CHAPTER 24 Technique of Radiofrequency Denervation

INTRODUCTION

History

The use of electrical current to create lesions for the treatment of pain began in 1931, when Kirschner introduced it for the treatment of trigeminal neuralgia.¹ In 1965, Mullan used direct current to perform percutaneous lateral cordotomy for unilateral malignant pain.² Later that year, Rosomoff modified this technique and used radiofrequency (RF) current for this treatment.³ An RF current was found to produce more predictable, circumscribing lesions. Several years later, Sweet and Wepsic described their technique for the use of RF lesions of the Gasserian ganglion in the treatment of trigeminal neuralgia.⁴ In spinal pain, the use of RF current was reported in a series of cases by Shealy (1975).⁵ He described RF lesioning of the medial branch to treat lumbar facet joint pain. Subsequently, more options for the use of RF lesioning in various spinal pain syndromes were introduced into clinical practice.⁶^–⁸ The introduction of a small-diameter (22-gauge) temperature monitoring electrode system by Sluijter and Mehta in 1981⁹ increased the safety of the current method used.¹⁰ Since then, RF procedures have been widely adopted by many pain practitioners. In addition there have been many new technical developments, such as the improvement of the quality of C-arm image intensifiers. All these developments have contributed to a widespread use of RF lesioning.

Recently, a modification of the traditional RF method was introduced by Sluijter et al.¹¹ They used the same output setting of the lesion generator that was used for making heat lesions as in RF lesioning, but they interrupted this output to allow generated heat to be washed out by thermal conductivity and circulation. This technique is called ‘pulsed radiofrequency’ (PRF) and is claimed to be nondestructive. Until now, there are no reports of any sensory or motor loss following this procedure. However, it is too soon to categorically recommend this technique because the mechanism of action is still unknown. Until several double-blind, randomized, controlled trials are conducted this new approach should not be routinely employed.

Physics

Radiofrequency current

An RF current is applied by an RF lesion generator through an electrode, which is insulated except for the most distal part. This exposed region is the active portion of the electrode. The RF current flows from the electrode tip to the dispersive ground plate, which is placed on the arm or leg of the patient and leads the current back to the RF lesion generator. RF current flows through tissue and results in an electric field. This electric field places an electric force on the ions within tissue electrolytes, causing them to oscillate at a high rate (i.e. 300 000 times per second).¹² Tissue heating is created by frictional dissipation of the ionic current within the fluid medium, which heats the electrode.

Lesion size

The size of the lesion not only depends on the diameter of electrode and the length of the uninsulated electrode tip, but also depends on tip temperature. This was confirmed in animal studies by Cosman et al. in 1984.¹³ Bogduk et al. studied not only the size of the lesions, but also the shape of the lesion.¹⁴ They performed experimental lesions in egg white and fresh meat. They found that RF lesions do not extend distal to the tip of the electrode, but extend radially around the active electrode tip in a spheroidal shape. Based on these observations, Bogduk suggested that the best placement of the electrode tip was parallel to the target structure.

Other research methods involving computerized or mathematical modeling of similar experiments have been used. For example, Moringlane et al. studied experimental RF coagulation with a computer-based on-line monitoring of temperature and power.¹⁵ Vinas et al. also studied lesion size in vivo using fresh eggs and in vitro using the subcortical white matter of rabbits.¹⁶ Both investigators obtained results that were similar to those reported by Bogduk et al. Once equilibrium temperature is reached (after 20–40 seconds), the size of the lesion does not increase. However, some variables such as circulation effect and the tissue heat conductivity may also produce a variation in lesion size, but are unpredictable.¹³ Consequently, the lesion size may therefore be variable at different locations within the body.

Radiofrequency lesion generator

The modern RF lesion generating system has different functions, including a pulsed RF mode. There is continuous on-line impedance measurement to confirm continuity of the electrical circuit and to detect any short circuits. One of the key components of the system is nerve stimulator function. It is used to confirm the proper position of electrode tip (it indicates the electrode-to-nerve distance) and to permit minor adjustments. To ensure the proximity of the active tip to the sensory fibers, stimulation is performed at 50 Hz. The 2 Hz stimulation is carried out to detect extraspinal muscle contractions that occur when the needle is placed too close to a nerve root motor fiber. Voltage, current, and wattage during an RF procedure are also monitored. Finally, a generator monitors temperature using a thermocouple. This is an important lesion parameter, but the temperature is only measured at the tip of the electrode and not in the more peripheral zones of the lesion area. The usefulness of tip temperature to monitor the lesion size is minimized because the lesion size is dependent on local blood circulation. There is a rapid drop in temperature over the first few millimeters from the electrode tip.^13,¹⁷

