Deep brain stimulation surgery under ketamine induced conscious sedation: a double blind randomized controlled trial

Study design

This study was conducted at Beilinson Campus of the Rabin Medical Center, Petah Tikva, Israel between September 2020 and January 2022 and was designed as a single center prospective, double-blind randomized placebo-controlled interventional trial. The study protocol was approved by the Institutional Ethics Committee and registered at https://clinicaltrials.gov (NCT04716296). All participants obtained and provided written informed consent.

Patient selection and inclusion criteria

Patients with Parkinson disease (PD), having movement disorder society - Unified Parkinson’s disease Rating Scale (MDS-UPDRS) part III improvement of >30% in levodopa-challenge test, and otherwise eligible and planned for DBS surgery targeting the subthalamic nucleus (STN), participated in the study. If any adverse event happened during surgery including patient’s extreme discomfort or anxiety – the anesthesia regimen was changed accordingly, and the patient was excluded from data analysis.

Outcomes

Sample size calculation

Sample size calculation was based on STN length identified with electrophysiology during DBS surgery. It usually fluctuated between 4 and 6 mm with a standard deviation (SD) of 1.6 mm.^16,17 The minimal STN length used for favorable DBS outcome is 4 mm.¹⁷ Using significance level of 5% (alpha) and power of 80% with SD of 1.6 mm and non-inferiority limit of 1.5 mm, a total 30 patients were required to prove no difference between placebo (saline) and experimental (ketamine) anesthesia protocols, 15 patients per group. Taking into account the follow-up period of 3 to 6 months, 15% of potential dropouts were allowed, and 35 patients were included in the study.

Randomization

Using block randomization technique, 17 patients were allocated to the experimental group and 18 to the control group. A computer-generated randomization list was prepared by a research assistant in opaque, sealed and numbered envelopes containing the group assignment. After inclusion, patients were given a number in the order of their participation in the study. On the morning of surgery, the neurosurgeon opened the corresponding numbered envelope to determine the patient’s group assignment (Group1 or Group2) and informed the anesthesiologist of the appropriate anesthesia protocol.

Blinding

Blinding of all the involved treatment providers (including the neurosurgeon and anesthesiologist) was not ethically feasible in our study. The neurosurgeon and the anesthesiologist were therefore not blinded. Patients, neurophysiologists, and the movement disorder neurologist were blinded throughout the duration of the study. Data analysis was made on blinded data, with the groups only being revealed upon completion of the analysis.

Surgery

DBS surgery targeting the STN, was performed as previously described.¹⁵ Briefly, the STN was identified with high resolution 3T T2-weighted preoperative MRI. Trajectory to target was planned using T1 with gadolinium, 3D volumetric 1 mm slices protocol. Parkinson’s medications were discontinued the evening before surgery. On the morning of surgery, a stereotactic frame (CRW, Integra) was fixated on the patient’s head under local anesthesia (lidocaine 2% and bupivacaine 0.5%, ∼20 cc) and mild sedation (IV midazolam 2 mg). A CT scan was then performed with a fiducial box attached to the stereotactic frame, and the CT and MRI images were fused using Medtronic Stealth 8 navigation system to extract precise coordinates for electrode implantation. The patient was transferred to the operating room (OR), positioned with his head elevated 15 to 30 degrees, and the frame was securely attached to the operating table. An intra-operative CT (O-Arm, Medtronic) was brought into the operating theatre and was located around the patient’s head throughout the procedure.

The surgical procedure consisted of three stages:

1. After skin incision, a burr hole was made in the skull at the desired entry point, then the lead fixation device was mounted to the skull and a brain cannula was lowered to a depth of 15 mm above the target.

