Table 3.
Results of logistic regression analysis having plasma glucose ≥ 140 mg/dl at discharge as the dependent variable in the training population
| Fasting plasma glucose at discharge | ||||
| < 140 | ≥ 140 | OR | 95% CI | |
| Age (years) | 76.8 ± 6.8 | 77.2 ± 6.7 | 1.01 | 0.99–1.03 |
| Gender (males) | 332 (44.9) | 245 (39.8) | 0.88 | 0.70–1.12 |
| Diastolic blood pressure < 80 mm Hg* | 101 (13.7) | 126 (20.5) | 1.63 | 1.19–2.22 |
| Plasma glucose on admission (mg/dl) | ||||
| <143 | 344 (46.5) | 110 (17.9) | 1.0 | **** – **** |
| 143–218* | 233 (31.5) | 218 (35.4) | 2.84 | 2.13–3.79 |
| >218* | 162 (21.9) | 287 (46.7) | 5.29 | 3.91–7.15 |
| White blood cells > 8970 cells/mm3 | 224 (30.3) | 225 (36.6) | 1.09 | 0.85–1.40 |
| Serum sodium < 138 mEq/l | 185 (25.0) | 210 (34.1) | 1.02 | 0.78–1.33 |
| Insulin or combined hypoglycemic therapy prior to admission* | 100 (13.5) | 151 (24.6) | 2.03 | 1.50–2.75 |
| Charlson's Index > 2* | 281 (38.0) | 273 (44.4) | 1.32 | 1.04–1.67 |
Data are means ± SD or number of cases with percent in parentheses. * Marks variables found to be independently correlated with the outcome.