Association between Baseline Cognitive Score and Postoperative Delirium in Parkinson's Disease Patients following Deep Brain Stimulation Surgery

Background Deep brain stimulation of the subthalamic nuclei (STN-DBS) is a standard treatment option for advanced Parkinson's disease (PD) patients. Delirium following DBS electrode implantation is common, by several studies, and cognitive impairment is a risk factor for developing postoperative delirium (POD). This prospective observational study was conducted to identify whether preoperative baseline cognitive status has an association with POD in PD patients undergoing DBS surgery. Methods Preoperatively, neuropsychiatric and neuropsychological assessments of the patients were performed including clinical dementia rating (CDR) score, instrumental activities of daily living (IADL) score, mini-mental state exam (MMSE) score, Montreal cognitive assessment (MoCA) score, Hamilton anxiety (HAMA) and Hamilton depression (HAMD) scores, and numerical cancellation test. POD was identified by the confusion assessment method (CAM) twice per day on postoperative day 1 until discharge. Results Twenty-seven (21.6%) of 125 patients developed POD. Among the variables screened, age, CDR score, MMSE score, and HAMA score were indicated to be independent influence factors of POD. The cutoff score, AUC, sensitivity, and specificity of age, CDR score, MMSE score, and HAMA score associated with POD was 58.5, 0.751, 92.6%, 52.0%; 0.5, 0.848, 77.8%, 91.8%; 27.5, 0.827, 88.9%, 62.2%; and 12.5, 0.706, 85.2%, 54.1%, respectively. Conclusions We observed age, CDR score, MMSE score, and HAMA score were independent influence factors of POD in PD patients who received DBS. It is necessary to assess the cognitive status of PD patients before surgery to identify high-risk patients.


Introduction
Parkinson's disease (PD) is a common neurodegenerative disease, with the main clinical symptoms being static tremors, muscle rigidity, and bradykinesia [1]. Deep brain stimulation of the subthalamic nuclei (STN-DBS) is a standard treatment option for advanced PD patients. Bilateral STN-DBS not only improves motor symptoms but also a variety of nonmotor symptoms [2][3][4], as well as health-related quality of life [2,5]. DBS could also reduce the levodopa medication dose and ameliorate the side efects associated with levodopa therapy [4].
However, delirium is one of the most common neuropsychiatric complications following DBS surgery [6], occurring in approximately .6% of patients [7,8]. Postoperative delirium (POD) is an acute disorder of attention and cognition in elderly people that is common, serious, costly, under-recognized, and often fatal [9]. POD has been independently associated with worsened clinical outcomes, increased costs, and increased mortality in patients [9]. For patients with Parkinson's, delirium is an increased risk factor for developing dementia, having a more severe motor impairment, and death [10].
Tis prospective observational study was conducted to identify whether preoperative baseline cognitive status has an association with POD in PD patients undergoing DBS surgery.

Study Design and Clinical Assessment.
Tis was a prospective, cohort study. Te research proposal has been approved by the Ethics Committee of Yuquan Hospital of Tsinghua University (20190014). All patients enrolled signed informed consent. Te clinical trial registration was completed before the frst patient is enrolled (https://www.chictr. org.cn, ChiCTR1900027210).

Subject.
A total of 128 consecutive PD patients from Tsinghua University Yuquan Hospital treated with bilateral STN-DBS were screened at baseline. All patients were diagnosed with PD according to the UK Brain Bank criteria [18]. Bilateral STN-DBS treatment was initiated according to the Movement Disorders Society guidelines [19].
According to Chinese deep brain stimulation therapy for Parkinson's disease expert consensus (Second Edition) [20], the inclusion criteria for performing DBS surgery are as follows: primary PD, hereditary PD or various genotypes PD, responds well to compound levodopa; drug efcacy has decreased signifcantly or obvious motor complications afect the patient's quality of life; adverse drug reactions that cannot be tolerated and afect the efcacy of drugs; and there are tremors that cannot be controlled by drugs. Contraindications for performing DBS surgery are as follows: signifcant cognitive impairment; severe (refractory) depression, anxiety, schizophrenia, and other mental diseases; and medical comorbidities that afect surgery or survival.
Preoperatively, neuropsychiatric and neuropsychological assessments of the patients were performed including the CDR score, IADL score, MMSE score, MoCA score, Hamilton anxiety (HAMA), Hamilton depression (HAMD) score, and numerical cancellation test.
Baseline information such as age, sex, body mass index, the highest level of education, American Society of Anesthesiologists (ASA) functional status, and preoperative complications was recorded.

