We included 189 patients (66 males and 123 females) into our study; the median age was 88 (range 85–99). One hundred and twenty-three (65.1%) cases were mortal in 28 days. Length of stay in hospital (p < 0.001) was significantly higher in the mortal cases than in the other patients. In addition, the coronary artery disease (p = 0.044) percentage was significantly higher in the mortal cases than in the other patients. We found no significant differences between groups in terms of age, sex, and other comorbidities.
APACHE II score (p < 0.001), SAPS II (p < 0.001), AST (p = 0.008), creatinine (p < 0.001), creatinine to ALT ratio (p < 0.001), neutrophil to lymphocyte ratio (p = 0.049), ferritin (p = 0.003) and CRP (p = 0.006) were significantly higher in the mortal cases than in the other patients. NIHSS severe (p < 0.001) percentage was significantly higher in the mortal cases than in the other patients. ALT (p < 0.001), lymphocyte (p = 0.009), and albumin (p < 0.001) were significantly lower in the mortal cases than in the other patients. We found no significant differences between groups in terms of glucose, hemoglobin, platelet, white blood cell, neutrophil, platelet to lymphocyte ratio, RDW, LDL, HDL, triglyceride, and total cholesterol (Table 1).
Table I. Summary of patients characteristics, risk scores and laboratory measurements with regard to 28-day mortality |
| | 28-day mortality | |
| Total (n = 189) | Yes (n = 123) | No (n = 66) | p |
Age | 88 (86–91) | 88 (86–91) | 88 (86–91) | 0.559 |
Sex | | | | |
Male | 66 (34.9%) | 43 (35.0%) | 23 (34.8%) | 1.000 |
Female | 123 (65.1%) | 80 (65.0%) | 43 (65.2%) |
Length of stay in hospital, days | 12 (5–27) | 8 (3–14) | 37 (20–58) | < 0.001 |
Comorbidities | | | | |
Hypertension | 106 (56.1%) | 67 (54.5%) | 39 (59.1%) | 0.542 |
Diabetes mellitus | 69 (36.5%) | 48 (39.0%) | 21 (31.8%) | 0.411 |
Hyperlipidemia | 32 (16.9%) | 18 (14.6%) | 14 (21.2%) | 0.344 |
Heart failure | 77 (40.7%) | 56 (45.5%) | 21 (31.8%) | 0.094 |
Atrial fibrillation | 72 (38.1%) | 48 (39.0%) | 24 (36.4%) | 0.840 |
Coronary artery disease | 59 (31.2%) | 45 (36.6%) | 14 (21.2%) | 0.044 |
APACHE II score | 27.02 ± 8.31 | 30.14 ± 8.15 | 21.21 ± 4.72 | < 0.001 |
NIHSS | | | | |
Mild | 30 (15.9%) | 7 (5.7%) | 23 (34.8%) | < 0.001 |
Moderate | 83 (43.9%) | 51 (41.5%) | 32 (48.5%) |
Severe | 76 (40.2%) | 65 (52.8%) | 11 (16.7%) |
SAPS II | 41.25 ± 9.34 | 45.03 ± 7.36 | 34.21 ± 8.55 | < 0.001 |
ALT, IU/L | 16 (11–32) | 13 (9–18) | 36 (23–46) | < 0.001 |
AST, IU/L | 24 (17–39) | 27 (17–56) | 21 (15–27) | 0.008 |
Creatinine ,mg/Dl | 1.24 ± 0.24 | 1.35 ± 0.11 | 1.03 ± 0.26 | < 0.001 |
Creatinine/ALT | 0.080 (0.037–0.126) | 0.107 (0.073–0.153) | 0.032 (0.022–0.042) | < 0.001 |
Glucose, mg/dL | 132 (111–169) | 135 (110–172) | 131 (113–168) | 0.604 |
Hemoglobin, g/dL | 12.9 (12.5–14.0) | 13.0 (12.6–14.0) | 12.85 (12.4–13.9) | 0.340 |
Platelet x103/mcL | 208.15 ± 75.41 | 200.89 ± 76.85 | 221.68 ± 71.27 | 0.071 |
White blood cell, x103/mcL | 10.0 (7.89–11.4) | 10.07 (7.7–11.4) | 9.83 (7.9–11.3) | 0.951 |
Neutrophil, x103/mcL | 6.1 (5.3–6.6) | 6.1 (5.25–6.6) | 6.14 (5.5–6.7) | 0.831 |
Lymphocyte, x103/mcL | 1.46 (1.1–2.39) | 1.27 (1.01–2.2) | 1.8 (1.2–2.46) | 0.009 |
Neutrophil/Lymphocyte | 3.82 (2.37–5.25) | 4.04 (2.39–5.38) | 3.10 (2.31–4.58) | 0.049 |
Platelet/Lymphocyte | 124.55 (90.5–176.71) | 128.89 (91.21–182.61) | 120.82 (88.93–172.22) | 0.531 |
RDW | 14.64 ± 0.93 | 14.63 ± 0.94 | 14.66 ± 0.92 | 0.788 |
Ferritin | 120.2 (74.7–334.5) | 144.5 (95.9–403.1) | 103.9 (50.0–198.0) | 0.003 |
LDL, mg/dL | 105.05 ± 35.61 | 101.86 ± 35.02 | 110.99 ± 36.21 | 0.093 |
HDL, mg/dL | 43.51 ± 9.98 | 43.15 ± 9.36 | 44.17 ± 11.08 | 0.504 |
Triglyceride, mg/dL | 108 (80–144.7) | 105 (78–144) | 112.95 (86–144.7) | 0.633 |
Total cholesterol, mg/dL | 153.65 ± 45.68 | 149.76 ± 46.10 | 160.92 ± 44.31 | 0.110 |
Albumin, g/dL | 3.05 ± 0.62 | 2.85 ± 0.51 | 3.43 ± 0.64 | < 0.001 |
CRP | 6.3 (4.5–8.7) | 6.6 (4.9–8.8) | 4.9 (3.9–8.1) | 0.006 |
Data are given as mean ± standard deviation or median (1st quartile − 3rd quartile) for continuous variables according to normality of distribution and as frequency (percentage) for categorical variables Serum creatinine/alanine transaminase ratio: sCr/Alt, The Acute Physiology and Chronic Health Evaluation II: APACHE II, SAPS II:Simplified Acute Physiology Score 2 : SAPS II, The National Institutes of Health Stroke Scale Scores: NIHSS, Alanine Transaminase : ALT, Aspartate Aminotransferase: AST, Serum Creatinine: sCr, Erythrocyte distribution width : RDW ,Low Density Lipoprotein: LDL, High Density Lipoprotein: HDL ,C-reactive protein: CRP ,Sarcopenia Index: SI, Receiver Operating Characteristic: ROC ,Acute ischemic stroke: AIS, Estimated glomerular filtration rate: eGFR |
When we evaluated the 28-day mortality prediction performance of the variables, we found creatinine to ALT ratio and albumin had the highest sensitivity, APACHE II score had the highest specificity, and positive predictive value, creatinine to ALT ratio had the highest accuracy, negative predictive value, and area under ROC curve (Table 2).
