Historical and Social Characteristics at Index ED Visit Arrival by EMS

44% (37)

44% (20)

44% (17)

p = 1

Symptoms >1 Week

35% (29)

33% (14)

39% (15)

p = 0.744

Any Change in Medication Within 7 days

32% (17)

47% (21)

15% (6)

p = 0.005**

Discontinued or Missed Antihypertensive (<=7 days)

26% (22)

38% (17)

13% (5)

p = 0.019*

Any Change or Missed Loop Diuretic(<=7 days)

25% (21)

36% (16)

13% (5)

p = 0.032*

Discontinued or Missed GDMT in HFrEF(<=7 days)

17% (14)

20% (9)

13% (5)

p = 0.557

Primary Care Visit in Prior 6 Months

38% (32)

44% (20)

31% (12)

p = 0.288

AHF Hospitalization in Prior 30 Days

26% (22)

22% (10)

31% (12)

p = 0.522

Pneumonia or UTI in Prior 6 Months

23% (19)

27% (12)

18% (7)

p = 0.490

Commercial insurance

12% (10)

13% (6)

10% (4)

p = 0.923


ED Loop Diuretic Dose (mg, furosemide equivalents)

40 (40-40)

40 (40-40)

40 (30-40)

p = 0.778

Total ED to Hospital Discharge Loop Diuretic Dose

120 (60-205)

120 (40-200)

160 (80-270)

p = 0.125

No Supplemental Oxygen Required in ED

44% (37)

47% (21)

41% (16)

p = 0.765

Positive Pressure Ventilation Required in ED

19% (16)

24% (11)

13% (5)

p = 0.283

Any ED or In-Hospital Positive Pressure Ventilation

24% (20)

29% (13)

18% (7)

p = 0.359

IV Nitroglycerin in ED

12% (10)

11% (5)

13% (5)

p = 1

Presented as median (interquartile range {IQR}) for continuous variables and %(n) for categorical; comparisons made with the Wilcoxon rank sum and chi-square tests, respectively. ECG = electrocardiogram, ED = emergency department, EMS = emergency medical services, GDMT = guideline directed medical therapy, HFrEF = heart failure with reduced ejection fraction, AHF = acute heart failure, UTI = urinary tract infection, IV = intravenous.

Gelman [19], where linear regression coefficients for continuous vari- ables and categorical variables in the same model approximate the same scale [19], facilitating better comparison of how much/how little each variable in the data set affected TAPSE. Collinearity was assessed by variance inflation factor (VIF), with no VIF > 3 covariates. The model was assessed by Nakagawa’s [20] R2 and statistical significance (p < 0.05).

Kaplan-Meier curves for event-free survival were fit for RVD, quar- tile LUS congestion severity, quartile LVEF, and quartile change in LUS severity from ED arrival to discharge (?LUS quartile). Group differences were assessed by the logrank test. Predictors were assessed separately for the ED timepoint and for repeated measures at all timepoints. A mul- tivariable Cox regression assessed survival by RVD, ?LUS quartile and LVEF adjusted for one another.

A multivariable GLMM of the primary clinical outcome was fit for the same three predictors to calculate area under the Receiver operating characteristic curve , with random intercept for patient and random slope for timepoint. R software (R Foundation, v3.6.1) was used for statistical analyses. Data is available upon request to the corre- sponding author (NH).

  1. Results

Fig. 1 reports inclusion, exclusion, and adjudication details. Initial ED echo and LUS were obtained in 120 eligible patients. After exclusions (n = 11), failure to obtain written consent within 24 h (n = 20) and di- agnostic adjudication (n = 5), 84 patients (197 exams) remained (0 h n = 84; 24 h n = 69; 72 h n = 35; 168 h n = 9). Median age was 62 (IQR 54-70), 79% were reduced LVEF, and 60% male. Baseline clinical characteristics, overall and for RVD vs. no RVD, are presented in Tables 1 and 2. Echo, LUS, and lab characteristics are presented in Table 3.

