Statistics from Altmetric.com
001 COMPARISON OF MECHANICAL ASYNCHRONY IN HEART FAILURE PATIENTS WITH WIDE AND NARROW QRS COMPLEXES USING REAL-TIME 3D ECHOCARDIOGRAPHY
E. Liodakis, O. Al Shareef, D. Dawson, P. Pearson, P. Nihoyannopoulos.Hammersmith Hospital, London, UK
Introduction: Electrical asynchrony is often observed in patients suffering from left ventricular dysfunction. Previous studies have shown that when electrical asynchrony is used as a selection criterion for cardiac resynchronization therapy (CRT), 30% of the patients do not respond to this treatment. Mechanical asynchrony expressed with the Dysynchrony Index (DI), which derives by calculating the standard deviation of the time for each of the 16 segments of the Left Ventricle, as defined by the American Society of Echocardiography (ASE), to reach its minimum volume, is also a major finding in patients suffering from heart failure. In this study we assessed the ability of real time 3D echocardiography (RT3DE) to quantify mechanical asynchrony, as well as the correlation between the DI and the QRS length.
Methods: We investigated 40 patients with any degree of left ventricular dysfunction. RT3D scanning was performed using the SONOS-7500 system and the X4 transducer from apical projections. The 4D full volume was then analysed offline using the TomTec software (4D LV Analysis 1.2) to derive the DI and the ejection fraction. A Marquette ECG machine was used to acquire an ECG from each patient. For the statistical analysis we used the SPSS 11 package.
Results: From the 40 patients investigated 40% had mild, 35% moderate and 25% severe heart failure. 50% of the subjects had a narrow and 50% had a wide (>140) QRS complex. A positive correlation was found between the QRS length and the DI (r = 0.44, p<0.001). We also observed that there was a group of outliers that presented with low mechanical and high electrical asynchrony. This group may represent the patients that are selected for CRT treatment, according to the electrical criteria but fail to benefit. Another group of patients was identified that showed high mechanical and low electrical asynchrony. These patients, although they do not fulfil the current criteria for CRT, could benefit from this treatment.
Conclusion: Although the selection criterion currently used for CRT is electrical asynchrony, we believe that measurement of the mechanical asynchrony could help us identify more efficiently the patients that will respond to this treatment.
002 AGREEMENT BETWEEN 3-DIMENSIONAL ECHOCARDIOGRAPHY AND CARDIAC MAGNETIC RESONANCE IMAGING IN MEASURING INDICES OF LEFT VENTRICULAR STRUCTURE AND FUNCTION
N. P. Nikitin, C. Constantin, J. Ghosh, S. Hurren, A. Bennett, E. I. Lukaschuk, A. L. Clark, J. G. F. Cleland.University of Hull, Hull, UK
Introduction: Accurate non-invasive assessment of (LV) structure and function is important in the diagnosis and management of patients with cardiac disease. Cardiac magnetic resonance (CMR) has shown high accuracy and reproducibility in measuring volumetric and functional LV indices. This study was designed to test the accuracy of 3-dimensional echocardiography (3DE) as compared with CMR in patients with good acoustic windows.
Methods: 64 subjects with good acoustic windows, including 40 cardiac patients with LV ejection fraction (EF) <45%, 14 patients with EF >45% and 10 volunteers with normal LV function underwent 3DE using a Philips Sonos 7500 scanner equipped with a x4 xMATRIX transducer using xSTREAM 3D architecture. The subjects also underwent CMR on a 1.5 T Signa CV/i scanner (GE Medical Systems) using ECG-triggered breath-hold gradient-echo FIESTA imaging. Volumetric assessment was performed using analytical 4D LV-Analysis software (Tomtec) for 3DE and MRI Mass software (Medis) for CMR.
Results: We found no significant difference in the mean values of the indices measured with 3DE and CMR: LV end-diastolic volume (EDV) (202±73 vs 195±72 ml, p = 0.60), LV end-systolic volume (ESV) (121±66 vs 118±68 ml, p = 0.78) and LV EF (43±15 vs 44±16%, p = 0.80) and there was an excellent correlation between the indices (r = 0.97 for EDV, r = 0.98 for ESV and r = 0.94 for EF). Bland-Altman analysis revealed bias of -7 ml for EDV, -3 ml for ESV and 1% for EF with limits of agreement (2SD) of 28ml, 22 ml and 10%, correspondingly.
