The predictive ability of Controlling Nutritional Status score on in-hospital mortality in patients admitted to coronary care unit.

Abstract

Objective: Controlling Nutritional Status score was previously described and has been used in predicting short- and long-term outcomes in different patient populations. The aim of this study was to test the relationship between Controlling Nutritional Status score and in-hospital mortality in coronary care unit patients (MORCOR-TURK population).

Methods: In this multicenter and national study, all patients with an available Controlling Nutritional Status score were included in the analysis. The Controlling Nutritional Status score was calculated according to previously described criteria. To be able to understand the significance of the Controlling Nutritional Status score, we constructed two models. Model 1 included age, heart failure, chronic kidney disease, hypertension, diabetes mellitus, and coronary artery disease history. Model 2 included the Controlling Nutritional Status score and Model 1. We then statistically compared the performances of the two models.

Results: A total of 1,018 patients with known Controlling Nutritional Status scores were included in the analysis. Demographic characteristics are shown. In Model 1, the -2 log-likelihood ratio was 395.995, Nagelkerke R2 was 0.133, and area under the curve was 0.739 (95%CI 0.67-0.81). In the second model to which the Controlling Nutritional Status score is added (Model 2), the -2 log-likelihood ratio was 373.743, Nagelkerke R2 was 0.191, and area under the curve was 0.787 (95%CI 0.72-0.85). The area under the curve value of Model 2 was statistically higher than Model 1 (DeLong p-value: 0.01). A statistically significant correlation was found between death and Controlling Nutritional Status score in Model 2 [OR 1.347 (1.193-1.521), p<0.001].

Conclusions: Our study showed that the Controlling Nutritional Status score may be a significant predictor of in-hospital mortality in coronary care unit patients.