    1. Prevalence, trend, and correlates of RVD in the cohort

At ED arrival 46% (n = 39) had RVD by TAPSE (<17 mm), decreasing to 38% (n = 26) at 24 h (p = 0.036). Patients with RVD in the ED had longer LOS (5.51 days, 95% confidence interval {95%CI}:4.21-6.82), p = 0.042) compared to those without (3.67 days, 2.45-4.88). Adjusted for LOS, RVD decreased at 24 h vs. ED arrival (adjusted odds ratio

{aOR} = 0.08, 95%CI:0.02-0.48), and 72 vs. 24 h (0.18, 0.03-0.99), but

not 168 vs. 72 h (0.85, 0.05-15.5).

Patients with RVD had lower (p < 0.05) systolic blood pressure, while brain-natriuretic peptide, oxygen saturation, and missed antihy- pertensive medications or loop diuretic changes 7 days before arrival were higher (Table 2). Right ventricular outflow tract (RVOT) velocity- time integral (VTI), RVOT acceleration time, LVEF, mitral E’, and Inferior vena cava collapsibility were lower with RVD (p < 0.05), while mitral E/A ratio, IVC max diameter, and congestion severity by LUS were higher (Table 3). Interrater reliability for TAPSE and the other echo variables were > 0.70.

Prior history of PHTN or RVD was not significantly different between those with/without RVD in the ED (36% vs. 29% p = 0.651, Table 1). Of 57 without a known history of either condition, 44% had RVD. 53% with- out known PHTN had RVSP >40 mmHg at ED arrival, the threshold above which the ACC and American Heart Association (AHA) recom- mend further evaluation [21]. Only 2/57 (4%) left the ED/hospital with documented clinical suspicion or workup for PHTN or RVD.

    1. Predictors of TAPSE from ED arrival to discharge

32 potential predictors of TAPSE through time were identified (13 a priori, 19 with p > 0.1 for RVD vs. no RVD, Tables 1-3). The 32 variable LMM explained 75.7% of variance in TAPSE across 197 scans and 84 patients (Supplemental Table). Significant multivariable-adjusted

Table 3

Echocardiographic, lung ultrasound, and lab measures – all timepoints Emergency Department (ED) to discharge

Measure {units} Right ventricular dysfunction (RVD) defined as TAPSE <17 mm

Total (n = 196)

No RVD at time of Measurement (n = 112)

RVD at time of Measurement (n = 85)


*p < 0.05

**p < 0.01

***p < 0.001

Echocardiography and Lung Ultrasound

Lung Ultrasound B-line Score {0-180}

46 (27-73)

40 (20-68)

54 (34-81)

p = 0.012*

Left Ventricular Ejection Fraction {%}

28 (20-37)

31 (25.5-41.5)

24 (19-30)

p < 0.001***

Right Ventricular Basal Diastolic Diameter (RVDD) {cm}

4.6 (4-5.3)

4.4 (3.8-5)

4.9 (4.3-5.6)

p < 0.001***

Left Ventricular Basal Diastolic Diameter (LVDD) {cm}

5.6 (5.1-6)

5.4 (5-5.9)

5.75 (5.3-6.23)

p = 0.003**

Right Ventricular – Left Ventricular Basal Diastolic Diameter Ratio (RV-LV ratio)

0.83 (0.72-0.95)

0.81 (0.71-0.91)

0.85 (0.74-0.98)

p = 0.048*

Right Ventricular Systolic Pressure (RVSP) {mmHg}

38 (31-48)

36 (29-47)

43 (33-50)

p = 0.017*

Right Ventricular Outflow Tract Velocity-Time Integral (RVOT VTI) {cm/systole}

10.8 (7.8-15.3)

13.0 (9.0-16.1)

8.3 (6.5-12.6)

p < 0.001***

RVOT Acceleration Time {ms}

89 (74-102.5)

91 (79-108.75)

84.5 (65.5-96)

p < 0.001***

Mean Septal-Lateral Mitral E’ {m/s}

0.05 (0.04-0.06)

0.06 (0.04-0.07)

