INTRODUCTION

For more than 70 years, the sacroiliac joint complex has been implicated as a significant and discreet structural source of low back pain. However, of the three major groups of spinal structures identified as contributing to chronic low back pain, the sacroiliac joint complex (SIJC) stands apart from the zygapophyseal joints and intervertebral discs as the least understood from a neuroanatomic, diagnostic, and therapeutic standpoint. The prevalence of low back pain arising from the SIJC has been estimated at 15–30%,^1,2 and is likely higher in the older population. The SIJC possesses a unique and complex three-dimensional anatomy with multisegmental sensory innervation. Unlike the zygapophyseal joints and intervertebral discs, the sensory innervation of the SIJC has been described in a detailed fashion only recently, and the definitive gold standard for successful treatment of SIJC pain has not yet been rigorously defined.

Numerous studies have demonstrated a lack of pathognomonic historical, symptom, physical examination, or imaging findings specific for the individual diagnosis of zygapophyseal joint, disc, or SIJC pain.^1,3–7 The identification of specific structural sources of low back pain must currently be made through validated interventional diagnostic means. A structure-specific diagnosis is ultimately essential for the eventual consideration of appropriate evidence-supported treatment options. A specific structural diagnosis also leads to a definable prognosis, as the long-term outcomes following treatment for zygapophyseal joint pain differ from disc pain, and from SIJC pain as well.^8–14

A rational, evidence-based diagnostic algorithm for the identification of the structural source of low back pain is thus invaluable (Fig. 116.1). The area of the SIJC is frequently involved with referred pain patterns associated with other lumbar structures,^2,3,15–17 and pain of SIJC origin frequently involves somatic referred pain to various areas of the low back, pelvis, abdomen, and lower extremities,¹⁸ with nearly one in seven patients experiencing pain radiating to the foot in a pseudoradicular pattern. Successful implementation of such a diagnostic algorithm is therefore ultimately dependent not only on the physician’s understanding of the technical aspects of diagnostic interventions, but also the overlapping distributions of referred somatic pain arising from different lumbar structures, and the defined sensitivities and specificities of each test considered.

Fig. 116.1 Sample SIJ diagnostic algorithm flowchart depicts an evidence-supported interventional diagnostic algorithm for the identification and treatment of SIJC pain.

The current gold standard for the diagnosis of SIJC pain involves controlled intra-articular anesthetic injections (IAI).^1,2,6,7,19 Pain may also arise from the deep interosseous ligament (DIOL), which is exquisitely innervated. The nociceptive potential of the DIOL is supported by anatomic and histologic findings. Although a technique for injection of the DIOL has been described, the clinical overlap in diagnostic sensitivity and specificity between controlled SIJC IAI and DIOL anesthetic injections has yet to be determined,^10,20–22 and represents an ongoing area of clinical and anatomical research. Once the diagnosis of SIJC pain has been established and the presence of other sources of lumbar spine pain excluded, definitive SIJC therapeutic intervention may be considered.

A definitive, highly successful, and durable intervention for the treatment of SIJC pain may be just around the corner. Limited reports of the efficacy of surgical fusion^23,24 and radiofrequency denervation^10,25,26 are available; however, long-term prospective, controlled or comparative studies are currently lacking or are under investigation. This chapter will focus on current evidence supporting therapeutic injection and radiofrequency SIJC denervation.

RELEVANT ANATOMY

The sacroiliac joint (SIJ) is the largest syndesmotic joint in the human body. Although possessing a cartilaginous articular surface and synovial component, the sacroiliac joint primarily functions as a stress-relieving joint, insulating the lumbar spine from transmitted shock associated with ambulation. The SIJ is not bounded by capsular tissue like other synovial joints. The joint capsule is composed of ligamentous fibers, and posteriorly, contiguous with the deep interosseous ligament. The joint as a whole is intimately associated with a multitude of surrounding ligaments, which provide structural integrity to the joint complex.

The synovial component of the SIJC is composed of two divergent joint planes, a larger lateral (or anterior) pole and the smaller medial (or posterior) pole. Percutaneous access to the medial pole traverses a minimum of ligamentous tissue, whereas access to the lateral pole must traverse the DIOL (Fig. 116.2). The total volume of the synovial component of the SIJC (1.5 mL) is small compared to its surface area.^15,27

Fig. 116.2 Medial and lateral poles, SIJ. An ipsilateral oblique image of SIJ arthrography, with needle penetrating the posterior joint capsule of the medial pole. The heavy dark line separates the medial from the lateral pole of the joint. Note capsular redundancy of the distal (caudal) medial pole (white arrow), and the small amount of dorsal capsular contrast extravasation (black arrow).

