Development

Medial branch blocks are local anesthetic blocks of the articular nerves (the medial branches) that transmit sensory information including nociceptive signals from ZJs (as described earlier under “Mechanisms”). Historically, the first method used to test a ZJ as a possible source of pain was injection of local anesthetic into the cavity of a lumbar ZJ⁸⁴; such tests are called intraarticular blocks (IABs). Comparable techniques were developed for cervical¹³⁵ and thoracic¹³⁶ ZJs as well. Later, less invasive methods were developed to achieve blockade of a specific joint by blocking the articular nerve(s) outside the joint^137,138; such tests are designated by the particular nerves involved as medial branch blocks (MBBs) and third occipital nerve blocks (TONBs). Because articular nerve blocks are less invasive and are performed more often to investigate the common sources of chronic spinal pain, they will be described first.

Cervical Medial Branch Blocks C3 to C6—Procedure

A cervical medial branch block procedure involves a lot more than the needling and injection commonly associated with the name. This is also true for other such procedures. Each diagnostic procedure involves three phases, a preoperative phase, an operative needling phase, and a postoperative observational phase. The full test process involves a fourth phase, the interpretation of individual block results and the integration of the information they provide in the overall process of diagnosis. If a medial branch block is thought of mainly as the needling part, the other phases may be given less emphasis than they deserve and the whole process may be compromised.

A cervical medial branch block test is intended to determine whether a particular cervical ZJ is a source of pain. The test is designed to be specific, so initially, only one joint should be tested at a time. If the results of specific tests suggest that two or more ZJs may be involved in generating the patient’s pain, on a subsequent occasion the combination of joints may be tested together to see if they are all contributing to the pain in that region.

Each cervical ZJ from C3-4 down to C6-7 is supplied by two medial branches, the medial branches of the cervical dorsal rami above and below the joint. Thus, the C5-6 ZJ is supplied by the C5 medial branch and the C6 medial branch. The procedure to test the C5-6 ZJ as a pain source involves blocking each of those two nerves. The courses of the medial branches have been plotted by cadaveric dissection studies and shown to lie more or less horizontally across the middle parts of the articular pillars that join the superior and inferior articular processes of each vertebra.¹³⁹ The courses of these medial branches are illustrated in Figure 36-12.

Figure 36-12 Plots of the courses of cervical medial branches across the articular pillars of the vertebrae.

(From Lord SM, Barnsley L, Bogduk N: Neurosurgery Quarterly 1998;8:288-308, with permission.)

These nerves are the targets at which cervical medial branch blocks (MBBs) are aimed. The nerves themselves are not seen on fluoroscopy, but by appreciating the ranges of their courses it can be understood how particular medial branches can be anesthetized. They are accessible for injection around them as they pass over the “waists” of the articular pillars. Needles can be placed onto them at that site readily and safely from lateral approaches, as depicted in Figure 36-13.

Figure 36-13 Target points for cervical medial branch blocks indicated by needles introduced via lateral approaches.

(From Bogduk N: Back pain: Zygapophyseal blocks and epidural steroids. In: Cousins MJ, Bridenbaugh PO [eds]: Neural Blockade in Clinical Anesthesia and Management of Pain, 2nd ed. Philadelphia, JB Lippincott, 1988, 935-954, with permission.)

The target point for each MBB injection is the centroid or geometric center of the relevant articular pillar (Fig. 36-14). Medial branch blockade is achieved by placing the tip of a spinal needle at that point and injecting a small volume of local anesthetic, just sufficient to reach the highest and lowest of the known courses of the nerve in question. The volume of the injectate must be large enough to achieve that coverage reliably but not so large as to spread to and block other nerves because that would compromise the specificity of the test. Local anesthetic preparations suitable for the purpose are lidocaine 2% or bupivacaine 0.5% and the volume of injectate sufficient to achieve the desired spread has been found to be no more than 0.5 mL.¹³⁷

Figure 36-14 The target point for a cervical medial branch block at the centroid of the articular pillar of the vertebra where the cervical medial branch runs across it.

(From Barnsley L, Bogduk N: Medial branch blocks are specific for the diagnosis of cervical zygapophysial joint pain. Regional Anesthesia 1993;18:343-350, with permission.)

The placing of a needle tip at a specific point on the spine requires particular facilities and equipment. MBBs and other such procedures should be performed in a procedure suite or operating room where infection control measures are undertaken regularly. The patient’s safety is a prime consideration and it is crucial for safety that the operator knows precisely where the needle is at all times. The procedure must be done with radiologic guidance, preferably that provided by a fluoroscope with a high-resolution image intensifier and a C-arm that allows the X-ray beam to be directed at different angles to provide anteroposterior (AP), oblique, and lateral views without moving the patient. A radiolucent x-ray table is also required, preferably one with a mobile plinth that can be moved to modify the fluoroscopic views. Also needed in the procedure suite are spinal needles, extension tubes, and syringes for the block- injections, local anesthetic agents, sterile layouts on which the equipment for the procedure can be placed, skin preparation trays with swabs, forceps and suitable antiseptics, sterile drapes, sterile towels and sterile gloves for the operator, and a scrub sink adjacent. The suite should also be equipped for resuscitation and supportive care in the rare event of a serious complication arising.

The test facility also requires a room where the patient can be prepared before the procedure and observed after the test has been performed. Equipment required in that part of the facility includes a device for measuring pain and charts on which pain scores can be recorded.

The procedure begins with the preoperative phase. When a patient arrives at the test facility, he or she is admitted by a nurse who is trained in pain assessment. The nurse answers any questions the patient may have, checks for any contraindications such as allergy, infection, anticoagulant therapy, or pregnancy, and makes sure the patient has given informed consent (as described in the earlier section “Indications for Interventions”). After that, the nurse asks the patient to describe the pain for which the test is to be done; that pain is designated the “index pain.” For example, if a patient has headache and lower neck pain, the lower neck pain may be identified as the index pain for that day’s test. The nurse asks the patient about the intensity and distribution of the index pain, and those attributes are recorded on appropriate charts. The intensity is assessed by the patient as a score out of 10 or 100, using a printed visual analog scale (VAS) or some similar device such as a pain ruler, and recorded by the nurse on a pain chart. The distribution is recorded by the patient marking it on a body chart. The selection of the joint to be tested is checked by comparing the pain pattern drawn by the patient with the pain maps shown in Figure 36-7 to 36-11. The patient then changes into a hospital gown and at the appropriate time is taken to the procedure suite, preferably for the patient’s safety and convenience in a wheelchair.

The operative phase begins when the patient reaches the procedure room. The patient is helped onto the x-ray table and positioned lying comfortably on their nonpainful side so that the painful side is uppermost (Fig.36-15).

Figure 36-15 The position in which a patient should lie on an x-ray table for a right cervical medial branch block.

Their gown is tucked away from the neck and shoulder on that side. The skin of the area is sponged with a suitable antiseptic such as povidone-iodine or chlorhexidine and alcohol, allowed to dry and then draped with a sterile fenestrated drape. The operator, who should be wearing operating room scrubs, a surgical cap, operating mask and protective leads, should scrub their hands and don sterile gloves. The needles and other equipment for the procedure should be laid out on a sterile set-up (Fig. 36-16).

Figure 36-16 A sterile set-up with skin preparation tray and equipment required for a cervical medial branch block.

(Image courtesy of Pendlebury Clinic, Newcastle, Australia.)

Next the operator and the radiographer confer to ascertain the first target nerve. For example, if the patient is to have C5-6 MBBs, the first target nerve will be the C5 medial branch and its target point is at the centroid of the C5 articular pillar. When the first target is agreed, the radiographer obtains a clear lateral view of the target region and cones the image to it. Coning is important to minimize the exposure to the patient, the operator, and the radiographer to ionizing radiation. In the directly lateral view, the left and right articular pillars of the target vertebra will overlie each other and the target point will have a wide margin of bone around it, which enhances the safety of the procedure. Having obtained this view, the operator and the radiographer stand or sit so they can see the monitors clearly and carry out the procedure comfortably as in Figure 36-17.

Figure 36-17 A fluoroscopy suite in which a patient is having a cervical medial branch block. Note the C-arm fluoroscope with image intensifier and high-resolution monitors.

(Image courtesy of Pendlebury Clinic, Newcastle, Australia.)

The MBB needling procedure should be performed in accordance with the protocol recommended by the International Spine Intervention Society (ISIS) and described fully in the relevant practice guidelines.^140,141 It is important for the operator to talk to the patient to explain what is happening at each stage of the procedure, including what might be felt. The operator selects a suitable needle (usually a 25- or 26-gauge spinal needle at least 3.5 inches or 88 mm in length) and with the plastic needle guard still on, places it on the skin of the patient’s neck to determine, by intermittent fluoroscopic images, the point on the patient’s skin directly over the target point; that will be the needle insertion site. Then the operator takes the guard off the needle, places its tip carefully on the patient’s skin at the insertion site and checks its position there with another intermittent image (Fig. 36-18).

Figure 36-18 Lateral fluoroscopic view of the right C5-6 zygapophysial joint with the needle tip on skin over the centroid of the C5 articular pillar, the target point for a C5 medial branch block.