2. Recording microelectrodes (Sonus, Alpha Omega Engineering, Ziporit, Israel, typical impedance at 1000Hz = 0.3-0.7 MΩ) were lowered through the cannula to 10 mm above target, and MER was performed by a neurophysiologist, aided by a Hidden Markov Model (HMM) algorithm for STN detection (HaGuide, Alpha Omega Engineering, Ziporit, Israel).^18,19 Monopolar cathodal stimulation (130 Hz, 60 μS pulse width, 0.25-4 mA) through a macro-contact (1.25 mm diameter, 1 mm length, typical impedance at 30Hz =1-3 kΩ) located 3 mm above the micro-contact, was applied at the lower border of the dorso-lateral oscillatory region (DLOR) of the STN to evaluate the TW.

3. If MER recording reveled >4 mm STN length and the TW was found to be acceptable, the recording electrodes were replaced by a permanent DBS lead (3389, Medtronic Inc.), usually with the bottom of the second contact from below at the stimulation depth, i.e., lower border of the DLOR. The lead was fixated to the skull-mounted lead fixation device (Stimloc, Medtronic) and intra operative CT (O-arm, Medtronic) confirmed lead location. Upon confirmation, the skin was closed.

The procedure was carried out for one or both hemispheres. Moderate sedation with propofol was resumed for the lead fixation and skin closure. Directly after, the stereotactic frame was removed from the patient’s head, the patients underwent intubation and general anesthesia and the lead extenders were connected to the brain leads and the internal pulse generator was connected and implanted on the patient’s chest.

Anesthesia protocols for experimental and control groups

Upon entering the OR, patients were given IV paracetamol (1 g) and ondansetron (4 mg), an 18-gauge IV line was established in addition to the previously inserted 20-gauge IV line and cannulation of radial artery (20-gauge catheter) was performed for blood pressure management. All patients were monitored in accordance with The American Society of Anesthesiologists (ASA) guidelines (ECG, Sp02, non-invasive and invasive blood pressure monitoring, and nasal EtCO₂ monitoring (iMDsoft’s MetaVision® and Mindray^©)). Additionally, a BIS monitor (BIS Vista, Covidien Medtronic Dublin, Ireland) was utilized. Antibiotic prophylaxis was administered 30 minutes prior to surgery (either cefazoline 2 g or clindamycin 900 mg).

Initial (before sedation) BIS values were between 92-96. Throughout the first and third stages of surgery, all patients received a continuous propofol infusion to maintain moderate sedation (BIS 60-80). Fifteen minutes prior to the second stage of surgery (MER and clinical assessment), when the burr hole was ready, the propofol infusion was discontinued, and patients received either a placebo (normal saline) or an interventional drug (ketamine), depending on their assigned group:

• The ketamine group received a continuous 0.1% ketamine infusion, (Renaudin, France), at a dosage of 0.25 mg/kg/hr (rate of 0.25ml/kg/hr).

• The saline group received a continuous normal saline (NaCl 0.9%) infusion at a rate of 0.25 ml/kg/hr.

Upon completion of the second stage of surgery, the placebo/interventional drug was discontinued and propofol was resumed until the end of the third stage. Intraoperative blood pressure was closely monitored and maintained within accepted limits (mean arterial pressure 65-85 mmHg, systolic blood pressure < 150 mmHg). In cases of hypertension, treatment was administered in the form of nicardipine, hydralazine or labetalol.

Pulse generator implantation was performed using a standard general anesthesia technique. Anesthesia induction consisted of propofol (1-2 mg/kg), fentanyl (0.1-0.2 mg) and rocuronium (30-50 mg) followed by maintenance with sevoflurane (minimum alveolar concertation 1-1.2). After surgery completion, patients were extubated, transferred to the post-anesthesia care unit (PACU) and transferred to the neurosurgical department the following day. PACU management included administration of IV fluids, antibiotics, analgetic and antiemetic drugs as needed, along with anti-parkinsonian treatment.

