High cervical spinal cord stimulation (SCS) is used in the management of disorders of consciousness (DOC) [1, 2, 3], especially in patients in a minimally conscious state. It has been shown to have an important role in modulating the complexity of neural activities and promoting functional recovery [4, 5]. Its target stimulus area is the dorsal columns of the C2-4 vertebrae. Traditionally, C5 laminectomy has been performed in the prone position to advance the electrode array craniad to the target zone, where a pulse generator is implanted in the lateral abdominal region [6]. However, one limitation of this procedure we observed in clinics is the challenge of ensuring that the electrode has completely covered the midline of the spine.

When high cervical stimulation is used in pain management, its most common indication, the clinical determination of the electrode position is accomplished under local anesthesia. However, intraoperative sensory tests cannot be performed in patients with DOC as they cannot effectively respond to stimuli. Thus, to increase the effectiveness of therapy, accurate implantation of electrodes at predetermined locations is a prerequisite. However, literature on the technical aspects of cervical SCS implantation is scarce, and few studies have documented the accuracy of DOC-SCS implantation.

In this study, we report our findings on an improved paddle-type C2-4 spinal cord stimulator implantation technique in patients with DOC operated in the lateral decubitus position. A C2-3 dural exposure procedure was adopted to overcome the limitation of previous surgical methods, such as the inability to ensure complete midline coverage. Then, we compared the rate of complete midline coverage by the electrode array in patients with DOC who underwent conventional surgery with those who received the proposed C2-3 dural exposure procedure and further investigated the determinants of complete midline coverage by the electrode array.

In this study, we proposed a C2-3 dural exposure procedure and compared its efficacy with conventional surgery. Our findings showed that although the caudal side of the electrode could be fixed in the midline, the electrode tip would still be skewed to one side in some patients, as viewed on early postoperative CT scans. These contacts that deviated from the midline would often produce a jolting stimulation corresponding to dorsal root involvement, resulting in patients’ discomfort and limiting the selection of program control parameters [8]. Specifically, according to the overlapping percentage of the contacts, complete midline coverage could be divided into two cases: (1) the contacts pair is either symmetrically distributed on both sides of the midline or (2) only one contact in a pair covers the midline (Fig. 3). Symmetrical stimulation is easier to achieve with the former, while the latter requires unilateral contact stimulation to ensure bilateral effects.

Fig. 3.

Diagrammatic coronal view depicting the electrode array placement. (A) The electrode array enters the epidural space through the laminectomy level, while the cephalic contacts deviate from the midline. (B) Dural exposure between C2-3 ensures full coverage of the midline by the electrode array.

The proposed C2-3 dural exposure procedure was associated with a significant reduction in cephalic contact deviation (Fig. 3) and solved difficulties in programming. Compared with the traditional method of C5 laminotomy and intraoperative adjuncts, including fluoroscopy, the C2-3 dural exposure procedure has the following advantages: (1) the electrode array can be implanted under direct vision, thereby avoiding parallax and visual alignment errors during photography; (2) it is easier to separate the epidural space at the C2-3 level to adjust the electrode array if the surgeon is not satisfied with the initial placement site. Such adjustment is difficult when using the traditional method because the first attempt often forms a misguided channel within the epidural ligaments, and; (3) potential dural laceration could be avoided by dissecting the dorsal meningovertebral ligaments under direct vision.

A correlation between electrode height and complete midline coverage by the electrode array was observed in this study. The height of the electrode was calculated from the lower edge of C2 because a force of lateral slippage would often be experienced as the electrode array passes below the C2 spinous process. In the prone position, the same obstacle can be encountered. This is not attributable to pathological spinal stenosis but to the complexity of high cervical SCS [9], which can often lead to complications and revisions of electrode placements. As upper cervical segments contribute substantially to the rotational movements of the neck, the epidural space contains intricate connections that merge and blend together [10]. Meningovertebral ligaments connecting the posterior dural sac to the ligamentum flavum or the lamina are strong and maintain the integrity of this region. Cadaver studies showed that the ligaments were high in collagenous tissues [10] and were thickest at the C1 and C2 vertebrae [11]. Our observation concurs with these findings. The issues related to epidural resistance at the C2 level could be largely resolved by implementing the C2-3 dural exposure procedure.

Our findings also showed that surgical exposure adjustments did not significantly prolong the operation time nor increase the amount of bleeding. Also, no surgical site infection was observed between the 2 treatment groups. Epidural resistance was encountered around C2, but this was physiologic rather than scarring. With the C2-3 dural exposure technique, blind exploration of the epidural can be avoided, thereby increasing safety. Despite extending the wound cephalically, there were no wound healing problems, and no significant difference was found in postoperative incision pain. Compared to techniques that require additional laminectomies in the presence of epidural resistance or adhesions [12], the proposed C2-3 dural exposure procedure was associated with less damage to bony structures and retained most muscle attachment points, especially at the C2 spinous process. This can help maintain neck stability and reduce the impact of surgery on muscle tone. Interestingly, considering that a significant proportion of patients with DOC often suffer from spasm problems, in this study, we observed a short-term relief of neck or upper limb spasms in both surgery groups. Only one patient who underwent conventional surgery developed a prolonged weakness in neck extension. The symptoms improved a lot during the subsequent six months of rehabilitation therapy.