0.05 (0.04-0.06)

p < 0.001***

Mitral E/A Ratio

2.67 (1.32-4.04)

1.63 (1.07-3.44)

3.41 (2.43-4.55)

p < 0.001***

Inferior Vena Cava Max Diameter {cm}

2.3 (1.8-2.7)

2.1 (1.6-2.6)

2.4 (2.2-2.8)

p < 0.001***

Inferior Vena Cava inspiratory collapse {%}

44 (25-64)

55 (33-70)

27 (18.75-48.5)

p < 0.001***

Tricuspid Annulus Plane Systolic Excursion (TAPSE) {mm}

18 (13-23)

22 (19-25)

13 (11-15)

p < 0.001***


Brain Natriuretic Peptide

1098 (569-1910)

689 (320-1729)

1553 (840-2801)

p < 0.001***

Blood Urea Nitrogen

23 (17-36)

22 (17-40)

28 (20-41)

p = 0.072

Estimated Glomerular Filtration Rate

58 (37-74)

55 (41-79)

49 (29-70)

p = 0.091


138 (136-140)

139 (137-141)

138 (136-140)

p = 0.090


4.1 (3.7-4.5)

4.0 (3.7-4.4)

4.1 (3.8-4.6)

p = 0.092


11.8 (10.1-13.5)

10.9 (9.2-12.9)

10.9 (9.5-12.8)

p = 0.877

Troponin >99th percentile

70.7% (82)

69% (38)

81% (33)

p = 0.306

Change (?) from ED to Last Timepoint Before Discharge ED to Discharge ?TAPSE {mm}

0 (-1 to 3)

0 (-1 to 2)

1 (-1 to 5)

p = 0.104

ED to Discharge ?LUS/B-lines

-13 (-28 to 0)

-14 (-30 to 0)

-12 (-26 to 2)

p = 0.614

ED to Discharge ?LVEF {%}

0% (-3% to 5%)

-0% (-4% to 3%)

2% (-3% to 5%)

p = 0.154

Presented as median (interquartile range) for continuous variables and %(n) for categorical; comparisons made with the Wilcoxon rank sum and chi-square tests, respectively. LUS = lung ultrasound, LVEF = left ventricular ejection fraction.

predictors of TAPSE through time (p < 0.05, Supplemental Table), in order of standardized effect size, were missed antihypertensive medica- tion (associated with +4.6 mm TAPSE, standardized effect size {SES}

+0.38), PPV in the ED (+3.5 mm TAPSE, SES +0.29), COPD history (-2.4 mm TAPSE, SES -0.20), LVEF (+1.0 mm TAPSE per 11% increase

in LVEF, SES +0.18), LUS severity (-1.0 mm TAPSE per 41 unit increase in LUS severity, SES -0.15), and RVSP (-1.0 mm TAPSE per 16 mmHg increase in RVSP, SES -0.13).

    1. RVD, LUS, and LVEF vs. 30-day death or AHF readmission

24% (20/84) of patients were readmitted for AHF or died (n = 3) within 30 days. Event-free survival was lower with RVD at ED ar- rival (hazard ratio {HR} 3.31, 95%CI:1.28-8.53, Fig. 2A) and every timepoint from ED to discharge (HR = 3.81, 95%CI:2.08-6.96, Fig. 2B).

The absolute value of LUS severity, separated by quartiles, was not associated with event-free survival at ED arrival (Fig. 3A, p = 0.300) or ED through discharge (Fig. 3B, p = 0.1). However, cumula- tive change in LUS severity from ED to discharge (?LUS, Fig. 3C) showed a significant survival difference between Q1 (largest reduc- tion in congestion) vs. all others (HR = 0.13, 95%CI:0.017-0.97, p = 0.020).

LVEF quartile in the ED (p = 0.3, Supplemental Fig. S1A) and through all timepoints (p = 0.060, S1B) were not significantly associ- ated with survival. In a Cox regression sensitivity analysis, both RVD (p = 0.003, Supplemental Fig. S2A) and ?LUS (p = 0.004, S2B) were as- sociated with survival after multivariable adjustment, while LVEF quar- tile was not (p = 0.5474, Supplemental Fig. S2C).