The ‘capsule’ of the sacroiliac joint is frequently incompetent, even in asymptomatic patients. In one series, 61% of intra-articular injections demonstrated extracapsular extravasation,²² and 25% of asymptomatic volunteers undergoing intra-articular injection demonstrated ventral capsular insufficiencies,²⁷ with contrast extravasating in the region of the traversing lumbosacral plexus. Detailed dissection of the ventral joint capsule in cadavers suggests that these ventral capsular ‘defects’ often appear as small foramina rather than traumatic capsular rents.²⁸

Neuroanatomy of the sacroiliac joint complex

Early studies of the sacroiliac joint suggested a combination of ventral and dorsal innervation,^29,30 but recent investigation has demonstrated a predominant dorsal innervation in humans^21,22 arising from the lateral branches of the L5 dorsal ramus and S1–4 dorsal rami, and composed of a wide range of sensory fiber types.^21,30 The lateral branch nerves arise from a lateral dorsal sacral plexus, and divide into multiple smaller branches, some of which enter the DIOL whereas others do not. The segmental location and number of lateral branch nerves innervating the SIJC is extremely variable. On the posterior aspect of the sacrum, these branches are often closely related (Fig. 116.3).²⁰ Unlike the lumbar medial branch nerves of the dorsal rami supplying innervation to the zygapophyseal joints, the complex three-dimensional anatomy of the lateral branch nerves innervating the SIJC combined with their close proximity to the dorsal sacral foramina renders consideration of selective anesthetic blockade futile (Fig. 116.4).²⁷

Fig. 116.3 Dorsal sacral plexus and lateral branch nerves. A representative illustration of the lateral branch nerves supplying sensation to the dorsal sacral joint complex (arrows) from L5 through S3. Note the variability in lateral branch topography between right and left. LDSIL, long dorsal sacroiliac ligament; PSIS, posterior superior iliac spine; ST, sacral tubercle; L5–S1, L5–S1 zygapophyseal joint (superior articular process).

(Courtesy of Frank Willard, PhD.)

Fig. 116.4 Incomplete lateral branch staining. Dissection in nonembalmed cadaver of lateral branch nerves, left dorsal sacral plexus following injection of 0.5 mL of colored dye lateral to S1 and S2 dorsal sacral foramina. Note incomplete staining of lateral branch nerves at S2 (black arrows). Unlike similar diagnostic injection techniques to identify painful zygapophyseal joints, it is not possible to selectively anesthetize the lateral branch nerves supplying the sacroiliac joint.

(Courtesy of Paul Dreyfuss, M.D. and Frank Willard, PhD, used with permission.)

The DIOL is richly innervated with nociceptive mechanoreceptors. Sensory fibers must traverse the DIOL to reach the posterior SIJ capsule, and understanding the topographic neuroanatomy of the SIJC becomes invaluable in consideration of the design and performance of diagnostic injection techniques and therapeutic interventions.

DIAGNOSTIC INJECTION TECHNIQUES

With the exception of imaging findings on magnetic resonance imaging (MRI) characteristic of sacroiliitis (and in the absence of tumor, fracture, or infection), there are no pathognomonic imaging findings specific for the diagnosis of SIJC pain. As previously discussed, there are also no physical examination findings that accurately identify patients with SIJC pain, although a combination of pain below L5 and localized just medial to the posterior superior iliac spine (PSIS) is suggestive. The diagnosis of SIJC pain therefore rests on a combination of interventional tests to exclude other lumbar sources of pain, and objectively verify the presence of SIJC pain.

Once other structures of low back pain have been excluded, consideration of SIJC-specific diagnostic interventions may be entertained. The current gold standard involves blinded, controlled intra-articular local anesthetic injections of the SIJC, or saline placebo-controlled blocks. Because the synovial component of the SIJC is rarely accessed successfully without image guidance, injection of the joint must be performed under fluoroscopy, or possibly computed tomography (CT) guidance.^31,32 However, it must be remembered that the sacroiliac joint capsule is frequently incompetent, and extravasation of local anesthetic into surrounding structures, especially from the ventral joint capsule in the region of the traversing lumbosacral plexus, will render such an injection non-specific, and therefore nondiagnostic. Additionally, since the false-positive rate of single anesthetic injections may approach 40%,^33–36 no therapeutic decisions regarding ablation or surgery should be made based on a single response to an isolated analgesic test. Patient responses to controlled, comparative or placebo-controlled anesthetic testing must be rigorously assessed and documented. As with any diagnostic procedure, analgesic tests generate negative, indeterminate, and positive results, and have associated false-negative and false-positive rates. Understanding what constitutes a negative, indeterminate, and positive result is essential. The methodology of comparative and placebo-controlled analgesic testing has been extensively validated elsewhere, and should be considered mandatory reading for all interventional spine diagnosticians.^{34,35,37–39} It is further recommended that any analgesic test be performed without sedation. Patient responses in the immediate postprocedural period should be rigorously documented, preferably with a q15 minute or q30 minute pain log. Sedation is an inadequate excuse for poor injection technique.

