Noninvasive predictive value of preoperative perfusion index for intraoperative hemodynamic fluctuations in patients undergoing lower-extremity orthopedic surgery under regional anesthesia: a retrospective cohort study.

OBJECTIVE: To investigate the predictive value of the preoperative perfusion index (PI) for intraoperative hemodynamic fluctuations in patients undergoing lower extremity surgery. METHODS: A retrospective analysis was conducted on 247 patients who underwent orthopedic lower extremity surgery at Wuchang Hospital from May 2022 to September 2024. Based on the occurrence of intraoperative hemodynamic fluctuations, patients were divided into a hemodynamically unstable group (n = 87) and a hemodynamically stable group (n = 160). Clinical data and preoperative PI levels were compared between the two groups. Multivariate logistic regression analysis was used to identify factors significantly associated with hemodynamic fluctuations. The predictive value of these factors was assessed using receiver operating characteristic (ROC) curves with DeLong's test, followed by internal validation. Decision curve analysis (DCA) was performed to evaluate clinical net benefit. Patients were additionally divided into high- and low-PI groups based on the optimal cutoff value derived from the ROC curve. RESULTS: Among the 247 patients, 87 (35.22%) experienced intraoperative hemodynamic instability. Compared with the hemodynamically stable group, the unstable group had significantly lower preoperative PI levels (2.52 ± 1.02 vs. 3.96 ± 1.41, P < 0.05). ROC analysis showed that preoperative PI predicted intraoperative hemodynamic fluctuations with an area under the curve (AUC) of 0.798, an optimal cutoff value of 3.28, a sensitivity of 80.46%, and a specificity of 69.37%. Using this cutoff, the incidence of hemodynamic fluctuations was significantly higher in the low-PI group (≤3.28) than in the high-PI group (58.82% vs. 13.28%, P < 0.05), with a relative risk of 4.43 (95% CI: 2.78-7.07). Multivariate logistic regression analysis identified elevated preoperative heart rate (OR = 1.080, 95%CI: 1.031-1.132), diabetes mellitus (OR = 7.652, 95%CI: 2.295-25.515), and hypertension (OR = 4.325, 95%CI: 1.446-12.937) as independent risk factors for hemodynamic fluctuations. In contrast, elevated preoperative mean arterial pressure (OR = 0.839, 95%CI: 0.778-0.905), elevated preoperative diastolic blood pressure (OR = 0.803, 95%CI: 0.738-0.873), elevated preoperative systolic blood pressure (OR = 0.896, 95%CI: 0.854-0.939), and elevated preoperative PI (OR = 0.296, 95%CI: 0.182-0.483) were protective factors. The combined prediction model based on preoperative PI achieved an AUC of 0.968 (95%CI: 0.937-0.986) for predicting intraoperative hemodynamic fluctuations, which was significantly superior to PI alone (all P < 0.001). After 1,000 bootstrap resampling for internal validation, the bias-corrected AUC was 0.954 (95%CI: 0.917-0.986). DCA confirmed that the combined model based on preoperative PI provided a favorable clinical net benefit within the threshold range of 0.1-0.9. CONCLUSION: Preoperative PI can effectively predict intraoperative hemodynamic fluctuations in patients undergoing lower extremity surgery. When combined with hemodynamic parameters and underlying comorbidities, preoperative PI offers significantly enhanced predictive performance and clinical net benefit for hemodynamic fluctuations.