Conclusions: 3DE provides accurate quantification of LV volumetric and functional data as compared with current golden standard of non-invasive cardiac imaging, CMR. 3DE can be recommended for the assessment of LV structure and function in subjects with good acoustic windows.
003 QUALITATIVE AND QUANTITATIVE ASSESSMENT OF MITRAL REGURGITATION BY REAL-TIME 3D ECHOCARDIOGRAPHY
S. Kapetanakis, N. Corrigan, J. DeGuzman, S. Owen, O. Murray, M. J. Monaghan.King’s College Hospital, London, UK
Background: Real-time 3D echo (RT3DE) has brought 3D colour Doppler to the bedside, and provides a novel approach to assessment of mitral regurgitation (MR). We examined correlation with 2D quantification in patients attending for routine echocardiography at our institution.
Methods: 78 consecutive patients with any degree of mitral regurgitation and no other significant valvular disease underwent 2D and RT3DE. Severity of MR was graded by an expert interpreter on a 6-point scale based on 2D morphology, as part of routine assessment of mitral regurgitation. 2D quantification was performed with the area method and, where possible, PISA calculations to derive peak regurgitant flow and effective regurgitant orifice area (EROA). For RT3D quantification, 2 indices were created: the ratio of cardiac at the mitral and LVOT positions calculated by 3D flow quantification, and the ratio of peak regurgitant volume to LA volume.
Results: 43% had grade 1/6 MR, while 14% had grade 6/6 MR on 2D echo. Flow quantification was performed in under a minute for each patient. ANOVA of the ratio MVCO/LVOTCO showed that this could differential all grades of MR, with only the difference between mild and mild-moderate MR not being statistically significant. There was very good correlation with the 2D area method (R = 0.77, p<0.0001). The ratio MRVOL/LAVOL showed almost equally good correlation with the area method (R = 0.73, p = 0.021). PISA calculations could be performed in 17 patients. There was a moderate correlation between MVCO/LVOTCO and PISA calculations (R = 0.47 and 0.39 for peak regurgitant flow and EROA respectively).
Conclusions: RT3DE provides simple and quick quantification of mitral regurgitation, which correlates well to 2D assessment of MR. 3D flow quantification may be useful in cases requiring serial monitoring or where the MR jet cannot be seen, such as prosthetic MR.
004 CAN BRAIN NATRIURETIC PEPTIDE BE USED TO REDUCE PRESSURE ON THE ECHOCARDIOGRAPHY IN-PATIENT SERVICE?
T. Robinson, L. O’Toole1, R. Stokes2, B. Morris2, A. Al-Mohammad.Cardiology, Northern General Hospital, Sheffield, 1Cardiology, Royal Hallamshire Hospital, Sheffield, 2Clinical Chemistry, Northern General Hospital, Sheffield, UK
Introduction: Demand for echocardiography in hospitalised patients places a great burden on cardiology departments, prolongs in-patient stay and delays clinical decision-making. Brain natriuretic peptide (BNP) is used as a “rule out” test in patients with suspected heart failure. Can BNP level perform a similar role in patients referred for in-patient echocardiography?
Aims: Assess (1) proportion of referred patients with BNP levels below the cut-off level of 100 pg/ml, (2) frequency of major echocardiographic abnormalities (MEA) that may alter in-patient management.
Methods: Blinded BNP measurement in 100 consecutive in-patients attending for echocardiogram at a University Teaching Hospital. Patient demographics, echo indication and results were recorded. Sub-group analysis performed in those referred for LV function assessment or suspected valvular heart disease (LV/Valve group). MEA recorded were (a) moderate or severe LV systolic dysfunction (LVSD) (b) LV dilatation (c) moderate or severe left heart valvular stenosis or regurgitation, (d) estimated pulmonary artery systolic pressure >50 mmHg (e) severe right heart lesion (f) large pericardial effusion.
Results: 36% of all tested and 37% of LV/Valve group had BNP levels <100pg/ml. MEA found in 2 (27%) with BNP<100pg/ml cf. 35 (43%) with BNP >100pg/ml (p<0.0001). Mild LVSD detected in 6 (17%) patients with BNP <100pg/ml cf. 15 (37%) with BNP >100pg/ml (p = 0.07). In LV/Valve group; MEA found in 2 (7%) patients with BNP <100pg/ml cf. 23 (53%) with BNP >100pg/ml (p<0.0001).
