Skip Navigation

European Journal of Heart Failure 2007 9(5):491-501; doi:10.1016/j.ejheart.2006.11.003
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (6)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Khan, N. K.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Khan, N. K.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© 2006 European Society of Cardiology

Prevalence of ECG abnormalities in an international survey of patients with suspected or confirmed heart failure at death or discharge

Nasrin K. Khana, Kevin M. Goodea, John G.F. Clelanda,*, Alan S. Rigbya, Nick Freemantleb, Joanne Eastaughb, Andrew L. Clarka, Ramesh de Silvaa, Melanie J. Calvertb, Karl Swedbergc, Michael Komajdad, VIu Mareeve, Ferenc Follathf and EuroHeart Failure Survey Investigators

a University of Hull Kingston upon Hull, United Kingdom
b University of Birmingham Edgbaston United Kingdom
c Department of Medicine Sahlgrenska University Hospital Göteborg Sweden
d CHU Pitié-Salpetrière Paris, France
e Cardiology Research Complex A L Myasnikov Institute of Clinical Cardiology ul. Tretiya Cherepkovskaya Moscow, Russia
f Department of Internal Medicine, University Hospital Zurich, Switzerland

* Corresponding author. Department of Academic Cardiology, University of Hull, Castle Hill Hospital, Castle Road, Kingston upon Hull HU16 5JQ, UK. Tel.: +44 1482 624084; fax: +44 1482 624085. E-mail address: j.g.cleland{at}hull.ac.uk


   Abstract
Top
 Abstract
1. Introduction
 2. Methods
 3. Results
 4. Discussion
 References
 
Background: Most patients suspected of having heart failure (HF) will get a 12-lead electrocardiogram (ECG) but its utility for excluding HF or assisting in its management has rarely been investigated.

Methods: The EuroHeart Failure survey identified 11,327 patients hospitalised with a suspected diagnosis of HF from 115 hospitals in 24 countries. ECGs were obtained from 9315 patients, of whom 5934 had cardiac imaging tests. The utility of the ECG was assessed for excluding or diagnosing major structural heart disease (MSHD) or major left ventricular systolic dysfunction (MLVSD) and for therapeutic decision making.

Findings: MSHD was present in 70% and MLVSD in 54% of patients overall but in only 21% and 5%, respectively, if the ECG was entirely normal. However, <2% of patients had a normal ECG. No single ECG characteristic identified a probability <25% of MSHD or <20% of MLVSD. Patients with QRS width ≥120 ms or anterior pathological Q-waves had a probability >80% of MSHD and >70% of MLVSD. Diagnostic models suggested that electrocardiographic criteria alone were not accurate for the diagnosis or exclusion of important heart disease in this population. However, 2468 patients (42%) had an electrocardiographic finding that should be used to guide the choice of therapy.

Conclusions: A normal ECG is rare in patients with suspected HF but has limited diagnostic value in this setting. The ECG has an important role in guiding therapy.

Key Words: ECG • Heart failure • Prevalence • Diagnosis

Received June 5, 2006; Revised October 5, 2006; Accepted November 13, 2006


   1. Introduction
Top
 Abstract
 1. Introduction
2. Methods
 3. Results
 4. Discussion
 References
 
The 12-lead ECG is a widely available, low-cost, non-invasive test. Guidelines from the European Society of Cardiology [1,2] and the National Institute for Clinical Excellence (UK) [3] suggest that patients with suspected heart failure (HF) should have an ECG included in the diagnostic process.

Screening programmes for [13] and surveys of patients [4-12] with suspected heart failure suggest that between 1 and 10% of patients with left ventricular systolic dysfunction (LVSD) will have a normal ECG. No study has attempted to quantify how normal the ECG should be to exclude serious heart disease and there has been little attempt to categorise and grade the spectrum of ECG abnormalities that may occur in clinical practice [8]. Similarly, little has been done to assess the value of the ECG for the confirmation of HF or its utility for therapeutic decision-making. The ECG may help identify the cause of heart failure. It may also identify (or trigger the need for further investigations to identify) those patients that might benefit from specific therapies, such as anticoagulation for atrial fibrillation (AF), cardiac resynchronisation devices or defibrillators.

The EuroHeart Failure survey [16,17] collected clinical data on a large cohort of patients with a suspected diagnosis of HF during a hospital admission, which allowed the utility of the 12-lead ECG in this setting to be assessed.


   2. Methods
Top
 Abstract
 1. Introduction
 2. Methods
3. Results
 4. Discussion
 References
 
2.1. Patients
Consecutive hospital deaths and discharges from 115 hospitals in 24 countries belonging to the European Society of Cardiology were screened. Patients were enrolled into the survey if they had a new or prior diagnosis of HF or if they were receiving loop diuretics for reasons other than renal dysfunction or if they were receiving other treatment for HF or major ventricular dysfunction within 24 h of death or discharge. Data were recorded in a specifically designed case report form. The statistical justification of the sample size has been discussed elsewhere [17]. Demographic and clinical characteristics and investigation of enrolled patients were ascertained by case note review.