TAPSE from ED to discharge and ?LUS predicted 30-day death or AHF readmission with AUROC = 0.807 {95%CI:0.742-0.872} (Fig. 4); the addition of LVEF to TAPSE and ?LUS had no impact on outcome pre- diction (i.e., no change in AUROC, Fig. 4).

  1. Discussion

To our knowledge, REED-AHF is the first study to evaluate compre- hensive echo or LUS at the earliest stages of AHF treatment (<1 h from first ED treatment) and examine changes through hospitalization. We found RVD to be present in nearly half of patients at ED arrival with no significant relationship with past history of PHTN despite a high prevalence in the sample (32%). Among those without a PHTN history, half had elevated pulmonary pressures in the ED, yet only 4% left the hospital with any new diagnosis or documented clinical suspicion for PHTN or RVD. Rapid change in RV function and pulmonary pressures from ED treatment to just 24 h later may help explain the latter finding. We also found that RVD was significantly less common at 24 h, even be- fore adjusting for the fact that patients with ED RVD stayed in the hos- pital longer. After this adjustment, there was an even more pronounced and significant downtrend in RVD from ED arrival through 72 h. None- theless, RVD in the ED, and from ED to discharge, was associated with longer LOS and higher 30-day death or AHF rehospitalization. Adjusting for LVEF did not mitigate this effect (Supplemental Fig. S2, Fig. 4). Rather, LVEF added virtually no predictive value beyond TAPSE and ?LUS (Fig. 4).

Given such findings, there may be value in routine measurement of

RVD early in the course of treatment. Because clinical assessment by the inpatient team may take up to 24 h after ED evaluation, and inpatient echocardiograms at many institutions (including our own) can take even longer, a large proportion of this early RVD would be clinically si- lent to current standards of care. Other reports have identified RVD as an important correlate of adverse outcomes [2,7,9,12-14,17], but the earliest-enrolling investigation to our knowledge examined patients up to 12 h after ED disposition [12]. Thus, not only does RVD appear quite common in patients with AHF, it is also likely unrecognized.

This presents important questions not only clinically, but also for public health. Among 30 patients who both met ACC/AHA criteria

Fig. 1. Screening, Inclusion, Exclusion, Enrollment, and Diagnostic Adjudication in REED- AHF. An overview of the study procedures, inclusions, and exclusions for the prospective REED-AHF study (“Right Ventricular Echocardiography in Emergency Department Acute Heart Failure”). CC = chief complaint, ED = emergency department, AHF = acute heart failure, IV = intravenous, SL = sublingual, LUS = lung ultrasound, LAR = legally authorized representative, h = hour, COPD = chronic obstructive pulmonary disease, ESRD = end stage renal disease, GI = gastrointestinal.

[21] for further PHTN evaluation and were without prior known PHTN or RVD, only 2 cases were recognized before discharge. An op- portunity to identify these patients early, refer them for outpatient specialty care after their acute illness, and hopefully prevent down- stream health consequences may exist. For instance, it has previ- ously been shown that 30% of ED patients undergoing negative com- puted tomography evaluation for pulmonary embolism have unrec- ognized PHTN and/or RVD, and such patients have high ED recidi- vism [22].

Assessing correlated clinical factors for RVD, we identified missed antihypertensive medication, PPV in the ED, COPD history, LVEF, LUS se- verity, and RVSP as significant predictors of RV function through time (section 3.3, and Supplemental Table). Given an observational design, these should be seen as hypothesis-generating and representative of as- sociation and not causation. For instance, missed antihypertensive med- ications within 7 days predicted higher TAPSE, but it lacks face-validity to suspect that medication non-adherence caused patients’ RV function