The operator stands the needle up in alignment with the fluoroscopic beam, warning the patient it will be felt, and inserts the needle through the skin as gently as possible. When the needle has purchase (i.e., when it has been inserted deeply enough to stand up without being held) its position is checked on another lateral view. Then the operator guides the needle through the patient’s neck muscles toward the target point, steering it by moving its hub, shaft and/or bevel in ways that are learned in practical training. Before each adjustment of the needle its position is checked in an intermittent fluoroscopic view; continuous screening is avoided to keep the x-ray exposure to a minimum. When the needle tip reaches bone at the target point, its position there is checked and recorded on a fluoroscopic image, as in Figure 36-19.

Figure 36-19 Lateral fluoroscopic view of the right C5-6 zygapophysial joint with the needle tip on bone at the target point for a C5 medial branch block.

When the needle is in position with its tip at the target point, a syringe loaded with local anesthetic solution is connected to it by an extension tube and, after aspirating to check for blood, 0.5 mL of the local anesthetic is injected.

The patient is then told the first of two blocks has been done and the same procedure will be followed to block the second nerve. The radiographer adjusts the fluoroscope to obtain a lateral view of the second target. The insertion point is determined and another needle is inserted in a similar manner, until its tip is resting on bone at the second target point. The position of the second needle at its target is checked and recorded on a fluoroscopic image, as in Figure 36-20, and 0.5 mL of local anesthetic is injected over the second medial branch. The patient is told both blocks are in place and the needles are about to be withdrawn, then that is done as smoothly as possible to complete the operative part of the MBB procedure.

Figure 36-20 Lateral fluoroscopic view of the right C5-6 zygapophysial joint with the two needle tips on bone at the target points for C5 and C6 medial branch blocks.

After the needles are out, the drape is removed and the antiseptic is sponged gently from the patient’s skin. Dressings are not usually required if 25- or 26-gauge needles have been used because they do not usually leave any mark, but if there is any bleeding small adhesive dressings may be applied. While still lying down, the patient is asked if they feel lightheaded or dizzy. If they do, they can be reassured that such symptoms are quite normal after a cervical nerve block and they should be allowed to lie quietly on the table until the dizziness settles, which may take a minute or two. When the patient is not dizzy, they are allowed to sit up with their legs over the side of the table, being supported by the operator and/or a nurse in case the change of posture makes them lightheaded. When they feel ready, the patient is helped from the table to a wheelchair and taken from the fluoroscopy suite to a room where they can rest and be observed by a nurse who is trained in pain assessment.

The test procedure continues with the postoperative, observational phase. In the period after the injections a nurse who is trained in pain assessment asks the patient to score the intensity of the index pain at 30-minute intervals, beginning immediately after return to the observation area and continuing until 90 minutes later, and the resultant pain scores are recorded on a pain chart. When scoring the pain, the patient is advised to ignore any procedural pain from the needling and any pain they may have in other regions, and to focus on the index pain they scored before the block was done. To maximize the accuracy of the result, the procedure (and the observational phase in particular) should be undertaken to minimize potential biases that threaten the integrity of the process. These potential biases include reporting bias and observer bias.

Reporting bias occurs if the patient describes their response to a block in ways that are influenced by misunderstanding of the purpose of the test or by their expectations or hopes, or if they wish (consciously or unconsciously) to please the recording nurse or the treating physician. Reporting bias may also occur if the patient is asked to describe their response some time after the test, such as the next day, when their memory of the postblock pain levels is less clear: this is termed recall bias.

Observer bias occurs if the recording nurse-observer is influenced by knowing when a patient’s pain might be expected to be relieved and when it might be expected to return. It may also occur if the nurse-observer wishes (consciously or unconsciously) for the patient to have a particular response to the block.

These potential threats are minimized if the procedure includes a double-blind protocol for selection of local anesthetic agents, if the patient and the nurse-observer have clear ideas about the test process and the potential biases it may involve, and the postblock pain scores are recorded during what is called “real-time assessment.” A double-blind protocol for selection of local anesthetics means the agent to be injected on any occasion is allocated randomly by the operator. Neither the patient nor the recording nurse are aware of the agent used, so they cannot predict when the pain might be expected to be relieved and when it might be expected to return. A double-blind protocol controls the procedure for reporting and observer biases. If the patient and the nurse-observer are clear about the index pain, and postblock scores refer to it only and do not include any pain the patient may have elsewhere, the method is controlled for other potential reporting and observer biases. Real-time assessment (meaning the postblock pain scores are recorded at the actual times they occur) controls the process for recall bias.

During the observational phase, the treating physician should stay away from the patient to avoid any interaction that might influence their responses. At the end of the observation period, the pain chart may be reviewed to determine the result of the test (Fig. 36-21).

Figure 36-21 Pain chart showing the initial intensity of the index pain and the visual analog scale (VAS) scores recorded after a medial branch block test that produced an analgesic response.

(Courtesy of Pendlebury Clinic, Newcastle, Australia.)

Cervical Medial Branch Block—Interpretation of Results

Accurate interpretation of test results depends on understanding and applying the standards set out in the literature and specified in the relevant practice guidelines.^140,141 A single MBB test result is deemed positive only if the index pain is relieved completely (i.e., its VAS score goes down to zero) in the postblock period, as shown on the pain chart in Figure 36-21. Any other result showing the index pain was still present, even if at a reduced level of intensity, must be considered negative or inconclusive. A treating physician may be tempted to try to interpret a reduced pain score, say if a patient’s index pain is reduced from 86/100 to 14/100, but rational interpretation of such a result will be confounded by the possibilities. The reduced score may be because the joint tested is contributing to, but not wholly responsible for, the pain or because the pain is actually stemming from an adjacent joint and one of its medial branches was blocked in the test, or because of some bias in recording.

Another factor to be taken into account in interpreting MBB responses is the possibility of false-positive and false-negative results. A study of cervical MBBs showed 27% of responses to single blocks are false-positive and 5% of responses are false-negative.¹⁴² That means even if the index pain is relieved completely after an MBB, the pain is only 73% likely to be stemming from the joint blocked and if the index pain is not relieved after an MBB, the pain is only 95% likely to be not from that joint. The same study considered the positive predictive values (PPVs) of positive single block responses. PPVs vary with the prevalence of the condition; if the prevalence is as high as 70%, the PPV of a positive block response is only 89%.

The liabilities of false-positive single block results are overcome by performing comparative blocks.¹⁴³ If a patient has a positive response to an MBB, a second block is performed on a separate occasion to test whether the first result was true-positive. The second block is done following the same procedure but using another local anesthetic agent with a different duration of action. If the patient again has complete relief of the index pain after the second block, the durations of the responses can be compared. If the duration of pain relief was longer with the longer acting anesthetic (bupivacaine) than with the shorter acting (lidocaine), that joint is positively identified as the source of the index pain and the result is described as “comparative block positive and concordant,” which is the criterion standard for definitive diagnosis of cervical ZJ pain.^140,141 Any other combination of results makes a comparative block test negative or inconclusive.

C7 Medial Branch Block—Procedure

Blockade of the C7 medial branch involves a slightly different technique because the nerve runs over a different part of its vertebra and has a wider range of anatomic variation in its course (see Fig. 36-12).¹³⁹ It usually passes forward over the bony surface of the tip of the superior articular process of the C7 vertebra, somewhere between its peak and the root of the transverse process, but in some individuals it is superficial to the bone rather than immediately adjacent to it because the nerve is separated from the bone by a slip of the semispinalis capitis muscle. To ensure blockade of the nerve, wherever it happens to lie, requires injections of 0.3 mL of local anesthetic at each of two target points, one on the lateral aspect of the curved surface of the articular process up near its peak and the other 4 mm superficial to that. The first part of the procedure is similar to that followed at other cervical levels. When the needle tip reaches bone on the C7 vertebra at the initial target site, its position is checked (first) in a lateral view (Fig. 36-22).

Figure 36-22 Lateral fluoroscopic view of the right C6-7 zygapophysial joint with the needle tips on bone over the centroid of the C6 articular pillar and the peak of the C7 articular pillar, the target points for C6 and C7 medial branch blocks, respectively.

The C-arm is then rotated 90 degrees for a direct AP view. In that view, the needle position is checked again to ensure its tip is on the “skyline,” (i.e., on the lateral edge of the articular process, not around on the anterior or posterior parts of the bone’s curved surface) and that the tip is above the transverse process (Fig. 36-23). If the needle tip is in the correct position, the first injection of 0.3 mL of local anesthetic is done while the needle tip is on bone. The needle is then withdrawn 4 mm and the position of its tip is recorded again off bone in an AP view before the second injection is made. The rest of the procedure is as described for the C3-4 to C5-6 levels.

Figure 36-23 Anteroposterior (AP) fluoroscopic view of the right C6-7 zygapophysial joint with the needle tips on bone at the target points for C6 and C7 medial branch blocks.

Third Occipital Nerve Block—Procedure

The C2-3 ZJ is supplied by a single nerve, the third occipital nerve (TON) on that side which is the medial branch of the dorsal ramus of the C3 spinal nerve. Accordingly, the test procedure at the C2-3 level that corresponds to MBBs at lower cervical levels is called a third occipital nerve block (TONB). The TON is thicker than other medial branches and that must be taken into account in consideration of procedures directed at it. Also, the TON carries some sensory fibers from a small patch of skin over the occipital region behind the ear on that side. Other medial branches generally do not carry skin sensation and this makes a difference to the observations after a TONB.

