Transoral approach using a tubular retractor system in the treatment of atlantoaxial Pott’s disease: a novel method of surgical decompression

Abstract

The use of tubular retractors in minimally invasive spine surgery has been described extensively in the literature. However, there are only select reports discussing their use specifically in a transoral approach for upper cervical pathologies such as an epidural abscess and a dystrophic os odontoideum.

We report the first use of a transtubular transoral approach in a rare case of an adult presenting with severe neck pain with progressing neurologic signs secondary to atlantoaxial Pott’s disease. Debridement and partial odontoidectomy performed using this approach allowed less trauma to surrounding soft tissues without sacrificing an excellent visualisation of the operative field. A halo vest was applied to provide temporary stability. The patient transitioned out of his halo vest after 6 months while completing medical therapy after a year without need for further intervention. The transtubular transoral approach is a suitable alternative to the conventional approach in patients with atlantoaxial Pott’s disease.

Background

The application of tubular dilators in spine surgery has been around for almost two decades.¹ However, its use specifically in a transoral transtubular approach in upper cervical pathologies has been limited to only two reports describing a similar technique in excision of a dystrophic os odontoideum and pyogenic abscess drainage.^{2 3} From our knowledge, this is the first reported case of atlantoaxial tuberculosis (TB), a disease occurring in less than 1% of all tuberculous spondylitis,⁴ treated by debridement and odontoidectomy using this approach. A tubular retractor system provides a much less traumatic approach to the anterior occipitocervical region as compared with the conventional transoral approach without sacrificing exposure. Given the rarity of the disease, the peculiarity of this patient’s presentation and the novel approach to his surgical treatment, we have documented these unique aspects in detail.

Case presentation

This is the case of a 20-ear old man with a 4-month history of progressive neck pain with no neurologic deficits, associated with weight loss, general body malaise and persistent cough. He eventually became bedridden due to the debilitating neck pain and was finally brought to the emergency department, where on physical examination he appeared cachectic but was afebrile with stable vital signs. The patient was stretcher borne and supported his chin with the palms of both hands all throughout his awake time. His neck was fixed and guarded in a slightly flexed position. There was tenderness on the posterior cervical area but no masses appreciated. A baseline neurologic examination showed no cranial nerve deficits or other sensorimotor deficits. Due to the inability of our patient to stand nor tolerate a sitting position, we were unable to have the patient perform a Romberg’s test to assess his balance. The posterior column was assessed by testing proprioception and his sense of vibration, both of which were intact. Pathologic reflexes were notably absent as well. The patient had decreased breath sounds on auscultation of the right upper lung field. The remainder of the physical examination was normal.

Investigations

Cervical MRI findings (figure 1) were consistent with osteomyelitis of C1 and the C2 with an extensive soft tissue abscess of the anterior and posterior aspects of C1 and C2. A cervical CT scan also revealed lytic areas over the dens with atlantoaxial subluxation. Chest imaging showed some minimal pleural effusion on the right haemithorax. Basic infection parameters that included a white cell count, erythrocyte sedimentation rate and C reactive protein were all elevated.

Figure 1.

T2-weighted MRI (sagittal cuts) showing bright marrow changes on most of the dens and C1 vertebra and surrounding abscess formation with cord compression.

An ultrasound-guided biopsy of the posterior soft tissue abscess appreciated on MRI (figure 2) was done yielding grossly thick purulent material with few Gram-negative bacilli and Gram-positive cocci found on gram stain but eventually no growth on culture media. While the patient’s blood culture results yielded no growth, Mycobacterium tuberculosis-specific PCR results from the biopsy specimen came back positive, hence the patient was treated as a case of atlantoaxial Pott’s disease with pleural effusion secondary to pulmonary TB with superimposed bacterial infection.

Figure 2.

T2-weighted MRI (axial cuts) reveal the extent of the soft tissue abscess within the anterior and posterior elements at the atlantoaxial level. Ultrasound-guided aspiration was performed over the posterior fluctuant area.

Treatment

After a definitive laboratory diagnosis, the patient was started on quadruple antituberculous therapy and empiric vancomycin and ceftriaxone to cover for a superimposed bacterial infection. Patient’s neck was temporarily immobilised on a Miami J collar. Over the course of admission, no clinical improvements were noted with the aforementioned strict antibiotic regimen. Infection parameters showed no definite downward trend, and on day 14, after initiation of medical therapy, the patient developed upper motor neuron signs (positive Tromner’s and Hoffman’s test). A repeat MRI was initially requested; however, due to operational delays in our hospital at the time, an MRI would take at least 3–5 days to be scheduled and performed. Given the urgency of his clinical deterioration, the lack of improvement with medical therapy, rather than delay intervention, an immediate decision was made to decompress the atlantoaxial spine. A transoral surgical drainage was planned to decrease the collected caseation and bacterial load of the atlantoaxial region.