2.2. Assessment and classification of cardiac dysfunction
Objective information on cardiac dysfunction by radiological, echocardiographic or other imaging technique was obtained. Investigators were asked to provide information on left ventricular ejection fraction (LVEF) and to rank left ventricular (LV) systolic and diastolic dysfunction, specific cardiac chamber dilatation and valve stenosis or regurgitation as ‘none, mild, moderate or severe’.

Cardiac dysfunction was classified as being due to LVSD if the LVEF was recorded as <50% or when LVSD or LV dilatation was reported by the investigator. LVSD was further classified as mild (LVEF 41-49% or investigator-reported mild LVSD or LV dilatation), moderate (LVEF 31-40% or investigator-reported moderate LVSD or LV dilatation) or severe (LVEF <30% or investigator-reported severe LVSD or LV dilatation). Valve disease (VD) and LV diastolic dysfunction (LVDD) were classified by the investigator as absent, mild, moderate or severe. When more than one valve abnormality was reported, the patient was classified according to the most severe abnormality.

The nature of cardiac dysfunction underlying HF is often complex. The likely principle cause of HF defined by echocardiography was classified in a hierarchical fashion (see footnote to Table 3). Moderate or severe LVSD, LVDD and VD were considered to be major abnormalities and termed MLVSD; major VD and major LVDD respectively and collectively were termed major structural heart disease (MSHD). Patients who were reported not to have any LVSD, LVDD or VD were considered to have normal cardiac function. Other patients were considered to have minor cardiac dysfunction.

2.3. Electrocardiography
A copy of the last available 12-lead ECG prior to death or discharge was requested for each patient. ECGs were sent to a core laboratory where they were read by one cardiologist (NKK). Most ECGs were recorded at 25 mm/s. ECG intervals were measured primarily using hand-held digital callipers to an accuracy of 0.2 mm (8 ms). When ECG intervals were recorded automatically by the recording device (<30% of cases), these data were used to confirm manual readings. If discrepancies were noted then manual measurements were re-done and taken as the true reading.

Ventricular rate, PR, QRS and QT intervals were measured from the 12-lead ECG. QT intervals were corrected for heart rate using Bazett's formula [18]. LV hypertrophy (LVH) was defined by the Sokolow-Lyon criteria [19] using the depth of S in V1 and the height R wave in V5 or V6. If either R or S was >25 mm or the sum of S and R waves was ≥35 mm this was taken as evidence of LVH. Septal q-waves and the presence and site of pathological Q-waves were recorded. ST-T-wave abnormalities were classified as minor (meaning flat T-waves), moderate (meaning biphasic or minor inversion) or severe (meaning deeply inverted). Inter-ventricular conduction abnormalities (bundle branch block=BBB) were classified as right (RBBB), left (LBBB) or indeterminate.

ECGs were then classified into groups showing no abnormality, minor abnormality and major abnormality. Normal ECGs had to have normal PR, QRS, QTc (using Bazett's formula) [18] and JTc intervals, without abnormal axis-deviation, bradycardia (<60 bpm), tachycardia (>100 bpm), ST-T-wave abnormalities, pathological Q-waves and voltage criteria for ventricular hypertrophy. The heart rate-corrected JT interval (JTc) was derived by subtracting the QRS duration from Heart rate-corrected QT interval (JTc=QTc–QRSd) [14]. Minor ECG abnormalities included sinus brady- or tachycardia, axis deviation (<–30 or >90°), PR 200-250 ms, QRS 100-120 ms, JTc 350-400 ms [14] and minor T-wave abnormalities. Major abnormalities included abnormal rhythm (including paced rhythm), PR>250 ms, QRS interval ≥120 ms, JTc>400 ms, pathological Q waves, moderate or severe ST-T abnormalities or voltage criteria for LVH. Septal q-waves, a sign of normal septal activation, were reported but not used for classification of degree of ECG abnormality.

Two hundred randomly selected ECGs were re-read by NKK and 400 were re-read for validation by co-workers (JGFC, RdeS) for major discrepancies, for example heart rhythm, conduction pattern, QRS width difference >20 ms or presence of pathological Q-waves. No major discrepancies were found.

2.4. Diagnostic algorithms
A normal ECG is stated to be useful in excluding a diagnosis of HF and therefore we tested the hypothesis that the ECG made an important contribution to identifying patients with either no cardiac dysfunction, no MLVSD or no MSHD. The value of classification of ECGs into normal, minor abnormalities and major abnormalities for the identification of the above subsets was also assessed. The hypothesis that the ECG can help identify patients with MLVSD (whether or not valve disease is present) or MSHD was also tested.