to improve. More likely, this association could reflect that patients with the low-risk [23] hypertensive-AHF phenotype had relatively preserved RV function compared to those with a different AHF precipitant [24,25]. Negative correlations between COPD and RVSP with TAPSE likely reflect hypoxic vasoconstriction and other causes of acute increase in pulmo- nary artery pressure, which in turn lead to acute worsening of RV func- tion. Higher TAPSE in patients receiving ED PPV could reflect a thera- peutic benefit of PPV for RV function. pulmonary congestion severity measured by LUS also had negative association with TAPSE; a finding not surprising as it reflects confluent physiological factors including el- evated LA pressure, LV dysfunction, and increased pulmonary pressures. An association between LUS severity and TAPSE, like PPV and TAPSE, could indicate a point of potential intervention. With RV function and B-line severity correlating through time, it is possible LUS-guided de- congestion could be used to improve RV function. The recent BLUSHED randomized trial [15] showed targeted reduction in LUS con- gestion severity as feasible.

While a positive correlation with LVEF and TAPSE is well described [5], the fact that RVD and LUS but not LVEF predicted 30-day readmis- sion or death warrants further consideration. LV and RV function inter- twine physiologically with both ventricles sharing the same cardiac out- put (assuming absence of Intracardiac shunt), and likewise share a pro- portion of systolic function through their common wall. LV dysfunction and elevated LA pressure are also conveyed backwards on the RV through hydrostatic pressure. Thus, prognostic implications of LV dys- function may already be captured by RVD as a variable. Further, there are critical components of AHF pathophysiology related to RV function but not LV function. Acute cardiorenal dysfunction is largely not related to LV dysfunction as long thought, and instead is now known to be a consequence of elevated central venous pressure [28]. Hypoxia, due to congestion, causes hypoxic vasoconstriction and subsequent elevation of pulmonary pressures. These two major facets of AHF directly impact RV preload and afterload, respectively, but affect the LV only down- stream and indirectly. Pulmonary congestion, moreover, is undertreated roughly half the time [29-33] by hospital discharge and inadequate de- congestion confers a 6 times higher rate of death or readmission [34- 37]. Our observations in the cohort align closely with the prior literature [34-37] in this regard: the change in congestion by LUS predicted death or AHF readmission (Fig. 3C), but not absolute severity of congestion at any single timepoint on its own (Fig. 3A and B). Being a result of LV dys- function rather than its cause, congestion may linger after LV function improves while still imposing backwards pressures on the upstream RV, in turn causing RVD and worse clinical outcomes independent of LV function. Lastly, each chambers’ resilience to conditions of acute de- compensation may be relevant. Elevated intracardiac pressures are the sine qua non of AHF physiology, beginning weeks before clinical signs or symptoms [38]. The LV’s thick walls accommodate a wide range of pressures before an acute change in LVEF, while even mild changes in filling pressures may have large effects on systolic function in the highly pressure-sensitive RV [39,40]. As such, the RV is directly and often dras- tically affected by the pathologic processes of AHF, including and not limited to acute LV dysfunction, thereby making it perhaps an ideal sur- rogate for overall cardiopulmonary decompensation.

    1. Limitations

This was a secondary analysis of REED-AHF related to, but separate from, the study’s primary aims. As such, though data collection was pro- spective, the analysis plan was generated retrospectively which could introduce bias.

We did not adjust TAPSE or LUS severity for every important variable with regards to clinical outcomes, nor assess for incremental prognostic significance compared to other risk measures. This is one of two pri- mary, a priori, aims of REED-AHF for which preliminary results during the study’s interruption for COVID-19 were presented [2] and for which a full report is forthcoming. Our goal with the comparison to

Image of Fig. 2

Fig. 2. Right Ventricular Dysfunction (RVD) vs. 30-Day Death or AHF Readmission – Kaplan Meier Curves. Panel A: RVD vs. no RVD at the ED timepoint only. Panel B: RVD vs. no RVD at each timepoint (i.e., repeated measures). AHF = acute heart failure, RVD = right ventricular dysfunction, TAPSE = tricuspid annulus plane systolic excursion, ED = emergency department.

clinical outcomes, was simply to show three things: 1. LUS and TAPSE predict adverse outcomes independent of one another 2. LVEF is not a confounder for the association of LUS and TAPSE to outcomes, and 3. A model with strong predictive value (AUROC >0.8) could be fit with just these two variables alone.

