This prospective study revealed a relationship of preoperative MDF with POD. Specifically, lower MDF values strongly predicted POD occurrence. Additionally, patients with POD showed a sharp decrease in the MDF on the 5th postoperative day, which persisted until 1 month after surgery. Follow-up MDF and TICS measurements suggest that the cognitive dysfunction might persist until 1 year after surgery.
There have been recent studies on the predictive utility of EEG for dementia. Alpha power is positively correlated with hippocampal atrophy and cognitive function19,20. Compared with other modalities, EEG can promptly detect subtle changes in brain function, especially early-stage cognitive decline. Since prefrontal EEG markers outperform MMSE scores in predicting dementia, combining both measures can yield more prediction accuracy18. Therefore, EEG has wide potential applications as a screening tool for cognitive impairment, including POD prediction. However, to our knowledge, there have been no studies on the predictive utility of preoperative EEG for POD.
Previous studies have demonstrated a negative correlation between prefrontal MDF and brain ageing in healthy participants, with the resting-state eyes-closed MDF in younger and older adults being > 10 Hz and < 9 Hz, respectively16. Further, in the elderly population, the prefrontal MDF was positively correlated with the MMSE scores. The MDF in patients with MCI and dementia is approximately 8.3 Hz14 and 6.6 Hz18, respectively. We observed a significant between-group difference in the MDF (delirium [8.65 Hz] vs. non-delirium [9.02 Hz]).
We only included patients with normal cognitive function at baseline based on the MMSE score. However, the delirium group showed a lower preoperative MDF than the non-delirium group, which suggests that the patients with delirium could have had preoperative MCI. This is further indicated by our MoCA findings. The MMSE21 is a quick and easy measure of cognitive functioning that has been widely used in clinical and research evaluation of patients with dementia. However, it has a relatively low sensitivity for MCI, with increased false-negative rates. Contrastingly, the MoCA was designed to screen for both MCI and dementia22,23. The preoperative MoCA scores in our patients with delirium were consistent with those in previously reported patients with MCI (average MoCA scores for MCI and mild AD were 22 [range 19–25] and 16 [11–21], respectively)24. Furthermore, the mean preoperative MoCA score in the delirium group in our study was 22 points, which suggests that the delirium group could have had preoperative MCI.
On the 5th day after surgery, the MDF of the delirium group was especially lower than that previously reported in patients with MCI14. At 1-month follow-up, the MDF in patients with delirium recovered to the level observed in patients with MCI (Supplementary Table 1). Moreover, the delirium group showed low MoCA scores at 1 month and 1 year after surgery, with the MoCA scores recovering to preoperative levels only one year after surgery (supplementary table 1). Although we could not obtain statistical significance given the small follow-up sample size, our findings suggested that cognitive impairment in patients with delirium could last until 1 year after surgery. Similarly, at the 1-month and 1-year postoperative follow-ups, the delirium group showed lower TICS scores than the non-delirium group (supplementary table 1). This is consistent with previous reports of reduced TICS scores in patients with MCI than in normal controls (32.81 and 37.28, respectively)25. These results suggest that regular postoperative EEG evaluations involving measurement of the prefrontal MDF, in combination with MoCA and TICS evaluations, can be used to postoperatively track MCI in the elderly and inform early interventions, which could attenuate the progression of the postoperative neurodegenerative disease after POD.
POD is associated with several adverse clinical effects, including dementia, distress, increased hospitalization duration and costs, and higher mortality26. We observed significant between-group differences in the IADL, MNA, and CCI scores.
The MNA is used to measure the nutritional status of elderly individuals27. Nutrition deprivation is associated with POD after on-pump coronary artery bypass graft surgery28. Since it is possible to improve the preoperative nutritional status, the nutritional status of elderly patients is a modifiable risk factor that can be addressed in POD prevention strategies.
Postoperative pain severity is among the most important postoperative factors that affect delirium. Delirium can be managed by controlling pain using appropriate analgesics29. Therefore, we performed between-group comparisons of the pain severity and total analgesic doses until 1 postoperative week. Generally, patients do not require analgesics at 1 postoperative week; accordingly, opioid analgesics are not prescribed from 5 days after surgery. Although there were no between-group differences in the analgesic doses, the delirium group showed persistent pain on movement (supplementary figure 1). Further research is warranted to determine whether pain could affect the severity and duration of delirium as well as whether anti-nociception could prevent POD.
In conclusion, a low MDF on preoperative EEG was associated with POD in elderly patients undergoing spinal surgery. Furthermore, perioperative EEG may facilitate screening of patients at a high risk of delirium and determination of the prognosis after POD. Our findings demonstrate the predictive utility of preoperative MDF for POD.