Experimentally Induced Peripheral Venous Congestion Exacerbates Inflammation, Oxidative Stress, and Neurohormonal and Endothelial Cell Activation in Patients With Systolic Heart Failure

BACKGROUND

Venous congestion (VC) is a hallmark of symptomatic heart failure (HF) requiring hospitalization; however, its role in the pathogenesis of HF progression remains unclear. We investigated whether peripheral VC exacerbates inflammation, oxidative stress and neurohormonal and endothelial cell (EC) activation in patients with HF with reduced ejection fraction (HFrEF).

METHODS AND RESULTS

Two matched groups of patients with HFrEF and with no peripheral VC vs without recent HF hospitalization were studied. We modeled peripheral VC by inflating a cuff around the dominant arm, targeting ∼ 30 mmHg increase in venous pressure (venous stress test [VST]). Blood and ECs were sampled before and after 90 minutes of VST. We studied 44 patients (age 53 ± 12 years, 32% female). Circulating endothelin-1, tumor necrosis factor-α, interleukin-6, isoprostane, angiotensin II (ang-2), angiopoietin-2, vascular cell adhesion molecule-1, and CD146 significantly increased after the VST. Enhanced endothelin-1 and angiopoietin-2 responses to the VST were present in patients with vs without recent hospitalization and were prospectively associated with incident HF-related events; 6698 messenger ribonucleic acid (mRNA probe sets were differentially expressed in ECs after VST.

CONCLUSIONS

Experimental VC exacerbates inflammation, oxidative stress, neurohormonal and EC activation and promotes unfavorable transcriptome remodeling in ECs of patients with HFrEF. A distinct biological sensitivity to VC appears to be associated with high risk for HF progression.

Intraoperative Electroencephalography Alpha-Band Power Is a Better Proxy for Preoperative Low MoCA Under Propofol Compared With Sevoflurane

BACKGROUND

Preoperative abnormal cognitive status is a risk factor for postoperative complications yet remains underdiagnosed. During propofol general anesthesia, intraoperative electroencephalography (EEG) variables, such as alpha band power (α-BP), correlate with cognitive status. This relationship under sevoflurane is unclear. We investigated whether EEG biomarkers of poor cognitive status found under propofol could be extended to sevoflurane.

METHODS

In this monocentric prospective observational study, 106 patients with intraoperative EEG monitoring were included (propofol/sevoflurane = 55/51). We administered the Montreal Cognitive Assessment (MoCA) scale to identify abnormal cognition (low MoCA) 1 day before intervention. EEG variables included delta to beta frequency band powers. Results were adjusted to age and drug dosage. We assessed depth of anesthesia (DoA) using the spectral edge frequency (SEF 95 ) and maintained it within (8-13) Hz.

RESULTS

The difference in α-BP between low and normal MoCA patients was significantly larger among propofol patients (propofol: 4.3 ± 4.8 dB versus sevoflurane: 1.5 ± 3.4 dB, P = .022). SEF 95 and age were not statistically different between sevoflurane and propofol groups. After adjusting to age and dose, low α-BP was significantly associated with low MoCA under propofol (odds ratio [OR] [confidence interval {CI}] = 0.39 [0.16-0.94], P = .034), but not under sevoflurane, where theta-band power was significantly associated with low MoCA (OR [CI] = 0.31 [0.13-0.73], P = .007).

CONCLUSIONS

We suggest that intraoperative EEG biomarkers of abnormal cognition differ between propofol and sevoflurane under general anesthesia.

Prevalence, clinical characteristics and outcomes of heart failure patients with or without isolated or combined mitral and tricuspid regurgitation: An analysis from the ESC-HFA Heart Failure Long-Term Registry

AIM

Mitral regurgitation (MR) and tricuspid regurgitation (TR) are common in patients with heart failure (HF). The aim of this study was to investigate prevalence, clinical characteristics and outcomes of patients with or without isolated or combined MR and TR across the entire HF spectrum.

METHODS AND RESULTS

The ESC-HFA EORP HF Long-Term Registry is a prospective, multicentre, observational study including patients with HF and 1-year follow-up data. Outpatients without aortic valve disease were included and stratified according to isolated or combined moderate/severe MR and TR. Among 11 298 patients, 7541 (67%) had no MR/TR, 1931 (17%) isolated MR, 616 (5.5%) isolated TR and 1210 (11%) combined MR/TR. Baseline characteristics were differently distributed across MR/TR categories. Compared to HF with reduced ejection fraction, HF with mildly reduced ejection fraction was associated with a lower risk of isolated MR (odds ratio [OR] 0.69; 95% confidence interval [CI] 0.60-0.80), and distinctly lower risk of combined MR/TR (OR 0.51; 95% CI 0.41-0.62). HF with preserved ejection fraction (HFpEF) was associated with a distinctly lower risk of isolated MR (OR 0.42; 95% CI 0.36-0.49), and combined MR/TR (OR 0.59; 95% 0.50-0.70), but a distinctly increased risk of isolated TR (OR 1.94; 95% CI 1.61-2.33). All-cause death, cardiovascular death, HF hospitalization and combined outcomes occurred more frequently in combined MR/TR, isolated TR and isolated MR versus no MR/TR. The highest incident rates were observed in isolated TR and combined MR/TR.

CONCLUSION

In a large cohort of outpatients with HF, prevalence of isolated and combined MR and TR was relatively high. Isolated TR was driven by HFpEF and was burdened by an unexpectedly poor outcome.

Impact of lockdown on cardiovascular disease hospitalizations in a Zero-COVID-19 country

OBJECTIVES

There are concerns about the potential effect of social distancing used to control COVID-19 on the incidence of cardiovascular diseases (CVD).

STUDY DESIGN

Retrospective cohort study.

METHODS

We examined the association between lockdown and CVD incidence in a Zero-COVID country, New Caledonia. Inclusion criteria were defined by a positive troponin sample during hospitalization. The study period lasted for 2 months, starting March 20, 2020 (strict lockdown: first month; loose lockdown: second month) compared with the same period of the three previous years to calculate incidence ratio (IR). Demographic characteristics and main CVD diagnoses were collected. The primary endpoint was the change in incidence of hospital admission with CVD during lockdown compared with the historical counterpart. The secondary endpoint included influence of strict lockdown, change in incidence of the primary endpoint by disease, and outcome incidences (intubation or death) analyzed with inverse probability weighting method.

RESULTS

A total of 1215 patients were included: 264 in 2020 vs 317 (average of the historical period). CVD hospitalizations were reduced during strict lockdown (IR 0.71 [0.58-0.88]), but not during loose lockdown (IR 0.94 [0.78-1.12]). The incidence of acute coronary syndromes was similar in both periods. The incidence of acute decompensated heart failure was reduced during strict lockdown (IR 0.42 [0.24-0.73]), followed by a rebound (IR 1.42 [1-1.98]). There was no association between lockdown and short-term outcomes.