Follow-up

All patients underwent a preoperative evaluation (2 months before the surgery) and were subsequently monitored 3-6 months after surgery. The post-op evaluation included UPDRS part III and L-dopa equivalent daily dose (LEDD). Cognitive impairment screened with MoCA test was assessed preoperatively (1-2 months before the surgery) as well as on the second postoperative day to measure early cognitive decline. To gain a comprehensive understanding of patients’ experience during the “awake” (experimental) part of their surgery, we used the ISAS (Iowa Satisfaction Anesthesia Scale),²⁰ a validated questionnaire that measures patient satisfaction with monitored anesthesia care during awake surgeries. Patients were asked to complete the questionnaire within 24-48 hours postoperatively and to refer only to their experience of the experimental part of surgery.

Electrophysiological data collection

Data acquisition for neurophysiological signals (MER, EEG and EMG), was performed utilizing the Neuro-Omega navigation system (Alpha Omega Engineering. Ziporit, Israel), offering precise spatial control of the micro-electrode position and real-time monitoring.

Multi-unit recordings were obtained through the microelectrodes (average impedance at 1000Hz = 0.5 ± 0.12 MΩ), which were inserted into the target area. Recording began 10 mm above target, advancing in increments of 0.4 mm before STN and 0.1 mm within the nucleus. Each site was recorded for at least 4 seconds after 2 seconds of stabilization. EEG electrodes were located at Fp1, Fp2, TP9 and TP10 locations according to the 10-20 system. EMG sensors were attached to the patient’s forearm to collect signals from the flexors (flexor carpi radialis) and extensors (extensor carpi radialis) muscles on both the right and left arms.

Data analysis

All intraoperative data (except spiking data that were recorded only in the experimental stage), was divided into 4 time periods: baseline (patient awake in the OR before propofol infusion), propofol infusion during the first surgery stage (propofol-1); experimental stage during MER (ketamine/saline), and propofol administration during third surgery stage during lead fixation and scalp closure (propofol-2). See Supplementary Fig. 1 for the details.

Spiking data (sampled at 44kHz, 300-6000 Hz SW band-pass filtered, following HW 0.1-9000Hz band-pass filter), recorded from the microelectrode trajectory, was extracted from the Neuro-Omega system. The borders and subdomains of the STN were determined using a previously published algorithm.^19,21 Briefly, the power spectrum density (PSD) and RMS were calculated for each recording site. The RMS was then normalized by the first 6 pre-STN (internal capsule white matter) sites (baseline RMS), resulting in the normalized RMS (NRMS). The peak value (peak RMS) and area under the curve (RMS AUC) of NRMS were computed for each trajectory. Spectrograms of the discharge rate (following absolute operator to expose low-frequency oscillations) of individual patients were normalized by the total power per site and calculated. The length of the STN was extracted and the NRMS and the spectrogram were normalized by the patient’s STN length and then grouped and averaged for each patient group. Percentage STN designated DLOR (DLOR %) by the HMM algorithm and confirmed by an expert electrophysiologist was calculated as: DLOR length/STN length*100. The ratio of beta power (intensity of beta oscillations) within the STN to beta power outside the STN was calculated for each trajectory.

EEG data was sampled at 1000Hz, software band-passed between 1 and 150 Hz with 4^th order Butterworth filter. 50 Hz notch filter was applied to remove electrical noise. The difference between frontal and temporo-parietal electrodes (Fp1 - TP9 and Fp2 - TP10) was taken from each side. Power spectral density was calculated using Welch’s method with Hamming window of 3 seconds (yields a resolution of 1/3 Hz), and 2/3 overlap. EEG spectra were compared using 75% spectral edge frequency (SEF75) and also with area under the curve (AUC) for each EEG frequency band including delta (1-4 Hz), theta (4-8 Hz), alpha (8-13 Hz), beta (13-30 Hz), gamma (30-70 Hz), high frequency oscillations (HFO, 70-150 Hz).

