Abstract
Aims
Remote dielectric sensing (ReDS) enables quick estimation of lung fluid content. To examine if ReDS is superior to other methods in detecting acute heart failure.
Methods and results
We included consecutive patients with dyspnoea from the emergency departments at Bispebjerg Hospital, Copenhagen, and performed ReDS, low-dose chest computed tomography (CT), echocardiogram, lung ultrasound, NT-Pro-brain natriuretic peptide (NT-proBNP), and a Boston score evaluation (chest X-ray and clinical signs). ReDS values >35% were used as a cut-off to diagnose pulmonary congestion. Acute heart failure was adjudicated by experts’ review of health records but independently of ReDS values. Sub-analyses investigated ReDS in acute heart failure patients with congestion on CT. We included 97 patients within a median of 4.8 h from admittance: 25 patients (26%) were ReDS-positive and 39 (40%) had adjudicated acute heart failure (21 with and 18 without CT congestion). Heart failure patients had median ReDS 33%, left ventricular ejection fraction 48%, and NT-proBNP 2935 ng/L. A positive ReDS detected heart failure with 46% sensitivity, 88% specificity, and 71% accuracy. The AUC for ReDS was like the Boston score (P = 0.88) and the lung ultrasound score (P = 0.74). CT-congested heart failure patients had higher ReDS values than patients without heart failure (median 38 vs. 28%, P < 0.001). Heart failure patients without CT-congestion had ReDS values like patients without heart failure (mean 30 vs. 28%, P = 0.07).
Conclusion
ReDS detects acute heart failure similarly to the Boston score and lung ultrasound score, and ReDS primarily identifies the acute heart failure patients who have congestion on a chest CT.
Keywords: Remote dielectric sensing technology, Dyspnoea, Acute decompensated heart failure
Graphical Abstract
Graphical Abstract.
Introduction
Dyspnoea is a common complaint in the emergency department1 and half of the patients presenting with dyspnoea have congestive heart failure as the primary or co-primary diagnosis.1,2 A correct diagnosis is essential for swift therapy and recovery, but diagnosis can be challenging, especially among elderly patients with comorbidities.3,4 Diagnostic echocardiography in every patient is not possible in a busy emergency department. Approximately 20% of patients presenting with dyspnoea are initially misdiagnosed, and emergency physicians struggle to decide which patients should be referred for instant cardiologic evaluation.5,6 Thus, simple, immediate, and reliable methods are needed to guide emergency physicians in choosing the right diagnostic strategy.
Remote dielectric sensing (ReDS) is a recent technology that immediately and non-invasively estimates lung fluid content based on electromagnetic energy.7 The ReDS technology is approved by FDA and CE for lung fluid monitoring. In previous studies, ReDS was correlated to pulmonary wedge pressure8 and it could accurately discriminate heart failure patients with pulmonary congestion on a chest computed tomography (CT) from normal subjects.9 A recent study found that ReDS detects pulmonary oedema moderately in emergency patients.10
In the real world, patients with acute congestive heart failure present with varying degrees of vascular and interstitial pulmonary congestion. There are at least two phenotypes of pulmonary congestion: one with interstitial tissue congestion displaying radiographic signs of congestion, and one with vascular congestion, without radiographic signs of congestion, that can only be determined after echocardiographic confirmation of elevated left ventricular (LV) filling pressure.11,12 Because ReDS estimates lung fluid content,9 we anticipated higher ReDS values in patients with radiographic signs of congestion.
The primary aim was to examine the value of ReDS to detect or rule out acute heart failure among consecutive patients with dyspnoea in an emergency department. Furthermore, we aimed to compare ReDS with a lung ultrasound13,14 and Boston score.15 Finally, we examined how ReDS values are affected by radiographic congestion on a chest CT scan.
Method
Design
A prospective single-centre observational study was performed in the emergency department at Bispebjerg University Hospital, Copenhagen, Denmark. The National Ethics Committee approved the study protocol on Health Research Ethics in Copenhagen, Denmark (Project-id: H-17000869).
Population
On 110 randomly selected workdays from October 2018 to August 2019, we screened all consecutive patients with a same-day hospital contact between 2 a.m. and 1.30 p.m. The main inclusion criteria were acute dyspnoea and age of 50 years or above. Acute dyspnoea was defined as self-reported sudden onset of dyspnoea or worsening of chronic dyspnoea within 14 days, combined with at least one abnormal objective parameter supporting respiratory imbalance (Figure 1). We excluded patients requiring intensive care or ventilation, acute coronary syndrome requiring telemetry, and patients who refused or could not consent. The protocol demanded that all study examinations had to be done as fast as possible and no later than 12 h after admittance, otherwise, the patient was excluded (Figure 1).13,16
Figure 1.
CONSORT diagram of inclusion.
Patients underwent a clinical examination, arterial blood gas, and phlebotomy as part of the clinical routine. All patients provided initial informed written consent for the package of the protocolled study procedures that were performed on every patient: ReDS, lung ultrasound, chest X-ray, low-dose chest CT, and echocardiography.
Remote dielectric sensing technology
ReDS (FDA: K150095 and CE: 3900874CE01) provides an estimate of the percentage of lung fluid content within the range of 15–60%.7 Values within the range of 20–35% are normal by standard, and values above 35% were defined as positive for pulmonary congestion (positive ReDS).7,9 Patients were measured once on the right hemithorax in a sitting position. After applying the ReDS vest, ReDS each measurement lasted 30 s.