Effect of radiofrequency current on nerve tissue

The effect of RF lesions on nerve tissue is controversial. Letcher and Goldring, in 1968, investigated the effect of RF current and heat on the nerve action potential of the saphenous nerve in cats.¹⁸ Delta and C-fibers were blocked before the alpha-beta group by both RF current and heat. Uematsu,¹⁹ and subsequently Smith et al.,²⁰ reported quite disparate results. Smith et al. created RF lesions of different temperatures by placing an electrode into the lumbar intervertebral foramina of dogs.²⁰ Indiscriminate damage of both small and large fibers by RF current was observed. Uematsu conducted a histological analysis of feline sciatic nerves that sustained RF thermocoagulation.¹⁹ These studies have been criticized because large, 14-gauge electrodes were used; in addition, they were placed close to the dorsal root ganglion during open surgery. However, these methods are not comparable to the technique which is commonly used in clinical practice today, since 22-gauge needles where not available at that time.²¹ In response to this critique, de Louw et al., in 2001, tried to investigate morphological effects of RF lesions as they develop in normal clinical situations.²¹ A percutaneous RF lesion adjacent to the dorsal root ganglion (DRG) of goats was created with a 22-gauge electrode at a tip temperature of 67°C. Immunocytochemical changes indicative of proliferation and regeneration was observed 2 weeks after the RF lesion was created. These alterations occurred outside the DRG, without interfering with the function of large myelinated fibers.

Effect of pulsed radiofrequency current on nerve tissue

Recently, the importance of heat as the definitive mechanism of action RF lesioning has been questioned.¹¹ In 1998, Sluijter et al. investigated isothermic RF lesioning also termed pulsed RF (PRF) treatment.¹¹ Ostensibly, PRF generates its therapeutic effects independent of thermal factors. The pulses are given at a rate of 2 Hz and last 20 msec. The rest period is 480 msec, and this allows the generated heat to be washed out by thermal conductivity and circulation.

The mechanism by which pulsed RF treatment effects symptom relief remains unknown. In 2002, Higuchi et al. undertook a study to elucidate this mechanism. They attempted to identify spinal cord neuron activation following the exposure of the dorsal ganglion of rats to PRF currents.²² A significant increase in the number of cFos-immunoreactive neurons in the superficial laminae of the dorsal horn was observed. This finding indicates that PRF current activates pain-processing neurons in the dorsal horn and that this effect is not mediated by tissue heating.

This technique is considered to be a safer method of treatment because until now the observations showed no signs of neurodestruction and thus neurological side effects.^11,²³ Cahana et al., in 2003, compared the acute effects of pulsed versus continuous RF energy on impulse propagation and synaptic transmission in hippocampal slice cultures and on cell survival in cortical cultures of rats.²⁴ In both systems, they found an induction of distance-dependent tissue destruction under the stimulating needle (namely, an inhibition of evoked synaptic activity), but this was more pronounced in the continuous RF group. Thus, they concluded that the acute effects of PRF are more reversible and less destructive in nature than the acute effects of classic continuous RF mode.

Nevertheless, the safety, efficacy, and the mechanism of action of pulsed RF current remains unresolved at the present time. Clinical trials still are needed to prove its potential efficacy.

RADIOFREQUENCY DENERVATION IN THE CERVICAL REGION

Cervical percutaneous radiofrequency denervation of the facet joints

Relevant anatomy

The cervical zygapophyseal joints are innervated in the same way as the lumbar zygapophyseal joints (Fig. 24.1). The cervical segmental nerves divide into the primary anterior ramus and the primary posterior ramus on exiting the neuroforamen. Immediately, the posterior branch divides into a lateral and a medial branch in the cervical intertransverse space. The medial branches of C4 to C8 run dorsally and medially through the ‘waist’ of their respective articular pillars. They innervate the facet joints at the level of the exiting segmental nerve and at the level below. Because of this multilevel innervation of the facet joints, usually the medial branch procedure is performed at several levels. Furthermore, the medial branches give deep branches to multifidus muscles and superficial branches to semispinal cervical muscles.

Fig. 24.1 Anatomy of cervical facet joints and their innervation. The medial branches of the posterior primary rami of the segmental nerves C3, C4, and C5 are indicated by red lines, the vertebral artery by the blue line

The anatomy of the innervation of the C2–3 facet joint is distinct from the lower cervical joints. It receives innervation by branches of the greater occipital nerve, which is the continuation of the posterior primary ramus of the C2 segmental nerve. The C3 posterior primary ramus of the segmental nerve gives a medial branch that innervates the C2–3 facet joint, but also gives a branch that forms the third occipital nerve that runs cranially to innervate the skin of the lower part of the occipital region. There is no facet joint between the C1 and the C2 vertebrae.

Procedure

Sedation is not employed because of the need for continuous communication between practitioner and patient. Several approaches to reach the medial branch of the dorsal ramus at the upper and middle cervical area can be used.