The general course of the TON is forward over the C2-3 ZJ in the pericapsular fascia to join its dorsal ramus in the C2-3 intervertebral foramen. The course varies between individuals so the nerve passes somewhat horizontally across the joint at a level that ranges between the top and just below the bottom of the intervertebral foramen (see Figure 36-12).¹³⁹ To ensure blockade of the TON wherever it happens to lie in this range requires injections of 0.3 mL of local anesthetic at each of three target points, which lie in a vertical line over the middle of the joint—a high one at the level of the apex of the C3 superior articular process; a low one at the level of the bottom of the C2-3 intervertebral foramen; and a middle one halfway between the other two (Fig. 36-24).

Figure 36-24 The three target points for a third occipital nerve block(arrows).

(From Barnsley L, Bogduk N: Medial branch blocks are specific for the diagnosis of cervical zygapophysial joint pain. Reg Anesth 1993;18:343-350, with permission.)

In other respects, the procedure is similar to that used for cervical MBBs at lower levels, although in the operative phase, only one needle is inserted and it is moved to each of the target points in turn. The patient is positioned on the x-ray table lying on his/her side with the painful side up. The insertion site is determined by the needle being placed on the skin of the patient’s neck over the middle target point (Fig. 36-25).

Figure 36-25 Lateral fluoroscopic view of the right C2-3 zygapophysial joint with the needle tip on skin over the target area for a third occipital nerve block.

The operator inserts the needle and steers it to each of the three target points in turn. At each target, the needle tip position is checked and recorded on a lateral fluoroscopic image (Figs. 36-26–36–28) and 0.3 mL of local anesthetic is injected.

Figure 36-26 Lateral fluoroscopic view of the right C2-3 zygapophyseal joint with the needle tip on bone at the high target point for a third occipital nerve block.

Figure 36-27 Lateral fluoroscopic view of the right C2-3 zygapophysial joint with the needle tip at the middle target point for a third occipital nerve block.

Figure 36-28 Lateral fluoroscopic view of the right C2-3 zygapophysial joint with the needle tip on bone at the low target point for a third occipital nerve block.

After the patient is taken from the fluoroscopic suite and their postblock responses are being recorded, there is an extra observation to be made in the case of a TONB. The observer should ask the patient if any numbness is felt. If an effective TONB has been achieved, the patient will usually have numbness of the small area of skin behind the ear on the side that the TON supplies; a lack of numbness suggests the block may not be effective. As with other MBBs, if a single test relieves the index pain completely, a second comparative block should be done to determine if the first result is true-positive.

Cervical Medial Branch Blocks—Validity

Validity, in the biologic sciences, is concerned with legitimate interpretation of observed phenomena. The validity of a diagnostic test, like comparative MBBs, is the extent to which the test assesses what it is intended to assess. There are three domains of validity: face validity, construct validity, and predictive validity—all of which contribute to the overall validity of any diagnostic instrument.

Face validity refers to whether a test or instrument appears, on the face of it, to assess what it is meant to assess. Cervical MBBs depend on the effects of blocking specific target nerves (and no others). Their face validity depends on whether they appear to block the target nerves only, (i.e., on their target-specificity). A study of the target-specificity of cervical MBBs¹³⁷ showed that when the blocks are performed in the manner described, in accordance with the standard guidelines and injecting the volumes specified, the injectates will cover the target nerves but will not spread to any other nerves. Therefore, cervical MBBs have face validity.

Construct validity refers to whether the interpretation of a measurement or test result reflects the theoretical construct of the phenomenon to be observed. The theoretical construct on which cervical MBBs depend is that a cervical ZJ generating pain can be identified by blocking the nerve(s) along which pain is transmitted from the joint. The construct validity of cervical MBBs depends on the extent to which positive or negative block results reflect that the joint tested is, or is not, a pain source. The sections on interpretation of results makes it clear that single cervical MBBs do not, on their own, identify painful joints definitively but double-blind controlled, comparative cervical MBBs, with positive and concordant results, do have construct validity.

Predictive validity refers to whether an instrument such as a diagnostic test is of value in predicting responses to treatment. It is directly related to therapeutic utility and of the greatest significance in the clinical application of diagnostic procedures like MBBs. The predictive validity of cervical MBBs depends on whether their results predict the likelihood of treatment being effective for relieving conditions diagnosed by the test. Positive and concordant, comparative MBB or TONB results identify pain stemming from a specific cervical ZJ on the basis of that pain being transmitted via the medial branch(es) blocked in the test procedure. Percutaneous radiofrequency cervical medial branch neurotomy is a therapeutic procedure established as effective for the treatment of cervical pain mediated via specific medial branches by the data of studies^139,144,145 that used positive and concordant, comparative MBB, or TONB results as the indications for the treatment procedure. Hence, comparative cervical MBBs and TONBs, performed in those studies as described, have predictive validity. Because they have face validity, construct validity, and predictive validity all based on sound, published data, cervical medial branch blockade is described in the ISIS practice guidelines¹⁴¹ as an established procedure.

There are many other publications on cervical blocks that touch on aspects of validity, including papers on procedures that sound like MBBs but are done in ways that differ from the standard guidelines, and the data from all those publications are extensive. Authors have described using placebo injections as an additional arm of a comparative block protocol¹⁴⁶; studies of the efficacy of cervical radiofrequency neurotomy^139,144,145 show that the results of comparative cervical MBBs using two local anesthetics and those of placebo-controlled comparative blocks have the same predictive validity for the outcomes of medial branch neurotomy treatment. Other authors have described MBB procedures that include the injection of contrast medium to confirm injectate spread¹⁴⁷ but their results do not contradict those of the earlier study¹³⁷ showing the target-specificity of MBBs done as described earlier. Placebo controls and contrast injection may be of value in specific circumstances, such as in research or for medicolegal purposes, but MBB and TONB procedures as laid down in the standard guidelines^140,141 (and as described in this chapter) are of known, established validity for use in regular clinical practice.

The data on validity in the foundation literature, on which the guidelines are based, are applicable to MBBs and TONBs performed as described. Those data cannot be generalized to blocks done in other ways, so such blocks cannot be assumed to be valid on that basis. In the absence of specific scientific evidence, whether blocks that are not performed in accordance with the standard guidelines are valid or not is simply unknown. Confusion about this has given rise to controversies about the reliability and validity of diagnostic block procedures in general; such controversies would diminish if block procedures were evaluated on the basis of the specific scientific evidence that applies to them.

Thoracic Medial Branch Blocks—Procedure

The thoracic medial branches run forward from above and below the thoracic ZJs they supply and over the transverse processes of the joint partners to join their respective dorsal rami in the intervertebral foramina. The numbering of the thoracic medial branches is one less than the vertebrae over which they run as explained in the earlier under “Anatomy”. This numeric variation need not be confusing if its pattern is appreciated. The T3-4 ZJ is supplied by the T2 and T3 medial branches, which run over the T3 and T4 transverse processes, respectively. The procedure to test the T3-4 ZJ as a source of pain involves blocking each of those two nerves. Even if they appreciate the numeric relationships between nerves and joints, it is important for clinicians to realize the potential for confusion and to be consistent in describing any procedure by the joint, or nerves, involved. The ISIS convention for designating joints by the use of a hyphen¹⁶ (e.g., the left T7-8 ZJ), is recommended for all written records.

The courses of the thoracic medial branches have been plotted by cadaveric dissection studies.¹⁴⁸ The upper and lower thoracic medial branches are reliably found on the bony surfaces of the thoracic transverse processes but the midthoracic medial branches are usually suspended in the soft tissues of the intertransverse spaces, as shown in Figure 36-29.

Figure 36-29 Plots of the courses of thoracic medial branches across the transverse processes of the vertebrae.

(From Chua WH: Clinical Anatomy of the Thoracic Dorsal Rami [thesis]. Newcastle, Australia, University of Newcastle, 1994, with permission.)

These thoracic medial branches are accessible for injection around them as they pass over the transverse processes. The target points are, generally, near the superolateral corners of the transverse processes, on the backs of the costotransverse joints. A needle can be placed onto the target point for a thoracic MBB most readily from a posterior approach. The facilities and equipment required are the same as specified for cervical MBBs.

A thoracic MBB procedure should be performed in accordance with the protocol described in the relevant practice guidelines.^140,149 The needling part of the procedure must be undertaken with particular care because thoracic needling carries the risk of pleural puncture and pneumothorax. To minimize that risk, the guidelines must be followed meticulously.

The procedure involves three phases similar to those described for cervical MBBs. The patient is positioned lying prone on the x-ray table and the skin of the thoracic region is swabbed and draped. The operator and the radiographer confer to ascertain the level to be tested and the first target nerve. For instance, if the patient is to have medial branch blocks to test the T3-4 ZJ, the first target nerve will be the T2 medial branch and its target point is near the superolateral corner of the T3 transverse process, on the back of the T3 costotransverse joint. When the first target is agreed, the radiographer obtains an AP view of the region. On that view, both operator and radiographer should count down the ribs from above to confirm the target level. When they agree, the radiographer cones the image to the region of the target joint and the adjacent transverse processes on that side. In the directly AP view, the ZJ itself will not be seen but its position can be inferred as being between the adjacent pedicles and immediately posterior to the corresponding intervertebral foramen. The transverse processes will be seen with the ribs in front of them, the top of each rib projecting above the top of the corresponding transverse process.