EMG data with initial sampling rate of 11 kHz was resampled to 1.1 kHz and band-passed between 10 and 200 Hz with 4^th order Butterworth filter and rectified. PSD was calculated for each EMG electrode using Welch’s method with Hamming window of 3 seconds, and 2/3 overlap (1/3 Hz resolution). The maximal power of the frequency band typical for parkinsonian tremor (4-6 Hz) was compared between the groups.

BIS measurements were extracted from the BIS device and resampled to 1/60 Hz (1 measure per minute). BIS values from separate patients were linearly interpolated to the same time length for each time period. Averaged BIS values for each of the periods were compared between experimental and control groups.

Hemodynamics measurements consisted of systolic blood pressure (SBP), mean arterial pressure (MAP) and heart rate (HR). Blood pressure values measured by arterial-line and heart rate – by ECG, were extracted from MetaVision® system and underwent similar preprocessing as for BIS values.

Because of possible sympathomimetic properties of ketamine,²² it was decided to measure hemodynamics variability and hypertensive/tachycardic events in the study cohort. Variability was assessed with two common methods:^23,24 coefficient of variation (CV; standard deviation divided by the corresponding mean) and average real variability (ARV; average of the absolute differences between all consecutive BP/HR measurements). Hypertensive/tachycardic event was defined as an increase in BP/HR more than 2 standard deviations of averaged BP/HR in one minute. The frequency of these events was calculated for every patient (number of events divided by length of each time period in minutes). All coefficients were calculated separately for each time period as well as for the whole experiment for each patient.

ISAS questionnaires were analyzed in accordance with the instructions (mean response of all 11 questions), with maximal score of 3 (maximal satisfaction) and minimal of −3.

The accuracy of lead implantation at the target was evaluated by measuring the radial error. The radial error was defined as the distance between the center of the implanted lead at the target plane and the planned target point. This was measured in probe’s eye view at 90° to the trajectory plane on the navigation system, using the fused images of the intra-operative CT and pre-operative MRI.

All analysis was performed using custom Matlab (R2017a or R2021a) scripts.

Statistical analysis

After data preparation, the statistical assumptions were examined. For data that followed a normal distribution, parametric tests were used, while non-parametric tests were employed for non-normally distributed data. The results were presented as mean and standard error of mean (SEM) for normally distributed continuous data, median (interquartile range) for non-normally distributed data, and count (percentage) for categorical variables. To compare groups, the Welch test or Mann-Whitney U test was used for continuous data, and the Chi-squared test for categorical data. Significance was determined as a p-value of less than 0.05 (two-tailed) except for ISAS, where right-tailed Mann-Whitney test was used (since ketamine has anxiolytic and anti-depressive properties) to test for more anesthesia satisfaction in ketamine group. Bonferroni correction was used in case of multiple comparisons (p-value of 0.05 divided by number of comparisons).

Data availability

The data that support the findings of this study are available from the corresponding author, upon reasonable request.

Primary outcomes

STN electrophysiology is similar under ketamine and saline

In all cases, we succeeded to identify the electrophysiological characteristics of the STN, including entry and exit from the nucleus, and motor and limbic subdomains, in both anesthesia protocols. Supplementary Figure 1E and F show examples of spiking activity and STN features of two STN trajectories. Figure 2 shows the average STN features of both groups. NRMS and PSD signatures showed similar patterns, while beta oscillation intensity showed a trend towards higher beta power in the ketamine group (Fig. 2A-B); however, this trend did not reach statistical significance. STN length was similar between ketamine and saline groups (0.2 (−0.415, 0.815) mm - absolute difference (95% CI); Fig. 2C, D.). There was no significant difference in NRMS peak and AUC, percentage of DLOR, DLOR length (Fig. 2C) and power spectrums, including theta, alpha, gamma and HFOs (Fig. 2D).

• Download figure
• Open in new tab

Figure 2 STN electrophysiology under ketamine sedation was noninferior to STN electrophysiology under saline.

