Abstract
Background:
The 2015 European Society of Cardiology non-ST-elevation myocardial infarction (NSTEMI) guidelines recommend angiography within 24 h in high-risk patients with NSTEMI. An organized STEMI-like approach with pre-hospital or immediate in-hospital triage for acute coronary angiography (CAG) may be of therapeutic benefit but it remains unknown whether the patients can be properly diagnosed in the pre-hospital setting. We aim to evaluate whether it is feasible to diagnose patients with NSTEMI in the pre-hospital phase or immediately upon admission.
Methods and results:
We randomized 250 patients to either acute or subacute CAG (i.e. <72 h of admission). Pre-hospital electrocardiogram acquisition and point-of-care troponin-T measurement ensured that 148 (59%) patients were identified already in the ambulance, whereas the remaining 102 (41%) patients were identified immediately after hospital admission. An acute coronary syndrome was verified in 215 (86%) and NSTEMI in 159 (64%) patients. The CAG rate was significantly higher in the acute CAG group (98% vs. 87%, p<0.001). A culprit lesion was identified in 74% and 64% of the patients underwent coronary revascularization: acute CAG group: 53% percutaneous coronary intervention, 5% hybrid, 7% coronary artery bypass grafting; conventional treatment: 48% percutaneous coronary intervention, 2% hybrid, 14% coronary artery bypass grafting, p=0.32. In patients randomized to acute CAG, time from randomization to CAG was 1.1 h; in patients randomized to subacute CAG it was two days. Time from randomization to initial revascularization was 1.3 h versus 2.4 days, and the median hospital stay was 4.0 days versus 4.5 days. Among patients randomized to subacute CAG, 17% crossed over to acute CAG and 5% developed STEMI before catheterization.
Conclusion:
Diagnosing NSTEMI patients in the pre-hospital phase or immediately upon hospital admission is feasible. Acute CAG may impact the mode of revascularization and is associated with earlier revascularization and shorter hospital stay. The clinical benefit of acute CAG in NSTEMI patients remains to be clarified.
Keywords
Introduction
The 2015 European Society of Cardiology guidelines for management of acute coronary syndrome (ACS) in patients presenting without persistent ST-segment elevation advise coronary angiography (CAG) within 24 hours in high-risk patients with non-ST-segment elevation myocardial infarction (NSTEMI). 1 The recommendation relies on data from meta-analyses showing that an early invasive strategy is associated with an improved outcome.2–4 Yet, it remains unknown whether high-risk patients with NSTEMI would benefit from an acute invasive strategy. Regardless of the future strategy, it is of great interest to investigate whether patients with NSTEMI may be identified acutely and admitted directly to invasive centres rather than to local non-CAG capable referral hospitals. This approach requires an ST-elevation myocardial infarction (STEMI)-like system with a 24/7 CAG service but its implementation is challenged by ambiguous electrocardiogram (ECG) findings in patients with NSTEMI. However, recent studies indicate that combining symptoms, ECGs and pre-hospital point-of-care troponin T (POC-cTnT) measurement may enable diagnosis of NSTEMI patients in the pre-hospital phase or immediately upon admission.5,6
The acute versus subacute angioplasty in patients with NON-ST-Elevation Myocardial Infarction (NONSTEMI) trial (clinicaltrials.gov: NCT01638806) was launched to investigate the application of a primary percutaneous coronary intervention-like strategy compared with standard care in patients with NSTEMI. In this paper we report the results of phase I of the NONSTEMI trial with the objectives to investigate: a) the feasibility of diagnosing patients with NSTEMI in the pre-hospital phase or immediately upon admission through combined use of point-of-care biomarker analysis and ECG; b) whether acute CAG changes the rate and mode of revascularization; c) whether acute CAG reduces the duration of the admission.
