Biochemical variables as predictors of in-hospital mortality in patients with acute coronary syndrome

Abstract Background There is a constant search of easily obtainable variables that can be put in relation with in-hospital mortality, enabling creation of prediction risk models in patients treated for acute coronary syndrome. Purpose The main goal of this study was to unravel what is the profile of the patients with acute coronary syndrome in risk of in-hospital mortality, based on simple biochemical variables acquired at the time of hospital admission. Methods This was a single-center cross-sectional cohort study on 3007 pts. treated for ACS (STEMI and NSTEMI). Venous blood sample at hospital admission was acquired for every patient. Patients were comparatively analyzed for their biochemical variables derived from venous blood sample, based on occurrence of in-hospital mortality. Results Out of 3007 pts, 988 (28.2%) were females, and 2519 (71.8%) males, at mean age 62.6 +/- 11.2 y. In-hospital mortality rate was 4.8% (145 out of 3007). We subjected to comparative analyzed hemogram, markers of renal function including sodium and potassium, glycemic profile, complete lipid panel (including ApoA1 and ApoB), and highly sensitive troponin. Significant association with in-hospital mortality was found for: hsTn, stress glycemia and glicated hemoglobin, BUN, creatinite, eGFR, sodium, potasium, full lipid panel [except for HDL-C and Lp(a)]. In the second step, markers proven to be significantly different were subjected to multivariate binary logistic regression (backward conditional). We designed two models, one that included ApoA1 and ApoB (including 1816 pts. for whom we had data), and the second without these two parameters (including 3007 pts.). With the first model we identified: Advanced age, stress glycemia, WBC, eGFR, HDL-C and ApoA1 as independent predictors. Surprisingly, HDL-C was found to be strong independent predictor in multivariate analyzis. While with second modeling we identified advanced age, hsTn, stress glycemia, RBC, WBC, BUN and eGFR as independent predictors. Conclusion Advanced age is a strong independent predictor in any modelling. When we have a possibility to obtain a complete lipid panel, beside stress glycemia and WBC as markers of the degree of neuro-hormonal activation and inflammation, and renal function, HDL-C and ApoA1 seems to be strong negative independent predictors. In the absence of this data beside aforementioned markers of neurohormonal activation and inflammation, and renal function, the degree of myocardial injury plays a significant role. According to our data it seems reasonable to obtain complete lipid panel at hospital admission.Multivariate logistic regression modelsAge distribution across genders