The most commonly used approach by these authors is the posterolateral approach.²⁵ For this technique, the patient is positioned supine on the operating table, which is tolerated well. The occiput is placed on a headrest and the cervical spine is slightly extended. The C-arm is positioned in a moderately oblique (±30°) location to secure a safe distance between the electrode tip and the exiting segmental nerve. In this position the gantry angle is parallel to the axis of the intervertebral foramen that is upward and slightly caudal. In this position the segmental nerves exit in a plane approximately perpendicular to the monitor screen. The degree of obliquity should be such that the projection of the pedicles is seen a little anterior to 50% of the vertebral body (Fig. 24.2). In the frontal plane, there should be a small angle of the C-arm with the transverse plane. This provides clear visibility of the intervertebral discs and the neuroforamina. In this projection, the medial branch runs over the base of the superior articular process, which is easily viewed. Entry points are marked on the skin, somewhat posterior and caudal to the target points as seen on the monitor, i.e. dorsal from the posterior border of the facet column and slightly caudal. Care should be taken to ensure that the entry point is not too anterior in the neck to avoid important vascular structures such as the carotid artery. The first needle (preferably the most cranial) is inserted in a horizontal plane and in a slightly cranial direction not deeper than 2 cm, so that the needle point is in line with the target point. Then the needle is carefully advanced anteriorly and cranially until bone contact is made with the facet column at the target point (see Fig. 24.2).

Fig. 24.2 Oblique fluoroscopic view of cervical facet joints with three needles aiming for the medial branches of the C3–4, C4–5, and C5–6 facet joints. The red line indicates the position of the pedicles projection a little anterior to 50% of the vertebral body. The green dots indicate the target points.

During the first step of the procedure, several things can go wrong. The needle tip may be projecting over or anterior to the target point without any bone contact. Should the needle then be introduced further and more anteriorly to an imaginary line connecting the posterior aspects of the neural foramina it could make contact with the segmental nerve or even with the vertebral artery. To avoid this, each advancement of the needle should be checked and, if necessary, should be redirected more posteriorly. As the needle is advanced it becomes increasingly difficult to reposition the needle tip to an accurate location. If the needle tip is projecting posterior to the bone of the facet column and is advanced further, there is the possibility that it will pass between the laminae and make contact with the spinal cord. This can occur if the entry point of the needle is too far posterior in the neck. To avoid this, an anteroposterior (AP) check should be made to confirm an ideal starting point.

The position of the C-arm in the AP direction should confirm the position of the needle tip adjacent to the concavity (‘waist’) of the articular pillars of the cervical spine at the corresponding level (Fig. 24.3). Once the first needle is in the proper position, the other needles are introduced in the same way as described above. The authors prefer to take advantage of the first needle serving as an indicator for the direction and depth for subsequent needles, which accounts for their recommendation of placing all needles before applying RF rather than lesioning a single medial branch and then placing the next needle. The technique is identical for the facet joints from C3–4 until C6–7. The direction of the needle at the C2 level is different. It should be oriented toward the small branches that innervate the C2–3 facet joint and not toward the medial branch, which is the greater occipital nerve. For the RF treatment of this facet joint, the needle should be placed at the arch of C2 at the level of the upper border of foramen C3.

Fig. 24.3 Antero-posterior fluoroscopic view of the needles aiming for the medial branches of the cervical facet joints C3–4, C4–5, and C5–6. The target points are indicated with green dots. Notice that the needle aiming for the C3–4 medial branch has to be redirected more cranially.

When optimal anatomical localization of the needles transpires, an electrical stimulation is performed to confirm correct needle position, i.e. parallel to the medial branch. This can be accomplished by identifying the stimulation thresholds. First, an electrical stimulation rate of 50 Hz should be given and should elicit a response (tingling sensation) in the neck at less than 0.5 volts. Next, a stimulation at 2 Hz is performed to confirm accurate needle position. Contractions of the paraspinal muscles will be noticed. Muscle contractions in the arm indicate needle placement too close to the exciting nerve root. In that case the needle should be repositioned more posteriorly. Once proper positioning of the needle is confirmed, the medial branch of the dorsal ramus is anesthetized with 1–2 mL local anesthetic solution (lidocaine 1–2%). An 80°C radiofrequency thermal lesion is made for 60 seconds at each level. To date, there have been no reported complications from cervical facet joint denervation when the procedure is methodically performed.²⁶

Some postoperative burning pain is described in >30% of patients.²⁷ It disappears spontaneously after 1–3 weeks. One should always bear in mind that there is risk of puncture of the vertebral artery when the needle is positioned anterior to the foramen in the posterolateral approach.

A second approach, which is more popular in the United States, is the posterior approach of the facet joint. This was first introduced by Lord et al. in 1995.²⁸ In this technique, the patient is positioned prone on the operating table, with the head flexed (about 5–10°) and with the face resting on a padded ring. This is a ‘tunnel vision’ technique, in which the target points are the posterior aspects of the waists of the articular pillars at the levels to be blocked and are the same as the entry points. The needle is introduced from posterior of the neck to make contact with each of the two nerves supplying the painful joints. When bone contact is made, the C-arm is turned to give a lateral view in which the needle tip should be seen in the posterior aspect of the waist of the articular pillar. A deviation toward the midline should be avoided because of risk of penetration into the epidural and subarachnoid space. Subsequently, the needle is incrementally advanced until the tip is projecting at the center of the pedicle. Then, a slightly lateral deviation should be made so that the needle tip is in the most lateral aspect of the articular pillar in the posteroanterior view. When this position is confirmed, stimulation is performed as described above to determine the 50 Hz and 2 Hz thresholds. If these are acceptable, local anesthetic is injected and an RF lesion (80°C for 60 sec) is made. At each location two lesions are made: the second one after a rotation of the RF needle over 90° from its first position.