Subthalamic Deep Brain Stimulation under General Anaesthesia for Parkinson’s Disease: Institutional Experience and Outcomes

Since the approval of deep brain stimulation (DBS) for the treatment of Parkinson’s disease (PD) back in 2002 [1], the field has experienced various substantial improvements [2]. Among them, direct targeting has had a remarkable impact in the patient’s experience through the surgery and in the overall workflow and duration of the procedure [3, 4].

Traditionally, DBS surgery has been performed with indirect preoperative targeting and intraoperative neurological evaluation in an awake patient, under local anaesthesia (LA) [5, 6]. Recently, the procedure has shifted towards improving the patient’s comfort with the “asleep” technique, under general anaesthesia (GA) which relies on direct targeting coupled with intraoperative imaging for verification of stereotactic accuracy. Compared to indirect targeting, in which lead placement is based on universal stereotactic coordinates obtained from anatomical atlases, direct targeting relies on high resolution magnetic resonance imaging (MRI) for preoperative planning, and in intraoperative three-dimensional imaging devices for co-registration during electrode implantation. Thanks to the image quality provided by 3-Tesla MRI, the subthalamic nucleus (STN) can be readily identified and selected as a target in the neuro-navigation workstation. This direct targeting method has the advantage of accounting for individual variations in the anatomy of the basal ganglia [7]. When the individualised plan is co-registered and fused with a 3D image acquired intraoperatively, the electrodes can be safely placed under GA [4, 8‒13]. A final control scan can be used to confirm the adequate position of the electrodes within the STN, according to the preplanned trajectory. Yet, as no neurological exploration is done intraoperatively, an assumption is required in the asleep DBS that accurately placing the lead at an anatomical target will correlate with improved outcomes [14].

It has been reported and strongly supported that, under these conditions, microrecording and neurological examination may not be necessary during the surgery for subthalamic DBS placement [4]. This fact has significantly impacted the surgical procedure, as it can now be done under GA. However, how this applies to a general practice scenery (outside clinical trials and in a moderate caseload centre) has been scarcely reported [15‒23].

In this prospective study, we have evaluated the clinical outcomes of a cohort of patients treated in our institution with subthalamic DBS placement under GA for patients with PD, without intraoperative microelectrode recording. The primary aim was not to compare the GA and the LA procedure, but to illustrate that with a multidisciplinary dedicated team with sufficient experience, patients with PD can be operated under GA with acceptable safety and effective outcomes. Our results were then compared to those reported in the literature.

This study has been approved by the Institutional Bioethics Committee and complies with national legislation and the Declaration of Helsinki.

The sample included a prospective cohort of all patients operated for PD with bilateral subthalamic electrode implantation at our institution between January 2021 and December 2022. Only patients who were operated for DBS placement under GA for the first time were included in the analysis. The procedures were performed by two specialised functional neurosurgeons, with more than 10-years’ experience in the field (J.R., P.R.). We excluded patients who had already been operated on (i.e., electrode replacement or contralateral electrode implantation).

All patients underwent a multidisciplinary evaluation before the indication of DBS treatment. This included neurology evaluation, neurosurgery consultation, neuro-anaesthesia consultation, neuropsychological testing, and physical therapy. Clinical evaluation included off medications and on stimulation examinations. Patient factors that favoured selection of the STN for DBS target included significant medication intolerance or medication side-effects (dyskinesia, wearing off symptoms, delayed-on medication effect, etc.).

All patients received directional eight-contact electrodes (Medtronic B33005 or Boston Scientific Cartesia DB2202). A generator (Medtronic Activa RC 37612 or Boston Scientific Vercise genus DB1216) was placed under the right clavicle. The direct targeting strategy and trajectory planning remained constant during the whole period. The plan was designed by either one of the two surgeons, and always cross-checked and reviewed by the other. The surgical procedure followed the same steps already published by our group [24]. Importantly, all the procedures were carried out under GA. Of note, anti-parkinsonian medication was maintained even in the morning before the intervention, as an intraoperative neurological exploration was not necessary.

3T MRI with FLAIR sequence was used for target selection, and T1-contrast sequence was used to select the optimal entry point and trajectory. Direct visual anatomical targeting followed the method previously described by Bejanni. In the operating room, an O-Arm^®2 fluoroscopic device (Medtronic Inc.) was used for intraoperative stereotactic sequence acquisition and fusion with preoperative images. Preplanned coordinates (referenced to AC-PC line) were then transferred to the Leksell G frame and adjusted to 0.5 figures. Then a single step direct electrode placement with ventral contact on the target was performed under GA. After electrode placement, a final control O-Arm sequence was acquired for control purposes. The accuracy of the electrode position was calculated intraoperatively, in the neuro-navigation station (Medtronic), by measuring the radial error between the desired trajectory of the preoperative planning and the actual trajectory in the co-registered intraoperative image acquisition. A threshold of >2 mm deviation from the desired location was set for the need to reposition the electrode.