2.5. Statistical methods
Likelihood ratios (LRs) were calculated for each component of the 12-lead ECG. The LRs were converted into ‘weights of evidence’ using the natural logarithm of the product of LR and 10 [20,21]. A positive weight of evidence indicates that the particular diagnosis (LVSD, for example) is more likely; a negative weight of evidence indicates the reverse. Weights of evidence that are zero provide little discriminatory value. Standard errors for the weights of evidence were calculated [22].

Logistic regression was used to identify age and sex corrected candidate predictors for MLVSD, MSHD and MSHD in the absence of MLVSD. Each candidate predictor was compared against a model containing age, sex and a constant. Those with a significant reduction in log likelihood ({Delta}-2LL) were selected for inclusion into a multivariable model. In order to avoid missing continuous predictors that have a non-linear relationship with the probability of disease, all continuous variables were also categorised using 9 equal percentiles. Significant age/sex corrected predictors were added to a multivariable logistic regression model using a forward likelihood ratio step-wise procedure. It has been shown that stepwise logistic regression can produce unstable models [23]. Consequently, bootstrapping [24] was used to test the stability of the step-wise models by calculating the 95% confidence intervals of each variable's p value. The predicted probabilities for the models were saved for comparison using ROC curve analysis.

Plots of the relative probability of disease compared to the probability at a pre-specified reference QRS width (i.e. 80 ms) were constructed using a moving average estimate. The size of the averaging window was set to 10% of the number of available test values (e.g. QRS width). The resulting curve was smoothed using a best-fit polynomial equation.

ROC curves were generated for each of the diagnostic outputs to compare the performance of individual predictors, age/sex corrected predictors and multivariable models. The cut-off threshold, accuracy, sensitivity (SN), specificity (SP) and odds ratio (OR) at the point of maximum Youden index (SN+SP–1) [25] and at a point where the maximum allowable false-negative rate was 1 in a 100 (SN=99%) were recorded.

Logistic regression was performed using SPSS. Probability plots, ROC curves and cut-point analyses were performed using MATLAB– programmes written in-house. An arbitrary level of 5% statistical significance (two-tailed) was used throughout.


   3. Results
Top
 Abstract
 1. Introduction
 2. Methods
 3. Results
4. Discussion
 References
 
3.1. Screening procedure
Over a 6 week period, 46,788 deaths and discharges for any reason were screened of which 11,327 patients (24% of screened patients) fulfilled the criteria for suspected heart failure, representing approximately 68 patients screened and 16 patients enrolled per hospital per week in the survey. Detailed clinical information and a 12-lead ECG were available in 9315 patients (87%) of whom 5934 (64%) also had an objective assessment of cardiac function reported, mainly by echocardiography. Patients for whom both a 12-lead ECG and a report of cardiac dysfunction were available constituted the principal data set. The reasons for enrolment and clinical characteristics of all patients who had an ECG and those that formed the principal data-set for analysis are shown in Table 1.


View this table:
[in this window]
[in a new window]

  		Table 1 Clinical characteristics of the patients with an analysed ECG compared to the principal data set (i.e. those also assessed for cardiac function)

 
3.2. Patients' clinical characteristics
Most patients were enrolled due to a previous or new diagnosis of heart failure. Their mean age was 71 (range 17-107) years and 4079 (46%) were women. 51% of women and 29% of men were aged >75 years. Characteristics of the patients with and without left ventricular assessment are shown in Table 1.

3.3. Evidence for and nature of cardiac dysfunction (Tables 2 and 3)
An assessment of LV function, in patients with suspected HF, using an imaging technique was reported in 5934 patients, usually by echocardiography (4509 patients) but also by radionuclide ventriculography (194 patients) or angiocardiography (414 patients). An additional 817 patients had LV function reported but did not indicate the method of assessment. LVEF was reported in 4605 patients (78%). 4213 patients (71%) were classified as having LVSD of which 3201 (54%) were MLVSD (Table 2). Major valve disease, predominantly mitral regurgitation, was present in 2165 cases (36%) most of whom also had MLVSD. LV diastolic dysfunction was reported in 1291 patients but major LVDD in the absence of MLVSD or major valve disease occurred in only 109 patients. No MSHD was identified in 1806 patients (30%) of whom 523 cases (9%) were considered to have entirely normal heart function.