Ketamine group is indicated in light orange and saline group is indicated in light blue. (A) Average normalized root mean square (NRMS) of ketamine (left) and saline (right) groups along the recorded subthalamic nucleus (STN) trajectory. Data is normalized by the total STN length of each patient. (B) Spectrogram of STN trajectory in ketamine (left) and saline (right) groups. Data is normalized by total power at each depth. X-axis is normalized as in A. Color code indicate the fraction of power of frequency bin out of the total power. (C) From left to right: STN length in mm; DLOR length in mm; DLOR percentage out of total STN length; peak NRMS; NRMS area under curve. (D) From left to right: Noninferiority test for STN length between ketamine and saline groups; relative power for 4 frequency bands: 4-12 Hz, 13-30 Hz, 31-70 Hz, 71-150 Hz. Significance assessed with Mann-Whitney test. N.S – p-value >0.05. RMS – root mean squared; DLOR - dorso-lateral oscillatory region.

Anesthesia level and patient cooperation during the surgery

The patients’ anesthesia depth and level of consciousness were continuously monitored through clinical observation and the utilization of the Bispectral index, as shown in Fig. 3A. At baseline, all patients exhibited BIS values consistent with an awake stage (ketamine group 92.6 (91.5-96.2) vs. saline group 92.8 (91.4-94.3), p = 0.6). During the first and second propofol sedation sections, BIS decreased to the level of light/moderate sedation (79.15 (73-81.72)), and there was no difference between the two groups (p = 0.3). During ketamine/saline administration, BIS values increased again to awake level (ketamine group 88.82 (87.31-90.55) vs. saline group 89.14 (88.06-89.95), p = 0.8). Patients of both groups were alert enough to fully cooperate during MER and clinical assessment. Some patients from the ketamine group reported a pleasant, dream-like experience during surgery, such as boat trips or odyssey-like journeys. None of the patients reported a negative hallucinatory experience either during or after surgery. Most patients had mild euphoria under ketamine and were relaxed.

• Download figure
• Open in new tab

Figure 3 EEG and EMG analysis showed similar patterns in the ketamine and saline groups.

Ketamine group is indicated in orange and saline group is indicated in blue. Experimental stages: baseline (beige), propofol-1 (green), ketamine/saline (pink), propofol-2 (purple). (A) BIS values during the whole experiment, time normalized. (B) EEG analysis. From top to bottom: spectrograms for ketamine and saline groups; mean power spectral density comparisons for experimental stages, from 1 to 150 Hz - delta, theta, alpha, beta, gamma and high-frequency oscillations; spectral edge frequency 75% comparisons calculated for frequency range 1-150 Hz. (C) EMG analysis. From top to bottom: spectrograms for ketamine and saline groups; mean power spectral density comparisons for experimental stages, shown from 0.1 to 30 Hz; maximal power comparisons from frequency range between 3 and 6 Hz for different stages. P-values are assessed using 2way-ANOVA with Bonferroni correction. N.S – p-value >0.05; *** - p-value < 0.001. PSD – power spectral density; SEF75-spectral edge frequency 75%.

Secondary outcomes

Hemodynamic stability is superior under ketamine

Hemodynamics analysis is shown at Fig. 4 and Supplementary Figs. 3,4. In both groups, baseline BP was hypertensive (average SPB = 157 mmHg, average MAP = 104.3 mmHg) followed by a decrease to normal ranges during the surgery, probably due to the effect of antihypertensive therapy and reduction of anxiety by the propofol treatment. There was no difference between the groups in average BP values during baseline, propofol-1 and ketamine/saline stages. However, at the propofol-2 stage there was a larger decrease in BP in the saline group compared to ketamine (For SBP: 120.23 (113.68-131.64) mmHg vs 110.47 (101.08-120.77) mmHg, p = 0.013; for MAP: 78.67 (74.52-85.77) mmHg vs 73.07 (69.39-76.79) mmHg, p = 0.05, ketamine vs saline groups, respectively).