Methods
Study population and study region
In an open-label, randomized, two-centre study, we included patients with a prominent suspicion of NSTEMI based on a history of ongoing typical chest pain and either significant ST-depression or elevated POC-cTnT (above 50 ng/l). The study was initiated in June 2012 and is still on-going. We pre-specified a review of the first 250 included patients to clarify feasibility of diagnosis, duration of hospital admission and use of coronary revascularization. Patient number 250 was enrolled in July 2014. The pre-hospital emergency medical services included 120 pre-hospital emergency ambulances, one helicopter emergency medical service and 11 physician-staffed mobile emergency care units. They routinely performed the POC-cTnT analysis using the cardiac T assay on the Cobas h232 instrument (Roche Diagnostics, Mannheim, Germany) and recorded a 12-lead ECG in patients suspected of acute myocardial infarction (AMI).6,7 The POC-cTnT assay has a lower detection limit of 50 ng/l. Standard operating criteria for ECG acquisition and biomarker analysis included ongoing or recent chest pain, recent onset dyspnoea without known lung disease, or other clinical suspicion of AMI. The ECG was transmitted to and interpreted by the cardiologist on call at Aarhus University Hospital or Aalborg University Hospital, who performed a phone interview with the patient and the pre-hospital personnel. Alternatively, the POC-cTnT and ECG were obtained immediately upon hospital arrival. Patients fulfilling the inclusion criteria and giving written informed consent were randomized to either acute CAG or to routine medical treatment and CAG within 72 h (24 h if Global Registry of Acute Coronary Events (GRACE) score > 140) (Figure 1). During the last year of the study period, all patients randomized to subacute CAG were scheduled for CAG within 48 h as required by the guidelines of the Danish Health and Medicines Authority. Enrolment was made using permuted block randomization (one block per centre), in a 1:1 ratio, using web-based clinical trial management system ‘Trial Partner’ with stratification according to presence of diabetes and place of randomization (ambulance versus hospital). The patients were recruited from the catchment area of the two invasive cardiology centres serving the Central Denmark region (13,000 km2, 1.3 million inhabitants, Aarhus University Hospital) and North Denmark region (8000 km2, 0.6 million inhabitants, Aalborg University Hospital).
Patient flow. A point-of-care cardiac troponin T measurement was used to determine elevated cardiac biomarker levels. No patients withdrew consent before angiography or revascularization therapy. Preregistration of eligibility was not performed. No patients were lost to follow-up.
Patients randomized to acute CAG received 300 mg aspirin orally and 10,000 IE unfractionated heparin intravenously (7500 IE if weight was below 70 kg). Treatment with an adenosine diphosphate (ADP) receptor blocker before CAG was left at the discretion of the attending cardiologist. Patients assigned to routine treatment were treated according to local standards: aspirin (300 mg loading dose, 75 mg daily), oral ADP receptor blocker treatment (loading dose/daily: ticagrelor 180 mg/90 mg × 2, or clopidogrel 600 mg/75 mg × 1), and subcutaneous fondaparinux (2.5 mg daily). All other treatment followed routine care standards. Among patients randomized to conventional therapy, cross-over for acute CAG was allowed in patients with ongoing chest pain despite intravenous nitroglycerin infusion, unstable haemodynamic status or signs of STEMI. CAG was cancelled if an alternative diagnosis was evident.
In phase I of the NONSTEMI trial, the primary endpoint was in-hospital diagnosis of NSTEMI. AMI was diagnosed in accordance with the third universal definition of myocardial infarction using the 99th percentile as decisive cut-point to detect elevated levels of cardiac troponin (cTn), and relative changes of 20% (50% if the first measurement was below the 99th percentile) to detect a rise and/or fall.8,9 Unstable angina pectoris (UAP) was diagnosed in patients with obvious signs of ischaemia but without cTn level dynamics. The secondary endpoints were coronary revascularization and admission duration. We estimated that 140 patients should be randomized to document a reduction in the duration of hospitalization of 1.5 days assuming a five days admission (standard deviation: three days) in patients assigned to conventional therapy. The endpoints were adjudicated by an Endpoint Committee (see Acknowledgements) who had full access to the individual patient files.
Criteria for eligibility
Patients with a suspected AMI over 18 years but below 80 years of age having ECG diagnostics performed by telemedicine or immediately after hospital arrival, and who could be randomized within 60 min of hospital admission, were eligible for inclusion if they had ongoing, typical chest pain and: 1) ST-depression of at least 0.1mV in at least four contiguous leads/0.2 mV in at least two contiguous leads and/or 2) an elevated POC-cTnT above 50 ng/l. Patients were not eligible for inclusion if they had a heart rate above 120 beats/min, an alternative reason for elevated POC-cTnT, ST-depression due to left ventricle myocardial hypertrophy, renal dialysis treatment, previous coronary artery bypass grafting (CABG) or pregnancy. The study was performed in accordance with the Helsinki Declaration and approved by the Central Denmark Regional Ethical Committee and the Danish Data Protection Agency. The study was supervised and monitored by an internal data and safety committee.
