Effect of BNP on risk assessment in cardiac surgery patients, in addition to EuroScore II

Patients' prognostication around cardiac surgery is key to better assess risk-benefit balance. Preoperative brain natriuretic peptide (BNP) biomarker has been associated with mortality after cardiac surgery, but its added value with EuroScore 2 remains to be confirmed. In a prospective registry cohort of 4,980 patients undergoing cardiac surgery, the prognostic performance of EuroScore 2 and preoperative BNP was assessed regarding postoperative in-hospital mortality. Discrimination feature was evaluated using receiver-operator-characteristics analysis with area under curve (AUROC). Calibration feature was assessed using Hosmer-Lemeshow test. Multivariable analysis was performed to assess the association between covariates and in-hospital mortality. In-hospital mortality was 3.7%. The AUROC of EuroScore 2 was 0.82 (95% confidence interval (95%CI) 0.79-0.85, p < 0.0001). The AUROC of BNP was 0.66 (95%CI 0.62-0.70, p < 0.0001). The combined model with an AUROC of 0.67 (95%CI 0.63-0.71, p = 0.0001) did not yield better AUROC than EuroScore 2 alone (p < 0.0001 in disfavor of the combined model), nor BNP alone (p = 0.79). In multivariable analysis, EuroScore 2 remained independently associated with mortality (adj.OR of 1.12 (1.10-1.14), p < 0.0001), but BNP was not. Preoperative BNP was not an independent risk factor of postoperative mortality and did not add prognostic information, as compared to EuroScore 2 alone.Clinical trial registry Registry for the Improvement of Postoperative OutcomeS in Cardiac and Thoracic surgEry (RIPOSTE) database (NCT03209674).

[EuroSCORE underestimate the mortality risk in cardiac valve surgery of Mexican population]

OBJECTIVE

The EuroSCORE (European System for cardiac operative risk evaluation) stratifies cardiac risk surgery in easy and accessible manner; it was validated in North America with good results but in many countries of Latin America is used routinely without prior validation. Our objective was to validate the EuroSCORE in patients with cardiac valve surgery at the Instituto Nacional de Cardiología Ignacio Chávez (INCICh) in México.

METHODS

EuroSCORE additive and logistic models were used to predict mortality in adults undergoing cardiac valve surgery from march 2004 to march 2008. The goodness of fit test of Hosmer-Lemeshow was used to evaluate the calibration. The area under the ROC curve was calculated to determinate discrimination.

RESULTS

We included 1188 patients with ages of 51.3±14.5 years, 52% women. There were significant differences in the prevalence of risk factors among the INCICh and the EuroSCORE populations. Total mortality was 9.68% versus 5% and 5.6% predicted by additive and logistic EuroSCORE. According to additive EuroSCORE the risk was low in 11.3%, intermediate in 52.9% and high in 35.9%; for these groups the mortality was .7%, 6.34% and 17.4% against those predicted of 2%, 3.9% and 7.64%. Hosmer-Lemeshow test had a P<.001 for both models and the area under the ROC curve was .707 and .694 for additive and logistic EuroSCORE.

CONCLUSION

In the INCICh 88.7% of patients with cardiac valve surgery had intermediate to high risk and EuroSCORE underestimated the risk of mortality.

Validation of two clinical prognostic models in patients with acute symptomatic pulmonary embolism

OBJECTIVE

The aim of this study was to compare the predictive value of two clinical prognostic models, the Spanish score and the simplified Pulmonary Embolism Severity Index (sPESI), in an independent cohort of patients diagnosed of acute symptomatic pulmonary embolism (PE).

METHODS

We performed a retrospective analysis of a cohort composed of 1447patients with acute symptomatic PE. The Spanish score and the sPESI were calculated for each patient according to different clinical variables. We assessed the predictive accuracy of these scores for 30-day mortality, and a composite of non fatal recurrent venous thromboembolism and non fatal major bleeding, using Cstatistic, which was obtained by means of logistic regression and ROC curves.

RESULTS

Overall, 138 patients died (9.5%) during the first month of follow-up. Both scores showed an excellent predictive value for 30-day all-cause mortality (Cstatistic, 0.72 and 0.74), but the performance was poor for the secondary endpoint (Cstatistic, 0.60 and 0.59). The sPESI classified fewer patients as low risk (32% versus 62%; P<.001). Low-risk patients based on the sPESI had a lower 30-day mortality than those based on the Spanish score (1.1% versus 4.2%), while the 30-day rate of non fatal recurrent VTE or major bleeding was similar (2.2% versus 2.3%).

CONCLUSIONS

Both scores provide excellent information to stratify the risk of mortality in patients treated of PE. The usefulness of these models for nonfatal adverse events is questionable. The sPESI identified low-risk patients with PE better than the Spanish score.

Valor predictivo del modelo EuroSCORE en la cirugía cardíaca de nuestro centro

OBJECTIVE

The use of risk prediction models in cardiac surgery makes it possible to compare and evaluate health care quality between different institutions in countries. This study aimed to assess the performance of the European System for Cardiac Operative Risk Evaluation (EuroSCORE) model in estimating the risk of mortality of cardiac surgery patients of our hospital.

PATIENTS AND METHODS

The additive and logistic EuroSCORE models were applied to all patients who underwent cardiac surgery with extracorporeal circulation from the time the cardiac surgery unit was opened in our center in November 2002 until February 2006. All data were obtained prospectively when the patients were admitted to the Intensive Care Unit. Mortality observed was compared with that estimated in the following subgroups: global cardiac surgery, isolated coronary surgery, isolated valvular surgery, combined valvular and coronary surgery and thoracic aorta surgery. Model discrimination was tested by determining the area under the receiver operating characteristic (ROC) curve.

RESULTS

We studied 1,053 patients who had several differences with the EuroScore model population. Overall observed mortality was 2.2% (95% CI 1.2 - 3.1). The EuroSCORE models overestimated mortality (additive predicted 5%, logistic predicted 4.6%). Mortality of coronary bypass graft surgery was 1.2% and both EuroSCORE models overestimated it. Discriminative power of both models was good with an Area under ROC curve for both models of 0.78 and 0.79.

CONCLUSIONS

The use of both EuroSCORE models overestimated overall observed mortality and that of the different surgical subgroups of cardiac surgery performed in our institution.