Radiology
CT scans were used as the reference standard for determining the presence or absence of pulmonary congestion. A multi-slice CT scan (Somatom Definitions Flash, Siemens Medical Solutions, Forschheim, Germany) was acquired during a single breath-hold without spirometry for detecting inspiration status. The mean CT-radiation dose was 1.3 mSv (95% CI: 1.2–1.4). The CT scans were reviewed by two specialized thoracic radiologists, independently of each other, and blinded to clinical information including chest X-rays and ReDS measurements. The radiologists assessed signs of pulmonary congestion on CT, and we defined CT with interstitial congestion based on the agreement between the two observers. In case of no agreement, the CT was classified as ‘without congestion’.17 In concert with the CT, a chest X-ray was also taken.
Symptoms, clinical signs of heart failure, and Nt-proBNP levels
We used modified Boston criteria as a clinical and radiographic measure of the signs and symptoms of heart failure.15 The New York Heart Association functional classification (NYHA) was used to grade the severity of dyspnoea in the Boston score and replaced the categories of ‘rest dyspnoea’, ‘dyspnoea while walking on level area’ and ‘dyspnoea while climbing’ with NYHA IV, NYHA III, and NYHA II, respectively. Orthopnoea and paroxysmal nocturnal dyspnoea were used accordingly to the Boston score. The score gave a total from 0 to 12, where 0–4 marks ‘unlikely heart failure’, 5–7 ‘possible heart failure’, and 8–12 marks ‘definite heart failure’ (Table 1).
Table 1.
Modified Boston criteria for heart failure in relation to ReDS
| Positive ReDS N: 25 |
Normal ReDS N: 72 |
Total N: 97 |
P-value | |
|---|---|---|---|---|
| History a | ||||
| NYHA II, N (%) | 8 (32.0) | 22 (30.6) | 30 (30.9) | | |
| NYHA III, N (%) | 12 (48.0) | 31 (43.1) | 43 (44.3) | 0.784 |
| NYHA VI, N (%) | 5 (20.0) | 18 (25.0) | 23 (23.7) | | |
| Orthopnoea, N (%) | 17 (68.0) | 31 (43.1) | 48 (49.5) | 0.032 |
| Paroxysmal nocturnal dyspnoea, N (%) | 14 (56.0) | 45 (62.5) | 59 (60.8) | 0.566 |
| Clinical examination | ||||
| Heart rate (per minute), mean (SD) | 87.2 (78.4–95.9) | 93.1 (87.8–98.4) | 90.0 (78.0–101.0) | 0.254 |
| Jugular venous elevation, N (%) | 1 (4.0) | 0 (0) | 1 (1.0) | 0.258 |
| Lung crackles basilar, N (%) | 5 (20.0) | 6 (8.3) | 11 (11.3) | 0.113 |
| Lung crackles more than basilar, N (%) | 3 (12.0) | 16 (22.2) | 19 (19.6) | 0.384 |
| Wheezing, N (%) | 2 (8.0) | 10 (13.9) | 12 (12.4) | 0.723 |
| Chest X-ray | ||||
| Alveolar pulmonary oedema, N (%) | 8 (32.0) | 8 (11.3) | 16 (16.7) | 0.017 |
| Interstitial pulmonary oedema, N (%) | 12 (48.0) | 12 (16.9) | 24 (25.0) | 0.002 |
| Accentuated flow shift, N (%) | 11 (44.0) | 13 (18.1) | 24 (24.7) | 0.011 |
| Bilateral pleural effusion, N (%) | 11 (44.0) | 8 (11.1) | 19 (20.0) | <0.001 |
| Enlarged heart (ratio >0.50), N (%) | 15 (60.0) | 22 (35.5) | 37 (38.1) | 0.034 |
| Boston score | ||||
| Total Boston score, median (IQR) | 8.0 (6.0–9.0) | 6.0 (4.0–8.0) | 7.0 (5.0–8.0) | 0.017 |
| Heart failure unlikely (score 0–4), N (%) | 2 (8.0) | 21 (29.2) | 23 (23.7) | | |
| Heart failure possible (score 5–7), N (%) | 10 (40.0) | 29 (40.3) | 39 (40.2) | 0.027 |
| Heart failure definite (score 8–12), N (%) | 13 (52.0) | 22 (30.6) | 35 (36.1) | | |
No patient had NYHA group I.
NT-Pro-brain natriuretic peptide (NT-proBNP) was measured in blood samples at admission and analysed with the Roche Elecsys technique. NT-proBNP results were available for the diagnosing adjudication panel.
Lung ultrasound
Certified investigators performed lung ultrasounds according to international recommendations using a 14-zone protocol, including 8 anterior zones and 6 posterior zones.14 We used a Sonosite X-Porte Ultrasound System device with a 2.5–3.5 MHz cardiac probe. One video sequence of 4 s was saved for all lung ultrasound zones, and two blinded certified reviewers assessed B-lines and pleural effusion according to guidelines.13,14 One zone with at least 3 B-lines on each hemithorax was considered positive for congestion. For this study, we created a continuous lung ultrasound score on a scale from 0 to 5 points based on the presence of bilateral B-lines and pleural effusion. A score of 3 corresponds to the traditional criterion of pulmonary congestion based on at least 3 B-lines in at least one zone on both sides (see supplementary material online, Table S1).13,14
Echocardiography
Dedicated specialists in cardiology performed the comprehensive transthoracic echocardiography, and readings were approved by at least two cardiologists, who were heart failure specialists and accredited in echocardiography by the European Society of Cardiology (ESC). Echocardiograms were performed according to the 2016 ESC guidelines,18 including a systematic evaluation of LV filling pressure and diastolic dysfunction.11,12 Cardiac dysfunction was classified as reduced left ventricular ejection fraction (LVEF) < 40% (HFrEF), mildly reduced LVEF from 40 to 49% (HFmrEF), and preserved LVEF 50% or greater (HFpEF).11 However, severe valve disease (HFvhd), was noted separately to distinguish it from HFpEF.
