Table 2.
Univariable and multivariable predictors of postoperative stroke
| Predictors | Univariable | Multivariable* | ||
|---|---|---|---|---|
|
|
|
|||
| Hazard Ratio (95% Confidence Interval) | P value | Hazard Ratio (95% Confidence Interval) | P value | |
| Demographics | ||||
| Age at surgery, per 5-year increase | 1.24 (1.13 – 1.36) | < 0.0001 | 1.18 (1.08 – 1.30) | 0.0006 |
| Race | ||||
| White | 1.0 | |||
| African American | 1.00 (0.62 – 1.62) | 0.9846 | ||
| Other | 0.46 (0.15 – 1.46) | 0.1893 | ||
| Female sex | 1.63 (1.11 – 2.39) | 0.0132 | ||
| Clinical risk factors | ||||
| Chronic obstructive pulmonary disease | 1.37 (0.86 – 2.19) | 0.1828 | ||
| Peripheral vascular disease | 1.79 (1.17 – 2.72) | 0.0067 | 1.66 (1.09 – 2.52) | 0.0193 |
| Cerebrovascular accident | 1.98 (1.21 – 3.24) | 0.0069 | ||
| Previous cardiac surgery | 1.22 (0.30 – 4.95) | 0.7786 | ||
| Preoperative serum creatinine > 2 mg/dL | 0.90 (0.40 – 2.05) | 0.8068 | ||
| Critical preoperative state | 2.95 (1.20 – 7.24) | 0.0180 | ||
| Unstable angina pectoris | 1.10 (0.68 – 1.76) | 0.7030 | ||
| Left ventricular function | ||||
| Normal | 1.0 | |||
| Moderate dysfunction | 1.58 (1.05 – 2.37) | 0.0283 | ||
| Severe dysfunction | 1.78 (1.00 – 3.16) | 0.0513 | ||
| Recent myocardial infarction | 0.76 (0.49 – 1.19) | 0.2312 | ||
| Pulmonary hypertension | 0.57 (0.14 – 2.30) | 0.4267 | ||
| Diabetes mellitus | 0.88 (0.60 – 1.30) | 0.5244 | ||
| Laboratory tests | ||||
| Preoperative: | ||||
| Platelet count, per 30 × 109/L decrease | 1.08 (0.98 – 1.16) | 0.1124 | ||
| Hemoglobin, per 1 g/dL decrease | 1.02 (0.93 – 1.12) | 0.6699 | ||
| Postoperative: | ||||
| Daily hemoglobin, per 1 g/dL decrease | 1.03 (0.89 – 1.19) | 0.6826 | ||
| Daily platelet count | ||||
| Normal, >150×109/L | 1.0 | 1.0 | ||
| Mild thrombocytopenia, 100–150×109/L | 1.64 (1.02 – 2.62) | 0.0406 | 1.40 (0.87 – 2.24) | 0.1661 |
| Moderate/severe thrombocytopenia, <100×109/L | 2.77 (1.66 – 4.61) | < 0.0001 | 1.99 (1.18 – 3.34) | 0.0093 |
| Preoperative medications | ||||
| Acetylsalicylic acid | 1.10 (0.74 – 1.63) | 0.6547 | ||
| Alpha-receptor blockers | 0.34 (0.08 – 1.37) | 0.1286 | ||
| Angiotensin converting enzyme inhibitors | 0.81 (0.55 – 1.20) | 0.2986 | ||
| Angiotensin II receptor antagonist | 0.95 (0.48 – 1.87) | 0.8779 | ||
| Beta-receptor blockers | 0.96 (0.66 – 1.40) | 0.8363 | ||
| Clopidogrel | 1.19 (0.70 – 2.02) | 0.5167 | ||
| Diuretics | 0.85 (0.55 – 1.32) | 0.4634 | ||
| Nitrates | 0.95 (0.64 – 1.40) | 0.7806 | ||
| Statins | 0.67 (0.46 – 0.98) | 0.0384 | 0.69 (0.47 – 1.00) | 0.0492 |
| Warfarin | 0.59 (0.15 – 2.41) | 0.4658 | ||
| Intraoperative characteristics | ||||
| Year of surgery, per year increase | 0.97 (0.92 – 1.01) | 0.1096 | ||
| Duration of cardiopulmonary bypass, per 30-minute increase | 1.10 (0.99 – 1.23) | 0.0880 | ||
| Duration of aortic cross clamp, per 30-minute increase | 1.03 (0.86 – 1.23) | 0.7388 | ||
| Intraoperative insertion of intra-aortic balloon pump | 2.14 (1.29 – 3.55) | 0.0031 | 2.10 (1.26 – 3.51) | 0.0043 |
| Intraoperative medications | ||||
| Aminocaproic acid | 0.91 (0.60 – 1.36) | 0.6342 | ||
| Aprotinin | 2.00 (1.27 – 3.15) | 0.0026 | ||
| Epinephrine | 1.40 (0.96 – 2.04) | 0.0794 | ||
| Norepinephrine | 1.61 (0.84 – 3.09) | 0.1488 | ||
| Vasopressin | 0.98 (0.64 – 1.50) | 0.9406 | ||
| Esmolol | 0.72 (0.18 – 2.92) | 0.6482 | ||
| Metoprolol | 0.97 (0.67 – 1.43) | 0.8952 | ||
| Tranexamic acid | 0.51 (0.24 – 1.09) | 0.0835 | ||
| Perioperative blood product use within 2 days | ||||
| Red blood cells | 2.20 (1.42 – 3.39) | 0.0004 | 1.62 (1.04 – 2.54) | 0.0338 |
| Fresh Frozen Plasma | 1.60 (1.07 – 2.40) | 0.0220 | ||
| Platelets | 1.81 (1.24 – 2.64) | 0.0220 | ||
| Cryoprecipitate | 1.11 (0.45 – 2.72) | 0.8226 | ||
The final multivariable Cox regression analysis was based on n = 6,067 due to missing information (n = 63) on preoperative medication use. Predictors of preoperative calcium-channel blocker use, and intraoperative nitroprusside and nitroglycerine use failed the proportional hazards assumption and could not be assessed via Cox regression analysis. The test of difference in survival function was performed via Wilcoxon test of the Kaplan-Meier estimate between strata, and indicated that the probability of remaining stroke free in patients with preoperative calcium-channel blocker use (P = 0.1918), and intraoperative nitroprusside (P = 0.1071) and nitroglycerine (P = 0.2739) use were not statistically not different compared to patients who did not receive these medications. The variables with P < 0.05 in the final multivariable model were selected by means of a forward stepwise technique.