The target point for each of the T1 to T4 medial branches is on the dorsal surface of the transverse process, below its upper border near the lateral end, so not quite the superolateral corner of the transverse process but slightly medial to it. The target point for each of the medial branches from T5 to T8 is not actually on bone but just above the superolateral corner of the transverse process and behind the dorsal surface of the rib that articulates with the transverse process. The target points for the T9 and T10 medial branches are like those for the T1 to T4 nerves, on the outer parts of the transverse processes. The target point for each of the T11 and T12 medial branches is at the junction of the root of the superior articular process and the root of the transverse process; this is best seen in an oblique “Scottie dog” view. In that view the target vertebra resembles a Scottish terrier, the superior articular process making the dog’s ear, the transverse process and pedicle making its head, the pars articularis its neck and the rest of the posterior elements making the dog’s body.

Having obtained a suitable view, the operator selects a suitable needle, determines the insertion point, inserts the needle, and directs it to the target site as described for cervical MBBs. It is important to keep the needle tip over the bone of the transverse process at all times, to avoid the danger of pleural puncture. Before each adjustment of the needle, its path toward the target point should be checked on intermittent AP views, and the depth of insertion should be checked on intermittent lateral views. When the needle tip reaches bone at the target point, or the point just above the transverse process and behind the rib for T5 to T8 medial branches, its position is checked and recorded first in a lateral view and then in an AP view (Fig. 36-30). The first injection, of 0.5 mL of local anesthetic, is made at that site. Then the second target point is injected similarly.

Figure 36-30 Anteroposterior (AP) fluoroscopic view of a needle at the target point for a thoracic medial branch block. Contrast medium (0.5 mL) has been injected to show the spread of injectate.

Thoracic Medial Branch Blocks—Interpretation of Results

As with cervical MBBs, a single thoracic MBB test result is positive only if the index pain is relieved completely (i.e., its VAS score goes down to zero) in the postblock period. Any other result, with postblock index pain scores of more than zero must be considered negative or inconclusive. If the first block is positive, the issue arises as to whether it is true-positive or false-positive. This is even more important than it is for cervical blocks, because a study has shown single thoracic MBBs have a false-positive rate of 58%.⁶⁴ Comparative blocks must be done before any sensible conclusion can be drawn from thoracic MBB results. Only two comparative block results that are both positive and concordant can be interpreted as identifying a specific thoracic ZJ as a pain generator.

Thoracic Medial Branch Blocks—Validity

Rational consideration of the validity of thoracic MBBs is problematic because there are no data on which a justifiable evaluation can be based. For this reason, the ISIS practice guidelines describe thoracic medial branch blockade as an “emerging” rather than an “established” procedure.

Thoracic MBBs may have face validity on the grounds that they appear prima facie to test what they are meant to test but there are no data to support this. There are no published studies of the target-specificity of thoracic MBBs and no legitimate assumption can be made on the grounds of target-specificity of MBBs in other spinal regions. The variable courses of thoracic medial branches, and particularly the fact that the midthoracic medial branches do not lie on bone but are suspended in the intertransverse spaces where they are hard to locate, weigh against the face validity of thoracic MBBs. In the absence of sound data on target specificity, thoracic MBBs cannot be said to have face validity.

The construct validity of thoracic MBBs would depend on the extent to which positive or negative block results reflect that the thoracic ZJ tested is, or is not, a pain source. Again, there are no sound data on which to base a rational evaluation but the high false-positive rate of thoracic MBBs militates against their construct validity.

Even if thoracic MBBs did have verifiable face and construct validity, they cannot be said to have predictive validity because there is no specific treatment proven effective, on the basis of published data, for the treatment of thoracic ZJ pain. Thus, no legitimate conclusion can be drawn about whether thoracic MBBs predict the likelihood of treatment being effective.

Lumbar Medial Branch Blocks—Procedure

Lumbar medial branch blockade is an established diagnostic procedure. It involves blocking the nerves that transmit nociceptive signals from lumbar ZJs so as to determine whether a specific ZJ is a pain generator.

Each lumbar ZJ has sensory innervation via two nerves. The joints from L1-2 to L4-5 are each supplied by articular fibers of two medial branches, numbered like the thoracic medial branches and for the same reason, as one less than the joint they supply. Thus, the L4-5 ZJ has sensory innervation via the L3 and L4 medial branches. The L5-S1 ZJ has its innervation by articular fibers of the L4 medial branch above it and articular fibers that go directly to the L5 dorsal ramus below. The L1 to L4 medial branches receive articular fibers from the tops and bottoms of adjacent ZJs and run through the mamillo-accessory notches, under the mamillo-accessory ligaments, and then across the bases of the superior articular processes near where they join the transverse processes. The numbering is such that the L1 medial branch runs across the superior articular process of the L2 vertebra near where it joins the L2 transverse process. The L5 dorsal ramus runs across the superior articular process of the S1 sacral segment near where it joins the ala of the sacrum (which is analogous to a transverse process). These nerves are accessible for blocking because they run across the bases of the superior articular processes, at the points marked by arrows in Figure 36-31. These points each correspond to the “eye” of a Scottie dog configuration as seen in an oblique fluoroscopic view. If the variation of joint and nerve numbering seems confusing, it may help to remember that the target points for a lumbar ZJ are near the roots of the transverse processes of the same designation (i.e., the target points for MBBs of the L4-5 ZJ are near the roots of the L4 and L5 transverse processes). As with thoracic structures, to avoid ambiguity when referring to a procedure targeting lumbar ZJs and medial branches, it is important to describe it by the joint or the nerves involved, explicitly.

Figure 36-31 Oblique diagram of the L3 and L4 medial branches, and the L5 dorsal ramus, running across the bases of the superior articular processes and accessible for injection at the points marked by arrows (the “eyes of the Scottie dogs” as seen in oblique fluoroscopic views).

The lumbar MBB procedure, like those of cervical and thoracic MBBs, involves a preoperative phase, an operative (needling) phase, and a postoperative (observational) phase, and when results are at hand, their interpretation which may be considered a fourth phase. The phases are much the same as those described for the other spinal regions. To ensure validity, the procedure should be carried out in accordance with the standard guidelines.^140,150

The patient is positioned lying prone on the x-ray table and the skin of their lumbar region is exposed, swabbed, and draped. The operator and radiographer ascertain the spinal level to be tested and the radiographer obtains an AP view of the region. On that view, both operator and radiographer should count down the lumbar vertebrae from the last thoracic vertebra that has a rib attached. If there seem to be six lumbar vertebrae because the first sacral segment is lumbarized, or if any other anatomic variation is noted, the joint to be tested should be confirmed. When the target level is agreed, the radiographer swings the C-arm along the patient to focus on the end-plates of the vertebrae of that segment in an AP view, and then moves it across to obtain an ipsilateral oblique (Scottie dog) view of the target region and cones to it. The target point will be the “dog’s eye.”

To block each nerve the operator takes a suitable needle, determines the insertion point, inserts the needle and directs it to the target site as described for cervical MBBs. It is important to keep the needle tip over bone at all times to avoid overshooting the target. Before each adjustment of the needle, its path toward the target point should be checked on intermittent oblique views. When the needle tip reaches bone at the target point, its position is checked and recorded in an oblique view (Fig. 36-32). Prior to the injection of local anesthetic, 0.5 mL of nonionic contrast medium should be injected to test for intravascular placement. This is done because a study has shown that at the target points for lumbar MBBs, the needle tip is intravascular in 8% of cases¹⁵¹ and if so, any injectate would be carried away in the blood and produce a false-negative result. If the contrast shows the needle tip is not intravascular, 0.5 mL of local anesthetic is injected.

Figure 36-32 Oblique “Scottie dog” view of the right L4-5 zygapophysial joint with needle tips on bone at the target points for blocking the L3 and L4 medial branches. From the lower needle, contrast is seen flowing into a regional vein (black arrows).

Lumbar Medial Branch Blocks—Interpretation of Results

As with other medial branch blocks, a single lumbar MBB test result is positive only if the index pain is relieved completely. Any other result must be considered negative or inconclusive.

If the first block is positive, the question arises as to whether that is a true-positive or false-positive result. One study showed that single lumbar MBBs have a false-positive rate of 38%¹⁵²; other studies have shown false-positive rates between 25% and 41%.^74,153 The practical effect of those rates is that if the true-positive to false-positive ratio is simplified conservatively as 2:1, of every three positive lumbar MBB results only two will be truly positive. When considered in conjunction with the relatively low prevalence of lumbar ZJ pain, the high false-positive rate renders the positive predictive value of a single test unacceptably low.¹⁵² To compensate for these liabilities, controlled, comparative blocks must be performed, as in other spinal regions, before valid conclusions can be drawn. For maximal diagnostic confidence based on lumbar MBB results, because of the high false-positive rate and the low prevalence, these comparative blocks must include placebo controls.¹⁵⁴ Placebo injections must only be given with informed consent but that is usually obtained readily if patients are told about the value of the extra control to the interpretation of the test results. Two comparative MBB results that are both positive and concordant, and a negative placebo control result, can be interpreted as identifying a specific lumbar ZJ as a pain generator.

Lumbar Medial Branch Blocks—Validity

The face validity of lumbar MBBs depends on whether they appear prima facie to test what they are meant to test. Evidence of their face validity is provided by a study of the target-specificity of lumbar MBBs which showed that 0.5 mL injectates at lumbar MBB target sites cover the medial branches and do not spread to any other diagnostically significant structures.¹⁵¹

The construct validity of lumbar MBBs depends on the extent to which positive or negative block results reflect that a lumbar ZJ tested by them is, or is not, a pain source. A randomized, controlled trial showed lumbar MBBs anesthetize painful lumbar ZJs in 89% of cases.¹⁵⁵ The issue of false-positive results is addressed by performing placebo-controlled, comparative blocks, as described earlier. False-negative results of lumbar MBBs have been shown to occur at a rate of 8% due to intravascular injection¹⁵¹ but that issue is addressed by preinjection of contrast medium. When performed as described, with contrast injection to check for intravascular needle placement and in comparative series with placebo controls, lumbar MBBs do have construct validity.

