Revista argentina de cardiología
versión On-line ISSN 1850-3748
BORRACCI, Raúl A. et al. Temporary Validity of Continuous Risk Stratification Mortality Monitoring Models in Cardiac Surgery. Rev. argent. cardiol. [online]. 2005, vol.73, n.5, pp. 341-345. ISSN 1850-3748.
Study objective To study the temporary usefulness of different models for the continuous monitoring of risk stratified mortality in cardiac surgery. Research design and methods An observational longitudinal bidirectional study was carried over a prospective series of 246 consecutive cardiac surgeries performed between January and December 2004, while the retrospective series belonged to two historical samples of 240 and 220 operations performed in 1993 and 1997, respectively. Date of surgery and mortality risk score was registered for each patient based on different risk stratification models. For the 1993-series, only Parsonnet's score was used; for the 1997-series, Parsonnet's and PACCN's (Provincial Adult Cardiac Care Network of Ontario) scores were used, and for the 2004-series, both previously mentioned scores plus EuroSCORE were employed. Outcomes were shown by using variable life-adjusted displays based on both actual mortality of consecutive surgeries and risk stratification score for each patient. Results Parsonnet's score prediction was accurate for 1993-series standards. However, this score overestimated risk in the 1997-series, showing an optimistic net life gain; on the contrary, PACCN's score plotted a negative net life gain, showing an outcome well under the new standards. Variable life adjusted display for 2004-series revealed again an overestimation with Parsonnet's score, an underestimation with PACCN's score and an accurate prediction with EuroSCORE with a net life gain near to zero. Conclusions Analysis of series of cardiac operations performed at different periods of time by using variable life adjusted displays demonstrated temporary validity of risk stratification scores. While Parsonnet's score showed accurate prediction in past series, risk for present series could be better predicted with EuroSCORE which constitutes a better choice to estimate the risk stratified cumulative mortality.
Palabras llave : Cardiac surgery; Mortality; Risk.