The senior consultant surgeon performed an anterior decompression, debridement and partial odontoidectomy via a transtubular transoral approach. The operative technique is detailed as follows:

At the operating room, a video-assisted asleep fiberoptic nasotracheal intubation with the patient supine was done by the anesthesiologist, after which the patient’s head was secured with a Mayfield head holder. The patient was positioned supine with the feet towards the anesthesiology machine to allow ample area around the head with a large space for the footprints of the image intensifier, monitor and endoscopic tower (figure 3). Preoperatively, the following measurements were obtained from sagittal images of the patient’s CT scan: distance (mm) of incisor to tip of abscess and the distance (mm) of incisor to anterior border of C2. These were measured as 75 mm and 95 mm, respectively. These measurements ensured that our tubular retractors would reach the oropharynx.

Figure 3.

Patient and C-arm positioning.

An oral prep was done including the perioral area followed by sterile draping. The nasotracheal tube was deflected on the side with the least bulge from the infection underneath. Using image guidance, the tubular retractor was positioned at the level of C1 on lateral view (figure 4A, D). The uvula was merely pushed laterally as the larger tubes (up to 18 mm) were serially inserted down into the posterior pharyngeal wall.

Figure 4.

Endoscopic setup using the MED and tubular retractor system. Serial insertion of the METRx-MD tubular dilators down into the posterior pharynx is done easily (A). The working tube is now attached to the table-mounted adjustable arm (B). Endoscopic camera and light source are then attached using the MED system (C). Intra-operative imaging lateral views are shown side by side with each stage of the endoscopic setup (D–F). MED, microendoscopic discectomy.

Once the largest possible tube size was inserted, the microendoscopic discectomy (MED) system allowed for our endoscopic camera and light source to be securely attached to the tube. The system itself is secured unto the siderail of the table (figure 4B–C). At this stage, the endoscopic operative setup is completed and the surgical incision is ready to be made.

An 8 mm midline incision was created in the posterior oropharynx, exposing the anterior arch of the C1 vertebra (figure 5B, C). We then proceeded with the initial debridement of caseation material that started to spill out followed by the exposure of the anterior arch of C1 exposing a good 6 mm of its osseous surface. Using a suction retractor to protect the delicate membrane of the oropharynx, we created a channel resection of the C1 arch using a high-speed Burr machine with a six-petal rosen burr tip until we reached the tip of the dens. This was done with uninterrupted irrigation and suctioning to maintain good visualisation. Intraoperatively, we were not satisfied with the expected amount of drainage we were expecting. On careful dissection with a micro Freer Penfield, we found chunks of caseation materials plugging the viscous pus behind the lysed dens. On removal of the caseation plug and careful removal of the lytic dens with a diamond burr tip, we were able to successfully drain about 80cc of pus. The slope of the base of the dens remained intact which we believed to be good enough to maintain stability of the atlantoaxial complex. Copious irrigation and drainage of the nasopharynx were followed. The posterior pharyngeal wall was then closed with three interrupted absorbable sutures. A halo vest was immediately applied to provide temporary stability. The total length of the procedure was 75 min with a 50 mL blood loss incurred. Prior to extubation, a nasogastric tube was inserted to allow feeding while waiting for the surgical site to heal. Examination of the teeth and soft tissues of the oropharynx on removal of the tubular retractors revealed no collateral tissue damages.

Figure 5.

Debridement and drainage. Endoscopic view of the posterior oropharynx is seen (A). After incising midline, part of the anterior arch is exposed and removed (B) with the use of a high-speed burr (C). Removal of necrotic bone fragments belonging to the tip of the dens as well as caseated material (D) allowed decompression and drainage of copious amounts of pus (E). Final inspection was done after adequate debridement (F).