Statistical analysis
The chi-squared test, Wilcoxon rank sum test and Kruskal–Wallis test were used to compare differences between the groups as appropriate. Categorical data are presented as absolute numbers (percentages). Continuous variables are presented as medians with interquartile ranges. We constructed Kaplan–Meier curves to investigate the association between treatment strategy and time to CAG, revascularization, and discharge. Comparison was made using log-rank statistics. Tests were considered statistically significant if p-values were below 0.05 (two-sided test). Data were analysed on an intention-to-treat basis. Statistical analyses were performed using Intercooled STATA version 14 (StataCorp LP, College Station, Texas, USA).
Results
Inclusion details
Among the 250 patients included, 122 were assigned to acute CAG and 128 to subacute, conventional treatment. Baseline characteristics are listed in Table 1. Admission cTn levels did not differ between groups. Aspirin treatment was the same in the two groups, but ADP receptor blockers (before or during CAG/percutaneous coronary intervention (PCI)), specifically treatment with ticagrelor, were more often avoided in patients in the acute CAG group. Randomization was performed while the patient was still in the ambulance in 148 patients (59%) while 102 patients (41%) were included within the first hour of arrival at the hospital (Figure 1). A total of 240 patients were included at Aarhus University Hospital (10 patients per month) and 10 at Aalborg University Hospital (one patient per month). The criteria used by the attending cardiologist for enrolment of patients in the study were: ST-depression in 81 patients (32%); ST-depression and an elevated POC-cTnT level above 50 ng/l in 42 patients (17 %); and an elevated POC-cTnT level alone in 127 patients (51%). However, admission high-sensitivity cTn levels were within the normal range in seven patients despite an elevated POC-cTnT measurement, and in three cases the POC-cTnT result was misinterpreted and falsely reported as elevated. In one patient, the admission cTn values were normal, but the original POC-cTnT result was not available for review from instrument data. In one patient, the result was falsely reported to be below 50 ng/l. In four patients, an elevated POC-cTnT was measured, but not registered by the enrolling doctor. Ultimately, at the time of enrolment, 163 patients (65%) patients had verified troponin T levels above 50 ng/l, 19 patients (8%) had verified troponin levels below 50 ng/l. In 68 patients (27%), POC-cTnT was not performed. We found no differences with regard to presence of elevated biomarker levels or presence of ST depression among the two treatment groups. The inclusion criteria used by the enrolling doctor were equally distributed in the two treatment groups (p = 0.98). However, the presence of both ST-depression and elevated POC-cTnT levels was more frequent upon arrival at the hospital than when enrolment took place in the ambulance with a borderline statistically significant p = 0.055 (Figure 1).
Baseline demographic data, clinical and biochemical characteristics at presentation, and treatment during admission. N = 250 patients, 122 assigned to acute coronary angiography and 128 assigned to conventional therapy.
p-values are shown for variables where the assigned treatment may alter the result: level of cardiac troponin; medical treatment.
IQR: interquartile range; ADP: adenosine diphosphate; LMWH: low-molecular-weight heparin.
Diagnoses
In total, 215 patients (86%) had ACS (acute CAG: 106 (87%); conventional: 109 (85%), p = 0.69). A final diagnosis of NSTEMI was confirmed by the Endpoint Committee in 159 patients (64%) with 84 patients (69%) in the acute CAG group and 75 patients (59%) in the conventional treatment group, p = 0.092 (Table 2). Another 29 patients (12%) were diagnosed with STEMI (acute CAG: 11 (9%); conventional: 18 (14%), p = 0.21) including eight patients (3.2%) who developed STEMI after inclusion while waiting for the scheduled CAG: two patients in the acute CAG group compared with six patients in the conventional treatment group, p=0.17. A total of 27 patients (11%) had UAP (acute CAG: 11 (9%); conventional: 16 (13%), p = 0.38). The remaining 35 patients (14%) had other diagnoses (Table 3). Figure 2 shows the final diagnoses grouped by inclusion criteria. An elevated POC-cTnT alone identified the highest proportion of patients with AMI (52%) and ACS (47%). Presence of both an elevated POC-cTnT and ST-depression had the best predictive value for AMI (86%), whereas ST-depression alone had the best predictive value for prediction of ACS (94%).
