The response to radiofrequency neurotomy of medial branches including a bipolar system for thoracic facet joints

1 Introduction

There are relatively few studies on thoracic facet joint denervation in the management of chronic thoracic spine pain as the thoracic spine in general is the least common site for back pain complaints, with a reported prevalence being only 13% in contrast to 43% for low back pain and 32% for neck pain [1]. However, when back pain is narrowed into regional pain disorders, as high as 48% of thoracic spine pain (upper and middle back) was due to thoracic facet joint pain [2].

The diagnostic criteria for thoracic facet joint pain remain inconclusive. The usual presentations include the following symptoms and signs: (1) history of thoracic paravertebral pain that worsens with prolonged standing, or rotation of the thoracic spinal column [3]; (2) back pain with a typical radiation pattern of thoracic facet joints [4, 5]; (3) reproduction of paravertebral tenderness upon pressure on the respective facet joint; (4) no significant neurological deficit; (5) normal or minimal changes in radiological images which exclude other sources of pain; (6) elimination of pain with intra-articular or medial branch injection of local anesthetic [3].

Radiofrequency (RF) neurotomy of the medial branch of dorsal rami has been used for the treatment of spinal pain of facet joint origin, at all spinal levels since 1970s [6, 7]. Despite being the treatment of choice for facet joint pain, there is still a paucity of literature on RF neurotomy of thoracic medial branch [8]. This could be due to inconsistency in the path of the thoracic medial branch across the transverse process before innervating the facet joint [4, 5], and a lower incidence of thoracic back pain compared to the cervical and lumbar regions [1].

Bipolar RF is an improvement of conventional RF neurotomy, as it creates a larger but more precise lesion during denervation [9, 10]. Bipolar RF neurotomy has been used in different pain disorders with variable efficacy, e.g. sacroiliac joint pain [11, 12], lumbar discogenic pain [13, 14] and ischemic pain syndrome [15, 16].

As there is no study on bipolar RF neurotomy of thoracic medial branch, the aims of this retrospective study are to describe the technique of bipolar RF neurotomy of thoracic medial branch and to evaluate the outcome of this technique in the treatment of chronic thoracic facet joint pain.

2 Methods

This is a retrospective, observational study of chronic pain patients undergoing bipolar RF neurotomy of thoracic medial branch at Orbis Medical Center (Zuyderland), Hospital Sittard Geleen (Heerlen), The Netherlands between January 2012 and December 2015. The study was approved by institutional research review board. Patients with treatment of facet joint pain involving two regions, e.g. cervico-thoracic and thoraco-lumbar regions were also included in this study. After identification of eligible cases using computerized clinic registry, medical records, procedure charts and follow-up questionnaires of all patients were reviewed. Confidential and de-identified data collected included patient demographics such as age, gender, body mass index and occupation; and clinical data such as diagnosis, duration of pain, previous treatment, adverse effects and complications from the procedure.

During patient evaluation, pain pattern and distribution suggestive of thoracic facet joint pain and local tenderness over facet joint area (spinal paramedian area) were identified and other pathologies of the thoracic region, such as vertebral fracture, stenosis, malignancy, infection and radiculitis were excluded, both by clinical or radiological examinations. Once thoracic facet joint pain was suspected, patients received a diagnostic thoracic medial branch block using 0.5 mL of 0.5% levobupivacaine to confirm the diagnosis, according to the standard of the SIS Guidelines [17]. The patients with more than 50% pain reduction lasting longer than 3 h after the diagnostic block were offered bipolar RF neurotomy of thoracic medial branch subsequently. Informed consent of the treatment was obtained after discussing the benefits and the possible adverse effects of bipolar RF neurotomy of thoracic medial branch including infection, allergic reaction, transient radicular pain and pneumothorax.