As the patients spent more time in bed, stimulator displacement in the upright position was not of concern in this study. Hence, we preferred placing the pulse generator into the chest wall rather than the flank or upper buttock area to reduce the length of the extension wire and potential injuries during passive turning over movements.

The lateral decubitus position was adopted in this study. This position allows a one-time skin disinfection followed by electrode implantation in the neck and pulse generator implantation in the chest wall, which can ultimately reduce the operation time and risk of infections. Another reason for choosing the lateral decubitus was to facilitate the anesthesiologist to check the tracheotomy cannula, which is not readily visible in the prone position. The neck position and operator’s perspective could make the electrode deviate to the decubitus side. We, therefore, emphasize that the incision should be perpendicular to the dermatoglyph of the neck to avoid lateral flexion and torsion of the neck, and the bed should be adjusted before the operation to ensure that it is at the surgeon’s eye level for incision. Thus, most deviation from the midline could be corrected with the proposed C2-3 dural exposure.

Compared to the single-column paddle, a multi-column paddle increases the coverage area and the flexibility of program control. In our previous attempts, we noted that the Medtronic Specify 2 $\times$ 8 was more suitable because the wider 5-6-5 paddle would encounter more resistance when passing through the epidural space at C2 and was likely to cover the nerve roots. After electrode implantation, conventional anchoring was adopted at the end of the electrode. Some studies have proposed a non-anchored method [13]. In our cases, no postoperative displacement or electrode revision was needed, which could be associated with the lower frequency and amplitude of neck rotation in most patients with DOC.

In addition to direct observation, there are two methods for evaluating the location of the electrode array during general anesthesia. One is intraoperative imaging. Fluoroscopy is widely used to locate the spinal segment and plan the skin incision. It also allows for some minimal access to spinal technologies that can be used when the patient is conscious [12, 14]. However, this requires an experienced surgeon to accurately determine intraoperative midline implants because parallax and visual alignment errors during photography cannot truly reflect the spatial relationships between the electrodes, and the spinal canal of patients with brain damage frequently has cervical lateral flexion and torsion [15]. Therefore, intraoperative CT is the preferred imaging modality. It can also accurately measure the implantation location of the electrodes.

For patients with a deviation of the spinal cord from the midline of the canal, a more desirable method of verification would be neurophysiological monitoring. This includes electromyography and somatosensory-evoked potential collision studies (SSEP-CS), which have recently been used in patients with chronic pain [16, 17, 18]. Electromyography is more sensitive and helps guide device programming [16, 17, 18]. SSEP-CS is safe for the intraoperative monitoring of midline placement of C1-2 paddle leads [19]. The common unilateral brain insufficiency in the patients of this study cohort required caution if SSEP-CS was to be used as an indicator of laterality. Thus, electromyography was preferred for intraoperative monitoring and device programming. C2-3 dural exposure facilitated the adjustment of electrodes according to neurophysiological monitoring. Two weeks after surgery, EMG was used to assess the efficacy of midline coverage, as determined by CT. The electrophysiological results matched with the imaging results, showing no obvious electrode displacement during this time interval.

GOSE was used to assess the long-term recovery of consciousness in this study. Under this scoring system, a score greater than 3 refers to full recovery of consciousness, while a score greater than 4 refers to partial recovery of self-care. Although there was a significant difference in midline coverage between the 2 treatment groups, no significant difference in conscious outcome was observed. There are many prognostic factors for consciousness disorder, including etiology, preoperative consciousness level and others. Long-term prognosis is closely related to the nursing level after discharge. Larger sample sizes may yield more reliable results. Thus, there are ample reasons to necessitate improved midline coverage based on current outcomes. One limitation of this study is that electromyography was performed on only 45 patients at the device programming stage. Therefore, we could not demonstrate the advantages of the C2-3 exposure procedure in terms of electrophysiological monitoring. Another limitation is that the lack of a sufficient number of long-term CT reviews of the electrodes to understand the incidence of long-term electrode displacement, which warrants further research in the future.

This study showed that the proposed C2-4 spinal cord stimulator implantation in patients with DOC under general anesthesia could achieve high midline coverage by: (1) using the lateral decubitus position, (2) exposing the dura between C2-3 to ensure that the electrode array proceeds upwards along the midline, and (3) ensuring that the electrode tip is not too far above the C2 level. These enabled the proposed C2-3 exposure procedure to have lower rates of midline deviation. We hypothesize that this could contribute to easy device programming and reliable efficacy evaluation.

YD, JH and JZ contributed to the study conception and design. YD, XX, YY, RH, JH, JZ collected the data and performed the data analysis. All authors contributed to the interpretation of the data and the completion of figures and tables. All authors contributed to the drafting of the article and final approval of the submitted version.

The study protocol was approved by the Ethics Committee of the Seventh Medical Center of the PLA General Hospital (approval number: 2018-58). The data of all patients were anonymized, and no additional patient consent was required due to the retrospective nature of the study.

Not applicable.

This research received no external funding.

The authors declare no conflict of interest.