Data on angiography, coronary anatomy, coronary revascularization and final diagnoses in 250 patients.
TIMI: Thrombolysis In Myocardial Infarction; PCI: percutaneous coronary intervention, CABG: coronary artery bypass graft, NSTEMI: non-ST-segment elevation myocardial infarction; STEMI: ST-segment elevation myocardial infarction.
Diagnoses of 35 patients without acute coronary syndrome. Twenty-three patients had a condition with an associated troponin release, 12 patients had other diagnoses without troponin release.
Final diagnoses in 250 patients as confirmed by the Endpoint Committee, grouped by the inclusion diagnostic criteria used by the enrolling doctor.
Revascularization
A total of 231 patients (92%) underwent angiography, with a significantly higher rate in the acute CAG group (120 (98%)) than in the conventional group (111 (87%)) (p = 0.001). A culprit lesion was identified in 185 patients (74%), with 92 patients (75%) in the acute CAG group compared with 93 (73%) in the conventional treatment group, p = 0.59. Coronary revascularization of culprit lesions was performed in 161 (64%) patients with no difference between the groups (acute CAG: 79 (65%); conventional treatment: 82 (64%); p = 0.91) (Table 2). Among these, 126 patients (50%) had PCI performed (acute CAG: 65 (53%); conventional treatment: 61 (48%)), nine (4%) had a hybrid procedure performed (acute CAG: 6 (5%); conventional: 3 (2%)) and 26 (10%) underwent CABG (acute CAG: 8 (7%); conventional treatment 18 (14%)), p=0.32. Among the patients randomized for acute CAG, 69 (57%) underwent direct PCI. In the conventional treatment group, PCI of a culprit lesion was performed in relation to the scheduled angiography in 59 patients (46%). Significantly more patients had Thrombolysis In Myocardial Infarction (TIMI) flow grade 0 in the acute CAG group (24 patients; 20%) than in the conventional therapy group (13 patients; 10%) (p = 0.034). Performance of CAG before the intended time was done in 22 (17%) patients assigned to conventional therapy; 12 of these with NSTEMI (55%), eight with STEMI (36%) and two with other cardiac conditions causing elevated cTn levels (9%).
Timing of angiography
Time variables stratified according to assigned treatment and place of inclusion are shown in Table 4. Figure 3 shows the temporal cumulative rate of CAG, first revascularization procedure (PCI or CABG) and discharge from hospital in the total cohort. The median time from randomization to CAG in the acute CAG/conventional group was 1.1 h (0.8–1.5) versus 49.0 h (23.2–69.8), p < 0.001; and the median time to first revascularization procedure was 1.3 h (0.9–1.9) versus 57.4 h (22.8–84.8), p < 0.001. The median duration of the hospitalization in patients with NSTEMI, UAP or STEMI was reduced by 0.73 days (16%) in the acute CAG group compared with the conventional group (time to discharge 3.9 days (2.9–4.9) versus 4.7 days (4.0–5.9), p < 0.001).
Timing data grouped by randomized allocation and whether the patient was enrolled in the pre-hospital phase or upon admission to the hospital.
Temporal cumulative rate of the angiography procedure (a), first revascularization procedure (PCI or CABG) (b) and discharge from hospital (c).
Discussion
We report a number of important new findings in this open-labelled, two-centre, randomized feasibility study assessing an acute CAG strategy in high-risk patients with NSTEMI. First, the combination of chest pain and either significant ST-segment depression and/or elevated pre-hospital/first-admission POC-cTnT levels facilitated early diagnosis of patients of whom 87% had ACS, 74% had a culprit lesion, 64% underwent coronary revascularization. Sixty per cent of these were identified while still in the ambulance. Next, among those randomized to conventional therapy, 5% developed STEMI, and almost one in five crossed over and had acute CAG performed due to an unstable cardiovascular condition. Finally, we demonstrated that a significant acceleration of the entire hospital course could be achieved by employing an acute CAG strategy. Notably, patients were revascularized a median of 2.4 days earlier and the duration of admission was reduced by 16%.
Accelerated strategies have been tested in four previous studies of patients with NSTEMI.10–13 However, none of these studies used a STEMI-like referral approach with randomization for acute CAG in the pre-hospital phase. Thus, it might not be appropriate to compare the results from those studies with the results for the acute CAG group in our study. Nonetheless, some points relating to the current study are still relevant for comparison and will be discussed below.