For bipolar RF neurotomy of thoracic medial branch, two 22 gauge 10 cm, sharp and straight RF insulated cannulas with an active tip of 10 mm, two 10 cm CSK-TC10 electrodes and a Cosman G4 RF generator (Cosman Medical, Inc., Burlington, MA, USA) were used. The procedure was performed with the patient in prone position on a radiolucent table. The C-arm of the fluoroscope was positioned in the axial plane and the proper level was identified by a radiopaque marker. The C-arm was turned cephalocaudal in order to square the end plates of the vertebrae. After identifying the transverse process, the skin was anesthetized with 0.5–1.0 mL of 1% lidocaine superficially and a RF cannula was inserted at the superior and lateral most portion of the transverse process. It was advanced in a tunnel view method until bone contact with the transverse process was made. A second similar RF cannula was inserted 5–10 mm medial and inferior to the first RF cannula in a similar manner as the first cannula until bone contact was made at the angle of the transverse and superior articular processes. A lateral fluoroscopic view was done to ensure that the tips of both cannulas were posterior to the line that connected the posterior aspects of the intervertebral foramina. Satisfactory final placement for both cannulas was confirmed with anteroposterior and lateral fluoroscopic views (Fig. 1). Neurostimulation was carried out first with 50 Hz which reproduced the pain, then followed with 2 Hz. The 2 Hz motor stimulation elicited contraction of the multifidus muscles at intensities below 0.5 V [3]. Subsequently, 0.5 mL levobupivacaine 0.5% was injected through both RF cannulas. Two RF electrodes were then inserted through the two cannulas, and the electrode tip temperatures were raised to 80°C for 90 s. One bipolar RF lesion was made for the medial branch of each thoracic ramus. As for treatment of facet joint pain involving two regions, e.g. cervico-thoracic and thoraco-lumbar regions, RF neurotomy of cervical and lumbar medial branches were carried out based on the conventional monopolar technique previously described [18, 19]. Patients were discharged home with oral celecoxib 400 mg daily, pantoprazole 40 mg daily and paracetamol 1 g 3 times daily for 5 days to relieve post-procedural needle soreness.

Fig. 1:

Anteroposterior and lateral fluoroscopic images of bipolar RF cannulas at target points (transverse process of thoracic vertebra). Note the cannulas are parallel and 10 mm apart to each other within the area of the transverse process in anteroposterior view. In lateral view, the tips of both the cannulas are located posterior to the line that connects the posterior aspects of the intervertebral foramina.

The primary outcome measurements were analgesic efficacy of bipolar RF neurotomy using Numerical Rating Scale (NRS) which measures pain from no pain to worst possible pain on a 0–10 scale and Pain Disability Index (PDI) with scores ranging from 0 to 70, with the higher index reflecting the person’s greater disability from pain [20]. Both NRS and PDI were done during clinic visit prior to treatment (baseline) and at 3rd month after treatment. For the study purpose, telephone calls were made during data collection to enquire NRS at 12th month after treatment. Reduction in medication use after treatment, as well as the number of complications such as pneumothorax, flare-up pain, intercostal neuritis, hematoma formation, infection and spinal cord trauma were recorded.

Data were analyzed using SPSS software version 12.0. Descriptive statistics of the data were presented as a mean and standard deviation. Frequency and descriptive statistics were calculated to check all relevant characteristics of the data. Differences in NRS and PDI during subsequent follow-up visits with baseline measurements were measured with Wilcoxin Signed Ranks test.

3 Results

Between January 2012 and December 2015, 116 patients were treated with bipolar RF neurotomy of thoracic medial branch. As a tertiary referral pain center, with many of our patients coming from great distances, we had a number of patients lost to follow-up. Consequently, the findings presented in this study were derived from complete data of 71 patients (Fig. 2). Those who lost to follow ups or with missing data were excluded from the study.

Fig. 2:

Flow diagram of the study.

The patient population consisted of 45 male and 26 female patients with a mean age of 57.9±11.2 years. The mean BMI of patients was 27.14±2.72 kg/m². The mean duration of pain was 23±10.5 months. According to our patient data, 56 (79%) procedures were carried out on the cervicothoracic region, three (4%) procedures on the thoracolumbar region and 16 (17%) on the thoracic region alone. Thirty-eight patients (54%) had the procedure performed on the right side, 29 patients (41%) on the left side, and four patients (6%) bilaterally. During the period of observation, none of the patients required repeat bipolar RF neurotomy. Demographic data and clinical presentation of patients are summarized in Table 1.

Eighty-two percent of patients (58 out of 71) reported more than 50% decrease in NRS both at 3 and 12 months after bipolar RF neurotomy of thoracic medial branch. Pain relief was statistically significant when comparing baseline NRS of 7.75±1.25 with NRS at 3 months (2.86±1.53) and 12 months (2.82±1.29) post procedure, p<0.001 (Table 2). A similar finding was shown in the difference between baseline PDI of 40.9±12.2 with PDI at 3 months post procedure (24.2±9.8), p<0.001 (Table 2). Serious adverse effects of the procedure such as pneumothorax, infection, hematoma formation, nerve root trauma or spinal cord injury were not reported in this study.