CONCLUSIONS

Our study showed that lockdown was associated with a striking reduction in CVD hospitalizations, independently from viral spread, and a rebound of acute decompensated heart failure hospitalizations during looser lockdown.

Exploratory proteome profiling in patients with peripartum cardiomyopathy – a biomarker study on the EORP cohort

Background

The diagnosis of peripartum cardiomyopathy (PPCM) remains challenging as heart failure symptoms may also occur during normal pregnancy. This is further aggravated by the absence of biomarkers specific for diagnosis or prognosis of women with PPCM. Indeed, current evidence from the EURObservational Research Programme (EORP) Registry, an ongoing prospective, international, multicentre, observational registry for women with PPCM, report that the time to diagnosis after symptom onset varies from 19.4 to 38.3 days.

Aims

We performed exploratory serum proteome profiling on patients with PPCM, as compared with healthy postpartum mothers, to uncover novel protein biomarkers that would further our understanding of the pathogenesis of the disease and enhance diagnostic evaluation.

Methods

Demographic and clinical data, as well as serum samples were collected from 84 patients with PPCM from seven EORP participating countries and 29 healthy controls (HC) from South Africa. Serum proteomic profiling was conducted using DIA-based label-free quantitative (LFQ) LC-MS at the time of diagnosis from depleted serum samples. Mass spectrometry data were analyzed by Spectronaut v15 using a study-specific spectral library. Proteomic statistical analysis was performed using Perseus version 2.0.3.0 (FDR=0.05; S=0.1).

Results

Patients with PPCM had advanced heart failure (50% had New York Heart Association functional classes III/IV, mean left ventricular ejection fraction [LVEF] of 33.5%±9.3 [vs 57.0±8.8 in HC, p<0.001]). Amongst the 329 proteins that were identified in the serum samples, 17 proteins were significantly differentially upregulated and 18 downregulated in patients with PPCM as compared to the HC (all p<0.05; Figure 1). Adiponectin (log fold change 1.378, p=0.001), pregnancy-specific beta-1-glycoprotein 1 (1.207, p=0.022), disintegrin metalloproteinase domain-containing protein 12 (1.185, p=0.039), peptidyl-prolyl cis-trans isomerase (1.182, p=0.031) and sulfhydryl oxidase 1 (1.101, p=0.004) were among the upregulated proteins, whilst immunoglobulin kappa variable 2–29 (0.856, p=0.029), ficolin-3 (0.898, p=0.001), platelet basic protein (0.917, p=0.006) and thrombospondin-1 (0.930, p=0.043) were among the downregulated. Gene ontology indicated that thrombospondin receptor activity, fibronectin-binding, and vascular endothelial growth factor receptor 2 binding among the most significant regulated molecular functions. The area under the curve (AUC) of the top 10 up-regulated biomarkers ranged from 0.61–0.68 (p<0.05).

Conclusion

Salient biological themes related to immune response proteins, inflammation, fibrosis, angiogenesis, apoptosis, and blood coagulation were identified to be predominant in PPCM versus HC. This indicates the complex pathophysiological mechanisms of PPCM. The newly identified proteins warrant further studies to evaluate their potential use as diagnostic and prognostic markers for PPCM.

Funding Acknowledgement

Type of funding sources: Public grant(s) – EU funding. Main funding source(s): 1. EurObservational Research Programme in conjunction with the Heart Failure Association of the European Society of Cardiology Study Group on Peripartum Cardiomyopathy2. Cape Heart Institute, University of Cape Town, South Africa

Elevated plasma Galectin-3 is associated with major adverse kidney events and death after ICU admission

Background

Galectin-3 (Gal-3) is a proinflammatory and profibrotic protein especially overexpressed after Acute Kidney Injury (AKI). The early renal prognostic value of Gal-3 after AKI in critically ill patients remains unexplored. The objective was to evaluate the prognostic value of plasma level of Gal-3 for Major Adverse Kidney Events (MAKE) and mortality 30 days after ICU admission across AKI stages.

Methods

This is an ancillary study of a prospective, observational, multicenter cohort (FROG-ICU). AKI was defined using KDIGO definition.

Results

Two thousand and seventy-six patients had a Gal-3 plasma level measurement at ICU admission. Seven hundred and twenty-three (34.8%) were females and the median age was 63 [51, 74] years. Eight hundred and seven (38.9%) patients developed MAKE, 774 (37.3%) had AKI and mortality rate at 30 days was 22.4% (N = 465). Patients who developed MAKE had higher Gal-3 level at admission compared to patients without (30.2 [20.8, 49.2] ng/ml versus 16.9 [12.7, 24.3] ng/ml, p < 0.001, respectively. The area under the receiver operating characteristic curve of Gal-3 to predict MAKE was 0.76 CI_95% [0.74–0.78], p < 0.001. Gal-3 was associated with MAKE (OR 1.80 CI_95% [1.68–1.93], p < 0.001, non-adjusted and OR 1.37 CI_95% [1.27–1.49], p < 0.001, adjusted). The use of Gal-3 improved prediction performance of prediction model including SAPSII, Screat_adm, pNGAL with a NRI of 0.27 CI_95%(0.16–0.38), p < 0.001. Median Gal-3 was higher in non-survivors than in survivors at 30 days (29.2 [20.2, 49.2] ng/ml versus 18.8 [13.3, 29.2] ng/ml, p < 0.001, respectively).

Conclusion

Plasma levels of Gal-3 were strongly associated with renal function, with an increased risk of MAKE and death after ICU admission.

Trial registration ClinicalTrials.gov NCT01367093. Registered on 6 June 2011.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s13054-021-03878-x.

Myocardial tissue function and characterization in patients with idiopathic cardiac magnetic resonance myopericarditis: evolutionary analysis

Introduction

The identification of prognostic markers related to the occurrence of events and recovery of ventricular function may be important in patients with acute myopericarditis (AMP). There is still a lack of data related to tissue characterization by cardiac magnetic resonance (CMR) of AMP, evolution and definition of possible long-term prognostic markers.

Purpose

To evaluate the myocardial tissue characterization of CMR related to the occurrence of combined events (death from all causes, heart failure and AMP recurrence) and the increase in left ventricular ejection fraction (LVEF) in patients with AMP.

Methods

Inclusion criteria were chest pain and/or electrocardiographic changes associated with elevated troponin (above the 99th percentile) in the absence of coronary stenosis and diagnosis of AMP by CMR &lt;48 hours of admission confirmed by the presence of edema and/or late enhancement. After a follow-up of up to 24 months, 100 patients remained and in the assessment of the increase in LVEF (increase &gt;5%), 36 cases remained, recalled for a new CMR between 6 and 18 months from the initial event.