Table II. Performance of the variables to predict 28-day mortality |
| Cut-off | Sensitivity | Specificity | Accuracy | PPV | NPV | AUC (95.0% CI) | p |
APACHE II score | ≥ 28 | 61.8% | 90.9% | 72.0% | 92.7% | 56.1% | 0.817 (0.758–0.875) | < 0.001 |
NIHSS | Severe | 52.8% | 83.3% | 63.5% | 85.5% | 48.7% | 0.739 (0.664–0.814) | < 0.001 |
SAPS II | ≥ 40 | 75.6% | 74.2% | 75.1% | 84.5% | 62.0% | 0.826 (0.764–0.888) | < 0.001 |
Creatinine/ALT | ≥ 0.059 | 85.4% | 84.8% | 85.2% | 91.3% | 75.7% | 0.909 (0.867–0.952) | < 0.001 |
Neutrophil/Lymphocyte | ≥ 3.94 | 54.5% | 65.2% | 58.2% | 74.4% | 43.4% | 0.587 (0.502–0.671) | 0.049 |
Ferritin | ≥ 76 | 82.9% | 42.4% | 68.8% | 72.9% | 57.1% | 0.630 (0.547–0.713) | 0.003 |
Albumin | < 3.41 | 85.4% | 60.6% | 76.7% | 80.2% | 69.0% | 0.753 (0.673–0.833) | < 0.001 |
CRP | ≥ 4.9 | 80.5% | 48.5% | 69.3% | 74.4% | 57.1% | 0.622 (0.534–0.710) | 0.006 |
PPV: Positive predictive value, NPV: Negative predictive value, AUC: Area under ROC curve, CI: Confidence intervals |
We performed multiple linear regression analysis to determine the best predictive factors of the 28-day mortality. We found, patients with high APACHE II score (≥ 28) had 11.090-fold higher risk of death than other patients (OR: 11.090, 95% CI: 1.658–74.178, p = 0.013). Patients with higher SAPS II (≥ 40) had 63.305-fold higher risk of death than other patients (OR: 63.305, 95% CI: 5.911–678.009, p = 0.001). Patients with high creatinine to ALT ratio (≥ 0.059) had 426.149-fold higher risk of death than other patients (OR: 426.149, 95% CI: 36.672–4952.097, p < 0.001). Patients with low albumin (< 3.41) had 15.885-fold higher risk of death than other patients (OR: 15.885, 95% CI: 3.542–71.249, p < 0.001). Patients with high CRP (≥ 4.9) had 6.300-fold higher risk of death than other patients (OR: 6.300, 95% CI: 1.409–28.157, p = 0.016) (Table 3).
Table III. Best predictive factors of the 28-day mortality, multiple logistic regression analysis |
| β coefficient | Standard Error | p | Exp(β) | 95.0% CI for Exp(β) |
APACHE II score (≥ 28) | 2.406 | 0.970 | 0.013 | 11.090 | 1.658 | 74.178 |
SAPS II (≥ 40) | 4.148 | 1.210 | 0.001 | 63.305 | 5.911 | 678.009 |
Creatinine/ALT (≥ 0.059) | 6.055 | 1.251 | < 0.001 | 426.149 | 36.672 | 4952.097 |
Albumin (< 3.41) | 2.765 | 0.766 | < 0.001 | 15.885 | 3.542 | 71.249 |
CRP (≥ 4.9) | 1.841 | 0.764 | 0.016 | 6.300 | 1.409 | 28.157 |
Constant | -7.854 | 1.561 | < 0.001 | | | |
Dependent Variable: 28-day mortality; Nagelkerke R2 = 0.859; Correct prediction = 92.6% |
CI: Confidence Interval |
Other variables included in the model, age (p = 0.414), sex (p = 0.519), coronary artery disease (p = 0.072), NIHHS (p = 0.253), neutrophil to lymphocyte ratio (p = 0.447) and ferritin (p = 0.935) were found to be non-significant.(Figüre 1, Figüre 2)