First, we aimed to investigate the potential for a STEMI-like system of care in patients with NSTEMI, necessitating inclusion already at the point where a strong suspicion of NSTEMI was raised, but with the diagnosis not yet confirmed. Inclusion of patients without ACS is an inevitable risk in such a setting. In comparison, in the four previous studies, patients were only eligible for inclusion at the point where ACS was confirmed.10–13 This is the explanation why all patients were not diagnosed with ACS in our study. The identification of eligible patients and enrolment was left to the discretion of the attending cardiologist, who routinely, via telemedicine, interprets the pre-hospital ECG and decides on primary PCI triage in STEMI patients. Considering the equivocal signs and symptoms in patients with non-ST-elevation ACS, we find that the 87% ACS rate among the included patients is acceptable. In our study, significantly more patients in the acute CAG group had angiography performed, including 11% who did not have ACS. We acknowledge that in these patients, an acute CAG could pose an additional risk in terms of CAG related complications, and might also delay important alternative diagnostic and therapeutic procedures. One way to reduce the rate of unnecessary CAG in patients with a strong suspicion of NSTEMI might include a thorough initial clinical evaluation including acute echocardiography in the ambulance or in the catheterization laboratory at arrival to the CAG centre. However, an acute CAG strategy could prove to be an advantage to misclassified patients who actually have a STEMI. In fact, in our study 29 patients were ultimately adjudicated a STEMI diagnosis and 21 of these patients had signs of STEMI at the time of inclusion according to the Endpoint Committee. This may raise concern about the qualifications of the enrolling physician. However, it is difficult to avoid missing borderline STEMI patients unintentionally in some cases. Introduction of an organized acute CAG strategy that includes both STEMI and NSTEMI patients implies that all potential patients would be triaged to the invasive centre, bypassing the referral hospital, which would reduce the delay for misclassified STEMI patients.
Second, we observed no difference in the revascularization rate among the two groups, but we found a tendency towards an altered revascularization strategy. In the conventional treatment group, the rate of CABG was twice as high as in patients assigned to acute CAG. This was also observed in the Leipzig Immediate Versus Early and Late Percutaneous Coronary Intervention Trial in Non-ST-Segment Elevation Myocardial Infarction (LIPSIA-NSTEMI) trial. Contrarily, the CABG rates were equal in the Timing of Intervention in Acute Coronary Syndromes (TIMACS) trial; the Intracoronary Stenting With Antithrombotic Regimen Cooling Off (ISAR-COOL) trial; and the Angioplasty to Blunt the Rise of Troponin in Acute Coronary Syndromes (ABOARD) study.10–13 An acute strategy will more likely facilitate the use of PCI because the vessels are more often occluded in the very early phase, as documented by the lower TIMI flow rate in the acute CAG group, whereas spontaneous re-canalization in the late group may facilitate more use of CABG in patients with three-vessel disease. Whether this altered revascularization strategy represents an advantage remains unknown.
Third, to our knowledge, our study is the first to investigate the association between timing of CAG and length of hospital stay. We observed a significant reduction in length of hospital stay in the acute CAG group and a highly significant reduction in time-to-CAG and -revascularization. The duration of the admission is not reported in the four previous studies. When our study was initiated, current guidelines advised subacute CAG in patients with NSTEMI within 72 h and CAG within 24 h when the GRACE-score exceeded 140, which was also the national strategy in Denmark. In May 2013, the Danish Health and Medicines Authority recommended that all patients should undergo CAG within 48 h following the establishment of a NSTEMI diagnosis. In addition, the majority of the patients were initially admitted to referral hospitals. These circumstances explain why time to CAG was 49 h in the conventional treatment group. In comparison, the reduced admission time of half a day associated with acute CAG may seem a small gain. However, in patients where ACS was the final diagnosis, admission time was in fact reduced by 0.75 days. The potential cost reduction associated with earlier discharge should be weighed against additional cost at the CAG centre. In a STEMI-like strategy in patients with NSTEMI, more patients would need CAG around the clock, which would challenge both personnel and healthcare budgets. We suggest the gained reduction of admission duration may be even higher if a dedicated discharge strategy 48 h after the acute CAG is implemented in patients with NSTEMI.