Results

Significant differences in CMR were found between patients who had combined events (n=26) versus no combined events (n=74) in the following characteristics evaluated: initial LVEF (OR=0.938; CI: 0.895–0.984, p=0.008), left ventricular (LV) systolic volume index (OR=1.034; CI: 1.005–1.062, p=0.019), LV diastolic volume index (OR=1.029; CI: 1.002–1.056, p=0.038), presence of hypersignal in T2 (OR=11.325; CI: 2.247–57.075, p=0.003), presence of late anteroseptal enhancement (OR=0.160; CI: 0.037–0.685, p=0.014), basal anteroseptal (OR=0.255; CI: 0.071–0.914, p=0.036) and lateral apical (OR=5.902; CI: 1.236–28.187, p=0.026). In relation to the increase in LVEF, significant differences were found in CMR in the following characteristics evaluated: LVEF (OR=0.870; CI: 0.758–0.988, p=0.047), end systolic volume of the right ventricle (OR=1.047; CI: 1.001–1.096, p=0.047), LV systolic diameter (OR=1.283; CI: 1.034–1.593, p=0.023), LV diastolic diameter (OR=1.225; CI: 1.012–1.482, p=0.038), LV systolic volume index (OR=1.340; CI: 1.066–1.685, p=0.012), LV diastolic volume index (OR=1.111; CI: 1.017–1.213, p=0.019) and right ventricular systolic volume index (OR=1.116; CI: 1.006–1.236, p=0.037).

Conclusion

We observed a significant association between combined events in the long-term follow-up with initial LVEF, LV systolic and diastolic volume indexes, T2 hypersignal and the presence of mid and basal anteroseptal and lateral apical late enhancement. Already related to the increase in LVEF in evolutionary CMR, we observed a significant association with initial LVEF, end systolic volume of the right ventricle, LV systolic and diastolic diameters, LV systolic and diastolic volume indexes and right ventricle systolic volume index.

Funding Acknowledgement

Type of funding sources: Public Institution(s). Main funding source(s): FAPESP

Heart failure and sepsis: practical recommendations for the optimal management

Acute heart failure (AHF) is a common clinical challenge that a wide spectrum of physicians encounters in every practice. In many cases, AHF is due to decompensation of chronic heart failure. This decompensation may be triggered by various reasons, with sepsis being a notable one. Sepsis is defined as a life-threatening organ dysfunction caused by the dysregulated host response to infection and is associated with a very high mortality, which may reach 25%. Alarmingly, the increase in the mortality rate of patients with combined cardiac dysfunction and sepsis is extremely high (may reach 90%). Thus, these patients need urgent intervention. Management of patients with AHF and sepsis is challenging since cornerstone interventions for AHF may be contraindicated in sepsis and vice versa (e.g., diuretic treatment). Unfortunately, no relevant guidelines are yet available, and treatment remains empirical. This review attempts to shed light on the intricacies of the available interventions and suggests routes of action based on the existing bibliography.

Growth differentiation factor-15 significantly improves prognostication of short-term mortality in acute settings

Introduction

Acute dyspnoea is one of the most common complaints in the emergency department with a diverse range of diagnoses and a vastly different prognosis. There is still a need for a reliable biomarker for prognostication in acute settings.

Purpose

To evaluate and compare the role of three biomarkers, which represent different pathophysiological pathways, in stratifying short-term mortality risk in patients admitted to the emergency department with acute dyspnoea.

Methods

Prospective observational cohort enrolled 1455 acute dyspnoea patients admitted to emergency departments of two university centres from 2015 to 2017. The present study included 867 (71 [62–79] years, 40% female) patients who had N-terminal pro–B-type natriuretic peptide (NT-proBNP), high-sensitive troponin T (hsTnT) and growth differentiation factor 15 (GDF-15) measured at admission. Clinical variables known to affect prognosis (age, gender, use of beta blockers, angiotensin-converting-enzyme inhibitors/angiotensin II receptor blockers at admission, history of chronic heart failure, diabetes and coronary artery disease, systolic blood pressure, heart rate, creatinine, sodium, haemoglobin and C-reactive protein levels) were used to build a baseline model for all-cause 3-month mortality risk prediction in acute dyspnoea [“Clinical score”]. The ability of each biomarker to improve risk prediction on top of the Clinical score was evaluated by area under the curve (AUC), comparing AUC of the model with the biomarker to the AUC of the clinical model alone using the DeLong test. The clinical benefit in risk prediction of adding a biomarker was further assessed by reclassification analysis, including the net reclassification improvement (NRI) and the integrated discrimination index (IDI). The statistical analyses were performed using R statistical software. P value of &lt;0.05 was considered statistically significant.

Results

Mortality at 3 months was 16% (143 patients). The AUC of the Clinical score was 0.787 [95% confidence interval (CI) 0.735 - 0.836]. GDF-15 was the only biomarker that significantly improved risk prediction on top of the Clinical score. The results of the individual biomarkers are displayed in Figure 1A. GDF-15 significantly reclassified 3-month mortality risk with an NRI of 32% [13%-51%] and an IDI of 0.026 [0.011–0.042]; other biomarkers had lower reclassification rates (Figure 1B).

Conclusions

This study demonstrates that the cellular stress biomarker GDF-15 taken at admission provides a substantial additive value to conventional clinical parameters for risk stratification in patients with acute dyspnoea.

Figure 1

Funding Acknowledgement

Type of funding source: Public grant(s) – National budget only. Main funding source(s): The work was supported by the Research Council of Lithuania (Lietuvos mokslo taryba), grant Nr. MIP-049/2015 and approved by Lithuanian Bioethics Committee, no. L-15-01. Roche Diagnostics supported the research with measurement of NT-proBNP, high-sensitivity troponin T and growth differentiation factor-15

Imbalance between procoagulant factors and natural coagulation inhibitors contributes to hypercoagulability in the critically ill COVID-19 patient: clinical implications

OBJECTIVE

Coronavirus Disease-2019 (COVID-19) predisposes patients to thrombosis which underlying mechanisms are still incompletely understood. We sought to investigate the balance between procoagulant factors and natural coagulation inhibitors in the critically ill COVID-19 patient and to evaluate the usefulness of hemostasis parameters to identify patients at risk of venous thromboembolic event (VTE).

PATIENTS AND METHODS

We conducted an observational study recording VTEs defined as deep vein thrombosis or pulmonary embolism using lower limb ultrasound (92% of the patients), computed tomography pulmonary angiography (6%) and both tests (2%). We developed a comprehensive analysis of hemostasis.