As mentioned above, our assessment of the predictors of RV func- tion are simply hypothesis generating and cannot assess causality. Moreover, while TAPSE is generally a good measure of RVD, patients with severe pulmonary-RV uncoupling and/or longstanding PHTN may have an artificially elevated TAPSE [18]. Echocardiographic

parameters such as the TAPSE/RVSP ratio have been proposed as more robust to this phenomenon [12] and are the subject of pending analyses.

Our model for the prediction of TAPSE initially had 32 variables com- pared to 196 observations. While overfitting is less of an issue in linear regression than logistic regression, this is still a large number of vari- ables. However, our goal was an explanatory model for TAPSE rather than prediction, so overfitting is arguably less prescient.

Finally, as in any study, we cannot guarantee our results apply to populations dissimilar from our own. Our population had a

Image of Fig. 3

Fig. 3. Lung Ultrasound vs. 30-Day Death or AHF Readmission – Kaplan Meier Curves. Reflects a 10-point LUS exam for congestion quantification, with possible scores ranging from 0 (no congestion) to 180 (most congestion), based on the LUS protocol of the BLUSHED-AHF trial [1,2]. Panel A: LUS quartile (absolute value) at the ED timepoint only. Panel B: LUS quartile (absolute value) at each timepoint (i.e., repeated measures). Panel C: Cumulative change in LUS severity by quartile (?LUS quartile) from ED arrival to last timepoint before ED or hospital discharge. AHF = acute heart failure, LUS = lung ultrasound, ED = emergency department.

Image of Fig. 4

Fig. 4. Receiver Operating Characteristic Curve for the Prediction of 30-Day Death or AHF Readmission, for TAPSE and Cumulative ED to Discharge Change in LUS severity, with/ without LVEF. Generalized Linear mixed models with a binary link (repeated measures logistic regression) were used to model the outcome for TAPSE and ?LUS from ED to discharge, with and without LVEF as an additional variable. The addition of LVEF did not add significant (p >= 0.05) predictive value (by AUC, sensitivity, or specificity) to TAPSE and ?LUS alone. AHF = acute heart failure, tricuspid annulus plane systolic excursion = TAPSE, LUS = lung ultrasound, LVEF = left ventricular ejection fraction, AUC = Area Under the Curve.

higher than average rate of hypertensive-AHF, but it is unclear how this would affect our results compared to a different AHF population.

    1. Conclusions

RVD is common and underrecognized in ED patients with AHF and, along with LUS is a strong predictor of 30-day death and AHF- readmission, independent of LVEF.


This work was supported by the Blue Cross Blue Shield of Michigan Declaration of interests”>Foundation (BCBSM).

Declaration of interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influ- ence the work reported in this paper.



Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi. org/10.1016/j.ajem.2021.11.024.