The day before surgery, patients were premedicated with a single dose of diazepam 5 mg; a second dose of diazepam 5 mg was administered 2 h before surgery. Their routine parkinsonian medication was maintained including the morning of the surgery.

Patients arrived at the operating room, wearing graduated compression stockings. Venoclysis and premedication with midazolam (depending on neurological status) were administered. Standard monitoring (electrocardiography, pulse oximetry, temperature and noninvasive blood pressure) (Carescape^® B850, GE Healthcare, Helsinki), together with Bispectral Index (BIS^® Complete 2-channel, Medtronic, Minneapolis, USA), neuromuscular block (“train-of-four” surface electrodes, TOF), and cerebral oximetry (near infrared spectroscopy, NIRS) (INVOS^® 5100C Oximeter, Medtronic) were placed.

Anaesthetic induction was performed with lidocaine 0.5 mg/kg, Target Controlled Infusion (TCI) system (Orchestra^® Base Primea, Moduls DPS, Fresenius Kabi, Bad Homburg, Germany) of propofol and remifentanyl, and rocuronium 0–5 mg/kg. Orotracheal intubation and mechanical ventilation (Primus, Drägerwerk AG & Co, Lübeck, Germany) were instituted. Monitoring EtCO₂ and ventilatory parameters were modified according to serial blood gas testing. PaCO₂ were maintained between 35 and 40 mm Hg to avoid changes in cerebral blood volume. Anaesthesia was maintained with propofol plus remifentanil and rocuronium infusion, in order to maintain a targeted BIS value between 40 and 60 and TOF monitor with 1–2 responses. Central temperature was kept between 35 and 36°C, using an air blanket, if necessary. Surgical positioning, including Leksell frame positioning, was performed according to the specific neurosurgical procedures. Strict haemodynamic management during surgery was carried out.

All patients had a basal 3D-intraoperative stereotactic O-Arm^®2 sequence acquisition at the end of the procedure to verify the final position of the electrodes. After the surgery, patients were awaked and controlled during 6 h in the postanaesthetic care unit. According to the in-house protocol, patients are transferred to the regular ward after 6 h of surveillance, provided they have been stable and without new neurological findings on examination. Before this transfer, patients undergo an additional CT exam in a conventional scanner to rule out postoperative complications, mainly intracranial bleeding. May a complication occur, they would be kept in the postanaesthetic care unit for an overnight monitoring or transferred to the intensive care unit, according to the severity of the complication and the underlying comorbidities.

The mean operative period included the duration of the whole surgical procedure that is patient positioning, intraoperative imaging acquisition, and co-registration, electrode placement, patient repositioning, and generator implantation. The surgical duration did not include the anaesthetic induction and awakening times. The mean length of hospitalization was recorded. Surgical-related complications were noted, both in the early and late postoperative period. The need of electrode replacement intraoperatively was documented; a threshold of >2 mm deviation from the preplanned target was set for electrode relocation. The need for a later reintervention to relocate the electrode was also noted.

A minimum of 6 months follow-up was allowed for programming optimization before clinical outcome evaluation. This was performed by two senior neurologists, dedicated to movement disorders (F.V., A.S.).

As part of their evaluation, all patients had baseline Unified Parkinson’s Disease Rating Scale (UPDRS) III ON/OFF testing [25], Parkinson’s Disease Questionnaire (PDQ-8) [26], and calculated Levodopa Equivalent Daily Doses (LEDDs) [27]. Patients programming started a week after the surgery and continued for up to 6 months after the intervention. Then, the same preoperative evaluation was performed, including UPDRS III, PDQ-8, and LEDDs. Several factors that could have impacted the benefit of the DBS therapy in terms of quality of life – namely premorbid quality of life (PDQ-8 score before surgery) – disease stage (years of disease evolution before surgery) and the amount of preoperative medication (LEDDs before surgery) were analysed.

Normality testing was performed with Kolmogorov-Smirnov test. The data were found to have a nonnormal distribution (p ≤ 0.01), and so a Wilcoxon test was used to compare the pre- versus post-repeated measures. A p ≤ 0.05 was set for statistical significance. Data were analysed with IBM SPSS version 27 (IBM Corp.).

To compare our results to those published by other groups for DBS in the STN under GA, a literature review was conducted. The research strategy was applied in Pubmed database with the following terms: (“General anesthesia” OR “asleep”) AND subthalamic AND dbs. Only articles with >10 cases of DBS in the STN were screened. Articles were included if they offered information on motor outcomes, and/or levodopa dependence, and/or surgical-related complications.