RESULTS

Ninety-two consecutive mechanically ventilated COVID-19 patients (age, 62 years [53-69] (median [25th-75th percentiles]); M/F sex ratio, 2.5; body-mass index, 28 kg/m2 [25-32]; past hypertension (52%) and diabetes mellitus (30%)) admitted to the Intensive Care Unit (ICU) from 03/11/2020 to 5/05/2020, were included. When tested, patients were receiving prophylactic (74%) or therapeutic (26%) anticoagulation. Forty patients (43%) were diagnosed with VTE. Patients displayed inflammatory and prothrombotic profile including markedly elevated plasma fibrinogen (7.7 g/L [6.1-8.6]), D-dimer (3,360 ng/mL [1668-7575]), factor V (166 IU/dL [136-195]) and factor VIII activities (294 IU/dL [223-362]). We evidenced significant discrepant protein C anticoagulant and chromogenic activities, combined with slightly decreased protein S activity. Plasma D-dimer >3,300 ng/mL predicted VTE presence with 78% (95%-confidence interval (95% CI), 62-89) sensitivity, 69% (95% CI, 55-81) specificity, 66% (95% CI, 51-79) positive predictive value and 80% (95% CI, 65-90) negative predictive value [area under the ROC curve, 0.779 (95%CI, 0.681-0.859), p=0.0001].

CONCLUSIONS

Mechanically ventilated COVID-19 patients present with an imbalance between markedly increased factor V/VIII activity and overwhelmed protein C/S pathway. Plasma D-dimer may be a useful biomarker at the bedside for suspicion of VTE.

Unusual arterial thrombotic events in Covid-19 patients

Introduction

COVID-19 infection is commonly complicated with pro-thrombotic state and endothelial dysfunction. While several studies reported a high incidence of venous thromboembolic events. The occurrence of arterial thromboses are yet rarely described and could be underestimated.

Objectives

To describe the clinical and biological characteristics of COVID-19 patients presenting with an associated arterial thromboembolic event.

Material and methods

We performed a retrospective multicentric study in 3 centers between France and Italy. All patients with a confirmed SARS-CoV-2 infection and arterial thromboembolic events were included in the analysis.

Results

From March 8th to April 25th 2020, we identified 20 patients (24 events) with arterial thromboembolic events over 209 admitted patients (9.6%) with severe COVID-19 infection. Arterial thrombotic events included acute coronary occlusions (n = 9), stroke (n = 6), limb ischemia (n = 3), splenic infarcts (n = 3), aortic thrombosis (n = 2) and occlusive mesenteric ischemia (n = 1). At the time of the event, 10/20 (50%) of patients received thromboprohylaxis, 2/20 (10%) were receiving treatment dose anticoagulation and 5/20 (25%) were receiving antiplatelet therapy.

Conclusion

Our observations suggest that serious arterial thrombotic events might occur in Covid-19 patients. However, the exact incidence of such events and the best way to prevent them yet remains to be investigated.

•

SARS-CoV-2 infection is commonly complicated with pro-thrombotic state and endothelial dysfunction.

•

High rate of venous thromboembolic events have been described among patients suffering from Covid-19. However, arterial thromboses are yet rarely described in this setting.

•

We report a total of 20 cases of COVID-19 patients suffering from arterial thromboses.

•

Our observations suggest that arterial thrombotic events might occur in Covid-19 patients and could be underestimated.

Infections fongiques invasives chez le patient admis en réanimation avec le COVID-19 : une cohorte rétrospective

Introduction

Matériels et méthodes

Résultats

Conclusion

Using velocity-pressure loops in the operating room: a new approach of arterial mechanics for cardiac afterload monitoring under general anesthesia

Cardiac afterload is usually assessed in the ascending aorta and can be defined by the association of peripheral vascular resistance (PVR), total arterial compliance (Ctot), and aortic wave reflection (WR). We recently proposed the global afterload angle (GALA) and β-angle derived from the aortic velocity-pressure (VP) loop as continuous cardiac afterload monitoring in the descending thoracic aorta. The aim of this study was to 1) describe the arterial mechanic properties by studying the velocity-pressure relations according to cardiovascular risk (low-risk and high-risk patients) in the ascending and descending thoracic aorta and 2) analyze the association between the VP loop (GALA and β-angle) and cardiac afterload parameters (PVR, Ctot, and WR). PVR, Ctot, WR, and VP loop parameters were measured in the ascending and descending thoracic aorta in 50 anesthetized patients. At each aortic level, the mean arterial pressure (MAP), cardiac output (CO), and PVR were similar between low-risk and high-risk patients. In contrast, Ctot, WR, GALA, and β-angle were strongly influenced by cardiovascular risk factors regardless of the site of measurement along the aorta. The GALA angle was inversely related to aortic compliance, and the β-angle reflected the magnitude of wave reflection in both the ascending and descending aortas (P < 0.001). Under general anesthesia, the VP loop can provide new visual insights into arterial mechanical properties compared with the traditional MAP and CO for the assessment of cardiac afterload. Further studies are necessary to demonstrate the clinical utility of the VP loop in the operating room.NEW & NOTEWORTHY Our team recently proposed the global afterload angle (GALA) and β-angle derived from the aortic velocity-pressure (VP) loop as continuous cardiac afterload monitoring in the descending thoracic aorta under general anesthesia. However, the evaluation of cardiac afterload at this location is unusual. The present study shows that VP loop parameters can describe the components of cardiac afterload both in the ascending and descending thoracic aorta in the operating room. Aging and cardiovascular risk factors strongly influence VP loop parameters. The VP loop could provide continuous visual additional information on the arterial system than the traditional mean arterial pressure and cardiac output during the general anesthesia.

2371Electrocardiographic features and their echocardiographic correlates in peripartum cardiomyopathy based on the EURObservational registry on PPCM

Background

Cardiac disease remains an important cause of maternal morbidity and mortality globally. Peripartum cardiomyopathy (PPCM), defined as heart failure secondary to left ventricular (LV) systolic dysfunction in previously healthy women towards the end of pregnancy or up to five months following delivery, can result in cardiogenic shock due to severe LV dysfunction or arrhythmias leading to sudden cardiac death. Cardiac electrical activity and its relationship to cardiac dysfunction have not yet been interrogated in large multi-centre studies.

Purpose

This study aimed to identify the ECG abnormalities associated with PPCM; their relationship with echocardiographic structural and functional abnormalities and explore regional and ethnic differences in ECG features.

Methods

We included the first 411 patients enrolled into the EURObservational PPCM registry (EORP). Baseline demographic, clinical and echocardiographic data were collected. ECGs were analysed for rate; rhythm; QRS width, axis and morphology; and QTc interval.