The most practical domain of validity is predictive validity or therapeutic utility. Lumbar MBBs have been shown to have predictive validity by studies of percutaneous radiofrequency lumbar medial branch neurotomy. These studies show that therapeutic procedure is effective for the relief of lumbar ZJ pain when it is performed on the basis of positive and concordant, controlled, comparative, lumbar MBBs.^156,157

Development

Intraarticular blocks (IABs) of ZJs were developed long before MBBs and for some years were used as diagnostic tests for ZJ pain. In the light of the current scientific literature, IABs of ZJs must now be considered obsolete for diagnostic purposes. Medial branch blocks are much more appropriate as diagnostic tests because they are easier to perform, less invasive, safer and validated by the comparative block protocol set out in the standard guidelines. By contrast, IABs of ZJs are more difficult to perform, especially when the joint space is narrowed or occluded by osteophytes. MBBs involve needling of skin, subcutaneous tissues, muscles, and periosteum only, whereas IABs involve intrusion into the joint where the needle may damage delicate intraarticular structures. In MBB procedures, the needle tip can be kept over bone, which prevents inadvertent intrusion into nontarget structures but when IABs of ZJs are done, it is possible for the needle to pass right through the joint and into the adjacent spinal canal, where it may damage the spinal cord or nerve roots. When comparative MBBs are performed in accordance with the standard guidelines they are of proven validity, whereas there are no data on the validity of ZJ IABs. Even if comparative IABs were performed, the differential effects of anesthetic agents inside joints are not known. For these reasons diagnostic IABs of ZJs are not recommended and they will not be described in this chapter.

The lateral atlantoaxial (C1-2) and atlantooccipital (C0-1) joints have nerve supplies quite different from those of the ZJs, as described under “Anatomy”. The lateral atlantoaxial (C1-2) and atlantooccipital (C0-1) joints are supplied via the ventral rami of their respective spinal nerves, not via the medial branches and dorsal rami as are the ZJs. Thus, the procedures of MBBs have no application for the joints of the C0-1 and C1-2 spinal motion segments. Diagnostic blocks of the ventral rami are not practical because of the other structures those nerves supply. Therefore, the lateral atlantoaxial (C1-2) and atlantooccipital (C0-1) joints are tested as sources of pain by intraarticular blockade. Diagnostic IABs of the LAAJs were first described in 1987.¹⁵⁸ LAAJ blocks are described in the ISIS practice guidelines as an established procedure¹⁵⁹ and will be described here. Diagnostic IABs of the atlantooccipital joints are not described at all in the guidelines, not even as an emerging procedure, because they are not supported by evidence of safety and validity; accordingly, they will be addressed only briefly in this chapter.

Lateral Atlantoaxial (C1-2) Joint Blocks—Procedure

The LAAJs have been shown to be injured in traumatic events⁶² and are reported as sources of cervicogenic headache^94,95,160 in the distribution illustrated in Figure 36-8. LAAJ blocks should be considered in the diagnosis of a patient with headache in that distribution and a history that suggests the possibility of LAAJ injury. Headache in that distribution may also stem from the C2-3 and C3-4 ZJs, and cervicogenic headache is believed to be more commonly associated with those upper ZJs, so it is prudent to test the C2-3 and C3-4 ZJs first. The main indication for diagnostic LAAJ blocks is for the investigation of cervicogenic headache when C2-3 and C3-4 blocks have proved negative, or when there is residual cervicogenic headache after C2-3 and/or C3-4 headache has been treated effectively.

Intraarticular blocks of the LAAJs are performed via a posterior approach. As for any invasive procedure, it is vital for the operator to have sound knowledge of the regional anatomy. In relation to an LAAJ block, the operator must be aware of two important anatomic structures, the vertebral artery and the C2 dorsal ramus, which lie very close to the target joint. The vertebral artery is usually located immediately lateral to the LAAJ but within the normal range of anatomic variation it may be posterior to the lateral part of the LAAJ. The dorsal ramus of the C2 spinal nerve passes superomedially across the back of the superior articular process of the axis (i.e., just behind the lower process of the LAAJ). To avoid inadvertent puncture of either of those structures, the operator must be meticulous in following the guidelines for the procedure.¹⁵⁹ The most appropriate local anesthetic for the block is 2% lidocaine. Longer-acting local anesthetics such as bupivacaine should not be used at the atlantoaxial level because if it should leak from the capsule, it could cause prolonged high spinal block.

Each LAAJ block procedure involves a preoperative phase, an operative needling phase, and a postoperative observational phase, followed by the diagnostic interpretation of test results (much as described for cervical MBBs). In the operative phase, the patient is positioned lying prone on the table and made comfortable with the neck flexed; the neck is swabbed with antiseptic and draped. Throughout the operative phase the patient should be kept informed of its progress because they are lying face down with their head covered and may feel vulnerable.

The radiographer obtains an AP view of the patient’s upper neck and moves the C-arm longitudinally until the LAAJs are seen as ellipsoid structures on either side of the odontoid process near its base. Both the operator and the radiographer should identify the image of the arch of the axis, which can lie over the LAAJs in an AP view. If necessary, the radiographer angles the x-ray beam so the image of the arch of the atlas is well above the target joint and there is a clear view of the upper margin of the LAAJ in its middle third, which is the initial target point. The radiographer then cones to that area. The operator takes a spinal needle and locates the insertion site on skin over the target (Fig. 36-33).

Figure 36-33 The anteroposterior (AP) fluoroscopic view required for an intraarticular block of the right lateral atlantoaxial joint. Note the image of the arch of the atlas, above and clear of the upper margin of the joint in its middle third (the initial target point) and the needle tip on skin over that point.

The needle is inserted and guided through the muscles as described for other blocks. Intermittent AP fluoroscopic views are used to keep the needle tip over the target zone and lateral views are used to check the depth of penetration (Fig. 36-34). When the needle tip reaches bone at the initial target point, on the upper margin of the LAAJ in the middle third of the inferior articular process of the atlas, its position is checked and recorded on AP and lateral fluoroscopic images (Figs. 36-35 and 36-36).

Figure 36-34 Lateral fluoroscopic view of the right lateral atlantoaxial joint showing the needle passing under the arch of the atlas and on track toward the initial target point for an intraarticular block.

Figure 36-35 Anteroposterior (AP) fluoroscopic view of the right lateral atlantoaxial joint with the needle tip on bone at the initial target point for an intraarticular block.

Figure 36-36 Lateral fluoroscopic view of the right lateral atlantoaxial joint with the needle tip on bone at the initial target point for an intraarticular block.

The needle is then withdrawn slightly, directed downward and inserted into the joint. Another intermittent lateral view should be obtained to check the needle tip is intraarticular. A syringe of nonionic contrast medium is connected to the needle by an extension tube and the needle tip placement confirmed by injection of about 0.2 mL of contrast medium to produce an arthrogram of the LAAJ (Fig. 36-37). The C-arm is then rotated 90 degrees and the confirmatory arthrogram recorded on an AP view as well (Fig. 36-38).

Figure 36-37 Lateral fluoroscopic view of a right lateral atlantoaxial arthrogram recorded to confirm correct placement of the needle tip for an intraarticular block.

Figure 36-38 Anteroposterior (AP) fluoroscopic view of a right lateral atlantoaxial arthrogram recorded to confirm correct placement of the needle tip for an intraarticular block.

Taking care to keep the needle in place, the operator then draws back with the syringe to aspirate the joint and remove as much contrast medium as possible from it; if left there the contrast will occupy space to be taken by the local anesthetic. After that, another syringe and a fresh extension tube filled with 2% lidocaine is fitted to the needle, again taking care not to dislodge the tip, and about 0.5 mL or a little more of the local anesthetic is injected slowly, the operator being careful to feel the pressure as the joint cavity is infiltrated so as not to inject too large a volume and rupture the joint capsule. Then the needle is removed.

Lateral Atlantoaxial (C1-2) Joint Blocks—Interpretation of Results

As with other diagnostic blocks, a lateral atlantoaxial joint block is considered positive only if the index pain is relieved completely (i.e., its VAS score goes down to zero) in the postblock period. Any other result must be considered negative or inconclusive. If the first block is positive, the issue arises as to whether it is true-positive or false-positive. There are no sound data on the false-positive and false-negative rates of LAAJ blocks, but such possibilities can be envisaged. Comparative blocks cannot be done safely as for MBBs because it is not safe to inject long-acting local anesthetic at the atlantoaxial level of the spine, so if confirmation of a result is required, the best that might be done is to repeat the block on another occasion and compare the two results.

Lateral Atlantoaxial (C1-2) Joint Blocks—Validity

Lateral atlantoaxial joint blocks have face validity on the grounds of the confirmatory arthrograms recorded during the procedure. Those arthrograms show that LAAJ blocks are target specific. The construct validity of LAAJ blocks would depend on the extent to which positive or negative block results reflect that the joint tested by them is, or is not, a pain source. There are no numeric data on these aspects of LAAJ blocks.