• Download figure
• Open in new tab

Figure 4 Systolic blood pressure patterns and antihypertensive drug dosage were similar between groups.

Ketamine group is indicated in orange and saline group is indicated in blue. Experimental stages: baseline (beige), propofol-1 (green), ketamine/saline (pink), propofol-2 (purple). (A) Systolic blood pressure dynamics during the experiment. From left to right: 10-minute baseline measurement – first 30 minutes of propofol administration; last 30 minutes of propofol sedation and first 30 minutes of ketamine/saline administration; last 30 minutes of ketamine/saline and first 30 minutes of propofol-2 sedation. First and last 30 minutes of each stage can overlap. Violin plots on the top – average SBP comparisons between the two groups according to stages. (B) Systolic blood pressure during whole experimental period after time normalization. (C-E) From top to bottom: Comparison of CV (coefficient of variation), ARV (Average Real Variability), frequency of hypertensive events (more than 2 standard deviations of SBP) in different experimental stages. (F) Comparisons of total dose of antihypertensive drugs for nicardipine, labetalol and hydralazine. Significance assessed with Mann-Whitney test. N.S – p-value >0.05; ** - p-value < 0.01.

Blood pressure variability analysis showed less consistent results. There was no difference in BP variability between the saline and the ketamine groups when separated into different surgical stages (Fig. 4C, D). On the other hand, when taking into account the whole experiment, there was an increase in BP variability for both variability indices in the saline group (For SBP: SV 0.1 (0.09-0.13) vs 0.13 (0.11-0.15), p = 0.018; ARV 10.49 (9.05-12.95) vs 11.77 (10.29-16.19), p = 0.047; for MAP: SV 0.09 (0.07-0.1), vs 0.12 (0.09-0.14), p = 0.025; ARV 5.82 (4.7-7.85) vs 8.47 (5.97-9.2), p = 0.028; ketamine vs saline groups, respectively). There was no difference in frequency of hypertensive events between the groups (Fig. 4E). Finally, no difference was found between the two groups in HR analysis (Supplementary Fig.4).

Clinical outcomes and surgical accuracy are similar in both groups

In both groups, there was no early cognitive decline after surgery (decrease from preoperative MoCA 0 (−4.71-4.77) %, p = 0.95 for ketamine group; 3.45 (−4.97-10.92) %, p =0.22 for saline group). Despite the trend towards more cognitive decline in the saline group, there was no significant difference between the groups (p = 0.35; Table 1, Fig. 5A). There was 44% decrease in medication in LEDD after surgery (62.5 (53.6-85.86) % vs 48.53 (36.92-57.31) %, p =0.13; ketamine vs. saline, respectively; Fig. 5B). Surgery led to 69% average reduction in UPDRS-III on medication and with stimulation on compared to preoperative off med UPDRS-III (Fig.5C) with no difference in the ketamine/saline groups (p = 0.5). There was no difference between the accuracy of microelectrode/lead trajectory (radial error, 0.8 (0.375-1) vs 0.6 (0.325-0.9), p = 0.6; ketamine vs saline groups, respectively). No complications were reported in the two study groups, except one patient allocated to the saline group that was subsequently excluded due to peri-lead edema²⁵.

• Download figure
• Open in new tab

Figure 5 Surgery outcomes were similar, yet patients under propofol-ketamine sedation were more satisfied with their anesthesia.

Ketamine group is indicated in orange and saline group is indicated in blue. (A) MoCA scales, preoperative and postoperative comparisons. (B) LEDD, preoperative and postoperative comparisons. (C) UPDRS III, preoperative off-medication, preoperative on-medication and postoperative on-medication and on-stimulation comparisons. (D) ISAS score comparisons. (E) Modified ISAS score comparisons. MoCA – Montreal Cognitive Assessment scale. LEDD - Levodopa equivalent dose in mg/day. UPDRS III - Unified Parkinson Disease Rating Scale. ISAS – Iowa Satisfaction with anesthesia scale. P-value – Mann-Whitney test, two-tailed for A-D; right-tailed for E, F. N.S – p-value >0.05; * - p-value < 0.05.