Results

Mean age of the women (from >40 countries) was 30.7±6.4 years. More than two thirds of patients presented with NYHA class III or IV (with no regional differences). The median QRS rate was 102bpm (IQR 87–117). More than half presented with sinus tachycardia (QRS rate >100bpm), whereas atrial fibrillation was rare (2.27%). The mean QRS width was 90.1ms ±21.5, with regional differences (ESC 93.8ms ±21.7 vs. non-ESC 86.8ms ±20.8, P<0.001). Left bundle branch block (LBBB) was reported in 9.30% with no regional or ethnic differences. Left ventricular hypertrophy (LVH) was present in a quarter of the cohort, and more prevalent amongst African (59.62%) and Asian (23.17%) than Caucasians (7.63%, P<0.001). The median QTc by Bazett was 456.7ms (IQR 409–490.7) and almost half (47.11%) had prolonged QTc (>460ms). The median LVEDD was 60mm (IQR 55–65) on echocardiography. Compared with their Asian and Caucasian counterparts, African patients were more likely to have LV dilatation (LVEDD>53mm: 70.11%, 79.31% and 89.42% respectively; P=0.004). The median LV ejection fraction (LVEF) was 32.50% (IQR 25–39) with no significant regional or ethnic differences. Sinus tachycardia predicted poor systolic function (OR 1.85 [95% CI 1.20–2.85], p=0.006). LVEF <35% was associated with a significantly higher QRS rate (median rate 107 vs. 98bpm, p=0.002). Women with LVEDD ≥53mm had a longer mean QRS duration (92.0±22.4 vs. 82.4±15.4ms, p<0.001) and frequency of LBBB (11.15% vs 1.54%, p=0.016). LBBB was a predictor of LVEDD >53mm (sensitivity 11.15%; specificity 98.46%; PPV 97.14%; NPV 19.10%; OR 8.02 [95% CI 1.08–59.66], p=0.042).

Conclusion

Patients with PPCM commonly present with sinus tachycardia, LVH, and/or prolonged QTc interval on their ECG. Wide QRS and/or LBBB, were associated with LVEDD>53mm. Sinus tachycardia, however, was associated with LVEF<35%. Risk of arrhythmia in those with prolonged QTc remains to be ascertained.

Acknowledgement/Funding

Heart Failure Association of the ESC

P1648Prevalence of increased filling pressure assessed by echocardiography in acute dyspnoea cohort

Introduction

Acute heart failure (AHF) is frequently associated with congestion leading to elevation of cardiac filling pressure. The present study investigates echocardiographic parameters of diastolic function in patients with AHF or non-AHF aetiology of acute dyspnoea.

Purpose

To determine the patterns of diastolic dysfunction in different profiles of acute dyspnoea.

Methods

Prospective multicentre observational study included 1455 acutely dyspnoeic patients in emergency departments from 2015 to 2017. Echocardiography was performed during the first 48 hours in 452 (31%) patients assessing left ventricular (LV) parameters. They were compared in four patient profiles based on dyspnoea cause and history of chronic HF (CHF): 1) AHF; 2) acute coronary syndrome without adjudicated AHF (Non-AHF+ACS); 3) non-AHF with CHF (Non-AHF+CHF); 4) other non-AHF patients (Non-AHF+other). Data were analysed using R statistical package.

Results

Significant differences in LV morphology and function were observed in the groups (Table 1). Increased LV filling pressure (E/E' >13) was found in most of AHF and Non-AHF+ACS patients, and in around 1/4 of Non-AHF+CHF group. Furthermore, more pronounced left-sided remodelling was observed in the first two groups. 1/3 of AHF patients had restrictive pattern of LV filling. Normal filling pressure dominated in Non-AHF+CHF and Non-AHF+other subgroups.

LV parameters in acute dyspnoea profiles Parameter AHF (n=291) Non-AHF + ACS (n=43) Non-AHF + CHF (n=44) Non-AHF + other (n=74) p value Age, years 71 [62–78] 72 [64–78] 71 [65–80] 68 [56–74] 0.045 LVEF, % 38 [25–55] 47 [32–55] 55 [45–55] 55 [50–55] <0.001 LV MMI, g/m2 126 [104.6–150.4] 99.1 [82.9–124] 94.4 [78.3–108.6] 79.6 [70.7–99.4] <0.001 LAVi, cm3 61.7 [50.9–81.1] 40.4 [35.5–46.8] 43.2 [39.8–58.9] 37.2 [32.6–43.8] <0.001 E/E' >13, % 57.7% 57.1% 23.1% 2.9% <0.001 E/E' <10, % 23.4% 38.1% 76.9% 70.6% <0.001 E/A >2, % 34.4% 14.8% 12.5% 3.9% <0.001 E/A <1, % 30.3% 59.3% 66.7% 74.5% <0.001 LVEF, left ventricular ejection fraction; LVdd, left ventricular diastolic diameter; LV MMI, left ventricular myocardial mass index; LAVi, left atrial volume index; AHF, acute heart failure; ACS, acute coronary syndrome; CHF, chronic heart failure.

Conclusions

Our data confirm the predominance of an increased cardiac filling pressure in acute heart failure patients, differently from chronic heart failure patients admitted due to other causes of dyspnoea. Patients with dyspnoea due to acute coronary syndrome frequently demonstrate elevated left-sided filling pressure.

Acknowledgement/Funding

The work was supported by the Research Council of Lithuania, grant Nr. MIP-049/2015 and approved by Lithuanian Bioethics Committee, Nr. L-15-01.

P785Unplanned readmissions after discharge increases risk of death in acute dyspnoea patients: non-cardiac is as severe as cardiac causes

Introduction

Acute dyspnoea is a major reason for admission to the emergency department and has been associated with high rates of readmission and mortality. However, the association of readmission with mortality risk has not been widely studied in patients with acute dyspnoea.

Purpose

To determine whether unplanned readmission during first 6 months after discharge is associated with greater risk of death within 1 year in patients with acute dyspnoea.

Methods

Derivation cohort consisted of 1371 patients from the prospective observational study, which enrolled acute dyspnoea patients admitted to emergency departments of two university centres from 2015 to 2017 and discharged alive from the hospital. Cox regression analysis compared 1-year risk of death between readmitted vs. non-readmitted patients in the first 6 months after discharge. In addition, we studied this association in 1986 patients from a multicentre validation cohort, which included acute dyspnoea patients from 2006 to 2014. Sensitivity analysis was done in the subgroups divided by cause of index admission (acute heart failure [AHF] and non-AHF) and cause of the first readmission (cardiovascular [CV] or non-CV). The statistical analyses were performed using R statistical software. P value of <0.05 was considered statistically significant.