  1. Collins SP, Storrow AB, Levy PD, et al. Early management of patients with acute heart failure: state of the art and future directions–a consensus document from the SAEM/HFSA acute heart failure working group. Acad Emerg Med. 2015;22(1): 94-112.
  2. Harrison NE, Ehrman R, Favot M, et al. Sonographic right ventricular dysfunction predicts acute heart failure outcomes independent of current emergency depart- ment risk measures. Ann Emerg Med. 2020;76(4):S95.
  3. Frea S, Pidello S, Bovolo V, et al. Prognostic incremental role of right ventricular func- tion in acute decompensation of advanced chronic heart failure. Eur J Heart Fail. 2016;18(5):564-72.
  4. Iglesias-Garriz I, Olalla-Gomez C, Garrote C, et al. Contribution of right ventricular dysfunction to heart failure mortality: a meta-analysis. Rev Cardiovasc Med. 2012; 13(2-3):e62-9.
  5. Guazzi M, Bandera F, Pelissero G, et al. Tricuspid annular plane systolic excursion and pulmonary arterial systolic pressure relationship in heart failure: an index of right ventricular contractile function and prognosis. Am J Physiol Heart Circ Physiol. 2013;305(9):H1373-81.
  6. Guazzi M, Naeije R, Arena R, et al. Echocardiography of right ventriculoarterial cou- pling combined with cardiopulmonary exercise testing to predict outcome in heart failure. Chest. 2015;148(1):226-34.
  7. Guazzi M, Dixon D, Labate V, et al. RV contractile function and its coupling to pulmo- nary circulation in heart failure with preserved ejection fraction: stratification of clinical phenotypes and outcomes. J Am Coll Cardiol Img. 2017;10(10 Pt B):1211-21.
  8. Ghio S, Guazzi M, Scardovi AB, et al. Different correlates but similar prognostic impli- cations for right ventricular dysfunction in heart failure patients with reduced or preserved ejection fraction. Eur J Heart Fail. 2017;19(7):873-9.
  9. Bilchick KC, Mejia-Lopez E, McCullough P, et al. Clinical impact of changes in hemo- dynamic indices of contractile function during treatment of acute decompensated heart failure. J Card Fail. 2018;24(1):43-50.
  10. Daley J, Grotberg J, Pare J, et al. Emergency physician performed tricuspid annular plane systolic excursion in the evaluation of suspected pulmonary embolism. Am J Emerg Med. 2017;35(1):106-11.
  11. Rutz MA, Clary JM, Kline JA, Russell FM. Emergency physicians are able to detect Right ventricular dilation with good agreement compared to cardiology. Acad Emerg Med. 2017;24(7):867-74.
  12. Palazzuoli A, Ruocco G, Evangelista I, De Vivo O, Nuti R, Ghio S. Prognostic signifi- cance of an early echocardiographic evaluation of right ventricular dimension and function in acute heart failure. J Card Fail. 2020;26(10):813-20.
  13. Harrison N, Favot M, Ehrman R, Levy P. Right ventricular dysfunction on arrival pre- dicts acute heart failure severity and recidivism. Acad Emerg Med. 2019;26(S1): S217.
  14. Harrison N, Lenning J, Ehrman R, et al. Pulmonary edema severity by lung ultrasound predicts central apnea in acute heart failure. Acad Emerg Med. 2021;28(S1).
  15. Pang Peter S, Russell Frances M, Ehrman R, et al. Lung ultrasound-guided Emergency Department Management of Acute Heart Failure (BLUSHED-AHF). JACC. 2021;0(0).
  16. Russell FM, Ehrman RR, Ferre R, et al. Design and rationale of the B-lines lung ultra- sound guided emergency department management of acute heart failure (BLUSHED-AHF) pilot trial. Heart Lung. 2019;48(3):186-92.
  17. Ojji DB, Lecour S, Atherton JJ, Blauwet LA, Alfa J, Sliwa K. Right ventricular systolic dysfunction is common in hypertensive heart failure: a prospective study in sub- Saharan Africa. PLoS One. 2016;11(4):e0153479.
  18. Rudski LG, Lai WW, Afilalo J, et al. Guidelines for the echocardiographic assessment of the right heart in adults: a report from the American Society of Echocardiography: endorsed by the European Association of Echocardiography, a registered branch of the European Society of Cardiology, and the Canadian Society of Echocardiography. J Am Soc Echocardiogr. 2010;23(7):685-713.
  19. Gelman A. Scaling regression inputs by dividing by two standard deviations. Stat Med. 2008;27(15):2865-73.