Adjudication of pulmonary and cardiac diagnoses
The final diagnoses of pulmonary and cardiac diagnoses were adjudicated by two cardiologists and two pulmonologists after patient discharge. They reviewed clinical data, medication, specifically intensification of intravenous or oral diuretics, routine clinical radiology reports, blood tests and echocardiography. CT and lung ultrasound images were not assessed by the adjudication panel.
A diagnosis of acute decompensated heart failure (ADHF) was ascertained by at least two cardiologists, and in the case of disagreement, the diagnosis was discussed with a third cardiologist. Criteria adhered to the ESC guidelines based on clinical signs, elevated NT-proBNP, a chest X-ray reading, echocardiographic cardiac dysfunction,11 and evidence of elevated LV filling pressure (Grade II or Grade III) as a sign of pulmonary vascular congestion.12 However, the adjudicators did not require Grade II or III elevated LV filling pressure in a few patients who fulfilled all criteria and already had received acute relevant therapy for triggers like a high blood pressure or fast a supraventricular tachyarrhythmia.
We defined interstitial congestion as radiographic congestion on a CT assessed by both independent radiologists. Correspondingly, we defined vascular congestion as an adjudicated ADHF diagnosis without radiographic congestion on chest CT.
Statistical analyses
The primary outcome of the study was the adjudicated ADHF diagnosis, including both patients with and without radiographic CT congestion. In sub-analyses, we examined ReDS in relation to ADHF patients with and without radiographic congestion on the chest CT. The main discriminatory exposure variable was ReDS dichotomized using 35% as the cut-point. However, we also examined ReDS as a continuous variable compared with the continuous Boston score and lung ultrasound score. Because we used NT-proBNP for adjudication, it was only examined as an exploratory exposure variable.
The diagnostic value of ReDS was evaluated against the primary outcome measure ADHF and was assessed by 2 × 2-tables and receiver operating characteristic (ROC)-curves with area under the curve (AUC). We compared the AUC of each modality against ADHF, and thresholds for indicating heart failure were ReDS >35%, lung ultrasound score of 3 or more (see supplementary material online, Table S1),13,14 and Boston score of 8 or more.15
As appropriate, we report data as means and standard deviations (SD), medians and interquartile ranges (IQR), or counts and percentages. Histograms and Shapiro-Wilks test of normality assessed distributions of variables. The independent samples t-test, the χ2 test/Fisher’s exact test, the Wilcoxon rank-sum, and the Kruskal Wallis test were used for statistical comparisons. Two-sided tests were used, and P-values of <0.05 were considered statistically significant. Interrater agreements were measured with kappa statistics. We used the ANOVA test to compare ReDS values in ADHF patients with and without CT congestion and non-ADHF patients and illustrated differences with boxplots. In the ANOVA test, the displayed P-values were adjusted for comparison of multiple groups. The Bonferroni correction was used for adjustment for multiple group comparisons.
The required sample size was based on an assumed prevalence of ADHF in 30% of dyspnoeic patients in the emergency department.1,2 The power was 0.80 with alpha 0.05 to detect a kappa-value of 0.85 against a null hypothesis corresponding to a kappa-value of 0.65. Thus, the minimum sample required was 90 patients.19 We performed the statistical analyses with SAS Studio version 3.7 (SAS Institute, Cary NC) and Rstudio version 3.6.0.20
Results
We included 97 patients among 134 eligible patients (Figure 1). Ineligibility reasons were the inability to provide informed consent, acute referral to intensive care unit due to critical respiratory and hemodynamic instability, or acute coronary syndrome. We excluded 37 patients because examinations were not performed simultaneously (Figure 1). A positive ReDS examination occurred in 25 (25.7%) patients and was associated with higher NT-proBNP levels, Boston score, chest X-ray signs of heart failure, lung ultrasound score (0–2, 3, and 4–5 points), and more abnormal echocardiographic parameters (Tables 1 and 2).
Table 2.