The predictive validity or therapeutic utility of LAAJ blocks is problematic because there is no simple, validated treatment for LAAJ pain. Intraarticular steroid injections are used by some but the evidence for their effectiveness is weak and based on practice audits^95,161; there are no data from randomized, controlled trials. The only other treatment described is surgical fusion of the atlantoaxial segment^162,163; although there is a place for such a procedure, there are no rigorous data to show that it is validly indicated by positive LAAJ blocks.

Atlantooccipital (C0-1) Joint Blocks

The atlantooccipital (C0-1) joints have been reported to be pain sources^94,161,164 and a pattern of cervicogenic headache associated with them has been described.⁹⁴ Consideration of the morphology and biomechanics of the atlantooccipital (AO) articulations suggests that they are less likely to be injured, and so to give rise to articular pain, than other spinal joints. The most likely mechanism of injury would be distraction, as the AO joints have powerful restraints to other biomechanical threats. Be that as it may, the prevalence of AO joint pain is unknown. Diagnostic intraarticular blocks of the AO joints are possible in theory but are seldom performed in practice. It is certainly not a procedure to be undertaken because it can be. The technique of injecting an AO joint is a highly specialized one carrying significant risks of damage to vital structures including the vertebral artery, the internal jugular vein, and the vagus nerve, all of which lie adjacent to the joint.

If an AO joint injection is considered necessary and the risks are worth taking, a lateral oblique approach is used with very careful control of the needle course on intermittent oblique fluoroscopic views and monitoring of the depth of insertion on open-mouth AP views. A posterior approach must not be used because the vertebral arteries cross the atlantooccipital joints posteriorly; any attempt to inject an AO joint from behind carries a serious risk of injuring the vertebral artery, with potentially disastrous consequences. Diagnostic AO joint blocks should not be attempted by anyone without rigorous practical training in the procedure. If needle position is confirmed by an arthrogram (Fig. 36-39) and a diagnostic block of the joint is achieved, the interpretation of the result is subject to the same liabilities as those of LAAJ joint blocks. The validity of diagnostic AO joint blocks is also limited like that of LAAJ blocks. If a confirmatory arthrogram is recorded, a block can be said to have face validity (target specificity). There are no data on the construct validity of AO joint blocks and very little data on their predictive validity or therapeutic utility—only two small data sets based on practice audits.^161,164

Figure 36-39 Oblique fluoroscopic view of a right atlantooccipital (AO) arthrogram recorded to confirm correct placement of the needle tip for an intraarticular block. The four arrowheads mark the margins of the target AO joint, the thinner arrow indicates the top of the dens, and the thicker arrow indicates the lateral margin of the contralateral lateral atlantoaxial (LAA) joint.

Development

ZJ pain can be diagnosed definitively, specifically and validly, by the careful application of medial branch block procedures. When it is, the question arises as to how to treat it. The treatment of pain of (or presumed to be of) ZJ origin has a long and colorful history. An extensive range of treatments has been put forward over the years. Many of treatments are still done by a variety of practitioners to relieve pain thought to stem from ZJ impairment of some sort, and variously believed to be due to injury, disease, mechanical dysfunction (including locking, displacement and/or hypermobility) and a host of other putative mechanisms. The forms of treatment still offered include physical modalities (heat, cold, infrared, ultrasound, laser, electrotherapy, massage, mobilization, manipulation), exercise programs, analgesic and anti-inflammatory medications, dry needling, “prolotherapy,” acupuncture, transcutaneous electrical nerve stimulation, other forms of neuromodulation such as spinal cord stimulation, spinal surgery (in particular arthrodesis), functional restoration, “back schools”, and psychological interventions such as cognitive behavioral therapy. These are only some of the “mainstream” treatments: the total range is vast and only limited by human imagination.

All these treatments have their places in the broader scheme of things and are justified to various extents by types of supporting evidence ranging from anecdotal testimonies to rigorous scientific data. The purpose of this chapter is to describe particular interventions used in the management of ZJ pain by practitioners of pain medicine or spine care. Even in those fields, individual practitioners differ in their approaches. To avoid drifting into uncertain areas, this chapter has an anchor in reality in the form of the scientific evidence base; it is limited (in the main) to interventions supported by scientific data for the diagnosis and treatment of ZJ pain. Two types of treatment for ZJ pain, intraarticular injections of corticosteroid and percutaneous radiofrequency medial branch neurotomy, will be described and the evidence relating to them will be reviewed.

Therapeutic Intraarticular Joint Blocks—Procedure

Intraarticular injections of corticosteroid are a traditional form of treatment used by physiatrists, rheumatologists, and others in the management of peripheral joint pain. The concept has been applied to spinal joints and therapeutic intraarticular joint block procedures have been developed. These procedures, designated by some as “facet blocks,” are readily performed by those with sound knowledge of regional anatomy and practical skills in fluoroscopically-guided injection techniques. A cervical ZJ is most easily injected via a lateral approach but the procedure must be performed cautiously, with intermittent lateral views to guide the needle path and careful control of the depth of penetration on intermittent AP views, to avoid the danger of the needle passing right through the joint and into the spinal cord. An upper or midthoracic ZJ is best injected via an approach from behind and below, so the needle tip enters the joint at the midpoint of its lower margin; the needle course must be controlled carefully with intermittent AP views to ensure the needle tip stays over bone and does not stray laterally, where it may cause inadvertent pleural puncture, or medially, where it may enter the intervertebral foramen. The lowest two thoracic ZJs, and the lumbar ZJs, are most easily injected from an oblique posterior approach controlled by intermittent oblique views of the Scottie dog projection.

The procedure for each of these intraarticular injections involves three phases like those described for cervical MBBs. In the operative phase the patient lies on the table in the appropriate position (laterally, with the painful side up for cervical injections or prone for thoracic and lumbar procedures). All precautions for aseptic technique are followed, as described for MBBs. The operator and radiographer confirm the target joint and the radiographer obtains a suitable view and cones to the target area. The operator introduces the needle and the guides it through the muscles, taking care to check its path and depth of penetration on intermittent fluoroscopic views, until its tip reaches bone near the target joint. The position of the tip on the joint margin is confirmed on AP and lateral views. Then the tip is guided into the joint cavity. Its position intraarticularly is checked on further AP and lateral views and contrast is injected for a confirmatory arthrogram as described for diagnostic blocks of the LAAJs. Typical cervical, thoracic, and lumbar ZJ arthrograms are shown in Figures 36-40–36–42.

Figure 36-40 Lateral fluoroscopic view of a right C2-3 zygapophysial joint arthrogram recorded to confirm correct placement of the needle tip for a therapeutic intraarticular joint block.

Figure 36-41 Anteroposterior (AP) fluoroscopic view of a right T6-7 zygapophysial joint arthrogram recorded to confirm correct placement of the needle tip for a therapeutic intraarticular joint block. Arrows indicate joint margins.

Figure 36-42 Anteroposterior (AP) fluoroscopic view of a right L5-S1 zygapophysial joint arthrogram recorded to confirm correct placement of the needle tip for a therapeutic intraarticular joint block.

After the confirmatory arthrogram has been recorded, the operator aspirates and then injects the joint with a mixture of local anesthetic and a depot corticosteroid preparation. As with diagnostic intraarticular blocks, the operator must inject slowly and monitor the pressure of the injection to avoid rupture of the joint capsule. When about 1 mL of the mixture has been instilled, more or less depending on the pressure felt during injection, the needle is withdrawn and the patient is taken from the fluoroscopy suite to rest and recover.

Therapeutic Intraarticular Joint Blocks—Effectiveness

Therapeutic intraarticular joint blocks have been used for the treatment of spinal joint pain for some years and remain in common use. The evidence of their effectiveness is not compelling. The reports of uncontrolled, observational studies,^165–168 mainly of intraarticular corticosteroid injection of lumbar joints, gave the impression that the treatment is effective for ZJ pain. These reports encouraged the use of corticosteroid injections for pain from spinal joints at all levels, including ZJs, LAAJs, and atlantooccipital joints.

The results of randomized, controlled trials do not confirm the utility of the treatment. A randomized, controlled trial of cervical intraarticular corticosteroid injections for ZJ pain¹⁶⁹ showed the treatment is not more effective than placebo. A randomized, controlled trial of lumbar intraarticular corticosteroid injections for ZJ pain¹⁷⁰ also showed no benefit of steroid injections when compared to saline controls. There are no sound data from randomized, controlled trials for the effectiveness of intraarticular corticosteroid injections at other spinal levels.

Radiofrequency Cervical Medial Branch Neurotomy—Development

Radiofrequency medial branch neurotomy (RFMBN) is a treatment that evolved from earlier concepts of trying to relieve ZJ pain by interrupting conduction in the nerves that transmit nociceptive information from the joints. Early attempts at denervation of ZJs by injection of phenol,¹⁷¹ by cryotherapy¹⁷² and by surgical neurectomy¹⁷³ either failed to achieve their purposes or had unacceptable unwanted effects, or both. In particular neurectomy, if it could be achieved, caused neuroma formation and deafferentation, both of which caused worse pain than that which the neurectomy was designed to relieve. In the 1970s, radiofrequency energy was explored as a means of treating nerves to prevent conduction of pain from ZJs.¹⁷⁴ Success was claimed in descriptive studies of the technique but it was called into question when its stated targets were found to be inaccurate anatomically.¹⁷⁵ Further research lead to the development of more useful techniques.