View this table:
[in this window]
[in a new window]

  		Table 2 Co-existing pathologies in patients with both ECG and an assessment of cardiac function (n=5934)

 
3.4. Electrocardiographic abnormalities (Table 3)
ECG abnormalities were common. Of the 9315 ECGs analysed, only 136 (<2%) were normal, 1505 (16%) showed only minor abnormalities and 7674 (82%) showed major abnormalities. AF±atrial flutter was present in 2486 patients (27%). Excluding patients with paced rhythm (533 patients), QRS prolongation ≥120 ms was observed in 2483 patients (28% of non-paced patients). Only 1662 (18%) ECGs had pathological Q-waves indicative of myocardial infarction but may have been obscured by bundle branch block or paced rhythm. Of patients with T-wave abnormalities, all but 55 had another associated ECG abnormality. Of patients with moderate to severe T-wave changes, 1698 (24%) patients were treated with digoxin, which was rather less than in the overall population (37%).

Of the 5934 patients for whom an assessment of cardiac function was available, only 75 ECGs were normal, 789 had minor and 5070 had major abnormalities (see Table 3). AF±atrial flutter was present in 1588 patients (27%) and was more common in those with major valve abnormality but no MLVSD (39%). Excluding those with paced rhythm, a broad QRS complex (≥120 ms) was observed in 1683 (28%) patients and was more common in those with MLVSD (37%) compared to those with a major valve abnormality but no MLVSD (19%), mild echocardiographic abnormality (19%) or no echocardiographic abnormality (13%).


View this table:
[in this window]
[in a new window]

  		Table 3 ECG classification according to mutually exclusive echocardiographic phenotype

 
3.5. Abnormalities that could guide treatment in the principal data set (Fig. 1)
Altogether, 2468 patients (42%) had an ECG finding that should guide the choice of therapy (Fig. 1). Not included in this figure are patients with a tachycardia, which does not require an ECG for diagnosis, but does indicate the need for improved heart rate control. AF±atrial flutter, indicating a need for anti-coagulation, was present in 1588 (27%) patients. High grade trifascicular heart block (RBBB with associated left anterior fascicular block and atrioventricular block), indicating a need for pacing, was present in 74 patients (2%) [33,34]. QRS ≥120 ms associated with MLVSD, indicating a need for evaluation for cardiac resynchronisation therapy, was present in 1193 patients (20%), of whom 419 had QRS ≥150 ms. Excluding patients with paced rhythm, there was a moderate negative correlation (r=–0.33 (95% CI=–0.29, –0.37) (p<0.05, n=4605) between QRS duration and LVEF, which improved further when patients with RBBB were excluded (Fig. 2)


Figure 01
View larger version (18K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 1 ECG findings (excluding paced rhythm) that might influence the choice of therapy.

 


Figure 02
View larger version (9K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 2 Correlation between QRS width and mean ejection fraction excluding patients with paced rhythm. Pearson's r=–0.33 (95% CI=–0.29, –0.37) (p<0.05, n=4605).

 
3.6. Utility of the components of the ECG for the detection of MSHD and MLVSD (Tables 4 and 5, Figs. 3 and 4)
The weights of evidence and the change in probability from that in the referred population as a consequence of the test result (i.e. the posterior probability) are shown in Table 4. A normal ECG predicted a very low probability of MSHD. However, this applied to very few patients and is therefore of limited clinical utility in patients with suspected HF. Minor ECG abnormalities were more common but did not exclude MSHD or MLVSD effectively.


View this table:
[in this window]
[in a new window]

  		Table 4 Weights of evidence for each ECG scoring category, demonstrating their relative ability to discriminate between patients with or without MSHD and MLVSD

 
Absence of Q-waves had little effect on the probability of MSHD or MLVSD being present, although anterior, but not inferior, Q-waves did indicate a higher probability of MLVSD. Patients with QRS width ≥120 ms, LBBB and anterior pathological Q-waves all had a >70% post-test probability compared to a pre-test probability of 54% of MSHD or MSLVSD. Apart from a normal ECG, there was no single ECG characteristic or its absence which identified a post-test probability of MLVSD of less than 10%.

On logistic regression analysis, after adjusting for age and sex, eight ECG variables were independently associated with MLVSD (see Table 5) including QRS duration, pathological Q-waves, the overall severity of ECG abnormalities, severity of T-wave abnormalities, depth of S in V1 and of height of R in V5/V6 (but not their sum), conduction pattern and heart rate. A cut-off of ≥15 mm for depth of S in V1 and ≤3.5 mm for height of R in V5/V6 optimised the odds ratio for the detection of MSLVD, and these dichotomised variables were included in the final multivariable model. Stability testing of the selected variables using bootstrapping found inferior/lateral/posterior pathological Q-waves, minor ECG abnormalities, severe T-wave abnormalities and LBBB failed to maintain significance.