Baseline characteristics for positive and normal ReDS-examinations
| Positive ReDS N: 25 |
Normal ReDS N: 72 |
Total N: 97 |
P-value | |
|---|---|---|---|---|
| Patient characteristics | ||||
| Age (years), mean (SD) | 73.3 (68.8–77.7) | 72.5 (70.1–75.0) | 72.7 (70.6–74.8) | 0.772 |
| Sex (male), N (%) | 18 (72.0) | 41 (56.9) | 59 (60.8) | 0.184 |
| BMI (kg/m2), median (IQR) | 27.2 (24.2–32.4) | 25.9 (22.3–28.7) | 26.0 (22.7–30.1) | 0.100 |
| BMI >39 kg/m2, N (%) | 1 (4.0) | 2 (2.8) | 3 (3.1) | 1.000 |
| BMI <22 kg/m2, N (%) | 3 (12.0) | 15 (20.8) | 18 (18.6) | 0.389 |
| History of heart failure, N (%) | 7 (28.0) | 18 (25.0) | 25 (25.8) | 0.768 |
| History of COPD, N (%) | 7 (28.0) | 44 (61.1) | 51 (52.6) | 0.004 |
| History of asthma, N (%) | 4 (16.0) | 15 (20.8) | 19 (19.6) | 0.773 |
| Known renal insufficiency, N (%) | 4 (16.0) | 7 (9.7) | 11 (11.3) | 0.467 |
| Diabetes Mellitus I/II, N (%) | 12 (48.0) | 13 (18.1) | 25 (25.8) | 0.003 |
| Peripheral pedal oedema, N (%) | 9 (36.0) | 18 (25.0) | 27 (27.8) | 0.290 |
| Fever, N (%) | 4 (16.0) | 11 (15.3) | 15 (15.5) | 1.000 |
| C-reactive protein (mg/L), median (IQR) | 11.0 (7.0–42.0) | 20.5 (6.5–56.5) | 16.0 (7.0–55.0) | 0.338 |
| NT-ProBNP (ng/l), median (IQR) | 2021.2 (800.0–4465.3) | 683.3 (225.8–2355.3) | 997.9 (257.1–2934.6) | 0.037 |
| Lung ultrasound | ||||
| No interstitial syndrome (score 0–2), N (%) | 16 (64.0) | 52 (72.2) | 68 (70.1) | | |
| Interstitial syndrome (score 3), N (%) | 1 (4.0) | 15 (20.8) | 16 (16.5) | 0.003 |
| Interstitial syndrome AND pleural effusion (score 4–5), N (%) | 8 (32.0) | 5 (5.0) | 13 (13.4) | | |
| Echocardiographic parameters | ||||
| Ejection fraction (%), median (IQR) | 50.0 (37.5–56.0) | 60.0 (45.0–60.0) | 55.0 (40.0–60.0) | 0.021 |
| Lateral e’ (cm/s), mean (SD) | 8.0 (6.9–9.1) | 9.3 (8.6–10.0) | 9.0 (6.9–10.2) | 0.067 |
| E/e’, median (IQR) | 13.6 (9.6–17.2) | 10.2 (7.5–13.0) | 10.6 (8.0–14.2) | 0.004 |
| Tricuspid regurgitation gradient max velocity (cm/s), median (IQR) | 2.9 (2.6–3.3) | 2.8 (2.3–3.1) | 282.0 (232.0–315.0) | 0.112 |
| Index LA-volume (mL/m2), median (IQR) | 43.0 (34.6–49.9) | 29.9 (22.0–38.6) | 32.3 (24.7–44.0) | 0.002 |
| LV filling pressure and diastolic dysfunction, N (%) | ||||
| Grade 1 or indeterminate/normal | 7 (28.0) | 44 (61.1) | 51 (52.6) | | |
| Grade 2 (elevated LV filling pressure) | 11 (44.0) | 17 (23.6) | 28 (28.9) | 0.001 |
| Grade 3 (elevated LV filling pressure) | 4 (16.0) | 1 (1.4) | 5 (5.2) | | |
We adjudicated ADHF in 39 (40.2%) patients (Table 3). ADHF was more frequent in ReDS-positive than ReDS-normal patients (72.0 vs. 29.2%, P < 0.001. Although most ReDS normal patients did not have ADHF (70.8%), still, 8.3% had ADHF with CT-congestion (Table 3). ADHF was also significantly correlated to the lung ultrasound score and the Boston score (respectively, P < 0.001) (see supplementary material online, Table S2).
Table 3.
Final diagnoses in relation to ReDS
| Positive ReDS N: 25 |
Normal ReDS N: 72 |
Total N: 97 |
P-value | |
|---|---|---|---|---|
| Adjudicated final diagnoses a | ||||
| All ADHF, N (%) | 18 (72.0) | 21 (29.2) | 39 (40.2) | <0.001 |
| CT-congested ADHF, N (%) | 12 (48.0) | 6 (8.3) | 18 (18.6) | <0.001 |
| CT-non-congested ADHF, N (%) | 6 (24.0) | 15 (20.8) | 21 (21.6) | 0.781 |
| No ADHF, N (%) | 7 (28.0) | 51 (70.8) | 58 (59.8) | <0.001 |
Assessed by an expert panel
At hospital admittance, ADHF patients had a higher frequency of atrial fibrillation, systolic blood pressure >140 mmHg, and elevated levels of NT-proBNP and troponin T (see supplementary material online, Table S3). Moreover, patients with chronic obstructive pulmonary disease (COPD) exacerbations and CT-verified emphysema had less ADHF (see supplementary material online, Table S3).
A positive ReDS detected ADHF with a sensitivity of 46% and specificity of 88% and with a kappa value of 0.36 (95% CI: 0.18–0.54, P < 0.001) (Table 4). Notably, ReDS diagnosed the subgroup with CT-congested ADHF better with a sensitivity of 67%, specificity of 84% and accuracy of 80% (P < 0.001) (Table 4).
Table 4.
Diagnostic value of ReDS
| N = 97 | All ADHF | CT-congested ADHF | ||
|---|---|---|---|---|
| Yes | No | Yes | No | |
| Positive ReDS | 18 | 7 | 12 | 13 |
| Normal ReDS | 21 | 51 | 6 | 66 |
| P-value | <0.0001 | <0.0001 | ||
| Diagnostic measurement | (95% CI) | Diagnostic measurement | (95% CI) | |
|---|---|---|---|---|
| Sensitivity (%) | 46 | (31–62) | 67 | (45–88) |
| Specificity (%) | 88 | (80–96) | 84 | (75–92) |
| Accuracy (%) | 71 | (62–80) | 80 | (73–88) |
| Negative predictive value (%) | 72 | (54–90) | 92 | (85–98) |
| Positive predictive value (%) | 71 | (60–81) | 48 | (28–68) |
| Odds ratio | 6.2 | (2.3–17.1) | 10.2 | (3.2–32.0) |
| Negative likelihood ratio | 0.6 | 0.4 | ||
| Positive likelihood ratio | 3.8 | 4.2 |
The CT-congested ADHF patients had higher ReDS values than patients without ADHF (median 38 vs. 28%, P < 0.001) (Figure 2). ADHF patients without CT-congestion had ReDS values that were not significantly different from non-ADHF patients (median 30 vs. 28%, adjusted P = 0.07) or CT-congested ADHF patients (mean 30 vs. 38%, adjusted P = 0.23).