Access techniques for the medial^40–44 and lateral poles^10,45 of the synovial component of the SIJC have been described, as has a technique for isolated injection of the DIOL.^10,45 Any image-guided injection of the SIJC must include the use of radiopaque contrast to obviate vascular uptake, and verify contained distribution of injected local anesthetic. Fluoroscopy offers several intrinsic advantages over CT, including the capability for real-time imaging (and detection of vascular uptake of contrast), although the superior contrast and spatial resolution of CT may offer potential advantages in documenting extracapsular or extraligamentous extravasation patterns. In patients with spinal or pelvic hardware, advanced fluoroscopic imaging technologies, such as digital subtraction algorithm (DSA) imaging permits imaging free of the metallic imaging artifacts common to CT.

THERAPEUTIC INTERVENTIONS

Injections of the SIJC have not been demonstrated to have therapeutic value except in the isolated instance of sacroiliitis associated with seronegative spondyloarthropathies.^46–50 No structure-specific diagnostic information can be gleaned from steroid injection; if a patient experiences relief beyond the duration of action of local anesthetic, all that can be ascertained is that a steroid response may be present. A prolonged steroid response implies that a component of inflammation may be present, but is not structure-specific; local anesthetic injection in and of itself may also provide prolonged analgesia, well beyond the expected duration of conduction block.³⁷ Once the diagnosis of SIJC pain has been established, therapeutic intervention may be considered.

Therapeutic options for SIJC pain may be categorized into surgical interventions, other interventions, and denervation procedures. Surgical intervention for SIJC pain is covered elsewhere in this text. Some religiously advocate other therapeutic options, such as manipulation and proliferative therapies. Regrettably, the efficacy of manual medicine techniques and proliferative treatments remain anecdotal.

Several denervation procedures for SIJC pain have described targeting the lateral branch nerves of the dorsal rami at various points between the dorsal sacral foramina and the posterior SIJ capsule. Among the first techniques is that described by Kline, involving the generation of a series of bipolar radiofrequency (RF) lesions along the posterior joint line (Fig. 116.5).⁵¹ Recognizing the multisegmental innervation of the SIJC, Kline reasoned that the lateral branch nerves must converge along the posterior joint capsule, and serial lesions along the posterior joint line should effectively denervate the joint. However, a study by Ferrante et al., utilizing uncontrolled single anesthetic diagnostic blocks, found that only 36% of patients experienced more than 6 months of >60% relief, and no patients experienced complete relief.²⁶ A more rigorous evaluation with controlled anesthetic blocks and rigorous postblock assessment may have yielded better results, as initial false-positive block responders would have been eliminated from consideration.

Fig. 116.5 Bipolar SIJ RF lesioning technique. Cone down anteroposterior radiograph of bipolar electrode placement over caudal dorsal sacroiliac joint capsule.

(Courtesy of Matthew T. Kline, M.D.)

Another approach toward denervation of lateral branch nerves at L5 and S1 has been described under CT guidance by Gevargez et al.²⁵ Like the Ferrante study, a single anesthetic block was used for patient selection. Thirty-five of 38 patients were followed for 3 months, with 34.2% reporting complete relief, and an additional 32% reporting ‘substantial relief.’

Based on a detailed neuroanatomic study of the dorsal sacral plexus and lateral branch topography, Yin et al. published results of a pilot study utilizing a sensory stimulation-guided approach to select pain-conducting from non pain-conducting lateral branch nerves for subsequent RF denervation.¹⁰ Dual analgesic injections of the SIJC DIOL were utilized to identify patients with probable SIJC pain after no relief was reported following anesthetic blocks of the lumbar zygapophyseal joints. In this study, provocative discography was not performed to rule out disc pain. Sensory stimulation-guided localization of the lateral branch nerves was performed to identify painful versus nonpainful nerves. The frequency of painful L5 and S1 lateral branch nerves identified in this group of patients was 100%, with 78% of patients demonstrating a painful branch arising from S2 and 42% with painful S3 branches. Although follow-up was limited to 6 months, preliminary results were encouraging, with 64% of patients experiencing longer than 6 months of >60% relief, and among responders, more than half (36% of total) experiencing 100% relief. No complications were noted. A prospective, long-term follow-up study is currently in progress. Ultimately, prospective, controlled studies will be required to identify a gold standard for the treatment of SIJC pain.

