© 2008 European Society of Cardiology
Validation of a risk score to estimate cardiac risk in subjects from the general population on cardioactive treatment
Internal Medicine department, University of Porto Medical School and Hospital S. João, Alameda Prof. Hernâni Monteiro Porto, Portugal
Internal Medicine Department, University of Porto Medical School and Hospital S. João, Porto, Portugal
Department of Hygine and Epidemiology, University of Porto Medical School and Hospital S. João, Porto, Portugal
E-mail address: joanamartinspiments{at}gmail.com.
Key Words: Heart failure • Diagnosis • Brain natriuretic peptide • Cardiovascular treatment
Received February 21, 2008; In a recent issue of the European Journal of Heart Failure, Nielsen et al [1] reported on a combined risk score based on raised BNP, history of myocardial infarction, atrial fibrillation and abnormal ECG that was useful to identify those who had cardiac dysfunction, defined as left ventricular ejection fraction <0.36 or significant valve disease, among general practice patients on cardioactive drugs. The authors concluded that this risk score allowed a more judicious selection of patients to undergo further cardiac evaluation.
It is of utmost importance to assess the external validity of such a score before its usefulness can be confirmed. We tested the performance of this score in a random sample of an urban Portuguese population aged
45 years (n=650, 380 (58.5%) women, mean (standard deviation) age=61.7 (10.6) years). In Table 1, we present the baseline characteristics of the 128 subjects of our sample on cardioactive drugs (loop diuretics, beta-blockers or ACE inhibitors/angiotensin receptor blockers) in comparison with that of Nielsen's report. In general, both populations were quite similar, the main differences being the higher prevalence of hypertension and lower prevalence of ischaemic heart disease in our sample. The prevalence of cardiac dysfunction in our patients on cardioactive drugs was 10.9% (14 patients). The combined risk score (1) detected cardiac dysfunction in our patients with an AUC of ROC curve of 0.90 (95%CI: 0.84-0.97). A risk score 2 had a sensitivity of 92.9%, specificity of 75.4%, and positive and negative predictive values of 31.7% and 98.9%, respectively, which were very similar to the values in the sample on which the score was developed (Table 2).
|
|
In conclusion, the risk score proposed by Nielsen et al. adequately identified patients with cardiac dysfunction in our population, whose characteristics are quite different from the British, typically at low risk of coronary heart disease and with very high prevalence of hypertension. These results add a strong argument supporting its external validity and its usefulness in identifying patients in primary care who should be referred for further cardiac evaluation.
 |
Reference |
|---|
 Top Reference |
|---|
- Nielsen O.W., Cowburn P.J., Sajadieh A., Morton J.J., Dargie H., McDonagh T. Value of BNP to estimate cardiac risk in patients on cardioactive treatment in primary care. Eur J Heart Fail (2007) 9(12):1178–1185.
[Abstract/Free Full Text]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||