Figure 2.
Boxplots of ReDS values in selected ADHF subgroups. Reported P-values are adjusted for pairwise comparisons between multiple groups. Median ReDS values (IQR) (%): No ADHF (− congestion) (N = 58): 28.0 (25.0–30.0)ADHF (− congestion) (N = 21): 30.0 (29.0–37.0) ADHF (+ congestion) (N = 18): 37.5 (29.0–43.0). *** Significant difference between the groups (P < 0.001).
In ROC analyses, ReDS detected ADHF with an AUC of 0.74 (Figure 3), not different from the Boston score or the lung ultrasound score (AUC = 0.77 and AUC = 0.71, P = 0.69 and P = 0.73 for direct comparison with ReDS) (Figure 3). NT-proBNP had a higher AUC than ReDS to detect ADHF (AUC = 0.89, P = 0.02). Sub analyses indicated that ReDS could improve the Boston score and the lung ultrasound score to detect ADHF (AUC 0.83 vs. 0.77, P = 0.03, and AUC 0.80 vs. 0.71, P = 0.06) (see supplementary material online, Figure S1).
Figure 3.
ROC curves of ReDS, Boston criteria and lung ultrasound against ADHF.
Echocardiography, radiology, and ReDS examinations were performed almost simultaneously within a median time of 1.2–1.6 h between comparative methods (see supplementary material online, Table S4). The median duration for one ReDS-measurement, including application, was 1.5 min (IQR: 1.0–1.8 min). During hospital admission, intravenous diuretics were given to 28 of the 39 ADHF patients and to 4 patients without ADHF (see supplementary material online, Table S4). Only five (5.2%) patients had diuretics administrated between study examinations. The median time from admittance to the intravenous diuretic administration was 4.9 h. Of the 18 ADHF patients with CT congestion, 17 (94.4%) had intravenous diuretics administrated within a median time of 3.9 h (see supplementary material online, Table S4).
Regarding reproducibility, for the overall estimation of lung ultrasound congestion based on either 3 or more B-lines or pleural effusion, the kappa value was 0.89.13 Similarly, the two radiologists agreed on adjudicated pulmonary congestion on CT with a kappa value of 0.83 (95% CI 0.72–0.94, < 0.0001).
Discussion
We examined whether ReDS could identify acute heart failure in consecutive emergency patients with dyspnoea and a high frequency of pulmonary comorbidity. We conclude that a positive ReDS examination has a high specificity for detecting ADHF. The objective ReDS method has an overall accuracy that parallels full clinical examination, chest X-ray, and lung ultrasound. Notably, in a sub analysis, ReDS primarily identifies those ADHF patients who have congestion on a chest CT scan rather than ADHF patients with sole vascular congestion and no radiographic congestion.
Previous proof-of-concept studies reported high accuracies of ReDS to detect CT measured lung fluid content in selected heart failure populations.7,9 ReDS also correlated well with pulmonary capillary wedge pressure and were found to have a high negative predictive value (94.9%) to exclude patients with PCWP >17 mm Hg.8 But these results were obtained in a mixed cohort of patients with heart failure and heart transplantation and cannot be directly adopted into an unselected emergency population.8 Our study is the first study to examine the value of ReDS to detect acute heart failure among emergency patients and to directly compare ReDS with other clinical methods to detect heart failure in the emergency department.
Only one other study has evaluated ReDS in an emergency department. Rafique Z, et al.10 found that ReDS values above 35% detected pulmonary congestion in dyspnoeic emergency patients with a moderate sensitivity of 79% and specificity of 62% and an accuracy of 68%. Similarly, we found a 71% accuracy for ReDS to detect ADHF. However, our study included more comorbid patients reflected in a 37% prevalence of pneumonia, and 50% had known COPD (Table 2; see supplementary material online, Table S3). In contrast to Rafique et al, we found a high specificity of 88% but a low sensitivity of 46% for ReDS to detect ADHF. The reason may be that our broad adjudicated ADHF diagnosis included those with definitely elevated LV filling pressure on a comprehensive echocardiogram. We found a better accuracy when restricting the outcome to ADHF patients with radiographic congestion (Table 4).
Thus, apart from testing ReDS in an emergency cohort with a high frequency of pulmonary comorbidity, our significant new finding is that a positive ReDS primarily detects the ADHF patients presenting with interstitial congestion, here demonstrated on a CT. In a comorbid emergency population, ADHF patients may present with merely pulmonary vascular congestion, but those patients had ReDS values similar to non-ADHF patients. Pulmonary disease may increase lung density in pulmonary fibrosis and decrease lung density in emphysema, which theoretically confounds the diagnostic value of ReDS.21 In this real-world scenario of consecutive patients with a high rate of comorbidity, a positive ReDS was associated significantly with ADHF patients who had pulmonary interstitial congestion on a simultaneous chest CT. It is a novel observation that the predefined threshold of ReDS 35%, was too high to detect ADHF patients with mainly vascular congestion but without interstitial or radiographic congestion on CT.