Conclusion: A high proportion of patients referred for in-patient echo have normal BNP levels. In these patients MEA are unlikely. The safety of BNP screening in this setting warrants large-scale study.
005 PERIPHERAL HAEMODYNAMICS ARE ALTERED IN DIASTOLIC HEART FAILURE
S. L. Cunningham1, K. H. Parker2, J. Mayet1, R. A. Foale1, S. A. McG Thom1, A. D. Hughes1.1St Mary’s Hospital, London, 2Physiological Flow Studies Unit, Imperial College, London, UK
Diastolic heart failure (DHF) is poorly understood and treatment options are limited. Wave intensity analysis (WIA) is a recently described technique that allows the study of wave travel in the circulation at any arterial site and the relationship between the heart and vessels. We have developed WIA non-invasively and used it to investigate the possible role of enhanced wave reflection in DHF. 14 patients with clinical heart failure in the presence of a normal ejection fraction and diastolic dysfunction on echocardiography were compared to 14 age-matched normal subjects (N). Pressure (P) and flow (U) were measured in the right common carotid artery using tonometry and pulsed wave Doppler. Brachial blood pressure was measured using sphygmomanometry. Forward and backward pressures and wave intensities were calculated, as was augmentation index (AI) and carotid wave speed. Data are expressed as means±SD, p was calculated by Student’s t test with log transformation where appropriate. SBP [122±16 (N); 137±23 (DHF) mmHg; p = 0.07], DBP [76±10 (N); 71±6 (DHF) mmHg; p = 0.10], and P [127±21 (N); 143±23 (DHF) mmHg; p = 0.051] were not significantly different in the two groups. Although U was dramatically reduced in DHF [0.73±0.12 (N); 0.57±0.05 (DHF) ms-1; p = 0.0004], ventricular wave intensity was not significantly lower [31±8 (N); 26±10 (DHF) mW/m2; p = 0.26]. AI [18±11 (N); 31±11 (DHF) %; p = 0.02], wave speed [12.7±3.7 (N); 19.2±7.8 (DHF) m/s; p = 0.003], and wave reflection from the body [1.1±0.9 (N); 6.6±9.5 (DHF) %; p = 0.04] and head [9.6±3.6 (N); 16.8±14.4 (DHF) %; p = 0.007] were significantly higher in DHF. DHF is associated with abnormal large artery function resulting in increased pressure augmentation and wave reflection. This increased reflection represents an additional load on the heart and may contribute to the pathophysiology of this condition.
006 RISK STRATIFICATION IN AORTIC STENOSIS USING DOBUTAMINE STRESS ECHO: LONG-TERM FOLLOW-UP OF 80 PATIENTS
P. Das, H. Rimington, R. Rajani, J. Chambers.Valve Study Group, Guy’s and St Thomas’ Hospitals, London, UK
Background: Risk stratification is important in aortic stenosis (AS). The value of dobutamine stress echo is untested in patients with preserved left ventricular function. The mean pressure drop/flow slope (P/FS) correlates to baseline grade of stenosis and may relate to outcome.
Methods: We studied 80 patients (59 male/21 female), median age 69 years, with asymptomatic AS and normal systolic function. Baseline transthoracic and low-dose dobutamine stress echo studies were made. Dobutamine was infused in stages of 5 mcg/kg/min to a maximum of 20 mcg/kg/min. PW Doppler of the outflow tract and CW across the aortic valve were recorded at each stage. P/FS was calculated for each patient in mmHg/mls-1 from a plot of mean pressure drop against transaortic flow. Patients were followed for up to 48 months, the end point was symptom onset.
Results: Baseline peak velocity was 3.7 (SD 0.7) ms-1 and effective orifice area (EOA) 0.92 (0.27) cm2. A linear relation was established between mean pressure drop and flow in 68 patients. Median PFS was 0.99 [0.06, 0.19] mmHg/mls−1. Mean follow-up was 27.6 months. 42 patients developed symptoms, 38 remained asymptomatic. Baseline EOA was 0.80 (0.21) vs 1.04 (0.27) cm2, peak transaortic velocity 4.0 (0.6) vs 3.3 (0.6) ms−1, mean pressure drop 37.0 (13.6) vs 25.9 (14.4) mmHg (all p<0.001). P/FS was 0.16 [0.07, 0.24] vs 0.08 [0.05, 0.13] mmHg/mls−1, p = 0.018. Symptom-free survival at 30 months for patients with P/FS >0.1 was 37% (95%CI 21-55) vs 75% (59-86) for P/FS <0.1 (p = 0.0013), relative risk 2.1.