Asleep or not asleep?

The major challenge in DBS surgery is precise lead localization to ensure wide therapeutic window and avoid long term stimulation induced side effects. As many centers move towards “asleep” DBS due to the ease of the procedure for both patient and neurosurgeon, MER and clinical evaluation of the stimulation effects are being abandoned, and anatomical navigation becomes the sole method to assure accuracy. Nevertheless, STN borders and functional subdomains are still not easy to define from standard MR images, and most centers do not use intraoperative-MRI techniques to compensate for brain shift during surgery. Other sedatives used so far in DBS surgery significantly affect MER quality, Parkinson’s disease motor symptoms, and patient cooperation.^2,26–28 The results of pure anatomic based approaches, or combined anatomic/physiological approaches under hypnotic sedation/anesthesia, are most probably, even if not proven yet, – reduced accuracy.

Our study reveals that low-dose ketamine conscious (dissociative) sedation has multiple advantages. It facilitates intra-operative electrophysiological identification of the STN borders and activity including beta oscillations, it improves patient experience, and increases cooperation by reducing anxiety, discomfort and pain. Moreover, it allows for clinical evaluation during stimulation, as tremor and rigidity are barely influenced by the drug, and therefore results in non-inferior (compared to awake procedure) of lead accuracy and clinical outcome. In addition, ketamine induces fast and stable conscious sedation, with a dissociation-like state, no respiratory depression and optimal hemodynamic stability. In the continuous low dose drip used in this study, it does not cause hallucinations or increase post-surgical complications (cognitive decline, seizures, peri-lead edema, etc.). These results are in agreement with our previous non-human primate and retrospective human studies showing similar STN neurophysiology recordings under ketamine and awake conditions.^14,15 Overall, it seems to be the most reliable, effective and safe option of all sedatives for DBS.

Ketamine’s mechanism of action

Ketamine, an N-methyl-D-aspartate (NMDA) antagonist, is a dissociative anesthetic agent discovered over 60 years ago, and has been in wide clinical use ever since. Its main characteristics are anesthesia, analgesia, and euphoria. Nowadays it is commonly used in pre-hospital care for pain²⁹ and anesthesia of trauma patients,³⁰ as well as in treatment of resistant status epilepticus,^31,32 despite past arguments citing increasing intra-cranial pressure and pro-epileptic effects.^33–36 More recent studies have even discovered its protective properties in severe traumatic brain injury patients.^37,38 In recent years, low dose ketamine has also been used for the treatment of major depressive disorder. Moreover, its anxiolytic, cognition enhancing, and anti-epileptic effects have garnered attention in ongoing research.^39–43

Ketamine acts via NMDA receptors and HCN1 channels to enhance neural activity and produce brain oscillations that are related to its psychotropic effects. Human intracranial recordings revealed that ketamine produces gamma oscillations in the prefrontal cortex and hippocampus, and delta oscillations in the posteromedial cortex. The latter was previously proposed as a mechanism for its dissociative effect.⁴⁴ The potential role of ketamine on depression might be achieved through modulation of AMPA (α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid) receptor signaling and function resulting in diminished long term potentiation in the nucleus accumbens of mice lasting at least 7 days.⁴⁵ In our study, we show an increase in low gamma and decrease in delta/theta oscillatory cortical activity in prefrontal EEG recordings during ketamine administration. These results are similar to previously published data.^46,47 The increase in cortical activity under ketamine is in contrast with robust attenuation of cortical and basal ganglia activity seen under hypnotic sedation (e.g., propofol and dexmedetomidine), even one hour after drug cessation.²⁸ In line with the changes in prefrontal EEG activity, most patients under ketamine sedation in our study were in a good mood, sometimes even slightly euphoric, during surgery. This is in agreement with the reported effect of ketamine on mood.⁴⁸ In addition, patients under ketamine did not experience cognitive decline after surgery and did not report amnesia for this part of the surgery. These findings are similar to previous results reported in the literature.⁴⁴

