Applicability of Ambler's risk score to patients who have undergone valve replacement with bovine pericardial bioprosthesis.

Objectives: This study aims to verify the applicability of Ambler's risk score to patients who have undergone implantation of bovine pericardial bioprosthesis at the Instituto de Cardiologia do RGS/FCU. This study also aims to quantify the risk factors.

Methods: Retrospective study with 703 patients who had undergone implantation of bovine pericardial bioprosthesis between 1991 and 2005 at the Instituto de Cardiologia do RS. Aortic implant occurred in 392 patients, mitral in 250 and combined in 61. Primary outcome was hospital mortality. Characteristics used to estimate risk were: gender, age, body mass index (BMI), NYHA functional class, left ventricular ejection fraction, valvular lesions, systemic arterial hypertension, diabetes mellitus, renal function, cardiac rhythm, previous cardiac operations, and surgical priority. Univariate and multivariate logistic regression was used to quantify preponderant risk factors by the odds-ratio (OR).

Results: The mortality rate was 14.3%, which was higher than the estimated mortality rate (3%, according to Ambler's mean score of 6, (p<0.01)). Patients who died presented a mean score of 8.26, which was higher than the survivors' average score of 5.68. Characteristics of increased risk were emergency surgery (OR=10.87), dialysis (OR=6.10), and age higher than 80 years (OR=6.10). Our sample indicates an area under the ROC curve of 72.9% (accepted value > 70%).

Conclusion: The mortality predicted in Ambler's score was not reproduced in the observed results. However, the ROC curve provides evidence that this model is applicable. Preponderant risk factors were individualizated.