Table 4.
In-Hospital Event Rates and Logistic Regression Models for Mortality by Prognostic Nutritional Index and Cox Proportional Analysis and Long-Term Mortality by Prognostic Nutritional Index.
| Prognostic Nutritional Index | |||
|---|---|---|---|
| T1 | T2 | T3 | |
| In-hospital mortality | |||
| Number of deaths | 19 | 6 | 2 |
| Mortality, % | 22.6 | 7.1 | 2.4 |
| Mortality, OR (95% CI) | |||
| Model 1: unadjusted | 11.8 (2.6-52.6) | 3.1 (0.6-15.9) | 1 (Reference) |
| Model 2: adjusted for age and sex | 11.9 (2.4-38.2) | 2.9 (0.8-12.5) | 1 (Reference) |
| Model 3: adjusted for all covariatesa | 8.1 (2.1-27.1) | 2.5 (0.9-10.4) | 1 (Reference) |
| Long-term mortality | |||
| Number of deaths | 17 | 10 | 4 |
| Mortality, % | 26.2 | 12.8 | 4.9 |
| Mortality, HR (95% CI) | |||
| Model 1: unadjusted | 5.5 (1.8-16.4) | 2.4 (0.7-7.9) | 1 (Reference) |
| Model 2: adjusted for age and sex | 5.2 (2.1-13.3) | 2.1 (0.9-6.8) | 1 (Reference) |
| Model 3: adjusted for all covariatesa | 4.6 (2.6-10.9) | 1.8 (1.0-5.5) | 1 (Reference) |
Abbreviations: CI, confidence interval; HR, hazard ratio; OR, odds ratio.
aDemographics (age and sex); first measurement of systolic blood pressure and heart rate; first measurement during hospitalization of the following laboratory values (admission glomerular filtration rate calculated by CKD-EPI, blood urea nitrogen, white blood cell count, hematocrit, and platelet count); creatine kinase–MB, troponin I, d-dimer, and brain-type natriuretic peptide; comorbidities (diabetes, chronic kidney disease, hypertension, stroke, heart failure, cancer, chronic lung disease, and atrial dysrhythmia), and medications (use of oral contraceptives, warfarin, and steroid).