Outcome and follow-up

Postoperatively, there was a significant improvement of neck pain reported by the patient from a preoperative visual analog scale score of 10/10 to a postoperative score of 4/10. On day 2 after surgery, patient was able to situp with assistance. Enteral feeding was discontinued on day 5. After removing the nasogastric tube, the patient slowly advanced from liquids to soft than regular food. With supervised physical therapy, the patient was able to transition to wheelchair ambulation within a week, after which he was discharged. The patient and family were thoroughly advised regarding a possible second-staged procedure for posterior fusion to address his residual subluxation and basilar impression.

On regular follow-up at the Philippine General Hospital - Spine Service Clinic, he had eventually progressed to walking without support within a month postoperatively. The patient continued to improve without any complications and on his sixth month after surgery, the halo vest was removed and replaced with a Minerva brace for 3 more months. A postoperative MRI done after halo removal (figure 6) showed complete resolution of the abscess. He was able to complete 12 months of antituberculous therapy. Despite the persistent atlantoaxial subluxation on CT scan 1 year postoperatively (figure 7), the regular follow-up evaluation of our patient showed no cervical pain or neurologic deficits. Hoffman’s and Tromner’s signs were consistently negative. The patient decided not to pursue further surgical intervention since he is very functional with no pain or residual neurologic deficits. At 18-month follow-up, he had only a minimal loss of cervical flexion-extension and an intact 70-degree arc of cervical rotation likely explained by the formation of a C1-C2 fusion mass on CT scan (figure 7).

Figure 6.

T2-weighted MRI taken 6 months postoperatively shows the resolution of abscess with decompression of the cord on both sagittal (A) and axial cuts (B).

Figure 7.

CT scan images taken 1 year postoperatively. On axial cuts, the residual dens remain eccentric to the lateral masses of C1 indicative of subluxation (A). An extensive fusion mass spanning the occiput to the lateral masses of C1 and C2 is evident on coronal cuts (B).

Discussion

TB remains an endemic disease in developing countries; however, atlantoaxial TB is extremely rare accounting for just 0.3%–1% of all tuberculous spondylitis.^{4 5} Based on unpublished data from the Philippine General Hospital—Spine Clinic experience from 2007 to 2017, only 17 cases of craniovertebral infections, were documented, all of which were TB in origin (From Dr Samuel Grozman’s Craniovertebral Infections Lecture presented at the AO Spine Meeting on 12 November 2018).

Cases of TB spondylitis at this level are typically diagnosed late due to its nonspecific presentation. Within that time frame, bony and ligamentous involvement can lead to joint subluxation or dislocation, epidural abscess and later on neurologic compression.⁶ In this particular case, the patient presented with chronic but severe neck pain, one of the most common features in atlantoaxial TB.^{4 7} However, one would expect neurologic deficits to accompany 63%–72% of cases.^{7 8} The absence of neurologic deficits presents a risk to any contemplated surgical procedure especially in the context of an illness in which medical therapy is the mainstay. Despite being documented since 1970, the literature on atlantoaxial TB lacks a set of guidelines that can clearly delineate between surgical and conservative treatment.^5–8 Observed recovery of several patients on antituberculous therapy while waiting for surgical decompression, known as the middle path regimen, is advocated by many.⁹ The decision, thus, to operate on this patient was based on multiple factors including a lack of response to medical therapy, the degree of instability causing debilitating pain and the development of upper motor neuron signs. A background of superimposed bacterial infection and the patient’s limited financial capacity were important reasons why we did not consider a fusion procedure at that point in time. The immediate goal was to debride, decompress and temporarily stabilise the atlantoaxial spine. Although the transoral approach and its modern variations remain to be the gold standard for exposure of the anterior craniovertebral junction,¹⁰ a less invasive procedure that would still allow adequate debridement would be considered best for this case. Hence, we decided to apply a transtubular transoral approach as opposed to the conventional or standard transoral approach classically described by Hadley et al.¹¹ Ideally, this is best combined with neuromonitoring when available, however, this is not routine in our institution since it is an out-of-pocket expense for patients.