Conclusion: Patients with moderate-severe aortic stenosis are at high risk of developing symptoms. The pressure drop/flow slope calculated using dobutamine stress echo could assist risk stratification of these patients.
007 ECHOCARDIOGRAPHIC FINDINGS IN 793 PATIENTS REFERRED TO A PROTOCOL-DRIVEN DIRECT-ACCESS ECHO SERVICE FOR GENERAL PRACTITIONERS
L. O’Toole, C. Loader, P. Beecher.Dept of Cardiology, Royal Hallamshire Hospital, Sheffield, UK
Direct access to echocardiography for General Practitioners is seen as a means of facilitating patient care and reducing demand on cardiological services but there have been concerns over the low yield of abnormal findings in such services in the past. We set up a protocol-driven service in 2001 allowing referral for investigation of (1) left ventricular systolic dysfunction in patients who had an abnormal ECG and/or cardiomegaly on chest X-ray (?LVSD group) or (2) had a murmur of uncertain cause (murmur ?cause group). All of the echos were reported by a Consultant Cardiologist.
Between February 2001 and August 2003, 793 patients were referred (∼26 pts/month), 363 (46%) for ‘?LVSD’, 367 (46%) for ‘murmur ?cause’ and 62 (8%) inappropriately. In the ‘?LVSD’ group 36% had significant echo abnormalities; 104 patients were found to have LV systolic impairment, 13 moderate or severe grade left heart lesions and 14 other clinically significant findings. In the ‘murmur?cause’ group 29% had significant echo findings; 88 had a moderate or severe left heart valvular lesion, 4 LV systolic dysfunction and 14 other significant findings. A further 72 patients had mild left heart valve lesions. Of those referred inappropriately 6 of 62 (10%) had significant echo findings. Overall, 31% (243/793) of patients referred had a significant echocardiographic abnormality.
These data compare favourably with findings from a truly open access service (19% significant echo findings; X2 = 12.5; p<0.001)1 and hospitalised patients referred for echocardiography in Sheffield (37% significant echo findings). The service has proved popular and has been used appropriately by General Practitioners with an acceptably high rate of significant echocardiographic findings using these criteria.
008 ROLE OF REAL-TIME TRANSTHORACIC 3D ECHOCARDIOGRAPHY IN THE ASSESSMENT OF MECHANICAL ASYNCHRONY
E. Liodakis, O. Al Shareef, D. Dawson, P. Pearson, P. Nihoyannopoulos.Hammersmith Hospital, London, UK
Introduction: Mechanical asynchrony assessed with Real-Time Transthoracic 3-D echocardiography (RT3DE) is a major observation in patients with left ventricular dysfunction. This asynchrony, observed in the 16 regions, as described by the American Society of Echocardiographers (ASE) can be measured using the Dysynchrony Index (DI), which derives by calculating the standard deviation of the time for each of the 16 segments to reach its minimum volume.
Hypothesis: Mechanical asynchrony measured using the DI may be correlated to the ejection fraction (EF) and to the length of the QRS complex in patients suffering from heart failure.
Methods: We investigated 35 patients with any degree of heart failure which were paired with 35 normal subjects. RT3D scanning was performed using the SONOS-7500 system and the X4 transducer from apical projections. The 4D full volume was then analysed offline using the TomTec software (4D LV Analysis 1.2) to derive the global EF and DI. A Marquette ECG machine was used to acquire a full ECG from each patient and measure the length of the QRS complexes. For the statistical analysis we used the SPSS 11 package.
Results: From the 35 patients that were investigated 38% had mild, 40% moderate and 22% severe systolic dysfunction. The DI was 9.5±1.1, 12.8±1.2 and 17.5±1.2 respectively which was statistically significant higher than that observed in the normal subjects (5.6±0.9 p<0.01). A strong negative correlation between the ejection fraction and DI was found with a higher systolic dysfunction associated with a higher degree of asynchrony. (r = -0.58 n = 35 p<0.01). The length of the QRS complexes was <120 ms in 29%, between 120-140 ms in 26.4% and >140 ms in 45% of the patients. The DI was 6.1±1.1, 13.2±1.3 and 16.8±1.09 respectively. There was a strong positive correlation between the length of the QRS complex and the DI. (r = 0.623 n = 35 p<0.001).