Results

In the derivation cohort 666 (49%) of patients were readmitted at 6 months and 282 (21%) died in 1 year. Readmitted patients died more frequently than non-readmitted patients (211 [32%] vs. 71 [10%], respectively, p<0.001). All-cause 6-month readmission was associated with an increased 1-year risk of death in a multivariate analysis in both the derivation cohort (adjusted hazard ratio (aHR) 3 [95% confidence interval (CI) 2.2–4], p<0.001) and the validation cohort (aHR 1.8 [95% CI 1.4–2.2], p<0.001). Moreover, deleterious effect of readmission on 1-year survival was equally observed in AHF and non-AHF patients, independent of whether the reason of first readmission was cardiovascular or non-CV, in both study cohorts. The results are displayed in Figure 1.

Figure 1. Main results of the study

Conclusions

Our data demonstrates that readmission is associated with a markedly increased risk of death within 1 year in patients presenting to the emergency department with acute dyspnoea. Furthermore, the detrimental relationship between outcomes is similar in non-cardiac and cardiac causes.

Acknowledgement/Funding

The work was supported by the Research Council of Lithuania, grant Nr. MIP-049/2015 and approved by Lithuanian Bioethics Committee, Nr. L-15-01.

P5412Prevalence of RV failure assessed by echocardiography in acute dyspnoea cohort

Introduction

Recently more attention has been placed on right ventricle (RV) parameters in acute settings. The present study investigates echocardiographic RV parameters in patients with acute heart failure (AHF) or non-AHF acute dyspnoea.

Purpose

To determine the patterns of RV injury in different profiles of acute dyspnoea.

Methods

Prospective multicentre observational study included 1455 acutely dyspnoeic patients from 2015 to 2017. RV focused echocardiography was performed during the first 48 hours in 452 (31%) patients. They were compared in three patient profiles based on cause of dyspnoea and history of chronic HF (CHF): 1) AHF; 2) Non-AHF with CHF (Non-AHF+CHF); 3) other non-AHF patients (Non-AHF+other).

Results

Significant differences in RV morphology and function were observed in the study groups (Table 1). RV global function assessed by tricuspid annular plane systolic excursion (TAPSE) and RV longitudinal shortening was mostly affected in AHF patients. This was accompanied by more enlarged RV and increased right atrial pressure (RAP), assessed by the inferior vena cava diameter and respiratory collapse. Less severely impaired RV function and increased RAP were also observed in non-AHF+CHF patients indicating RV involvement in the chronic disease. Normal RV parameters dominated in Non-AHF+other group, however pulmonary artery systolic pressure >40 mmHg was observed in all profiles, suggesting similar severity of pulmonary hypertension in cardiac or pulmonary causes of acute dyspnoea.

Table 1. RV parameters in acute dyspnoea profiles Parameter AHF (n=291) Non-AHF + CHF (n=73) Non-AHF + other (n=88) p value LVEF, % 38 [25–55] 50 [40–55] 55 [50–55] <0.001 RV basal diameter, cm 4.5 [3.9–5.2] 4 [3.5–4.5] 4 [3.5–4.55] <0.001 TAPSE, cm 1.5 [1.2–1.8] 1.8 [1.6–2] 2 [1.5–2.4] <0.001 RV free wall strain, -% −15.3 [−19; −11.24] −19.3 [−24.5; −15.78] −23 [−24.5; −19.69] <0.001 Entire RV strain, -% −12.03 [−15.17; −9.11] −16.4 [−19.31; −10.5] −18 [−18.75; −16.9] <0.001 PASP >40, % 66% 51% 50% 0.039 IVC diameter, cm 2.4 [2–2.8] 2 [1.7–2.4] 1.8 [1.4–2.3] <0.001 IVC collapse, % 34.9 [19.7–50.2] 44.1 [28.7–59.3] 52.6 [35–72.7] <0.001 LVEF, left ventricular ejection fraction; RV, right ventricle; TAPSE, tricuspid annular plane systolic excursion; PASP, pulmonary artery systolic pressure; IVC, inferior vena cava; AHF, acute heart failure; CHF, chronic heart failure.

Conclusions

Our data confirm more pronounced acute failure of right ventricle in acute heart failure patients than in chronic heart failure patients admitted due to other causes of dyspnoea. Pulmonary hypertension is present in a majority of the acute dyspnoea patients.

Acknowledgement/Funding

The work was supported by the Research Council of Lithuania, grant Nr. MIP-049/2015 and approved by Lithuanian Bioethics Committee, Nr. L-15-01.

P3506A novel urinary biomarker predicts 1 year mortality after discharge from Intensive Care

Rationale

Tools reflecting molecular processes predicting death after discharge from intensive care unit (ICU) are currently unavailable.

Objectives

To construct a urinary proteomic biomarker predicting 1-year post-ICU mortality.

Methods

In 1081 patients, enrolled in the French and European Outcome Registry in Intensive Care Unit (NCT01367093), the urinary proteome was measured at ICU discharge using capillary electrophoresis couple with mass spectrometry along with clinical variables; circulating biomarkers (NT-proBNP, hsTnT, biologically active adrenomedullin, soluble ST2, and NGAL) and urinary albumin were available in 886 patients.

Measurements and main results

In the discovery sample (60/256 deaths/survivors), support vector machine modelling identified ACM150, mainly consisting of collagen fragments, yielding an AUC of 0.676 (95% CI, 0.615–0.737). In the replication sample (151/608 deaths/survivors), AUC was 0.704 (0.659–0.750). While accounting for center and clinical risk factors, the hazard ratios in all available patients were 1.27 (1.18–1.37) for ACM150 (+1 SD), 1.20 (1.16–1.33) for the Charlson score (+1 point), and ≥1.30 (P≤0.0071) for the other biomarkers (+ 1 SD). Model performance assessed by adding ACM150 to a basic model including the aforementioned covariables, the Charlson score or any other biomarker confirmed the prognostic accuracy of ACM15 with significant increases (P≤0.0038) in integrated discrimination (≥ +0.50) and net reclassification improvement (≥ +53.7) and AUC (≥ +0.037). Interactome mapping (STRING) based on 72 sequenced peptides and 25 parental proteins gravitated around collagen nodes.

Conclusions

ACM150 is a urinary proteomic classifier predicting 1-year post-ICU mortality over and beyond other biomarkers and reflects dysregulation of collagen turnover as underlying pathophysiological process.

P6354Decreased beneficial effects of oral heart failure medications in patients with acute decompensated heart failure and hyperglycemia: results from an international observational cohort

Background

Hyperglycemia is common, regardless of diabetes mellitus (DM), and is associated with increased mortality in patients with acute heart failure (AHF). Current oral heart failure (HF) medications improve the outcome in patients with AHF. However, the relationships between HF medications, admission glucose levels, and prognosis in AHF patients remained unknown.