Anesthesia Method.
Te general anesthesia and surgery were operated by a team to avoid interfering factors. After the patients were transferred to the operating room, the electrocardiograph, noninvasive blood pressure, heart rate, saturation of pulse oximetry, and bispectral index (BIS) were monitored. 2 ml venous blood was collected when the peripheral vein was accessed. Te induction drugs were sufentanil 0.3 μg/kg, propofol 1.0-2.0 mg/kg, etomidate 0.2-0.3 mg/kg, and cisatracurium 0.2 mg/kg. After induction, a 7.5 # (female) or 8.0 # (male) endotracheal tube was intubated.
Te anesthesia time, operation time, intraoperative fuid volume, hypotension, bradycardia, and other side efects were recorded.

Operation.
All patients included in the study were diagnosed with PD and met DBS indications.
Patients underwent two surgeries in this study. In the frst surgery, patients underwent stereotactic implantation of DBS electrode in the subthalamic nucleus (STN). Te anesthesia method usually was local anesthesia with minimal sedation. Tey then underwent imaging examination to confrm the place of the electrodes.
Te second surgery was conducted after the imaging confrmation and the DBS batteries and leads were placed. Te second surgery was performed under general anesthesia, the DBS generator was implanted in the subclavicular region, and the extension wires were tunneled through the neck and connected to the DBS electrode. Te patient returned to the ward after extubation.
All the patients received an assessment of the unifed Parkinson's disease rating scale (UPDRS) score three times and in two phases: preoperative, 2 weeks after surgery, and 6 months after surgery and medication on and of phase. Te levodopa equivalent daily doses (levodopa equivalent daily dose) before, 2 weeks, and 6 months after surgery were also recorded. Te stimulation generator was switched on 2 weeks after DBS surgery. Te postoperative followup was operated 6 months after DBS surgery.

Clinical
Assessment. POD was identifed by the confusion assessment method (CAM) [9] which could be used for medical staf, caregivers, and family members. Te CAM was administered twice per day on postoperative day 1 until discharge. POD assessment is generally divided into two steps: frst, use Richmond agitation-sedation score (RASS) to evaluate patients' consciousness, and second, evaluate the content of consciousness. If someone's RASS score was −4 or −5, then the patient could not be evaluated. Te whole process needed about 5 min, including the following 4 questions: (1) acute onset fuctuated mental status; (2) attention defcit; (3) altered level of consciousness (RASS score); and (4) confusion. If 1 + 2 + 3/4 is matched, delirium was diagnosed [22].
Te assessment of POD was performed twice a day between 7 : 00 am and 7 : 00 pm after the surgery until discharge. Te occurrence of POD was assessed by using the CAM [9] and the severity of the POD was assessed by using the memorial delirium assessment scale (MDAS) [23].
Tree patients were excluded due to the second operation. Terefore, 3 patients were excluded from further analysis. Finally, 125 PD patients were analyzed.
2.6. Statistical Analysis. SPSS 26.0 software was used for data analysis. Te Kolmogorov-Smirnov test was used frst to test the normality of all of the variables. Mean ± standard deviation (x±s) was used for statistical description, and an independent sample t-test or one-way analysis was used for the normal distribution variables comparisons between groups. Median (interquartile spacing) was used for statistical description, and the Kruskal-Wallis test was used for abnormal distribution variables comparisons between groups. Percentages were used for statistical description, and the χ 2 or Fisher exact test was used for counting data. A binary logistic regression was used to estimate the odds ratio (OR) of maintaining independence. P < 0.05 was considered statistically signifcant.

Results
A total of 128 patients were screened, and 3 patients were excluded because of the second operation. Tus, 125 patients completed both preoperative and postoperative assessments and were entered into the fnal analysis. Te symptoms were improved after DBS surgery. Te UPDRS part II and part III score and levodopa equivalent daily dose (LEDD) were signifcantly decreasing than preoperative. Tere was no serve complication during the perioperative period (Table 1).
Patients were divided into two groups according to the occurrence of POD: POD group (27 cases) and non-POD group (98 cases).

Clinical Data of Patients with POD.
Twenty-seven patients were diagnosed with POD, and the incidence of POD was 21.6% (27/125). All 27 patients developed POD on the frst postoperative day. Delirium disappeared on the second day after surgery in 19 patients (70.37%).