Between January 2021 and December 2022, a total of 93 patients with PD were operated for DBS implantation under GA. Of those, 4 patients were excluded from the analysis, as they had been first operated outside the selected time period (2021–2022) and underwent a revision surgery for electrode replacement during this time. Therefore, 89 patients primarily operated were included in the study.

Table 1 summarises patient’s characteristics. Among them, 48% were female. Mean age was 59 (SD 8) years. And the mean duration of the disease prior to surgical intervention was 11 (SD 4) years.

Bilateral electrode placement under GA could be completed in all 89 cases. Intraoperative electrode replacement was not necessary in any of the cases, with an accuracy threshold of 5 mm from the planned target as observed in the intraoperative control CT scan.

Mean surgical duration was 217 (SD 44) minutes, and mean length of hospital stay was 3 (SD 1) days. There was one (1%) surgical-related complication, a late infection. There were three non-surgical-related complications in the postoperative period, namely a urine infection, a prostate-related acute urinary retention and an anaesthesia-related arrhythmia. One patient (1%) needed a late reintervention for electrode replacement due to suboptimal effects.

Information about the final programming parameters can be found in Table 1.

Preoperatively, the mean values of the evaluated scales were UPDRS III 35 (SD 16) and PDQ-8 12 (SD 5). After the intervention and the optimised programming, the mean values of these changed to UPDRS III 13 (SD 8) and PDQ-8 6 (SD 4) (Table 2).

Only 4 patients had worse UPDRS III scoring after DBS and none of them expressed a decrease in quality of life (PDQ-8). Overall, a significant and clinically relevant improvement in both parameters was observed after the DBS treatment. Mean decrease of UPDRS III was 61% (21 points) (p < 0.001) and that of PDQ-8 was 50% (6 points) (p < 0.001). Among the factors potentially implicated in the benefit obtained after DBS implantation, only the preoperative PDQ-8 value seem to have a significant influence in the magnitude of clinical effect, i.e., in the degree of reduction of the UPDRS III and PDQ-8 (p < 0.001).

The preoperative LEDDs was 1,049 (SD 608) mg and after the treatment decreased to 331 (SD 334) mg. None of the patients required an increase in medication during the first year after DBS implantation. Overall, the daily doses of medication were halved in our sample (mean decrease of 67%, 709 mg, p < 0.001).

In our centre, the DBS surgical protocol changed in 2013, when we began to transition towards a GA procedure. Ever since, it has been a journey of continuous learning and evolution. Not only did the whole team improve their knowledge and performance (anaesthesiologists, neurologists and neurosurgeons), but also the devices have significantly evolved. For instance, we know use eight-contact directional electrodes, which have remarkably changed the programming process, increasing its complexity, but more importantly, facilitating to overcome some side-effects and improving the overall motor performance. For these reasons, we decided to include just the last 2 years of experience with the GA DBS procedure, as these represent the plateau of our learning curve and provide information on a homogenously treated cohort of patients, in terms of anaesthetic management, surgical planification and performance, postoperative programming, and drug tapering schemes.

The aim of our study was not to compare GA versus LA, as the nature of the design would not have allowed to obtain definitive conclusions in this controversial topic. Rather, we wanted to illustrate that with a multidisciplinary team with sufficient experience, patients with PD can be operated under GA with acceptable safety and effective outcomes. According to the literature, DBS under GA seems at least as effective as that performed under LA with or without micro-recordings [28]. According to our data, a significant improvement was seen motor and functional outcomes, comparing the preoperative status with that after DBS placement and program optimisation. Concretely, a mean decrease of 61% in the UPRS III was attained after 6 months. These results are in line with those reported in larger series under GA [5, 29]. In fact, most patients could discontinue or reduce the levodopa medication, and overall LEDDs were decreased by 67% at 6 months.

The ability to maintain the effectiveness even when intraoperative neurological examination is not performed as a confirmation of the electrode position is possible due to the implementation of intraoperative imaging for co-registration with the preoperative planning and for final electrode location control. In fact, the increase in MR resolution and the emergence of intraoperative 3D imaging acquisition devices can be considered as a benchmark in DBS surgery [11, 30].

In our institution, we have established a 2 mm deviation threshold for intraoperative electrode reposition. This is based in our previous experience with directional electrodes, and in agreement with other groups [31]. It must be noted that the optimal cutting point is yet to be defined and that some groups have adopted a more restrictive accuracy goal. However, effectiveness should be balanced with security and each trajectory entails a non-negligible risk of haemorrhage, with potentially devastating results [32]. Moreover, when the deviation is rather small (<2 mm), the probability that the new trajectory follows the same route as the previous one is high, given the differences in the resistance properties of the brain tissue.