The study limitations were the single-centre design, the limited sample size, and that examinations were not performed in the evenings and nights. Still, it was a strength that all tests were finished within a median of 4.8 h because it warranted patients had a similar cardiorespiratory state for every examination. The 97-patient sample size was arbitrary set as there were no previous studies to guide a formal sample size calculation. For these reasons, it may be regarded as an important pilot study to guide further use and research of this modality. Still, we do not anticipate an excellent diagnostic value of ReDS if a larger study with the same type of patients had been performed. The inclusion rate was reduced because a high rate of acute patients were unable to provide informed consent. We speculate the ReDS method could be useful in these, and future studies should try to include such patients, perhaps in a cluster design. A limitation is that we did not exclude patients according to their BMI. Thus, we included 18 patients with BMI <22 and 3 patients BMI >39, and ReDS has not been validated in these patient groups.
Patients with severe acute pulmonary oedema or needing mechanical ventilation were excluded, and these rarely represent a diagnostic problem, so our results apply to the large contingency of patients with grey zone symptoms and a less obvious clinical presentation.
One strength of the study was the adjudicated reference ADHF diagnosis based on comprehensive echocardiographic evidence for cardiac dysfunction and elevated LV filling pressure11,12 and the expert panel that judged whether acute symptoms and signs were due to ADHF. We used the ESC/AHA recommended classification for LV filling pressure.12 It is a limitation that this echocardiographic classification of cardiac filling pressures has not been validated in acute patients. Still, it is currently the best renowned non-invasive method to estimate cardiac filling pressures. Another strength is that this study was investigator-initiated and -designed, and therefore not influenced by industry.
Clinical methods like lung ultrasound, and clinical examinations including chest X-ray detect ADHF as well as ReDS. This is a novel observation, but plausible because all methods indirectly estimate the lung fluid content. Each method has its advantages, but ReDS can be measured by less experienced medical staff in less than five minutes. Thus, ReDS can easily be combined with other diagnostic methods, and sub analyses also indicated that ReDS improves the Boston criteria and lung ultrasound to detect ADHF. Lung ultrasound detected ADHF with a lower diagnostic value than reported in other studies. However, our study differed as we examined patients above 50 years of age with a high rate of comorbidities, and we did not excluded patients with known pulmonary fibrosis or emphysema.13
NT-proBNP had a higher AUC than ReDS to diagnose ADHF. It is plausible that NT-proBNP detects ADHF with both vascular and interstitial congestion compared with ReDS which primarily identifies patients with interstitial congestion. Furthermore, the AUC for NT-proBNP was inflated because the adjudicating expert panel was aware of the NT-proBNP levels. Thus, we do not find that ReDS adds diagnostic value to current clinical practice, including NT-proBNP. However, ReDS can provide a rapid assessment before the complete clinical examination, chest X-ray, and NT-proBNP, thereby ensuring fast triage in the emergency department, where a positive ReDS examination should prompt cardiac workup. Acute heart failure with radiographic interstitial congestion will be picked up by ReDS, which is relevant because only 12% of the dyspnoeic patients had diuretics administrated within 3 h from admittance (see supplementary material online, Table S2). Hence, a positive ReDS examination could prompt diuretic treatment in more relevant patients.
Supplementary Material
Acknowledgements
We would like to thank all the involved departments for supporting this project.
Contributor Information
Anne Sophie Overgaard Olesen, Department of Cardiology, Copenhagen University Hospital, Bispebjerg and Frederiksberg, Bispebjerg Bakke 23, 2400 Copenhagen, Denmark.
Kristina Miger, Department of Cardiology, Copenhagen University Hospital, Bispebjerg and Frederiksberg, Bispebjerg Bakke 23, 2400 Copenhagen, Denmark.
Andreas Fabricius-Bjerre, Department of Cardiology, Copenhagen University Hospital, Bispebjerg and Frederiksberg, Bispebjerg Bakke 23, 2400 Copenhagen, Denmark.
Kathrine Dyrsting Sandvang, Department of Cardiology, Copenhagen University Hospital, Bispebjerg and Frederiksberg, Bispebjerg Bakke 23, 2400 Copenhagen, Denmark.
Ingunn Eklo Kjesbu, Department of Cardiology, Copenhagen University Hospital, Bispebjerg and Frederiksberg, Bispebjerg Bakke 23, 2400 Copenhagen, Denmark.
Ahmad Sajadieh, Department of Cardiology, Copenhagen University Hospital, Bispebjerg and Frederiksberg, Bispebjerg Bakke 23, 2400 Copenhagen, Denmark.
Nis Høst, Department of Cardiology, Copenhagen University Hospital, Bispebjerg and Frederiksberg, Bispebjerg Bakke 23, 2400 Copenhagen, Denmark.
Nana Køber, Department of Cardiology, Copenhagen University Hospital, Bispebjerg and Frederiksberg, Bispebjerg Bakke 23, 2400 Copenhagen, Denmark.
Jesper Wamberg, Department of Emergency Medicine, Copenhagen University Hospital, Bispebjerg and Frederiksberg, Bispebjerg Bakke 23, 2400 Copenhagen, Denmark.
Lars Pedersen, Department of Pulmonary Medicine, Copenhagen University Hospital, Bispebjerg and Frederiksberg, Bispebjerg Bakke 23, 2400 Copenhagen, Denmark.
Hans Henrik Lawaetz Schultz, Department of Pulmonary Medicine, Copenhagen University Hospital, Bispebjerg and Frederiksberg, Bispebjerg Bakke 23, 2400 Copenhagen, Denmark.
Annemette Geilager Abild-Nielsen, Department of Radiology, Copenhagen University Hospital, Bispebjerg and Frederiksberg, Bispebjerg Bakke 23, 2400 Copenhagen, Denmark.