Baseline Characteristics of POD and Non-POD Group.
Tere were no diferences between POD and non-POD groups on genders, body mass index (BMI), ASA status, preoperative complications, LEDD, the rate of receiving dopamine receptor agonists and MAO inhibitors preoperative, operation time, and general anesthesia maintenance method. However, the age of patients in the POD group was signifcantly elder than in the non-POD group. Te years of education in the POD group were signifcantly longer than in the non-POD group. Te anesthesia time of the POD group was signifcantly longer than the non-POD group ( Table 2).

Comparison of Preoperative Cognitive Status between the Two Groups.
Tere were no diferences between POD and non-POD groups on the correct elimination and elimination index. However, there was a signifcant diference between POD and non-POD groups on the CDR score, IADL score, MMSE score, MoCA score, HAMA and HAMD score, missing elimination, incorrect elimination, and elimination time (Table 3).

Univariate and Multivariate Logistic Regressions.
Univariate and multivariate logistic regressions were performed to identify the potential risk factors of POD. Each variable was screened using the univariate regression and the variables with P < 0.05were selected for multivariate logistic regression. Tose variables with P < 0.05in the multivariate regression were defned as having an association with POD. Te odds ratio (OR) and 95% confdence interval (95% CI) were used to illustrate the predictive power of certain characters.
Among the variables screened, age, CDR score, MMSE score, and HAMA score were indicated to be independent infuence factors of POD (Table 4).

Diagnostic Analysis of Indicators for the Prediction of POD.
Te AUC, cutof, sensitivity, specifcity, and Youden index of independent infuence factors of POD are given in Table 5.
A receiver operating characteristic (ROC) curve was used to determine the optimal cutof score for the diagnosis of POD. Te optimal score was calculated according to the Youden index (maximum of (sensitivity + specifcity−1)) [24]. Te total area under the curve (AUC), the sensitivity, and the specifcity were all used for this determination.
We obtained 58.5 as the optimal cutof score of age associated with POD. Tis cutof score of age led to a sensitivity of 92.6% and a specifcity of 52.0% for the association with POD. Te AUC was 0.751 (95% CI: 0.657-0.844, P < 0.001) (Table 5, Figure 1(a)). We obtained 0.5 as the optimal cutof score of the CDR score associated with POD. Tis cutof score of the CDR score led to a sensitivity of 77.8% and a specifcity of 91.8% for the association with POD. Te AUC was 0.848 (95% CI: 0.750-0.946, P < 0.001) (Table 5, Figure 1(a)). We obtained 12.5 as the optimal cutof score of the HAMA score associated with POD. Tis cutof score of HAMA score led to a sensitivity of 85.2% and a specifcity of 54.1% for the association with POD. Te AUC was 0.706 (95% CI: 0.601-0.811, P < 0.001) (Table 5, Figure 1(a)). We obtained 27.5 as the optimal cutof score of the MMSE score associated with POD. Tis cutof score of MMSE score led to a sensitivity of 88.9% and a specifcity of 62.2% for the association with POD. Te AUC was 0.827 (95% CI: 0.733-0.920, P < 0.001) (Table 5, Figure 1(b)).