Conclusions: RT3DE is an effective tool for quantifying mechanical asynchrony which is strongly correlated with systolic dysfunction and electrical asynchrony.
009 ECHOCARDIOGRAPHIC WALL MOTION SCORE INDEX: CAN WE MAKE IT MORE USER FRIENDLY?
A. Grabham, M. Nixon, T. Irvine.Department of Echocardiography, Royal Victoria Infirmary, Newcastle upon Tyne, UK
Assessment of left ventricular (LV) function is the commonest reason for echo referral in the UK. LV regional wall motion scoring (RWMS) is an integral part of these studies and quantifies global LV function through the generation of a wall motion score index (WMSI). Since most non echocardiographers are unfamiliar with WMSI we developed and assessed a simple method for the conversion of WMSI to the more widely recognised LV ejection fraction (LVEF).
As part of the assessment of global LV function in our Echo laboratory we routinely perform RWMS using the standard 16 segment model (American Society of Echocardiography). Among the range of scores used, normokinesia = 1 and akinesia = 3. We assessed a simple conversion of the resulting global WMSI to LVEF as follows. If we assume that a global WMSI of 1.0 equates with normal function, LVEF should also be normal (assumed to be 60%). In theory a global WMSI of 3.0 would equate with global akinesia or an LVEF of 0%. Assuming a linear relationship, the two entities can be linked by the following equation: y = −30x + 90 (where y = LVEF and x = WMSI).
Using this model the WMSI results (range 1.0 – 2.8) of 44 patients undergoing echo assessment of LV function were converted to LVEF. For each patient the result was compared to a reference LVEF calculated independently using the biplane Simpson’s rule method of disks (MOD). Operators were blinded to the WMSI result during these calculations. Imaging was performed on two GE Medical System 5 machines. WMSI and MOD LVEF were calculated offline using EchoPac analysis software (GE Medical systems). There was good correlation between MOD LVEF and WMSI derived LVEF (y = 0.9x + 1.3, r = 0.9). There was also good agreement between the two methods with a mean difference WMSI LVEF – MOD LVEF of –4.0 ± 6.8 (mean ± SD) LVEF points.
We propose that this simple conversion of WMSI to LVEF should yield sufficiently reliable data for general clinical use and facilitate interpretation of WMSI data by non echocardiographers.
010 PROGNOSTIC VALUE OF SYSTOLIC MITRAL ANNULAR VELOCITY MEASURED WITH QUANTITATIVE COLOUR-CODED DOPPLER TISSUE IMAGING IN PATIENTS WITH CHRONIC HEART FAILURE DUE TO LEFT VENTRICULAR SYSTOLIC DYSFUNCTION RECEIVING OPTIMAL PHARMACOLOGICAL TREATMENT
N. P. Nikitin, R. de Silva, J. Ghosh, C. Constantin, K. Goode, P. Jones, A. Rigby, F. Alamgir, A. L. Clark, J. G. F. Cleland.The University of Hull, Kingston upon Hull, UK
Aims: This study was designed to assess the prognostic value of various conventional and novel echocardiographic indices in patients with chronic heart failure (CHF) due to left ventricular (LV) systolic dysfunction receiving optimal pharmacological treatment.
Methods and results: We prospectively enrolled 185 patients (age 67±11 years) with CHF and LV ejection fraction < 45% despite optimal pharmacological treatment. Patients underwent echocardiography with tissue harmonic imaging and conventional Doppler studies. Systolic (Sm) and diastolic mitral annular velocities were measured with colour-coded Doppler tissue imaging (DTI). During a median follow-up of 32 months (range 24 to 38 months in survivors), 34 patients died and one underwent heart transplantation. Sm velocity (p = 0.011), diastolic arterial pressure (p = 0.015), serum creatinine (p = 0.023), LV ejection fraction (p = 0.024), age (p = 0.052), LV end-systolic volume index (p = 0.067), and restrictive pattern of transmitral flow (p = 0.074) predicted the outcome of death or transplantation on univariate analysis. On multivariate analysis, only Sm velocity (HR = 0.648, 95%CI (0.460-0.912), p = 0.013) and diastolic arterial pressure (HR = 0.966, 95%CI (0.938-0.994), p = 0.016) emerged as independent predictors of outcome.