Purpose

This study sought to investigate the effect of oral HF medications on relationships between hyperglycemia at admission and 1-year all-cause mortality in patients with AHF.

Methods

From the GREAT (Global Research on Acute Conditions Team) registry, 13840 patients presenting with AHF whose admission glucose levels were available were included and followed up for 1-year all-cause mortality. Hyperglycemia was defined as a glucose levels of ≥7 mmol/L for patients without history of DM and ≥10 mmol/L for those with history of DM. Patients with hypoglycemia (defined as a glucose levels of ≤4 mmol/L, n=193, 1.4%) were excluded in this analysis.

Results

There were 6418 (%) patients with hyperglycemia and 7229 (%) patients with normoglycemia. One-year mortality was higher in patients with hyperglycemia than those with normoglycemia (1911 [30%] and 1821 [25%], respectively). Even after adjustment, the risk for 1-year mortality was significantly higher in hyperglycemia (HR 1.14, 95%-CI 1.04–1.26, P=0.008) compared with normoglycemia. Detrimental effects of hyperglycemia on 1-year mortality were more severe in de novo AHF patients than in patients with history of HF (p for interaction 0.004). Oral HF medications (beta blockers and/or angiotensin converting enzyme inhibitors/angiotensin receptor blockers) at discharge were effective in AHF patients with normoglycemia regardless of history of HF. Oral HF medications at discharge are very effective in de novo AHF patients with hyperglycemia and less effective in acute decompensated HF patients with hyperglycemia (Figure).

HF medications and 1-year mortality

Conclusions

Hyperglycemia at admission is associated with increased risk for 1-year mortality. Current oral HF medications are effective in most of subgroups, though they were less effective in patients with acute decompensated HF and hyperglycemia. These patients might need more aggressive therapies to improve outcomes.

Acknowledgement/Funding

This work was supported by a research fellowship from Japan Heart Foundation (E.A.)

P3530Right ventricular remodelling as part of the cardiorenal syndrome

Introduction

The relationship of glomerular filtration rate (GFR) to the remodelling of right (RV) and left ventricle (LV) in acute settings is not well studied.

Purpose

To investigate the relationship of renal function to LV and RV echocardiographic parameters in acute dyspnoea patients.

Methods

Echocardiography was performed in 438 patients with acute dyspnoea in prospective multicentre observational study during the first 48 hours after presentation. Echocardiographic quantification included the following parameters: tricuspid annular plane systolic excursion (TAPSE), velocity of the tricuspid annular systolic motion (RV S'), RV fractional area change (FAC), RV basal diameter, entire RV strain, strain of free RV wall, inferior vena cava (IVC) and LV parameters. Patients were divided into groups according to the GFR (ml/min/1.73m2): Normal or mildly decreased GFR (≥60), moderately decreased GFR (30–59) and severely decreased GFR (≤29). Data were analysed using One-way ANOVA.

Results

Mean GFR in this cohort was 63±26 ml/min/1.73m2. The severity of RV remodelling and venous congestion was significantly different in three subgroups of renal dysfunction.

The distribution of parameters Normal or mildly decreased GFR Moderately decreased GFR Severely decreased GFR N=233 (52%) N=166 (38%) N=39 (9%) TAPSE (cm) 1.8±0.3 1.5±0.5 1.4±0.4* RV FAC (%) 41±15 38±13 29±12*§ RV S' (cm/s) 11.0±3.3 9.7±3.4 9.1±4.2*§ Strain of RV free wall (%) −19±7 −16±6 −13±5*§ Strain of entire RV (%) −15±5 −12±5 −11±5* IVC collapsibility (%) 46±26 38±25 36±28* PASP (mmHg) 44±16 48±15 51±16*§ LVEF (%) 43±15 38±14* 46±16§ LV internal diameter in diastole (cm) 4.6±2.1 4.3±2.3 4.4±2.0 LV myocardial mass index (g/m2) 119±40 122±39 114±25 NT-proBNP (ng/l) 1485 [471; 3409] 5282 [1499; 10695] 6677 [1971; 15522]*§ Troponin T (ng/l) 55±18 92±19* 95±12* Admitted due to AHF (%) 59 72* 67 Admitted due to pulmonary causes (%) 22 21 28 *p<0.05 compared to GFR ≥60 ml/min/1.73m2, §p<0.05 compared to GFR (30–59 ml/min/1.73m2).

Conclusions

The progressive impairment of renal function coupled with larger congestion is associated with a significant right-sided but not left-sided ventricular remodelling in acute dyspnoea patients.

Acknowledgement/Funding

The work was supported by the Research Council of Lithuania, grant Nr. MIP-049/2015 and approved by Lithuanian Bioethics Committee, Nr. L-15-01.

Pressure guided surgery of compartment syndrome of the limbs in burn patients

Compartment syndrome is a serious complication of high voltage electrical burns, limb carbonization and deep circular burns with delayed escharotomy. Without treatment, ischemic tissue damage leads to irreversible necrosis. Treatment is emergency surgical decompression. The burned patient is usually not searchable and cannot always be readily examined because of bulky dressings; diagnosis of compartment syndrome is always hard to make. The pressure transducer used in central arterial catheters is easy available. We used it to measure pressure in muscular compartments. We measured compartment pressure three times at different depths in all cases of electrical burn, carbonization and deep circumferential burns with delayed escharotomy. We also took the pressure in the uninjured limb. The pressure assessment device was composed of a blood pressure transducer commonly used in arterial catheters for arterial pressure monitoring with three connecting branches. The first branch was connected to the 'arterial pressure exit' in the monitoring device. The second, an IV tube, was connected to one litre of physiological serum in a pressure bag inflated to 200 mmHg. The third, also an IV tube with a sterile extension cable, was directly connected to an 18G standard straight needle to be inserted in the tissues for which interstitial pressure had to be measured. In patients with thermal burns, we measured pressure before and after escharotomy. Threshold intracompartmental pressure was 35 mmHg. We carried out pressure assessment of all muscular compartments during and at the end of surgery. The pressure transducer provides a pressure value in all muscular compartments with a time of installation and measuring of less than 5 minutes. Sensitivity is measured at +/- 1 mmHg. Operation is simple, non-operator dependent, and accessible to medical and paramedic teams. The pressure transducer allows accurate diagnosis of early or established compartment syndrome. It requires no additional equipment and its application does not delay therapeutic management. Its use helps with fasciotomy decision, especially after escharotomy, guides the surgeon in the exploration of different compartments and verifies the effectiveness of surgery.