Discussion
Parkinson's disease is a common neurodegenerative disease, which is more common in the elderly. Te prevalence of PD in people over 65 years of age is approximately 1.0-3.0% [25]. Deep brain stimulation of the subthalamic nuclei (STN-DBS) is a standard treatment option for advanced PD [19].
With the progress of population aging, more and more PD patients are expected to receive DBS surgery. So, perioperative optimization management of PD patients should attract the attention of anesthesiologists.
POD is an acute neuropsychiatric syndrome after surgery that is associated with an altered level of consciousness, confusion, and impaired attention [26,27]. POD in elder adults has been associated with both short-term and longterm adverse consequences [9], prolonged hospital stays [28], higher costs per patient [9,29], and an eight-fold increased risk of future dementia [30]. Fortunately, delirium may be preventable in one-third of cases [31,32]. Terefore, it is important to identify high-risk patients and optimize perioperative management to reduce the risk of POD.
Te prevalence of POD varies from 5.1 to 52.2% with diferent types of surgery [9]. Patients with PD are at an increased risk for delirium which may be underdiagnosed due to phenomenological overlap between delirium and chronic neuropsychiatric features of PD or side efects of dopaminergic medication [33]. Both Sakai's [34] and Pan's [35] studies showed that PD is an independent risk factor for POD. Oichi [36] reported that POD was more common in patients with PD (30.3%) than in controls (4.3%) following spinal surgery. Carlson's [37] study showed that POD following implantation of DBS electrodes was common (22% of patients). In this study, the incidence of POD in PD patients following DBS was 21.6%.
As a neurodegenerative disease, the previous focus on PD was mainly on motor function. However, cognitive impairment is a common nonmotor symptom of PD [38]. 80% of PD patients reported experiencing dementia after 8 years [39].
In this study, all patients received a preoperative cognitive examination by a specialized psychiatrist including a CDR score, IADL score, MMSE score, MoCA score, HAMA and HAMD score, and numerical cancellation test. Among them, age, CDR score, MMSE score, and HAMA score were independent infuencing factors of POD.
At present, age as an important independent risk factor for POD has become a consensus in the feld of neurocognitive research [40,41]. In this study, we also found that age is an independent risk factor for POD.
Tis study showed that PD patients had been combined with cognitive impairment before surgery. With an assessment of CDR, a score above 0.5 is considered suspicious dementia [42]. In this study, suspected dementia was found in 21.6% of patients. With an assessment of MMSE and MoCA, cognitive function was impaired in 28.8% and 73.6% of patients, respectively. Te cognitive impairment above all indicated that these patients were a susceptible population. According to Chinese deep brain stimulation therapy for Parkinson's disease expert consensus (Second Edition) [20], patients with mild cognitive impairments were included, but patients with mild dementia were excluded from this study.  Figure 1: (a) ROC analysis used for the determination of the diagnostic sensitivity and specifcity of the preoperative optimum cutof score of age, CDR score, and HAMA score. Te probability of the POD increases with the increase in age, CDR score, and HAMA score. (b) ROC analysis used for the determination of the diagnostic sensitivity and specifcity of the preoperative optimum cutof score of the MMSE score. Te probability of the POD decreases with the increase of the MMSE scores.
Khan et. al [43] also confrmed that cognitive defcits are predictive of POD and mortality following TAVI and emphasized the value of screening for geriatric risk factors before TAVI by demonstrating that screening increased the identifcation of at-risk patients.
Tis study also showed that PD patients had been combined with emotional abnormalities before surgery. With the assessment of HAMA and HAMD scores, 27.2% were likely to be anxious, 10.4% were anxious, 35.2% of patients were mildly depressed, and 6.4% were moderately depressed.
Depression and anxiety are among the most important nonmotor signs of PD [44]. Tere are close interactions between depression, anxiety, and PD itself. Some of these symptoms are induced by dopaminergic medications, while others rely on the disease pathophysiology itself [45][46][47]. Widespread dysfunction of the limbic system has been observed in PD patients [44].
PD-associated depression and anxiety are linked with many anatomical changes within the limbic system [44,[48][49][50]. Some studies showed the atrophic temporal cortex, particularly the amygdala and hippocampus, in PD patients could participate in mood/emotion learning defcits [51][52][53]. Anita's study also showed limbic systems were identifed by neuroimaging as putative substrates for delirium which support cognitive functions [54].
Tis study also suggested that the preoperative anxiety score was an independent risk factor of POD in PD patients. Although there was a statistical diference in preoperative depression score between the two groups, the fnal logistics analysis only suggested that the preoperative anxiety score was an independent infuencing factor of POD. In the future, a large sample study needs to confrm the relationship between preoperative depression and POD.
In this study, age, CDR score, MMSE score, and HAMA score were independent infuence factors of POD in PD patients who received DBS. Te cutof value was 58.5, 0.5, 27.5, and 12.5, respectively. So, if the PD patients are older than 58.5, CDR score higher than 0.5, MMSE score lower than 27.5, and HAMA score higher than 12.5 before DBS, it is necessary to pay attention to the risk of POD.
In summary, delirium may occur more frequently in PD patients. In this study, age, CDR score, MMSE score, and HAMA score were independent infuence factors of POD in PD patients who received DBS. Tis study highlights that it is necessary to assess the cognitive status of PD patients before surgery to identify high-risk patients. Ten, advanced interference treatment would be benefcial to reduce the prevalence of POD, or we could improve the outcomes of PD patients with POD through early identifcation and treatment.

Data Availability
Te data used to support the fndings of this study are available from the corresponding author upon request.

Additional Points
Delirium after implantation of DBS electrodes is common. Age, CDR score, MMSE score, and HAMA score were indicated to be independent infuence factors of POD. It is necessary to assess the cognitive status of PD patients before surgery.

Conflicts of Interest
Te authors declare that they have no conficts of interest.