Electronic Letters to:
|
|
E-letters: Electronic letters published:
-
![[Read eLetter]](/icons/misc/sectionDOWN.gif)
The Reliability of Scores
- Brigitte R. Osswald (13 October 2009)
-
RISK SCORES AND OVERESTIMATION OF THE OPERATIVE MORTALITY RISK IN AORTIC VALVE REPLACEMENT
- Marco Ranucci, Serenella Castelvecchio (13 October 2009)
|
|
|||
|
Brigitte R. Osswald, Assistant Professor University of Essen, Dept. of Thoracic and Cardiovascular Surgery, 45122 Essen
Send e-letter to journal:
|
Indeed, the surgical procedure itself bears a "procedure-dependent" risk. Furthermore, the neutralization of most "risk factors" by time is a well-known phenomenon. We appreciate the comments made by Marco Ranucci since they reflect and support our experience. Most physicians dealing with cardiac patients are meanwhile aware about the weak points of scores and the immense lack of reliability in terms of estimation of operative early mortality. Nevertheless, scores remain often the crucial factor for different therapeutic options, since they are easy to use and seemingly give an "objective" estimation of patient characteristics. Some colleagues meanwhile accept for patients with transcatheter aortic valve replacement (TAVI) even very low score values. This could reflect the knowingly ignorance of score values because of a turn towards a more individual decision based on the clinical appearance of each patient. However, generally decreasing proportions of patients with low ejection fraction combined with a turn towards younger patients could announce the era of TAVI without any preoperative risk assessment. Conflict of Interest:None declared |
|||
|
|
|||
|
Marco Ranucci, Head Dept cardiothoracic Anaesthesia and ICU IRCCS Policlinico S.Donato - 20097 San Donato Milanese (Milan) Italy, Serenella Castelvecchio
Send e-letter to journal:
|
We read with interest the article of Osswald and associates (1), who highlight the risk of including the EuroSCORE within the decision-making process in the selection of patients to be treated with transcatheter aortic valve implantation (TAVI) rather than with standard surgical techniques. The authors clearly demonstrate that both the additive and the logistic EuroSCOREs greatly overestimate the operative risk of this patient population, therefore providing a misleading information which may result in an unjustified greater rate of patients addressed to TAVI. The authors provide a set of possible explanations for the poor predictive ability of the EuroSCORE in this setting, basically related to the fact that the EuroSCORE development series is mainly comprising coronary artery bypass graft (CABG) operations and is more than 10 years outdated. We agree with this interpretation, but we think that another major bias may be attributed to the EuroSCORE if this tool is used to assess operative mortality risk in isolated aortic valve replacement (AVR). The EuroSCORE, as well as other risk scores, keeps into minimal consideration the role of the operation itself in determining the operative risk. Actually, only thoracic aorta operations and post – myocardial infarction ventricular septal defect repair are considered “per se” as operation-related risk factors, whereas isolated coronary operations are considered as the reference value; all the other operation fall into the limbo of “operation other than isolated CABG”. This condition attributes an additional (or exponential ) risk, but actually does not differentiate between operations at very low risk (isolated AVR) and operations at very high risk (CABG + mitral valve procedure, double/triple valve procedure…). As a consequence, it is not surprising that both the additive and the logistic EuroSCORE are seriously overestimating the operative risk of isolated AVR operations. In a recent article (2) we compared a 3-factors (age, creatinine, ejection fraction) score (ACEF), with the other risk scores currently available. In the subset of isolated AVR operations (872 patients), the observed mortality was 1.6% (95% confidence interval 0.7% - 2.4%); the predicted mortality was significantly overestimated by the additive EuroSCORE (5.2%), the logistic EuroSCORE (5.7%), and the the Parsonnet score (3.1%), whereas it was not significantly different from the ACEF prediction (2.2%) . We can therefore confirm that rather than using a bad assessment of the operation-related risk, is probably better to ignore this factor or, in alternative, to use risk models that are different for each operation subgroup as the STS models. In any case, risk models with an overestimation bias cannot be used as an easy way to deprive patients of the standard surgical approach in favour of TAVI. References 1. Osswald BR, Gegouskov V, Badowski-Zyla D, Toctermann U, Thomas G, Hagl S, Blackstone EH. Overestimation of aortic valve replacement risk by EuroSCORE: implications for percutaneous valve replacement. Eur Heart J 2009; 30: 74-80. 2. Ranucci M, Castelvecchio S, Menicanti L, Frigiola A, Pelissero G. Risk of assessing mortality risk in elective cardiac operations: age,creatinine, ejection fraction, and the law of parsimony. Circulation 2009; 119: 3053-3061. Conflict of Interest:None declared |
|||