4 Discussion

In recent years, interventional pain management has seen many new areas of advancement, with the aim of increasing the efficacy in alleviating symptoms, at the same time improving the safety of patients during procedure. The current evidence for conventional monopolar RF neurotomy of the cervical and lumbar medial branches in the treatment of respective facet joint pain is relatively strong. However, such evidence has not been demonstrated in the treatment of thoracic facet joint pain [3, 21, 22]. Current available treatment of thoracic facet joint pain includes thoracic facet joint injection, medial branch of dorsal rami block and RF neurotomy [8]. Studies by Manchikanti et al. showed that thoracic medial branch block with local anesthetics with or without steroid significantly reduced pain and improved functional status but the effect did not last long as most patients required repeated medial branch block with an average of 3.5 blocks per year [23, 24]. Stolker et al. reported a case series of 40 patients who underwent conventional RF neurotomy of medial branch of thoracic dorsal rami. In this study, 80% of the participants showed more than 50% pain relief at 12 month follow-up [25]. In another series which reported 17 monopolar RF neurotomy on thoracic medial branch, only 40% of patients had more than 50% pain relief [26]. Subset data of a more recent study by Speldewinde showed that 68% of 28 monopolar RF neurotomies of thoracic medial branch had a 50% pain reduction lasting longer than 2 months [27]. In contrast, the current study with bipolar RF neurotomy showed the highest success rate with 82% of 71 patients having more than 50% pain reduction even at 12 months after treatment. Cooled RF, another variant of conventional RF was recently shown to be able to produce neurotomy lesions larger than conventional monopolar RF on thoracic medial branches in human cadavers [28]. However, human studies with long term clinical outcome and safety profile utilizing this technique are still limited. One recent reported case of third degree skin burn in the treatment of thoracic facet joint pain with cooled RF raised the question of the safety profile of this technique [29].

The inconsistent association between the medial branch and a bony landmark (transverse process) makes the thoracic medial branch a challenging target for percutaneous interventions [4, 5]. It has also been demonstrated that the thoracic medial branch at some vertebral levels does not have bone contact with the transverse process [30]. Furthermore, the morphology of the thoracic transverse process changes quite significantly from T1 till T12 level. Based on the most recent human cadaveric studies, the width of unilateral transverse process increases from T1 till T5, subsequently decreases to T12 [31, 32].

Accurate placement of the electrode adjacent to the targeted neural structure is the most important factor in ensuring the successful outcome of RF neurotomy. This could be difficult to achieve in the case of thoracic facet joint denervation due to inconsistency of the neural pathway of medial branch of dorsal rami and changes in transverse process morphology of the thoracic spine. These shortcomings could be overcome by increasing the area of RF neurotomy lesion with bipolar RF, as the coverage are of two bipolar RF needles is larger compared to the single needle conventional monopolar RF technique [33].

As the transverse processes of thoracic vertebrae are relatively flat and superficial [29, 30], the insertion of the first RF needle at the superolateral aspect of the transverse process in tunnel view under antero-posterior fluoroscopic guidance is relatively straight forward and safe but similar to other interventional procedures done in the thoracic region, bipolar RF neurotomy of thoracic medial branches carries the risk of pleural puncture [3]. To prevent this complication, fluoroscopic visualization of the transverse process before needle insertion is crucial. The first needle should be inserted within 1–2 mm of the superolateral border of the transverse process in a tunnel vision. It is not necessary to place a needle at the superior and lateral most border of the transverse process, as the needle may slip pass the transverse process and puncture the pleura inadvertently. The second needle should be inserted strictly parallel to the first and the depth of insertion for the second needle should be the same as the first, in order to prevent diversion of the second needle. Anatomical studies of the average size of the transverse processes across the whole spine indicated that the transverse processes in the thoracic vertebrae can safely accommodate two RF cannulas 5–10 mm apart depending on the level of thoracic spine, when carrying out the bipolar RF neurotomy [31, 32].