Conclusions: In patients with CHF and LV systolic dysfunction despite optimal pharmacological treatment, the strongest independent echocardiographic predictor of prognosis was systolic mitral annular Sm velocity measured with quantitative colour-coded DTI.
011 SYSTOLIC DYSSYNCHRONY INDEX: A NOVEL 3D ECHO PARAMETER FOR EVALUATION OF LEFT VENTRICULAR FUNCTION
S. Kapetanakis, A. Siva, N. Corrigan, M. T. Kearney, M. Cooklin, M. J. Monaghan.King’s College Hospital, London, UK
Background: Left ventricular mechanical dyssynchrony (LVMD) is associated with left ventricular (LV) dysfunction and prolonged QRS duration and has emerged as an important factor in cardiac resynchronisation therapy. Real-time 3D echo (RT3DE) provides a novel assessment of LV function and synchronicity of segmental contraction. We investigated left ventricular LVMD defined by RT3DE in normal subjects and patients with cardiovascular disease.
Methods: 78 normal volunteers and 178 patients with cardiovascular disease and normal LV function or various degrees LV dysfunction underwent 2D and RT3D echocardiography. In RT3DE, volumetric analysis was performed to created time-volume curves for regional volumes, and we created a Systolic Dyssynchrony Index (SDI) based on the dispersion of times to minimum volume for each of the standard 16 segments, expressed as percent of cardiac cycle duration.
Results: In normal subjects and patients with normal LV systolic function, segmental contraction was highly synchronised (SDI 3.5±1.6). SDI was 5.4±3.5, 10±8.7 and 14.7±6.7 for mild, moderate and severe LV dysfunction respectively. SDI was found to have an inverse logarithmic correlation to LVEF (R = 0.82, p<0.0001), which persisted regardless of QRS duration (R = 0.79 and 0.77 for QRS duration of <120 or ⩾ 120 ms respectively). QRS duration of >120 ms could identify only 50% of patients with SDI > 3 SD above the range for normal subjects. Calculation of SDI demonstrated excellent inter-and intra-observer agreement with Interclass Correlation Coefficient of 0.95 in 20 RT3D studies. Using high-resolution polar maps, LVMD could be displayed graphically, providing excellent spatial localisation.
Conclusions: RT3DE provides a simple and reproducible method for quantification of LVMD. Reduced systolic function is associated with increasing prevalence of LVMD irrespective of QRS duration. This may be a valuable parameter in assessment of LV function, particularly in patients considered for cardiac resynchronisation therapy.
012 USE OF A PRE-TEST PREDICTION SCORE TO STRATIFY PATIENTS FOR DOBUTAMINE STRESS ECHOCARDIOGRAPHY (DSE)
H. Chaw, P. Wong, L. Overend, R. Akram, A. Rao, P. Osborne, A. Amadi.University Hospital Aintree, Liverpool, UK
Introduction: DSE is an established diagnostic test for coronary artery disease (CAD), however it is semi-invasive, relatively expensive and requires experienced staffs. We investigated the potential application of the European Society of Cardiology pre-test CAD score and the resting ECG in stratifying patients for DSE.
Methods: Consecutive patients underwent DSE between March 2000 and December 2002, of which 84 had no prior history of CAD. They had a mean age 58 years (SD ±13), 32 (37%) were men, and were followed up for 12 months after their DSE for occurrence of major cardiac events (death, myocardial infarction, revascularisation). Using the pre-test CAD score, we correlated the score and the ECG of each patient with the DSE findings.
Results: In this group 29, 28 and 27 patients had a pre-test score of <50%, >50% to <67% and ⩾67%, and a positive DSE result was found in 0(0%), 3(11%) and 8 (29%) respectively. There were no major cardiac events in patients with a pre-test score <50% and >50% to <67%, but 3 patients with a pre-test score ⩾67% had revascularization. Of the 11 patients with a positive DSE result, 3 (100%) and 6 (75%) patients with a pre-test score>50% to <67% and ⩾67% had abnormal ECG.
Conclusion: In our low event rate cohort, patients with a pre-test score ⩽50% do not require DSE for risk stratification. The extra resources should be directed to patients with a pre-test score>50% to <67% who have abnormal ECG, and those who have a pre-test score ⩾67% with or without abnormal ECG. This could reduce demand by up to 50%. It would result in both cost saving and enable the delivery of a faster and more efficient service.