View this table:
[in this window]
[in a new window]

  		Table 5 Multivariable analysis for the determination of moderate-severe LVSD including bootstrap test of significance

 
The relative probability of MSHD and MLVSD plotted against QRS duration is shown in Fig. 3. Excluding patients with RBBB increased the probability of having MLVSD as QRS duration increased. QRS broadening had a lower relative probability of detecting MSHD compared to MLVSD due to its inability to detect major abnormalities other than LVSD.


Figure 03
View larger version (2K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 3 Relative (referenced to QRS duration of 80 ms) and actual probability of disease according to QRS duration for detecting moderate-severe LVSD (—), moderate-severe LVSD excluding patients with RBBB (Figure 03), any MSHD (Figure 03) and MSHD excluding patients with moderate-severe LVSD (Figure 03). All curves were referenced to a nominal duration of 80 ms and produced using a 10% window moving average estimator with polynomial fitting.

 
The utility of QRS duration alone, after correction for age and sex and of the multivariable model for the identification of MLVSD and of MSHD is shown in the ROC curves of Fig. 4. QRS alone was of modest diagnostic utility for discriminating between the presence or absence of MSHD or MLVSD. Correction for age and sex improved performance slightly. The multi-variable model improved diagnostic performance further but accuracy for MLVSD was still only 70.3% at the optimal Youden Index point. This equates to a 29.8% false-negative rate and a 29.5% false-positive rate. Examination of the specificity at an accepted failure rate of 1 in 100 (SN=99%) gives a false-positive rate of 91.7%.


Figure 04
View larger version (1K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 4 ROC curves showing the comparative diagnostic accuracy of QRS width (—), QRS width corrected for age/sex (Figure 04) and multivariable model (Figure 04) to detect the presence of moderate-severe LVSD. The multivariable model included the variables: QRS width, age, sex, presence and type of pathological Q-wave, severity of ECG abnormality, severity of T-wave abnormality, depth of S in V1, presence and type of conduction abnormality, height R in V6 and heart rate.

 

   4. Discussion
Top
 Abstract
 1. Introduction
 2. Methods
 3. Results
 4. Discussion
References
 
The ECG appears to have limited utility in refuting the presence of HF in a secondary care setting. Very few patients had a normal ECG and therefore even though a completely normal ECG made a diagnosis of MLVSD very unlikely, in agreement with previous reports [4-12], it was of little practical assistance. Moreover, 20% of patients with a normal ECG had some form of MSHD. Although T-wave abnormalities were one of the most common abnormalities, they rarely occurred in isolation and did not seem more prevalent amongst patients on digoxin. Excluding T-wave changes from the classification of the ECG would have made little difference to our findings.

MSHD was common, as might be expected in this study population. However, no single ECG variable increased the probability of MSHD to >90% (a specific test) or reduced it to <20% (a sensitive test). Major ECG abnormalities, particularly broad QRS with a LBBB pattern and anterior Q-waves, increased the probability of finding MLVSD to >70% and multivariable models improved diagnostic accuracy further but still lacked precision. These results suggest that the ECG, used alone, is of limited practical clinical use as a screening tool for selection of patients with suspected heart failure for cardiac imaging in a secondary care setting

The prevalence of MSHD in patients with suspected heart failure in primary care is much lower than in our population and therefore the clinical utility of the ECG in that setting may be very different. Previous studies investigating the role of the ECG in primary care were not large, provided insufficient detail on ECG classification and focused solely on MLVSD [4,6,8,13]. However, a recent meta-analysis of studies in primary care concluded that the ECG was an inadequate screening tool because of variable sensitivity, possibly due to variations in ECG classification, and low specificity [26].

Screening for MSHD combining data from natriuretic peptides and the ECG may improve diagnostic accuracy [27,28]. The ECG is complex to interpret and requires years of training for expert evaluation. Simplifying interpretation by focusing on easily quantified components that carry important diagnostic information might enhance the diagnostic utility of the ECG. QRS duration is easy to identify, objective, quantitative and the ECG variable most strongly related to LVEF, especially if the conduction pattern is taken into account. Therefore, focusing on QRS duration rather than the whole ECG, may simplify analysis and lead to less operator-dependent results.

The 12-lead ECG appears also to have limited value for the identification of ischaemic heart disease as the cause of heart failure. Ischaemic heart disease was present in 64% of patients on the basis of angiography or a history of myocardial infarction but only 24% of patients with MLVSD had pathological Q-waves. In 36% of patients with possible myocardial infarction the diagnosis was confounded by the presence of conduction abnormalities. Other patients will have LVSD due to non-transmural infarction or myocardial hibernation [29].