Incidence, risk factors, and outcome of multidrug-resistant Acinetobacter baumannii acquisition during an outbreak in a burns unit

BACKGROUND

Multidrug-resistant Acinetobacter baumannii (MR-AB) can cause outbreaks in a burns unit.

AIM

To study the incidence, risk factors and outcome of MR-AB colonization during an outbreak.

METHODS

A prospective study was conducted from April to November 2014 in a burns unit in Paris. Weekly surveillance cultures of patients and their environment were performed. MR-AB acquisition, discharge, or death without MR-AB colonization were considered as competing events. To identify risk factors for colonization, baseline characteristics and time-dependent variables were investigated in univariate and multivariate analyses using Cox models. MR-AB strains were genotypically compared using multi-locus sequence typing.

FINDINGS

Eighty-six patients were admitted in the burns unit during the study period. Among 77 patients without MR-AB colonization at admission, 25 (32%) acquired MR-AB with a cumulative incidence of 30% at 28 days (95% CI: 20-40). Median time to MR-AB acquisition was 13 days (range: 5-34). In multivariate analysis, risk factors for MR-AB acquisition were ≥2 skin graft procedures performed [hazard ratio (HR): 2.97; 95% confidence interval (CI): 1.10-8.00; P = 0.032] and antibiotic therapy during hospitalization (HR: 4.42; 95% CI: 1.19-16.4; P = 0.026). A major sequence type of MR-AB (ST2) was found in 94% and 92% of patients and environmental strains, respectively, with all strains harbouring the blaOXA-23 gene. MR-AB colonization increased length of hospitalization (HR: 0.32; 95% CI: 0.17-0.58; P = 0.0002) by a median of 12 days.

CONCLUSION

A high incidence of MR-AB acquisition was seen during this outbreak with most strains from patients and their environment belonging to single sequence type. MR-AB colonization was associated with more skin graft procedures, antibiotic use, and prolonged hospitalization.

Norepinephrine reduces arterial compliance less than phenylephrine when treating general anesthesia-induced arterial hypotension

INTRODUCTION

During general anesthesia, arterial hypotension is frequent and may be an important contributor to perioperative morbidity. We assessed the effect of a 5 μg bolus of Norepinephrine (NA) when compared with 50 μg bolus of Phenylephrine (PE) administered to treat hypotension during maintenance anesthesia, on MAP, derived cardiac output and arterial stiffness parameters.

METHODS

Patients scheduled for a neurosurgical procedure under general anesthesia were prospectively included. Monitoring included invasive blood pressure, esophageal Doppler, and arterial tonometer used to estimate central aortic pressure with arterial stiffness parameters, such as augmentation index (Aix). After initial resuscitation, hypotensive episodes were corrected by a bolus administration of NA or PE in a peripheral venous line.

RESULTS

There were 269 bolus administrations of vasopressors (149 NA, 120 PE) in 47 patients with no adverse effects detected. A decrease in stroke volume (SV) was observed with PE compared with NA (-18 ± 9% vs. -14 ± 7%, P < 0.001). This decrease was associated with an increase in Aix, which was greater for PE than for NA (+10 ± 8% vs. +6 ± 6%, P < 0.0001), and a decrease in total arterial compliance greater for PE compared to NA (Ctot = SV/Central Pulse Pressure) (-35 ± 9% vs. -29 ± 10%, P < 0.001).

DISCUSSION

This study suggests that 5 μg of NA administered as a bolus in a peripheral venous line could treat general anesthesia-induced arterial hypotension with a smaller decrease in SV and arterial compliance when compared to PE.

Upcoming Challenges in Multidisciplinary Heart Failure Management: Active Role of Future Clinical Pharmacists

In their article "Role and Value of Clinical Pharmacy in Heart Failure Management" Stough and Patterson discuss the potential benefits that clinical pharmacists could bring to heart failure management. They mostly focused on chronic stable heart failure. The objective of this commentary is to provide future perspectives for clinical pharmacist involvement in multidisciplinary heart failure team in the management of patients admitted in emergency for decompensation of their heart condition.

European Society of Cardiology - Acute Cardiovascular Care Association position paper on safe discharge of acute heart failure patients from the emergency department

Heart failure is a global public health challenge frequently presenting to the emergency department. After initial stabilization and management, one of the most important decisions is to determine which patients can be safely discharged and which require hospitalization. This is a complex decision that depends on numerous subjective factors, including both the severity of the patient's underlying condition and an estimate of the acuity of the presentation. An emergency department observation period may help select the correct option. Ideally, during an observation period, risk stratification should be carried out using parameters specifically designed for use in the emergency department. Unfortunately, there is little objective literature to guide this disposition decision. An objective and reliable definition of low-risk characteristics to identify early discharge candidates is needed. Benchmarking outcomes in patients discharged from the emergency department without hospitalization could aid this process. Biomarker determinations, although undoubtedly useful in establishing diagnosis and predicting longer-term prognosis, require prospective validation for emergency department disposition guidance. The challenge of identifying emergency department acute heart failure discharge candidates will only be overcome by future multidisciplinary research defining the current knowledge gaps and identifying potential solutions.

Erratum to: 36th International Symposium on Intensive Care and Emergency Medicine: Brussels, Belgium. 15-18 March 2016

[This corrects the article DOI: 10.1186/s13054-016-1208-6.].

Propensity score estimators for the average treatment effect and the average treatment effect on the treated may yield very different estimates

Objective

Propensity score matching is typically used to estimate the average treatment effect for the treated while inverse probability of treatment weighting aims at estimating the population average treatment effect. We illustrate how different estimands can result in very different conclusions.

Study design

We applied the two propensity score methods to assess the effect of continuous positive airway pressure on mortality in patients hospitalized for acute heart failure. We used Monte Carlo simulations to investigate the important differences in the two estimates.

Results

Continuous positive airway pressure application increased hospital mortality overall, but no continuous positive airway pressure effect was found on the treated. Potential reasons were (1) violation of the positivity assumption; (2) treatment effect was not uniform across the distribution of the propensity score. From simulations, we concluded that positivity bias was of limited magnitude and did not explain the large differences in the point estimates. However, when treatment effect varies according to the propensity score (E[Y(1)–Y(0)|g(X)] is not constant, Y being the outcome and g(X) the propensity score), propensity score matching ATT estimate could strongly differ from the inverse probability of treatment weighting-average treatment effect estimate. We show that this empirical result is supported by theory.

Conclusion

Although both approaches are recommended as valid methods for causal inference, propensity score-matching for ATT and inverse probability of treatment weighting for average treatment effect yield substantially different estimates of treatment effect. The choice of the estimand should drive the choice of the method.