Following these premises, none of the electrodes needed a reposition during the primary surgical procedure in our selected cohort. We did, however, exclude patients (n = 4) who were having a revision surgery from this analysis, as explained in the methods section. In the period evaluated, there was 1 patient who needed a second surgery to reposition an electrode. She had first been operated in the year 2019 under GA condition. Her symptoms were refractory in the left side of the body, even after optimal programming. After a multidisciplinary discussion, she went on to have a revision surgery, under GA conditions, and her right electrode properly positioned within the STN. Motor outcomes improved after the second intervention.

According to the literature, the GA procedure may have lower overall complication rates [33]. However, there are also studies claiming a lower rate of side-effects following awake intervention [34]. In our sample of patients operated under GA, the rates of surgical complications were low (1/89), and in line with those reported in previous studies [15‒23]. This supports the idea that DBS can be pragmatically performed under GA in a safe manner by a dedicated neurosurgical and neuroanaesthetic team yet does not allow a direct comparison with a LA protocol. Previous experiences have demonstrated that the duration of the surgery under GA could be reduced compared to the “awake” procedure – i.e., with microelectrode recording and neurological exploration- [35], therefore both the brain shift due to pneumocephalus and the anaesthetic-related management are improved. In our centre, before the GA procedure was implemented, patients were taken off their medication at least 24 h before the procedure, so that they could be accurately evaluated during the intraoperative stimulation. This preoperative deprivation along with longer surgical times without anti-parkinsonian medication meant that the patients could enter a long OFF period, which may prolongate the awakening times after the surgery. In the GA procedure, the anti-parkinsonian medication is maintained in the morning of the surgery and resumed at midday, thus the awakening process may become smoother for the patient. Nonetheless, a two-staged procedure under a LA protocol could also be an option to minimise this problem in high-risk patients.

Conversely, since neurological exploration cannot be performed intraoperatively, DBS under GA requires an assumption that accurately placing the lead at an anatomical target will correlate with improved outcomes. In this sense, it has been noted that it may be associated with a higher rate of side-effects from the stimulation [33]. Indeed, one of the advantages of the LA procedure is that it might help increase the therapeutic/stimulation width by testing for high-amplitude stimulation side-effects intraoperatively; something that could impact the long-term outcomes in a progressive disease. Moreover, the accuracy at determining the final electrode position immediately after surgery may be hampered by brain shift due to pneumocephalus; this is another major concern about the GA procedure [31, 32]. Ultimately, it should be stressed that the literature available to date does not definitively support one technique over the other, and the choice should be based on the institutional experience and considering patient’s preference as much as possible.

Several authors have previously demonstrated that quality of life improves after the implementation of DBS in patients with PD, as a result of improved mobility and activity of daily living and reduced bodily discomfort. Several factors including presurgical quality of life, dopaminergic medication, disease stages, and depression seem to correlate with postsurgical changes in quality of life [36, 37]. In our patients, only the preoperative value of quality of life (PDQ-8) seems to have a significant influence on the postoperative outcomes.

A comparison in terms of efficacy and safety between our data and that reported in the main asleep DBS series can be found in Table 3.

STN DBS for PD can be safely performed under GA, in a pragmatic clinical environment. The effectiveness in ameliorating motor symptoms, the ability to reduce the drug load, and the improvement in quality of life demonstrated in clinical trials can be reproduced under more generalised conditions. The outcomes achieved and hereby presented seem to be in line with those reported in “awake” patients, while the safety parameter could be better with the patient under GA. Notably, multidisciplinary committees for patient selection, and a dedicated surgical and anaesthetic team are required to offer the best outcomes. This is the culmination of long journey of teamwork, from a strong background of “awake” DBS, through a progressive evolution towards refined direct targeting, meticulous electrode placement, custom anaesthetic management, and upgraded programming schemes.

This study protocol was reviewed and approved by the Institutional Ethics Committee of the Hospital Clínic of Barcelona (HCB/2023/1077). Patients gave their written informed consent.

The authors have no conflicts of interest to declare.

This study was not supported by any sponsor or funder.

Conceptualisation: P.R., J.R.A., F.V., and R.V. Data curation: P.R., A.M., A.S., M.G., and N.R. Formal analysis: P.R., J.R.A., A.M., A.S.G., and R.V. Investigation: P.R., D.A., A.S.G., and N.R. Methodology: P.R., A.M., and R.V. Project administration: P.R. Resources: J.R.A., F.V., and R.V. Supervision: J.R.A. and F.V. Validation: P.R. and A.M. Visualization: all authors. Roles/writing – original draft: A.M. Writing – review and editing: all authors.

All data generated or analysed during this study are included in this article. Further enquiries can be directed to the corresponding author.