The real clinical utility of the ECG in heart failure appears not to be for diagnosing the presence or aetiology of MSHD but rather for the identification of important complications that influence the choice of treatment. AF was present in 27% of patients. These patients should receive anti-coagulants, unless contraindicated, and require adequate ventricular rate control [30]. QRS duration was ≥120 ms in 37% of patients with MLVSD, a rate similar to that observed in large clinical trials [31] indicating the need for evaluation of device therapy, including implantable defibrillators and cardiac resynchronisation therapy [32]. Also, the ECG provides an accurate, objective and verifiable measure of heart rate that may be used to guide both pharmacological and device therapy.

Our analysis has a number of limitations. The method of data collection was designed to reduce patient-selection bias but may not always have been effective. However, the survey enrolled a large number of patients over a short time and their characteristics were close to epidemiological predictions. The study was designed to capture information after the hospital episode was complete to prevent interference with usual clinical practice. Consequently, many patients with suspected HF did not undergo cardiac imaging as recommended in guidelines, particularly those without gross ECG abnormalities and it seems reasonable to suppose that they would be less likely to have MSHD and MLSVD. Had they undergone imaging, it might have improved the diagnostic performance of the ECG. A prospective study design might have produced a different result, but would have brought its own biases.

One other observation deserves mention. The EuroHeart Failure survey is an agreement with many other hospital discharge surveys that suggest that only about half of patients have LV systolic dysfunction [35]. There is a widespread perception that the other half has diastolic heart failure. However, the clinical diagnosis of heart failure with a normal ejection fraction should not be equated with diastolic heart failure. Valve disease is another major cause of heart failure as are poorly controlled hypertension and atrial fibrillation. Many of these patients probably have abnormal long-axis systolic function. The classification of major structural heart disease adopted in this survey is, we believe, a useful template for use in future surveys. In this classification, diastolic LV dysfunction severe enough to cause heart failure appears rather infrequent. Ultimately, natriuretic peptides may be a better guide to diagnosis and outcome for this group of patients than echocardiography [36].

In conclusion, the 12-lead electrocardiogram is an important tool for the evaluation of patients with HF after the diagnosis is made. Its value as a screening or diagnostic tool in patients at low, medium or high risk of heart failure has yet to be proved.


   Acknowledgements
 
The authors would like to acknowledge the following sponsors who have supported the study.

Main sponsors: AstraZeneca, Medtronic Europe, Orion Pharma, Pfizer, GSK.

Other sponsors: Acorn, Agilent/Phillips Medical, Aventis UK, BMS, Guidant Europe, Menarini UK, MSD UK, Novartis UK, Pharmacia, Roche, Servier international, Wyeth-Ayerst, Orqis Medical, OrthoBiotech, Educational grant: Merck KgaA. British Heart Foundation (K M Goode was funded by BHF as a junior research fellow).


   References
Top
 Abstract
 1. Introduction
 2. Methods
 3. Results
 4. Discussion
 References
 