Hemodynamic effects of ketamine

Ketamine was reported in the past to have unwanted hemodynamic effects for neurosurgery including increase in blood pressure and intra-cranial pressure (ICP). Nevertheless, and in contrast to most neurosurgical procedures, DBS surgery requires average-high blood pressure and ICP to diminish inadvertent effects like pneumocephalus (intracranial air accumulation) and brain shift during the procedure, that affect accuracy of lead localization.

In our study, ketamine resulted in stable blood pressure and heart rate without a decrease in these vital measures, that can be seen with other common sedatives like propofol and dexmedetomidine.^49,50 The maintenance of hemodynamic stability is of particular importance in DBS surgery, where both increase and decrease in blood pressure may result in suboptimal clinical outcome, including intra-cranial bleeding and reduced lead accuracy,⁵¹ respectively.

Dissociative state under ketamine sedation

Another important difference between ketamine and previously used sedatives is the level of consciousness maintained during surgery. Due to the induction of a dissociative state with ketamine, the degree of “brain-wakefulness” of the patient is high (as is reflected from the EEG and high BIS levels). This allows for both clinical testing with a cooperative patient and airway securing without the fear of apnea events. Decreased level of consciousness causes CO₂ retention and therefore might enhance brain hyperemia and secondary inflammatory injury resulting from the implanted electrodes. These, in turn, might affect postoperative pulmonary complications (atelectasis, aspiration pneumonia, etc.) as well as neurological complications, such as acute delirium, peri-lead edema, and long term cognitive decline.^52,53 Therefore, the combination of ketamine with interleaving periods of propofol moderate sedation might help to mitigate these unwanted complications and is preferred over a propofol-only regimen.

Disadvantages of ketamine use

Ketamine itself might also have some unsavory effects, and caution needs to be taken when using it in DBS surgery. Higher ketamine dose might cause undesired hallucinations during surgery, in addition to the dissociation state induced.⁵⁴ This might hamper patient cooperation and result in an overall negative experience for the patient. It is probably a result of the non-selective effect of ketamine on neurotransmitter release in high, sedative dose, and can be avoided by keeping a continuous drip at a low dose.⁵⁵

Another issue is patient variability of response to the drug (using the same dose). Some patients might become sleepy/mildly sedated even when low ketamine dose is used, which, in turn, might affect their cooperation level with stimulation testing.⁵⁶ However, this phenomenon is common with other sedatives as well, and is reversible with dose reduction.

Finally, in some patients ketamine results in a relative decrease in tremor, due to either the decrease in anxiety or non-specific sedative effect. However, in most cases, it is still possible to evaluate tremor during stimulation testing. Nevertheless, this might be of special importance in tremor predominant PD patients and essential tremor patients.

Study limitations

Our study has a few limitations, mainly the small number of patients enrolled and inability to blind the anesthesiologist and the neurosurgeon. In addition, we used racemic ketamine, while some of the characteristic effects of ketamine were previously described with S-ketamine.⁵⁷ We did not evaluate possible long-term effects of ketamine use on cognition and short- and long-term effects on mood, even though short administration is not supposed to cause long term mood effects.⁵⁸ Also, we did not test differential effect of various doses of ketamine on consciousness level, quality of MER and patients’ subjective experience. Future studies should address all of these questions, as well as the effects of ketamine in DBS for other movement disorders like dystonia and essential tremor, and for other targets, like the globus pallidus and thalamus. Finally, future studies might test the possible use of this regimen in other awake neurosurgical procedures like tumor resection and epilepsy surgery.