Bipolar RF neurotomy is essentially an extension of conventional monopolar RF neurotomy. Although monopolar and bipolar RF neurotomies differ in their mechanisms of grounding (grounding pad vs. passive electrode, respectively), the complications and adverse effects are similar. Conventional monopolar RF neurotomy has been utilized in pain management for the past five decades and is proven to have a good safety profile [29, 34, 35, 36]. Bipolar RF neurotomy is utilized in the treatment of other pain conditions, especially those with inconsistent or complex neural anatomy, such as in sacroiliac joint denervation and lumbar sympathectomy [11, 12, 15, 16]. Based on the study by Cosman and Gonzalez bipolar RF with two RF needles with 10 mm active tips 10 mm apart will produce an elongated lesion, measuring approximately 15 mm by 8 mm, a lesion area which almost covers the whole surface of a thoracic transverse process [9]. With such a large area of electrical field coverage over the transverse process, it is not surprising to observe sensory stimulation of 0.4 V reproducing concomitant pain and motor stimulation of 0.9 V eliciting contraction of multifidus muscle during the first attempt of needle placement in most patients without further adjustment of needles positions. However, this finding should be investigated further to compare the efficiency of bipolar versus monopolar RF neurotomy in terms of the need for electrostimulation. The amount of radiation exposure prior to achieving satisfactory needle placement and the duration of the whole procedure should be the subject of further investigation. Pertaining to bipolar RF technique, Kim reported a series of nine patients who underwent bipolar RF thermocoagulation of thoracic facet joint by inserting two RF cannulas 5 mm apart directly into the inferior aspect of the facet joint. Assessment was carried out only at 1 month post-procedure and six patients had significant pain reduction [37].

Most of the bipolar RF neurotomies in this study took place at the cervico-thoracic junction. This is an area of transition from a mobile cervical to a rigid thoracic spine, and from cervical lordosis to thoracic kyphosis, and is an injury-prone area of the spine [38, 39]. The incidence of thoracic facet joint pain may be under reported, as patients who present with lower neck or upper back pain might be only receiving cervical facet joint treatment with thoracic facet joint pain being neglected, due to lower success rate with conventional RF neurotomy of thoracic medial branches.

In the present study, we hypothesized that post-procedural pain after bipolar RF neurotomy would be much greater than that of conventional monopolar RF neurotomy, as the larger lesion size would have caused a larger area of acute soft tissue injury. It is our routine practice to prescribe all patients oral analgesics, e.g. celecoxib and paracetamol, if not otherwise contraindicated in order to decrease post-procedural needle soreness and discomfort. We observed that all the participants in this study tolerated the procedure well and none of them came back complaining of localized flare up pain. However, the finding of post-procedural pain tolerance should be validated comparing bipolar with monopolar RF neurotomy.

The limitation of the current study is that it is retrospective in design, and therefore it has all the flaws and problems associated with such studies. Only data of 71 out of 116 patients were complete and available for analysis, which includes a risk of selection bias. A very high portion of patients (83%) received a combined treatment including monopolar neurotomy either at cervical (n = 56) or lumbar (n = 3) levels would have affected the outcome of the treatment in current study. Selection of patients with only thoracic facet pain, rather than patients with pain from two regions, such as cervicothoracic and thoracolumbar facet joint, might have been a more accurate reflection of the efficacy of bipolar RF neurotomy in the treatment of thoracic facet pain. This is because the monopolar RF neurotomy of both cervical and lumbar medial branch is well studied and the successful outcome is very high. Hence, combined treatment for facet pain of two regions as in most of the patients in this study, might have benefited from a more established technique, such as monopolar RF neurotomy of cervical and lumbar medial branch. We included all patients with facet pain of two regions as this is a common presentation among chronic pain patients in clinical practice. Other relevant information such as changes in post-procedural oral analgesic intake and a more detailed assessment on the long-term outcome and functional status of patients should be included in future studies. Even though this is a retrospective study, the importance of reporting such series in the pain medicine literature cannot be overlooked, as it is the only one reporting bipolar RF denervation of thoracic facet joint.

In conclusion, this case series suggests that bipolar RF neurotomy of thoracic medial branch is associated with a significant reduction in thoracic facet joint pain probably due to its ability to create a larger lesion to overcome the inconsistency of the neural path of thoracic medial branch. The promising findings from this case series merit further assessment with prospective, randomized controlled trial which will produce a more reliable and accurate finding for its clinical applications.