  1. Remme W.J., Swedberg K. Comprehensive guidelines for the diagnosis and treatment of chronic heart failure. Task force for the diagnosis and treatment of chronic heart failure of the European Society of Cardiology. Eur J Heart Fail (2002) 4:11–22.[Free Full Text]
  2. Remme W.J., Swedberg K. Guidelines for the diagnosis and treatment of chronic heart failure. Eur Heart J (2001) 22:1527–1560.[Free Full Text]
  3. The National Institute for Clinical Excellence. Management of chronic heart failure in adults in primary and secondary care. NICE Guideline (2003) vol. 5. Available at www.nice.org.uk.
  4. Davie A.P., Francis C.M., Love M.P., et al. Value of the electrocardiogram in identifying heart failure due to left ventricular systolic dysfunction. BMJ (1996) 312:222.[Free Full Text]
  5. Rihal C.S., Davis K.B., Kennedy J.W., Gersh B.J. The utility of clinical, electrocardiographic, and roentgenographic variables in the prediction of left ventricular function. Am J Cardiol (1995) 75:220–223.[CrossRef][Web of Science][Medline]
  6. Lindsay M.M., Goodfield N.E., Hogg K.J., Dunn F.G. Optimising direct access ECHO referral in suspected heart failure. Scott Med J (2000) 45:43–44.[Web of Science][Medline]
  7. Christian T.F., Miller T.D., Chareonthaitawee P., Hodge D.O., O'Connor M.K., Gibbons R.J. Prevalence of normal resting left ventricular function with normal rest electrocardiograms. Am J Cardiol (1997) 79:1295–1298.[CrossRef][Web of Science][Medline]
  8. Houghton A.R., Sparrow N.J., Toms E., Cowley A.J. Should general practitioners use the electrocardiogram to select patients with suspected heart failure for echocardiography? Int J Cardiol (1997) 62:31–36.[CrossRef][Web of Science][Medline]
  9. O'Keefe J.H. Jr., Zinsmeister A.R., Gibbons R.J. Value of normal electrocardiographic findings in predicting resting left ventricular function in patients with chest pain and suspected coronary artery disease. Am J Med (1989) 86:658–662.[CrossRef][Web of Science][Medline]
  10. Henzlova M.J., Croft L.B. The electrocardiogram as a predictor of left ventricular systolic function: correlation with gated SPECT imaging. Mt Sinai J Med (2003) 70:306–309.[Medline]
  11. Khan M.A., Sinha S., Hayton S., Fynn S., Henderson R.A., Bennett D.H. A normal electrocardiogram precludes the need for left ventriculography in the assessment of coronary artery disease. Heart (1998) 79:262–267.[Abstract/Free Full Text]
  12. Gillespie N.D., McNeill G., Pringle T., Ogston S., Struthers A.D., Pringle S.D. Cross sectional study of contribution of clinical assessment and simple cardiac investigations to diagnosis of left ventricular systolic dysfunction in patients admitted with acute dyspnoea. BMJ (1997) 314:936–940.[Abstract/Free Full Text]
  13. Mosterd A., de Bruijne M.C., Hoes A.W., Deckers J.W., Hofman A., Grobbee D.E. Usefulness of echocardiography in detecting left ventricular dysfunction in population-based studies (The Rotterdam Study). Am J Cardiol (1997) 79:103–104.[CrossRef][Web of Science][Medline]
  14. Crow R.S., Hannan P.J., Folsom A.R. Prognostic significance of corrected QT and corrected JT interval for incident coronary heart disease in a general population sample stratified by presence or absence of wide QRS complex: the ARIC Study with 13 years of follow-up. Circulation (2003) 108:1985–1989.[Abstract/Free Full Text]
  15. Cleland J.G., Swedberg K., Cohen-Solal A., et al. The Euro Heart Failure Survey of the EUROHEART survey programme. A survey on the quality of care among patients with heart failure in Europe. Eur J Heart Fail (2000) 2:123–132.[Abstract/Free Full Text]
  16. Cleland J.G., Swedberg K., Follath F., et al. The EuroHeart Failure survey programme—a survey on the quality of care among patients with heart failure in Europe: Part 1. Patient characteristics and diagnosis. Eur Heart J (2003) 24:442–463.[Abstract/Free Full Text]
  17. Bazett H.C. An analysis of the time relations of the electrocardiograms. Heart (1920) 7:353–370.[Web of Science]
  18. Antikainen R., Grodzicki T., Palmer A.J., et al. The determinants of left ventricular hypertrophy defined by Sokolow-Lyon criteria in untreated hypertensive patients. J Hum Hypertens (2003) 17:159–164.[CrossRef][Web of Science][Medline]
  19. Good I.J. Weight of evidence, corroboration, explanatory power and the utility of experiments. J R Stat Soc Ser B (1960) 11:311–322.
  20. Rigby A.S. Development of a scoring system to assist in the diagnosis of rheumatoid arthritis. Methods Inf Med (1991) 30:23–29.[Web of Science][Medline]
  21. Spiegelhalter D.J., Knill-Jones R.P. Statistical and knowledge-based approaches to clinical decision support systems, with an application to gastroenterology (with discussion). J R Stat Soc Ser A (1984) 147:35–77.[CrossRef]
  22. Austin P.C., Tu J.V. Automated variable selection methods for logistic regression produced unstable models for predicting acute myocardial infarction mortality. J Clin Epidemiol (2004) 57:1138–1146.[CrossRef][Web of Science][Medline]
  23. Efron B., Tibshirani R. Bootstrap methods for standard errors, confidence intervals, and other measures of statistical accuracy. Stat Sci (1986) 1:54–77.[CrossRef]
  24. Youden W.J. Index for rating diagnostic tests. Cancer (1950) 3:32–35.[CrossRef][Web of Science][Medline]
  25. Khunti K., Squire I., Abrams K.R., Sutton A.J. Accuracy of a 12-lead electrocardiogram in screening patients with suspected heart failure for open access echocardiography: a systematic review and meta-analysis. Eur J Heart Fail (2004) 6:571–576.[Abstract/Free Full Text]
  26. McDonagh T.A., Robb S.D., Murdoch D.R., et al. Biochemical detection of left-ventricular systolic dysfunction [see comments]. Lancet (1998) 351:9–13.[CrossRef][Web of Science][Medline]
  27. Kruger S., Filzmaier K., Graf J., et al. QRS prolongation on surface ECG and brain natriuretic peptide as indicators of left ventricular systolic dysfunction. J Intern Med (2004) 255:206–212.[CrossRef][Web of Science][Medline]
  28. Cleland J.G., Pennell D.J., Ray S.G., et al. Myocardial viability as a determinant of the ejection fraction response to carvedilol in patients with heart failure (CHRISTMAS trial): randomised controlled trial. Lancet (2003) 362:14–21.[CrossRef][Web of Science][Medline]
  29. Khand A.U., Rankin A.C., Kaye G.C., Cleland J.G. Systematic review of the management of atrial fibrillation in patients with heart failure. Eur Heart J (2000) 21:614–632.[Abstract/Free Full Text]
  30. Hawkins N.M., Wang D., McMurray J.J.V., Pfeffer M.A., Swedberg K., Granger C.B., et al. for the CHARM Investigators and Committees. Prevalence and prognostic impact of bundle branch block in patients with heart failure: evidence from the CHARM programme. Eur J Heart Fail (2007) 9(5):510–517.[Abstract/Free Full Text]
  31. Cleland J.G., Daubert J.C., Erdmann E., et al. The effect of cardiac resynchronization on morbidity and mortality in heart failure. N Engl J Med (2005) 352:1539–1549.[Abstract/Free Full Text]
  32. Gabriel G., Melvin D.C., Alicia C., et al. ACC/AHA guidelines for implantation of cardiac pacemakers and antiarrhythmia devices: executive summary. A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (committee on Pacemaker Implantation). Circulation (1998) 97:1325–1335.[Free Full Text]
  33. Kusumoto F.M., Phillips R., Goldschlager Nora. Pacing in the elderly. Am J Geriatr Cardiol (2002) 11(5):305–316.[CrossRef][Medline]
  34. Cleland J.G.F., Tendera M., Adamus J., et al. The perindopril in elderly people with chronic heart failure (PEP-CHF) study. Eur Heart J (2006) 27:2338–2345.[Abstract/Free Full Text]
  35. Cleland J.G.F., Adamus J., Taylor J., Polonski L., Freemantle N., Tendera M. Prognostic significance of NTproBNP in patients with a clinical left ventricular systolic function. ESC Online (2006) [Abstract].

Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 Eur J Heart FailHome page
D. J. van Veldhuisen, A. H. Maass, S. G. Priori, P. Stolt, I. C. van Gelder, K. Dickstein, and K. Swedberg
Implementation of device therapy (cardiac resynchronization therapy and implantable cardioverter defibrillator) for patients with heart failure in Europe: changes from 2004 to 2008
Eur J Heart Fail, December 1, 2009; 11(12): 1143 - 1151.
[Abstract] [Full Text] [PDF]

Home page
 BMJHome page
R. M Cubbon and K. K A Witte
Cardiac resynchronisation therapy for chronic heart failure and conduction delay
BMJ, April 28, 2009; 338(apr28_2): b1265 - b1265.
[Full Text]

Home page
 J Am Coll CardiolHome page
J. Cleland, N. Freemantle, S. Ghio, F. Fruhwald, A. Shankar, M. Marijanowski, Y. Verboven, and L. Tavazzi
Predicting the Long-Term Effects of Cardiac Resynchronization Therapy on Mortality From Baseline Variables and the Early Response: A Report From the CARE-HF (Cardiac Resynchronization in Heart Failure) Trial
J. Am. Coll. Cardiol., August 5, 2008; 52(6): 438 - 445.
[Abstract] [Full Text] [PDF]

Home page
 HeartHome page
J. G F Cleland, A. Tageldien, N. Maarouf, and N. Hobson
Patients with heart failure who require an implantable defibrillator should have cardiac resynchronisation routinely
Heart, August 1, 2008; 94(8): 963 - 966.
[Full Text] [PDF]

Home page
 Eur J Heart FailHome page
S. Chattopadhyay, M. F. Alamgir, N. P. Nikitin, A. G. Fraser, A. L. Clark, and J. G.F. Cleland
The effect of pharmacological stress on intraventricular dyssynchrony in left ventricular systolic dysfunction
Eur J Heart Fail, April 1, 2008; 10(4): 412 - 420.
[Abstract] [Full Text] [PDF]

Home page
 Eur J Heart FailHome page
K. M. Goode, A. L. Clark, J. A. Bristow, K. B. Sykes, and J. G.F. Cleland
Screening for left ventricular systolic dysfunction in high-risk patients in primary-care: A cost-benefit analysis
Eur J Heart Fail, December 1, 2007; 9(12): 1186 - 1195.
[Abstract] [Full Text] [PDF]

Home page
 Eur J Heart FailHome page
N. M. Hawkins, D. Wang, J. J.V. McMurray, M. A. Pfeffer, K. Swedberg, C. B. Granger, S. Yusuf, S. J. Pocock, J. Ostergren, E. L. Michelson, et al.
Prevalence and prognostic impact of bundle branch block in patients with heart failure: Evidence from the CHARM programme
Eur J Heart Fail, May 1, 2007; 9(5): 510 - 517.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (6)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Khan, N. K.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Khan, N. K.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?