Fourth, in the four previous studies, patients were recruited from high volume centres with a CAG facility. We recruited patients already in the pre-hospital phase with the majority of patients living in the catchment area of referral hospitals. We used pre-hospital POC-cTnT measurement for identification of eligible patients. The utility of pre-hospital POC-cTnT measurement has been described in detail in two previous studies documenting successful analysis in a large proportion of patients with suspected AMI.5,6 In the current study, 32% of patients did not have POC-cTnT performed. The paramedics were instructed not to spend time doing the analysis if significant ST-depression was observed in the pre-hospital ECG. In addition, blood sampling was omitted in cases with short driving distance to the admitting hospital. If a future STEMI-like system was adopted in patients with NSTEMI, blood sampling should be prioritized in all patients even if the scene of the event is near the hospital to avoid initial admission to a non-PCI capable centre. In the era of high sensitivity troponin assays, the time gained from point-of-care diagnostics may be limited, particularly in patients admitted directly to a PCI capable hospital. However, pre-hospital biomarker measurement advances the first blood sampling by more than one hour and reporting of the cTn level by more than two hours.5,6 Such a delay would not allow a true STEMI-like strategy. In addition in most countries patients with suspected NSTEMI are primarily admitted to referral hospitals or emergency departments. The logistics required for subsequent transfer is considerable and would delay CAG significantly. Pre-hospital POC-cTnT allows rerouting the patient directly to the PCI capable centre, which would obviate the initial diagnostic admission, reduce time and save precious health care resources. As high sensitivity cTn assays are not yet available for point-of-care testing, it remains to be investigated whether pre-hospital high sensitivity cTn measurement would improve the pre-hospital NSTEMI logistics.
This paper does not include outcome data which was pre-specified in the protocol for this first part of the NONSTEMI trial. Still, there are some observations that may be associated with outcome. Notably, 17% of the patients in the conventional therapy group had the initially scheduled CAG accelerated due to unstable cardiac conditions and five patients developed STEMI while awaiting CAG. Previously, only the TIMACS sub-group analysis on patients with high risk has been able to document an advantage associated with an early invasive strategy. The TIMACS subgroup analysis led to the current recommendation that CAG should be performed within 24 h in high-risk patients.1,10,14
The 250 patients were included over a period of two years and likely do not represent the full cohort of high risk patients with ACS. The study was initiated gradually, and full coverage throughout the study region was achieved only after March 2014. The current inclusion rate is approximately 10 patients per month at the centre in Aarhus and we believe this is an acceptable inclusion rate considering that patients above 80 years and previous CABG were excluded.
Conclusion
Diagnosing patients with NSTEMI in the pre-hospital phase or immediately at hospital admission and subsequent triage directly to invasive centres is feasible. This STEMI-like strategy is associated with earlier revascularization and shorter hospital stay. It remains unknown whether acute CAG in NSTEMI patients has beneficial clinical implications.
Footnotes
Acknowledgements
identifier: NCT01638806. This manuscript is not under consideration elsewhere. The data presented in this paper have never been published before. The authors would like to thank: supervising paramedics of Falck Denmark A/S and Responce A/S: Kim Witting Hedegaard, Bjørn Møller and Lars Borup; staff at the clinical trial unit: Karin Møller Pedersen, Christel Gry Aagren Nielsen, Helle Bargsteen, Helle Pedersen, Kasper Villefrance, Charlotte Skov, Jakob Hjort; Endpoint Committee members: Kristian Thygesen (Chairman), Hanne Maare Søndergaard and Henning Rud Andersen.
Conflict of interest
MR, LF, MG, KKD, TJ, TN, C-HR, TMH, ISR, JA, HS, HEB, CJT have no conflicts of interest to disclose. CS: consulting fees/honorarium: Roche Diagnostics, Thermo Fischer Scientific; research grants: Roche Diagnostics, The Medicines Company. JTS: Honorarium: Roche Diagnostics. SDK: honoraria: Medicines Company, Aspen, AstraZeneca. CJT: research grants: Roche Diagnostics, The Medicines Company, Terumo; consulting fees/honorarium: Astra Zeneca.
Funding
The work was supported by Aarhus University Hospital Spydspidspuljen at Aarhus University Hospital; Hjerteforeningen (grant R. No 14-R97-A5237-22813); Lundbeck Foundation (grant R. No. R126-2012-11480); Laerdal Foundation; The Medicines Company; and Karl G Andersen Foundation.