Mathilde Marie Winkler Wille, Department of Radiology, Nordsjaellands Hospital, Dyrehavevej 29, 3400 Hillerød, Denmark.
Olav Wendelboe Nielsen, Department of Cardiology, Copenhagen University Hospital, Bispebjerg and Frederiksberg, Bispebjerg Bakke 23, 2400 Copenhagen, Denmark; Copenhagen Center for Translational Research, Copenhagen University Hospital, Bispebjerg and Frederiksberg, Nielsine Nielsens Vej 4B, 2400 Copenhagen, Denmark; Faculty of Health and Medical Sciences, Copenhagen University, Blegdamsvej 3, 2200 Copenhagen, Denmark.
Lead author biography
Anne Sophie Olesen graduated from Copenhagen University of Medicine in 2020. During the medical study, Anne Sophie completed a research year at the Department of Cardiology, Bispebjerg University Hospital, where she included the study population, of which this article is applied on. Anne Sophie has clinical experience from working as a physician at the internal medical department at Gentofte Hospital, and in the emergency department at Bispebjerg University Hospital.
Anne Sophie is especially interested in examining and improving diagnostic and prognostic issues among acute dyspnoeic patients and will proceed this research under guidance from prof. Olav W. Nielsen.
Data availability
The authors have full control of all data, and some of the data underlying this article will be shared on reasonable request to the senior (last) author Prof. Olav W. Nielsen.
Supplementary material
Supplementary material is available at European Heart Journal Open online.
Funding
This work was supported by the research fund of Bispebjerg Hospital. Sensible Medical Innovations Ltd made the ReDS device available for free and provided an unrestricted grant to specifically collect the ReDS measurements. The writing of this paper was funded by Holger & Ruth Hesse’s Mindefond. The statistical analyses, study design, data collection, and writing of the paper were not affected by the sponsors.
Conflicts of interest: No conflicts to declare.
References
- 1. Mueller C. Acute dyspnoea in the emergency department. In: The ESC textbook of intensive and acute cardiovascular care. 3rd ed. Oxford University Press; 2021. Chapter 11. [Google Scholar]
- 2. Maisel AS, Kriwhnaswama P, Nowak RM, McCord J, Hollander JE, Duc P, Omland T, Storrow AB, Abraham WT, Wu AHB, Clopton P, Steg PG, Westheim A, Knudsen CW, Perez A, Kazanegra R, Herrmann HC, McCullough PA. Rapid measurement of B-type natriuretic peptide in the emergency diagnosis of heart failure. N Engl J Med 2002;347:161–167. [DOI] [PubMed] [Google Scholar]
- 3. Martindale JL, Wakai A, Collins SP, Levy PD, Diercks D, Hiestand BC, et al. Diagnosing acute heart failure in the emergency department: a systematic review and meta-analysis. Acad Emerg Med 2016;23:223–242. [DOI] [PubMed] [Google Scholar]
- 4. Lien CTC, Gillespie ND, Struthers AD, McMurdo MET. Heart failure in frail elderly patients: diagnostic difficulties, co-morbidities, polypharmacy and treatment dilemmas. Eur J Heart Fail 2002;4:91–98. [DOI] [PubMed] [Google Scholar]
- 5. Ray P, Birolleau S, Lefort Y, Becquemin MH, Beigelman C, Isnard R, et al. Acute respiratory failure in the elderly: etiology, emergency diagnosis and prognosis. Crit Care 2006;10:R82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Hunold KM, Caterina JM. High diagnostic uncertainty and inaccuracy in adult emergency department patients with dyspnea: a national database analysis. Acad Emerg Med 2019 Feb;26:267–271. [DOI] [PubMed] [Google Scholar]
- 7. Amir O, Rappaport D, Zafrir B, Abraham WT. A novel approach to monitoring pulmonary congestion in heart failure: initial animal and clinical experiences using remote dielectric sensing technology. Congest Hear Fail 2013;19:149–155. [DOI] [PubMed] [Google Scholar]
- 8. Uriel N, Sayer G, Imamura T, Rodgers D, Kim G, Raikhelkar J, Sarswat N, Kalantari S, Chung B, Nguyen A, Burkhoff D, Abbo A. Relationship between noninvasive assessment of lung fluid volume and invasively measured cardiac hemodynamics. J Am Heart Assoc 2018;7:e009175. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Amir O, Azzam ZS, Gaspar T, Faranesh-Abboud S, Andria N, Burkhoff D, Abbo A, Abraham WT. Validation of remote dielectric sensing (ReDSTM) technology for quantification of lung fluid status: comparison to high-resolution chest computed tomography in patients with and without acute heart failure. Int J Cardiol 2016;221:841–846. [DOI] [PubMed] [Google Scholar]
- 10. Rafique Z, McArthur R, Sekhon N, Mesbah H, Almasary A, Peacock F. Remote dielectric sensing for detecting pulmonary edema in the emergency department. Am J Emerg Med 2022;55:11–15. [DOI] [PubMed] [Google Scholar]
