Objective To analyze the risk factors of perioperative neurocognitive disorders (PND) in colorectal cancer patients, and explore the predictive value of serum S100β, MMP-9, and ADP for PND.

Methods Eighty-nine patients treated with elective colorectal cancer surgery were selected as research subjects. According to the postoperative Mini-Mental State Examination score, the patients were divided into the non-PND group and PND group. The age, gender, body mass index, education level, preoperative mean arterial pressure, ASA classification, comorbidities (hypertension, diabetes, myocardial infarction, cerebral infarction), anesthesia drug dosage during surgery, operation time, anesthesia time, intraoperative fluid infusion volume, blood loss, and postoperative 24-hour pain visual analog scale (VAS) scores were compared between two groups. The serum levels of S100β, matrix metalloproteinase 9 (MMP-9), and adiponectin (ADP) were detected before surgery and after 1 and 7 days of surgery. Multivariate logistic regression analysis was used to identify the independent risk factors of PND in colorectal cancer surgery patients. The predictive ability of model was validated using ROC curves and DCA curves. Pearson correlation analysis was conducted to study the relationships among S100β protein, MMP-9, ADP and MMSE scores in two groups.

Results Among 89 patients, 26 cases developed PND, with an incidence rate of 28.89%. The esultsof Univariate analysis showed that there was no statistical significance in the education level, gender ratio, ASA classification, preoperative MMSE score, remifentanil dosage, intraoperative colloid fluid dosage, intraoperative blood loss and extubation time between two groups (P ＞ 0.05). However, the proportion of patients aged over 65 years, body mass index, preoperative mean arterial pressure, history of diabetes, dexmedetomidine dosage, operation time, and anesthesia time in the PND group were higher or longer than those in non-PND group (P ＜ 0.05 to P ＜ 0.01). The Serum S100β protein, MMP-9 and ADP levels in the PND group were higher than those in non-PND group (P ＜ 0.01). The results of multivariate logistic regression analysis indicated that the history of diabetes, high serum S100β protein, high MMP-9 and low ADP levels after 7 days of surgery were the independent influencing factors of cognitive dysfunction (P ＜ 0.05).The evaluation of prediction model using the ROC curve and DCA curve showed that both the accuracy (Cox & Snell R² = 0.636, AUC = 0.994) and clinical application value (Nagelkerke R² = 0.908) of model were relatively high. The results of Pearson correlation analysis showed that the S-100β protein was positively correlated with MMP-9 (r = 0.776, P ＜ 0.01), negatively correlated with ADP (r = –0.348, P ＜ 0.01), and negatively correlated with MMSE score (r = –0.405, P ＜ 0.01).

Conclusions History of diabetes, and high serum S-100β protein, high MMP-9 and low ADP levels after 7 days of surgery are the independent risk factors for the occurrence of PND. The multivariate logistic regression prediction model for PND in colorectal cancer patients constructed based on this has good accuracy and high clinical application value.