In its evolution over the last 3 decades, RFMBN has appeared in several different forms. The most effective form, that set out in the ISIS practice guidelines¹⁷⁶ and described in this chapter, involves placing an RF electrode parallel to, and closely up against, each medial branch nerve to be treated. Radiofrequency energy is then applied in such a way as to “cook” (i.e., coagulate by heat) a length of the nerve in the accessible part of its course. The purpose of medial branch neurotomy (MBN) is to change the chemical structure of the nerve fibers so they do not conduct nociceptive signals. Nerves treated by thermal RF undergo chemical change but their structure remains as before (like a hard-boiled egg). The maintenance of neural structure obviates the risk of neuroma formation, which occurred after neurectomy. To perform RF cervical MBN properly, making multiple heat lesions over the course of each target nerve from different directions, takes about an hour for each nerve treated, so a two-nerve procedure such as is necessary to treat each of the cervical ZJs (other than that at C2-3) takes about 2 hours.

The equipment required for MBN includes, in addition to the facilities and equipment described for MBBs, an RF generator capable of producing a conventional, thermal RF field, a matched earth plate, connecting leads, and an RF electrode. A large-diameter (18 or 16 gauge) RF electrode should be used to avoid the inadequate effects that were reported in earlier times when smaller electrodes were used.¹⁷⁷

Radiofrequency Cervical Medial Branch Neurotomy—Procedure

The percutaneous RF cervical MBN procedure involves three phases, a preoperative phase, an operative phase, and a postoperative, recovery phase. In the preoperative phase the patient is admitted to the treatment facility and checked by the admitting nurse in the same ways as described for diagnostic block procedures, but with extra attention to their understanding of the procedure planned and its effects. The nurse should make sure the patient understands that RFMBN is a much longer procedure than the diagnostic blocks they will have had previously, and they will need to lie still at certain stages. Any questions the patient may have should be answered by the nurse or referred to the treating physician, so the patient has all relevant information before giving informed consent in writing. The patient is then changed into a hospital gown for transfer to the fluoroscopy suite.

For the operative phase the patient is positioned laterally on the x-ray table with the side to be treated up (see Fig. 36-15). They are made as comfortable as possible with padding under their lower shoulder and hip, and reassured that they will be kept informed of the progress of the procedure. Full aseptic precautions are observed to ensure a sterile field. The operator should scrub and put on a sterile gown and gloves, the image intensifier should be covered with a sterile cover, and the skin of the side and back of the patient’s neck should be swabbed widely with antiseptic and allowed to dry, then sterile drapes should be applied so the patient’s upper body is covered, except for the side and back of the neck, and the face. It is important to keep the face clear of drapes so the patient can see adequately, breathe easily, and communicate with the operator. The operator and the radiographer confer to confirm the nerves to be treated and the one to be treated first. Then the radiographer obtains a clear lateral view of the target joint and cones to it.

The operative phase begins with an MBB of the first target nerve, done as described earlier for a cervical MBB. When the needle is on bone, the operator injects through it a long-acting local anesthetic such as bupivacaine 0.5%, in multiple small volumes up to about 1 mL in total to cover the articular pillar around the nerve. The needle should be left in situ with its tip on bone to mark the position of the target nerve.

The C-arm is then swung though 90 degrees for an AP view in which the tip of the block needle is seen on the waist of the articular pillar of the target vertebra. A second spinal needle is inserted via a posterior approach so its tip meets that of the block needle. That second needle will lie along the path to be followed by the electrode in its first (sagittal) pass. A short-acting local anesthetic, such as lidocaine 1%, is injected through the second needle as it is withdrawn slowly to anesthetize the electrode path. Then the electrode is introduced and directed along a sagittal course, with the guidance of intermittent AP and lateral views, until it rests on the articular pillar as shown in Figure 36-43. The position of the electrode is confirmed on an AP view to check it is lying snugly against bone, as in Figure 36-44.

Figure 36-43 Lateral fluoroscopic view showing an electrode in place on the articular pillar for percutaneous radiofrequency cervical medial branch neurotomy.

Figure 36-44 Anteroposterior (AP) fluoroscopic view showing an electrode in place snugly against the articular pillar for percutaneous radiofrequency cervical medial branch neurotomy.

The first radiofrequency lesion is made at that site. The patient should be told the nerve is about to be “cooked” and they should lie still for the time it will take to make the lesion, but they must speak out to report any discomfort they feel during the lesioning process. The generator is then activated to heat the electrode tip to 85° C for 90 seconds. If the patient reports any pain while the energy is being applied, the RF generator should be switched off immediately, more local anesthetic injected via the block needle, and the electrode position readjusted if necessary and checked again before proceeding.

After the first lesion is made and the electrode has cooled sufficiently, it is moved to parallel positions above and below the first position, and further lesions are made at those sites. Then the electrode is withdrawn. Another spinal needle is introduced for an oblique pass, at 30 degrees to the course of the sagittal one. In the oblique pass, the electrode will reach further forward on the articular pillar. Three more lesions are made there at 82° C. Six RF lesions are made over the target nerve so they overlap to make sure they cover the area where the nerve may lie. In some cases, when the articular pillar is taller than usual and the nerve may lie outside the field covered by the standard number of lesions, more than three lesions are required in each pass. After the first nerve has been treated adequately, the procedure is repeated to treat the other target medial branch and then the electrode is removed. The insertion sites are covered with sterile dressings and the surgical drapes are removed.

When the operative phase is complete, the patient is transferred from the fluoroscopy suite for recovery. A cold pack is applied to the treatment site to minimize bruising. The patient should be warned to be careful because they may be unsteady while the local anesthetic is acting, told not to drive a vehicle that day, and advised to rest over the next few days. They should be provided with analgesic medication to relieve the soreness they will feel from the muscles through which the electrodes passed. When they feel ready, they can be allowed to go home. They should be followed up by telephone over the next few days until the postoperative soreness dissipates.

Radiofrequency Cervical Medial Branch Neurotomy—Effectiveness

The efficacy of RF cervical MBN has been demonstrated by the results of a rigorously designed, randomized, placebo-controlled clinical trial of the procedure done as described.¹⁴⁴ The trial involved a relatively small number of subjects (N = 49) but the design of the study gave it immense statistical power (100%) which is far more important than the sample size. The method used in that trial was judged by objective assessors as the standard for the procedure¹⁷⁸ and is the method described in the standard practice guidelines.¹⁷⁶ Percutaneous RF cervical MBN, when validly indicated and performed in accordance with those guidelines using 18- or 16-gauge electrodes to make multiple overlapping lesions over the medial branches from sagittal and oblique passes, is an effective treatment for established cervical ZJ pain. Most patients will have complete relief of pain from the joint treated and of any disability caused by that pain.^144,145 Any psychological distress associated with the pain can also be expected to resolve.¹⁷⁹ If the patient has any residual neck pain after the postoperative discomfort has dissipated, it will be from another pain generator, possibly another injured ZJ that can be investigated and treated similarly.

The effect of RF lesions made as described can be expected to last for an extended period. Data from studies with long follow-up show that after cervical MBN, total pain relief usually lasts for between 270 and 400 days (9 to 13 months).^{144,145,180,181} The treated nerves are coagulated but not destroyed anatomically and they will recover from the lesions over time. When they do, the pain will return but the RF treatment can then be repeated as needed to restore control of the pain and pain-related disability for as long as the problem persists. RF cervical MBN, when performed correctly and followed appropriately, can completely relieve patients of cervical ZJ pain.

No other report has been published of a rigorously-designed, randomized, controlled trial of MBN for cervical ZJ pain, although there have been publications describing cervical RF procedures done in other ways. As with other spine interventions, it is strongly recommended that RF cervical MBN be undertaken in accordance with the standard method described in the standard guidelines, the method that has been proven effective.

The establishment of a standard necessarily implies other methods are less acceptable. The data of the rigorous controlled trial of MBN¹⁴⁴ cannot be generalized to “RF neurotomy” performed in other ways, such as procedures using smaller gauge electrodes, those aimed at different target sites and/or those in which fewer lesions are made over each nerve. Such procedures are likely to achieve less satisfactory outcomes. Systematic reviews that include reports of diverse procedures, all called “radiofrequency neurotomy” but done in different ways, will inevitably show a diversity of outcomes which can be perplexing for anyone trying to gauge the effectiveness of proper MBN treatment. Failure to appreciate the differences between procedures that are given similar names, but vary in fundamental ways, can lead to confusion and controversy. If so, the problem lies with the understanding, not with the standard method. This issue is addressed further under “Radiofrequency Lumbar Medial Branch Neurotomy—Effectiveness”.

Radiofrequency Thoracic Medial Branch Neurotomy

The principles of RFMBN could, in theory, be applied for treatment of thoracic ZJ pain. The targets would be the thoracic medial branches whose anatomic positions are described in the section “Thoracic Medial Branch Blocks”. The relationships of thoracic medial branches to other structures, especially the pleura, present technical difficulties in placing thermal RF lesions over the nerves safely. Some have undertaken thoracic RF procedures but those procedures must be considered experimental. No standard method has been described for RF thoracic MBN and there is no evidence of the effectiveness of any such procedure.

Radiofrequency Lumbar Medial Branch Neurotomy—Procedure

Percutaneous thermal RF lumbar MBN is an established procedure for the treatment of lumbar ZJ pain. It is performed in a way similar to that described for RF cervical MBN, using large-diameter (18 or 16 gauge) RF electrodes placed parallel to, and very close to, each of the medial branches that supply the target joint. The treatment should be undertaken in accordance with the protocol laid down in the standard practice guidelines¹⁸² for maximal safety and effectiveness.

