Patient characteristics at baseline
Based on the inclusion and exclusion criteria, two hundred and twenty-one patients were included in this study. The excluded patients were as follows: 5 patients with prior neurological or psychiatric diseases, such as dementia, stroke, schizophrenia and depression; 3 patients with liver cirrhosis and uremia; 10 patients with preoperative stroke or brain malperfusion, and patients with preoperative shock or hemodynamic instanility due to cardiac tamponade ; 8 patients with liver enzymes greater than four times the baseline; 5 patients who were comatose after the surgery or who died within 24 h after the surgery; 3 patients receiving extracorporeal membrane oxygenation therapy; and 5 patients who did not agree to participate in this clinical study.
The baseline characteristics of the patients are shown in Table 1. There were no significant differences between the two groups in terms of sex, education, body weight, height, smoking, hypertension, diabetes mellitus, coronary artery disease, ejection fraction, New York Heart Association class, or American Society of Anesthesiologists class. The patients in the POD group were significantly older than the patients in the non-POD group (56.48±11.68 years vs 52.22±10.50 years, P=0.040), and the proportion of alcoholism in the patients in the POD group was significantly higher than that in the non-POD group (25.80% vs 6.84%, P=0.003).
The preoperative blood analysis results are shown in Table 1. There were no significant differences in the levels of hemoglobin, hematocrit, creatinine, albumin, glucose, lactate, Na+, procalcitonin, glutamate aminotransferase or glutamate aminotransferase between the two groups. The level of leukocytes in the POD group was higher than that in the non-POD group (14.65±5.64 vs 11.97±4.03, P=0.035). Compared with that in the non-POD group, the level of preoperative IL-6 in the POD group was significantly different (P<0.001).
Surgical-perioperative treatment
The perioperative and postoperative data of the two groups are presented in Table 2. There were no statistically significant differences in terms of surgery type between the two groups. No significant differences were observed for the surgical time, cardiopulmonary bypass time, aortic cross-clamp time, selective cerebral perfusion, lower-body arrest time, or lowest rectal temperature between the two groups. Although the results of our study showed that POD patients had longer ICU stays than non-POD patients (5.21±2.81 vs 4.94±3.03), this difference did not reach significance. The ventilation support time in the non-POD group was shorter than that in the POD group (P=0.033). In addition, the length of hospital stay was significantly different between groups, with 25.23±13.12 days in the POD group and 18.90±8.87 days in the non-POD group (P=0.014).
Plasma IL-6 concentrations
The plasma IL-6 concentrations in the patients are presented in Figure 2. Baseline plasma concentrations of IL-6 were significantly different between the two groups (45.3 (30.0, 82.8) vs. 119.2 (58.75, 170.4) pg/ml; P<0.001); compared with preoperative IL-6 concentrations, postoperative IL-6 concentrations were significantly elevated in both groups, except at 72 h after surgery. In most patients, plasma IL-6 levels increased immediately after surgical treatment, reached peak values at 24 h, and gradually decreased. Additionally, we found that plasma IL-6 concentrations were significantly higher in the POD group than in the non-POD group at two time points (329.9 (227.3, 538.5) vs 199.7 (132.1, 312.1) pg/ml after 24 h; 211.3 (103.2, 308.6) vs 141.7 (99.2, 203.1) pg/ml after 48 h; P<0.05 at both time points). Moreover, the levels of IL-6 72 h after the operation tended to be higher in patients who developed delirium, yet this difference was not statistically significant.
Plasma inflammatory cytokine measurement to predict POD
In Figure 3, using the ROC curve, an optimal cutoff value of preoperative IL-6 levels (95.45 pg/ml) was chosen, which distinguished patients developing POD with 61.3% sensitivity and 79.4% specificity values. The mean AUC was 0.73 (95% CI, 0.632-0.833) for preoperative IL-6. In addition, the ROC curve selected a suitable cut-off value of 24 h postoperative IL-6 level (266.80 pg/ml) as an indicator for predicting POD, which generated a sensitivity value of 71.0% and a specificity value of 69.8%. The AUC values corresponding to IL-6 at 48 and 72 h after surgery were 0.61 and 0.57, respectively. The best thresholds at 48 and 72 h were 0.548 and 0.419 for sensitivity and 0.762 and 0.798 for specificity, respectively. In terms of the AUCs, the discriminatory ability of preoperative IL-6 levels was significantly higher than that of IL-6 levels 72 h after the operation (P=0.001).