Cerebrovascular pulsatility indicates preoperative subcortical cognitive impairment and an increased risk for postoperative delirium in elderly patients undergoing elective spine surgery

Abstract

Introduction: Advances in spine surgery enable safe interventions in elderly patients, but perioperative neurocognitive disorders (pNCD), such as post-operative delirium (POD) and cognitive dysfunction (POCD), remain a serious concern. Pre-operative cognitive impairment is a major risk factor for pNCD. Comprehensive pre-operative cognitive assessments are not feasible in clinical practice, making effective screening methods desirable. This study investigates whether pre-operative cerebrovascular duplex sonography can assess subcortical (vascular) cognitive impairment and the risk for POD.

Methods: This prospective single-center study recruited patients aged ≥60 years scheduled for elective spine surgery at a German university hospital. Patients underwent pre-operative assessments including cognitive abilities (CERAD test battery), structural MRI, and cerebrovascular duplex sonography. POD screening was conducted three times daily for at least 3 days. The primary hypothesis, that the mean pulsatility index (PI) of both internal carotid arteries (ICA) predicts POD risk, was tested using logistic regression. Secondary analyses examined the association between POD risk and ICA flow (time-averaged peak velocities, TAPV) and correlations with cognitive profiles and MRI characteristics.

Results: POD occurred in 22% of patients (n = 22/99) within three postoperative days. Patients with POD were significantly older (75.9 ± 5.4 vs. 70.0 ± 6.9 years, p < 0.01) but did not differ by gender (p = 0.51). ICA PI significantly predicted POD risk (OR = 5.46 [95%CI: 1.81-16.49], p = 0.003), which remained significant after adjustment for age and duration of surgery (OR_adj = 6.38 [95% CI: 1.77-23.03], p = 0.005). TAPV did not inform the POD risk (p = 0.68). ICA PI Pre-operative cognitive scores were significantly associated with ICA PI (mean CERAD score: r = -0.32, p < 0.001). ICA PI was also significantly associated with total white matter lesion volume (τ = 0.19, p = 0.012) and periventricular white matter lesion volume (τ = 0.21, p = 0.007).

Discussion: This is the first study to demonstrate that cerebrovascular duplex sonography can assess the risk for POD in elderly spine surgery patients. Increased ICA PI may indicate subcortical impairment, larger white matter lesion load, and lower white matter volume, predisposing factors for POD. Pre-operative cerebrovascular duplex sonography of the ICA is widely available, easy-to-use, and efficient, offering a promising screening method for POD risk. Increased ICA PI could supplement established predictors like age to adjust surgical and peri-operative procedures to individual risk profiles.

Keywords: carotid ultrasound; cognitive impairment; delirium; risk prediction; spine surgery; white matter lesions.

Figure 1

ROC curve illustrating the diagnostic performance of CVDS measures in predicting POD. The data indicate that the PI, in contrast to TAPV, is effective in predicting the occurrence of POD. The diagnostic accuracy is moderate, with a sensitivity of 73% and a specificity of 72% at a PI threshold of 1.68. The AUC for the PI ROC curve is 0.74, with a Gini index of 44%. Given the high prevalence, PI is more effective at identifying patients with a lower rather than a higher risk of developing POD.

Figure 2

Exemplary case of a patient with high PI and white matter changes. FLAIR images depict a significant white matter lesion load and atrophy. The PI automatically estimated from the flow velocity envelope is enhanced in line with group results. The mean CERAD score of this patient was z = −1.38 indicating cognitive impairment lower than one standard deviation than the reference population.

Figure 3

Exemplary case of a patient with low PI and absent white matter pathology. FLAIR images do neither indicate a relevant white matter lesion load nor atrophy. The PI automatically estimated from the flow velocity envelope is low in line with group results in patients without MRI abnormalities. The mean CERAD score of this patient was z = 0.46 indicating normal cognitive function compared to the reference population.