The RF lumbar MBN procedure involves three phases like those of the cervical treatment, a preoperative phase, an operative phase, and a postoperative, recovery phase. In the operative phase, the patient is positioned lying prone on the table and is made comfortable with suitable padding. Aseptic conditions are observed to ensure a sterile field. Local anesthesia is established and the block needle is left in place to mark the target zone. An RF electrode is introduced in an oblique direction from below and lateral to the target nerve, and aimed toward the target indicated by the block needle. The electrode is directed carefully through the muscles until its tip lies at the base of the superior articular process that the medial branch courses over, and on the posterior part of the adjacent transverse process, with the shaft of the electrode parallel to the nerve. The position and orientation of the active tip of the electrode is critical to the success of the procedure.¹⁸³ When the electrode is in place at the target site, its position there is checked and recorded on AP, oblique and lateral views, as shown in Figures 36-45 and 36-46.

Figure 36-45 Oblique fluoroscopic view showing an electrode in place at the junction of the superior articular process and transverse process for percutaneous radiofrequency lumbar medial branch neurotomy.

Figure 36-46 Lateral fluoroscopic view showing an electrode in place at the junction of the superior articular process and transverse process for percutaneous radiofrequency lumbar medial branch neurotomy.

With the electrode in the correct position, the patient should be advised that the nerve is about to be “cooked” and told to report any pain, feeling of heat or any other unusual sensation during the process. Then the generator is turned on to produce a thermal RF field that will coagulate the target nerve in a way similar to that described for cervical RF treatment. In the lumbar RF procedure only one pass of the electrode is required at each spinal level and two or three overlapping lesions over the course of each medial branch are enough to produce the desired therapeutic effect.

Radiofrequency Lumbar Medial Branch Neurotomy—Effectiveness

The therapeutic value of percutaneous RF lumbar MBN is controversial. The controversy, like that associated with cervical MBN, is related more to the diversity of RF procedures described in the literature than to the outcomes of valid lumbar MBN treatments. There are many more publications on lumbar RF procedures than on cervical ones, so the potential for confusion and controversy is much greater when lumbar treatments are being considered.

When properly indicated (by positive, controlled, comparative MBBs) and performed in the manner described in the standard practice guidelines,¹⁸² percutaneous RF lumbar MBN is an effective treatment for chronic lumbar ZJ pain. In most cases, thermal RF treatment will relieve the pain from the target joint completely and free the patient of any disability related to that pain. If the patient has any residual ZJ pain after the procedure, it will be from another source that may be treated similarly. The published evidence of the effectiveness of thermal RF lumbar MBN is compelling but that is clear only when it is separated from the large amounts of data relating to other RF procedures. These may seem similar superficially but they are essentially dissimilar if they are done to treat different conditions, and/or are based on different indications, and/or involve different techniques, and/or are aimed at different targets. The potential for confusion is considerable.

The effectiveness of RF lumbar MBN was demonstrated first by the report of a randomized, controlled clinical trial published in 1999,¹⁸⁴ which showed substantial benefits of the RF procedure over placebo. Although the pain relief was significantly greater in the treatment group (showing that coagulating nerves is better than a sham treatment), the results of the trial are compromised by two factors. First, it included patients on the basis of 50% or more relief of pain after single MBBs (so the study population would have been mixed) and it used a technique in which the electrodes were placed perpendicular to the target nerves, not parallel to them (so only short lengths of nerve were likely to have been coagulated and, in many cases, the relief was only of brief duration). The trial data count as evidence of the effectiveness of medial branch coagulation but do not count as evidence for the full RF lumbar MBN procedure as described in the standard practice guidelines.

Another controlled trial published in 2007,¹⁸⁵ used a valid RF technique with electrodes placed parallel to the medial branches. It again showed that conventional, thermal RF coagulation is more effective than placebo, and also more effective than a “pulsed RF” technique which was also tested in the study. The inclusion criteria for this trial were clinical features and a single positive MBB only, so again the study population would have been mixed. The thermal RFMBN outcomes of this study were better and lasted longer than those of the earlier controlled trial. These data again show the effectiveness of medial branch coagulation but because of the inclusion criteria, they do not count as evidence for the full RF lumbar MBN procedure.

A further randomized, controlled clinical trial of RF lumbar MBN published in 2008,¹⁵⁷ used positive, comparative MBBs as the diagnostic criterion and employed the RF technique recommended in the standard guidelines. It showed clinically (and statistically) significant relief of pain and improvements in quality-of-life indicators after the active treatment. These data are evidence of the effectiveness of the RF lumbar MBN procedure done in accordance with the standard guidelines. Another controlled clinical trial of RF lumbar MBN published in 2009,¹⁸⁶ used a valid, parallel electrode technique and produced further evidence of the effectiveness of the treatment but it used a combination of a positive MBB and a positive intraarticular block as the indication for the treatment, so in that respect it did not conform to the standard criteria.

The data from these four randomized, controlled trials are reinforced by the results of two descriptive outcome studies^156,187 of RF lumbar MBN using positive, comparative MBB results as the indication for the treatment and valid, parallel electrode techniques in the therapeutic procedure. Both these studies showed that RF lumbar MBN, when properly indicated and performed according to the standard guidelines, is an effective treatment for lumbar ZJ pain. These data are supported by the conclusions of two systematic reviews of the literature published in 2002¹⁸⁸ and 2009,¹⁸⁹ which both found evidence in favor of the effectiveness of the procedure.

This clear evidence of the effectiveness of RF lumbar MBN is overshadowed by the reports of several other studies and “systematic reviews” (including one review¹⁹⁰ within the framework of the Cochrane Collaboration) which seem to show that the weight of evidence is against the effectiveness of lumbar RF procedures. As intimated earlier, confusion about different types of RF treatment has lead to controversy. Both the confusion and the controversy are easily resolved by considering the different methods used in lumbar RF treatments with specific focus on their validity (i.e., whether those methods achieve what they are intended to achieve). In this regard, the significant elements of any lumbar RF procedure are (1) its target, (2) its indications and (3) how that target is treated. In evaluating the evidence of the effectiveness of RF lumbar MBN for the treatment of ZJ pain, it is important to consider all reports of the outcomes of procedures that (1) target the lumbar medial branches, (2) were indicated by positive, controlled, comparative MBBs and (3) used a valid technique involving multiple lesions made by thermal RF fields applied with 18 or 16 gauge electrodes correctly positioned parallel to the target nerves. The 2008 randomized, controlled trial¹⁵⁷ and the two descriptive studies^156,187 showing effective RF treatment met all those criteria. Significantly, there are no other published studies that meet those criteria and have unsatisfactory outcomes. Thus, the evidence of the effectiveness of RF lumbar MBN for the treatment of ZJ pain is incontrovertible.

The evidence on the effectiveness of lumbar RF treatment provides a good example of how controversies about interventions arise and how they are resolved. From a scientific point of view, skepticism is always to be encouraged and any doubts raised should be dispelled by considering all relevant scientific evidence. That is the rationale of evidence-based medicine. Current principles of evidence-based medicine flowed from the ideas of Archie Cochrane,¹⁹¹ a Scottish physician who practiced from the 1930s until the 1980s. Cochrane advocated the evaluation of medical interventions by consideration of the scientific evidence of their outcomes. His ideas lead to development of the systematic review, a process in which all relevant, scientific evidence is collated and analyzed to produce a summary of current knowledge of the topic, on which rational management decisions can be based so patients have the best treatment available. After his death, the Cochrane Collaboration, a movement to encourage systematic reviews for that purpose, was named in his honor.

The published literature on lumbar RF procedures includes numerous reports of case series, clinical trials, and systematic reviews showing various outcomes including total pain relief, partial pain relief, only short-term pain relief, and no relief after what are all described as “lumbar radiofrequency” treatments. None of the individual studies other than those described earlier^156,157,187 are of procedures performed in the manner recognized as the standard for RF lumbar MBN, so strictly speaking, they are irrelevant to it. These irrelevant publications include reports of procedures targeting the lumbar dorsal root ganglia rather than the medial branches, procedures using pulsed radiofrequency rather than thermal radiofrequency fields, and procedures done for various indications including partial relief of low back pain after single, uncontrolled intraarticular injections. Some clinicians fail to understand the difference between these studies and those of valid RF lumbar MBN, perhaps simply because the words “lumbar” and “radiofrequency” appear in their titles. The confusion is made worse by the publication of what purport to be systematic reviews, which include dissimilar studies as if they are comparable. The authors of such reviews are supposed to understand methodology and to be discriminating in selecting study reports for comparison, yet there are three so-called “systematic reviews”^190,192,193 of RF procedures for spinal pain that lump together the data of widely disparate interventions and draw unwarranted conclusions. These conclusions are essentially irrelevant to RF lumbar MBN but the confusion they engender results in some patients with chronic ZJ pain being denied the treatment they need for relief, contrary to what Archie Cochrane strove to achieve.

If anyone contended that apples and oranges are the same and then discussed fruit publicly as if the equality of apples and oranges is self-evident, their contention would be bound to cause controversy and the origin of the controversy would be obvious. When discussing spine interventions, including MBBs and RFMBN, only apples should be compared with apples. There are published standards for these procedures, standards set on the basis of sound, scientific evidence and agreed by clinicians experienced in their effective use. When the procedures are being evaluated for their validity and effectiveness, controversy can be avoided by considering only procedures performed in accordance with those standards, which are described fully in the ISIS practice guidelines.