The MED system has been in use since 1997 allowing spine surgeons to perform minimally invasive spine surgery (MISS) primarily for the lumbar spine.¹² Coupled with the introduction of tubular retractor systems like the METRx-MD (Medtronic, Memphis, Tennessee USA), MISS became widely effective among different surgical procedures including lumbar and thoracic discectomies and decompression, transforaminal interbody fusions, anterior odontoid screw fixations and anterior cervical debridement.^12–15 However, there are only two published case reports that have described the use of tubular retractors in the transoral approach.^{2 3} Ariffin et al in 2015 was the first to document a transtubular transoral approach in the excision of a dystrophic os odontoideum with the aid of an operating microscope for illumination and visualisation.² The authors concluded that the dreaded complications (rhinolalia, nasal regurgitation and palatal wound dehiscence) associated with the conventional transoral approach could be avoided by the use of the tubular retractor system. Zaninovich et al used the transtubular approach in an odontoidectomy procedure for epidural abscess decompression.³ In their technique, they used Brainlab, an image navigation software, and matched anatomical surface markers to guide the position of their tubular retractors. Both cases had an unremarkable postoperative recovery and were able to avoid the soft tissue trauma associated with the use of oral retractors.

Our technique differs slightly from Ariffin et al and Zaninovich et al in that neither a microscope nor Brainlab software was needed. The MED system was used to provide quality visualisation and illumination through an attached endoscopic camera and light source. Combined with the use of a readily available C-arm, we were easily able to position the tubular retractor at the desired level.

Table 1 summarises the differences between the two approaches based on the literature. In the transtubular approach, trauma to the perioral, oropharyngeal or nasopharyngeal area is avoided, whereas in the conventional approach, much care must be taken to avoid complications associated with oral retractors such as lip laceration, traumatic contusion or other mucosal injury.¹⁶ In both published reports by Ariffin et al and Zaninovich et al, they describe the setup using tubular retractors as quick and simple since the surgeon merely pushes the appropriate tubular retractor without need for additional soft tissue manoeuvring, splitting or padding. There have been documented cases of oropharyngeal morbidity following the conventional approach, such as dysphagia, nasal escape and regurgitation, and impaired wound healing especially when there is soft palate splitting.¹⁷ The conventional approach has also been associated with a potential risk of oropharyngeal bacterial contamination,¹⁸ however, the risk may be minimised by isolating the field of access using the deep and narrow transtubular retractors. A transtubular transoral approach would have a steeper learning curve compared with the conventional approach.¹⁹ Comparing the working fields between the two approaches also demonstrates the advantage of a transtubular approach wherein you have a clear sharp image that has excellent illumination due to the light of the endoscope situated very close to the surgical field, as well as an unobstructed, and surprisingly spacious operating field. In a conventional approach, one would normally require surgical loupes within a working area whose spaciousness depends on the intensity of the retraction around the perioral area. The illumination by an overhead lamp, or a head lamp, even those coming from operating microscopes cast shadows over the operative field in the conventional approach, because it is far from the surgical site, and this can often times be inconvenient to the surgeon. An operating microscope when available can be used to greatly augment the working room, visualisation, depth of field and safety for both conventional and transtubular approaches.¹⁹

Table 1.

Comparison between conventional transoral and transtubular transoral approaches based on literature2 3 15–19

	Conventional	Transtubular
Soft tissue trauma to: Lips	+/−	–
Teeth	+/−	–
Tongue	+/−	–
Palate	+/−	–
Uvula	+/−	–
Tonsils	+/−	–
Nasopharynx	+/−	–
Relative sterility	Potential risk of oropharyngeal bacterial contamination	Minimised risk
Ease of setup	Difficult	Simple and quicker
Ease of procedure	Easier	Steeper learning curve

We have described in detail the use of a transtubular transoral approach in the decompression and debridement of a rare case of atlantoaxial Pott’s disease presenting with progressing neurologic signs and demonstrated a successful outcome for this patient. The transtubular transoral approach is a suitable alternative to the conventional transoral approach for pathologies at the level of the atlantoaxial spine in the realm of minimally invasive surgery.

Learning points.

• Although uncommon, tuberculosis (TB) is still an important differential in the diagnosis of atlantoaxial pathologies in areas where it remains endemic.

• In the rare absence of neurological deficits on initial presentation, the decision to perform surgery for atlantoaxial TB can become a difficult balancing act between risk and benefit. Surgeons must take into consideration an evolving clinical picture, which will include response to medical therapy, degree of instability and even subtle neurological changes.

• Choosing which surgical procedures will be of most benefit is tailored to the patient’s problem list and surgeon’s experience rather than to a standard set of guidelines that are yet to exist in the management of atlantoaxial TB.

• The transoral transtubular technique is a very useful addition to the spine surgeon’s armamentarium of treating atlantoaxial pathologies. It is a novel minimally invasive anterior approach to the atlantoaxial complex. The learning curve often seen with minimally invasive spine surgery procedures is immediately offset by its advantages over the conventional transoral approaches.