- 11. McDonagh TA, Metra M, Adamo M, Gardner RS, Baumbach A, Böhm M, Voors AA, Anker SD, Bueno H, Cleland JGF, Coats AJS, Crespo-Leiro MG, Farmakis D, Gilard M, Heymans S, Hoes AW, Jaarsma T, Jankowska EA, Lainscak M, Lam CSP, Lyon AR, McMurray JJV, Mebazaa A, Mindham R, Muneretto C, Francesco Piepoli M, Price S, Rosano GMC, Ruschitzka F, Kathrine Skibelund A, de Boer RA, Christian Schulze P, Abdelhamid M, Aboyans V, Adamopoulos S, Anker SD, Arbelo E, Asteggiano R, Bauersachs J, Bayes-Genis A, Borger MA, Budts W, Cikes M, Damman K, Delgado V, Dendale P, Dilaveris P, Drexel H, Ezekowitz J, Falk V, Fauchier L, Filippatos G, Fraser A, Frey N, Gale CP, Gustafsson F, Harris J, Iung B, Janssens S, Jessup M, Konradi A, Kotecha D, Lambrinou E, Lancellotti P, Landmesser U, Leclercq C, Lewis BS, Leyva F, Linhart A, Løchen M-L, Lund LH, Mancini D, Masip J, Milicic D, Mueller C, Nef H, Nielsen J-C, Neubeck L, Noutsias M, Petersen SE, Sonia Petronio A, Ponikowski P, Prescott E, Rakisheva A, Richter DJ, Schlyakhto E, Seferovic P, Senni M, Sitges M, Sousa-Uva M, Tocchetti CG, Touyz RM, Tschoepe C, Waltenberger J, Adamo M, Baumbach A, Böhm M, Burri H, Čelutkienė J, Chioncel O, Cleland JGF, Coats AJS, Crespo-Leiro MG, Farmakis D, Gardner RS, Gilard M, Heymans S, Hoes AW, Jaarsma T, Jankowska EA, Lainscak M, Lam CSP, Lyon AR, McMurray JJV, Mebazaa A, Mindham R, Muneretto C, Piepoli MF, Price S, Rosano GMC, Ruschitzka F, Skibelund AK. 2021 ESC guidelines for the diagnosis and treatment of acute and chronic heart failure. Eur Heart J 2021;42:3599–3726. [DOI] [PubMed] [Google Scholar]
- 12. Nagueh SF, Smiseth OA, Appleton CP, Byrd BF, Dokainish H, Edvardsen T, Flachskampf FA, Gillebert TC, Klein AL, Lancellotti P, Marino P, Oh JK, Popescu BA, Waggoner AD. Recommendations for the evaluation of left ventricular diastolic function by echocardiography: an update from the American society of echocardiography and the European association of cardiovascular imaging. J Am Soc Echocardiogr 2016;29:277–314. [DOI] [PubMed] [Google Scholar]
- 13. Miger KC, Fabricius-Bjerre A, Maschmann C, Wamberg J, Wille MM, Abild-Nielsen AG, Pedersen L, Lawaetz Schultz HH, Damm Nybing J, Nielsen OW. Clinical applicability of lung ultrasound methods in the emergency department to detect pulmonary congestion on computed tomography. Ultraschall Med 2021;42:21–30. [DOI] [PubMed] [Google Scholar]
- 14. Volpicelli G, Elbarbary M, Blaivas M, Lichtenstein DA, Mathis G, Kirkpatrick AW, Melniker L, Gargani L, Noble VE, Via G, Dean A, Tsung JW, Soldati G, Copetti R, Bouhemad B, Reissig A, Agricola E, Rouby J-J, Arbelot C, Liteplo A, Sargsyan A, Silva F, Hoppmann R, Breitkreutz R, Seibel A, Neri L, Storti E, Petrovic T. International evidence-based recommendations for point-of-care lung ultrasound. Intensive Care Med 2012;38:577–591. [DOI] [PubMed] [Google Scholar]
- 15. Di Bari M, Pozzi C, Cavallini MC, Innocenti F, Baldereschi G, De Alfieri W, Antonini E, Pini R, Masotti G, Marchionni N. The diagnosis of heart failure in the community. Comparative validation of four sets of criteria in unselected older adult: the ICARe Dicomano Study. JACC 2004;44:1601–1608. [DOI] [PubMed] [Google Scholar]
- 16. Miger K, Fabricius-Bjerre A, Olesen AS, Sajadieh A, Høst N, Køber N, Abild A, Winkler Wille MM, Wamberg J, Pedersen L, Lawaetz Schultz HH, Torp-Pedersen C, Wendelboe Nielsen O. Radiologic signs of acute heart failure on a computed tomography: a prospective observational study in breathless patients. Cardiol J 2022;29:235–244. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Hansell DM, Bankier AA, MacMahon H, McLoud TC, Müller NL, Remy J. Fleischner society: glossary of terms for thoracic imaging. Radiology 2008;246:697–722. [DOI] [PubMed] [Google Scholar]
- 18. Ponikowski P, Voors AA, Anker SD, Bueno H, Cleland JGF, Coats AJS, Falk V, González-Juanatey JR, Harjola V-P, Jankowska EA, Jessup M, Linde C, Nihoyannopoulos P, Parissis JT, Pieske B, Riley JP, Rosano GMC, Ruilope LM, Ruschitzka F, Rutten FH, van der Meer P. 2016 ESC guidelines for the diagnosis and treatment of acute and chronic heart failure. Eur Heart J 2016;37:2129–2200.27206819 [Google Scholar]
- 19. Bujang MA, Baharum N. Guidelines of the minimum sample size requirements for Cohen’s Kappa. Epidemiol Biostat Public Heal 2017;14:e12267–1. [Google Scholar]
- 20. RStudio Team . RStudio: Integrated Development Environment for R [Internet]. Boston, MA; 2015. http://www.rstudio.com/
- 21. Cardinale L, Priola AM, Moretti F, Volpicelli G. Effectiveness of chest radiography, lung ultrasound and thoracic computed tomography in the diagnosis of congestive heart failure. World J Radiol 2014;6:230. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The authors have full control of all data, and some of the data underlying this article will be shared on reasonable request to the senior (last) author Prof. Olav W. Nielsen.