  20. Nakagawa S, Johnson PCD, Schielzeth H. The coefficient of determination R(2) and intra-class correlation coefficient from generalized linear mixed-effects models revisited and expanded. J R Soc Interface. 2017;14(134).
  21. McLaughlin VV, Archer SL, Badesch DB, et al. ACCF/AHA 2009 expert consensus doc- ument on pulmonary hypertension: a report of the American College of Cardiology Foundation task force on expert consensus documents and the American Heart As- sociation developed in collaboration with the American College of Chest Physicians; American Thoracic Society, Inc.; and the Pulmonary Hypertension Association. J Am Coll Cardiol. 2009;53(17):1573-619.
  22. Russell FM, Kline JA, Lahm T. High rate of isolated right ventricular dysfunction in pa- tients with non-significant CT pulmonary angiography. Am J Emerg Med. 2018;36 (2):281-4.
  23. Harrison N, Pang P, Collins S, Levy P. Blood pressure reduction in hypertensive acute heart failure. Curr Hypertens Rep. 2021;23(2):11.
  24. Fonarow GC, Abraham WT, Albert NM, et al. Factors identified as precipitating hos- pital admissions for heart failure and clinical outcomes: findings from OPTIMIZE-HF. Arch Intern Med. 2008;168(8):847-54.
  25. Arrigo M, Gayat E, Parenica J, et al. Precipitating factors and 90-day outcome of acute heart failure: a report from the intercontinental GREAT registry. Eur J Heart Fail. 2017;19(2):201-8.
  26. Rangaswami J, Bhalla V, Blair JEA, et al. Cardiorenal syndrome: classification, patho- physiology, diagnosis, and treatment strategies: a scientific statement from the American Heart Association. Circulation. 2019;139(16):e840-78.
  27. Gheorghiade M, Filippatos G, De Luca L, Burnett J. Congestion in acute heart failure syndromes: an essential target of evaluation and treatment. Am J Med. 2006;119 (12 Suppl 1):S3-s10.
  28. Gheorghiade M, Follath F, Ponikowski P, et al. Assessing and grading congestion in acute heart failure: a scientific statement from the acute heart failure committee of the heart failure association of the European Society of Cardiology and endorsed by the European Society of Intensive Care Medicine. Eur J Heart Fail. 2010;12(5): 423-33.
  29. Ambrosy AP, Pang PS, Khan S, et al. Clinical course and predictive value of congestion during hospitalization in patients admitted for worsening signs and symptoms of heart failure with reduced ejection fraction: findings from the EVEREST trial. Eur Heart J. 2013;34(11):835-43.
  30. Lucas C, Johnson W, Hamilton MA, et al. Freedom from congestion predicts good sur- vival despite previous class IV symptoms of heart failure. Am Heart J. 2000;140(6): 840-7.
  31. Testani JM, Chen J, McCauley BD, Kimmel SE, Shannon RP. Potential effects of aggres- sive decongestion during the treatment of decompensated heart failure on renal function and survival. Circulation. 2010;122(3):265-72.
  32. Coiro S, Rossignol P, Ambrosio G, et al. Prognostic value of residual pulmonary con- gestion at discharge assessed by lung Ultrasound imaging in heart failure. Eur J Heart Fail. 2015;17(11):1172-81.
  33. Cogliati C, Casazza G, Ceriani E, et al. Lung ultrasound and short-term prognosis in heart failure patients. Int J Cardiol. 2016;218:104-8.
  34. Rivas-Lasarte M, Maestro A, Fernandez-Martinez J, et al. Prevalence and prognostic impact of subclinical pulmonary congestion at discharge in patients with acute heart failure. ESC Heart Fail. 2020;7(5):2621-8.
  35. Gargani L, Pang PS, Frassi F, et al. Persistent pulmonary congestion before discharge predicts rehospitalization in heart failure: a lung ultrasound study. Cardiovasc Ultra- sound. 2015;13:40.
  36. Fudim M, Hernandez AF, Felker GM. Role of volume redistribution in the congestion of heart failure. J Am Heart Assoc. 2017;6(8):e006817.
  37. Konstam Marvin A, Kiernan Michael S, Bernstein D, et al. Evaluation and manage- ment of right-sided heart failure: a scientific statement from the American Heart As- sociation. Circulation. 2018;137(20):e578-622.
  38. Abel FL, Waldhausen JA. Effects of alterations in Pulmonary vascular resistance on right ventricular function. J Thorac Cardiovasc Surg. 1967;54(6):886-94.

Leave a Reply

Your email address will not be published. Required fields are marked *