DESCRIPTION OF SENSORY STIMULATION-GUIDED LATERAL BRANCH RADIOFREQUENCY NEUROTOMY TECHNIQUE

Background

Detailed neuroanatomic examination of the dorsal sacral plexus and associated lateral branch nerves innervating the SIJC demonstrates that these nerves follow a complex and variable course to the DIOL and posterior joint capsule. Penetrating numerous ligaments along their course, the segmental lateral branch nerves are inconsistent with regard to number and location in three dimensions soon after exiting the dorsal sacral foramina (Fig. 116.6).^10,28 Because consistent lateral branch contributions to the SIJC have been demonstrated to arise from the S2 and S3 dorsal rami, procedures that only target the lateral branch nerves at L5 and S1 may miss pain-bearing conducting from these lower dorsal rami. A description of the sensory stimulation-guided radiofrequency neurotomy targeting the L5 through S3 lateral branch nerves follows.

Fig. 116.6 Wire study of LB topography. Thin gauge wires have been placed directly overlying lateral branch nerves seen entering the dorsal sacroiliac deep interosseous ligament arising from the dorsal sacral foramina of S1 in this cadaver. Heavier gauge wire has been placed at the dorsal margin of the dorsal sacral foramina. Note the variability or number and location of lateral branch nerves entering the dorsal SIJC.

Procedural details

The patient is brought to the operating theater and placed in the prone position on a radiolucent operating room table. If unilateral, the side of the procedure is verified with the patient. Anesthetic monitors are applied, and supplemental oxygen is provided, if needed. Although the requirement for intraoperative analgesia or anesthesia is exceedingly rare, the presence of a qualified anesthesia provider is invaluable in the event of a significant vasovagal episode or other unexpected intraoperative occurrence.

The patient’s back and buttocks are sterilized with an antiseptic skin preparation (e.g. Betadine) and draped in a sterile fashion. Meticulous aseptic technique is observed. Fluoroscopic imaging is utilized to visualize the sacrum in real time, and the fluoroscope is angled parallel to the ring apophysis of S1. A skin entry site is marked and anesthetized with 1% lidocaine just medial to the posterior superior iliac spine on the operative side. To facilitate subsequent introduction of a blunt radiofrequency electrode, a 1.25, 16-gauge intravenous catheter is introduced percutaneously, through which a 100 mm, 5 or 10 mm active tip, blunt, curved radiofrequency electrode may be advanced. Through this single entry site, the lateral branch nerves from L5 through S3 may be accessed.

Initially, the RF electrode is advanced under fluoroscopic imaging to overlie the superior-medial aspect of the sacral ala, in the region of the medial dorsal ramus of L5. It is often helpful to provide the patient with an initial stimulation sensation that is expected to be nonpainful (and if L5–S1 zygapophyseal joint pain has been adequately excluded, stimulation over the medial dorsal ramus should not result in a painful response). Localization of the medial dorsal ramus of L5 is then achieved with application of stimulation (50 Hz, 1 msec pulse duration) at an initial ‘seeking’ voltage of 0.2–0.4 volts. The L5 medial dorsal ramus has an occasional motor component to the multifidus, but gross tetanic contraction of the multifidus at this level is often not seen, especially in older patients with multifidus atrophy. A strong sensory response to patient-blinded stimulation indicates successful localization of the L5 medial dorsal ramus. Nonpainful responses to stimulation include the sensation of ‘buzzing,’ ‘tingling,’ ‘vibration,’ or ‘pulsing.’ Occasionally patients will describe the sensation as ‘thumping.’ These nonpainful sensations are nearly always clearly differentiated from pain. Painful stimulation is most commonly described as sharp,’ ‘burning,’ ‘aching,’ or ‘stabbing.’ Stimulation threshold is then decreased to a minimum voltage necessary to elicit a sensation, as excessive stimulation of a nonpain-bearing nerve may be incorrectly perceived as painful. Because the proximity of the electrode to the target nerve is inversely (and nonlinearly) proportional to the voltage (or, more accurately, current) required for a sensory response, finely manipulating the electrode to achieve the lowest stimulation threshold will maximize the juxtaposition of electrode to target nerve. Ultimately, this close apposition of electrode to nerve is important for RF lesioning, as the size of an RF lesion is relatively small.

To minimize the possibility of a false-positive or false-negative response to stimulation, patient-blinded stimulation, including faux stimulation is performed at the minimum threshold stimulation voltage (ideally <0.2 volts). If patient responses to stimulation are absolutely reproducible, the particular nerve under examination may be considered ‘negative’ if nonpainful stimulation is elicited, and ‘positive’ if painful stimulation is reported.

Once the L5 dorsal ramus has been eliminated as a potential source of pain, the lateral branch is localized. The lateral branch of L5 typically lies 5–10 mm lateral to the lateral aspect of the superior articular process of S1 (Fig. 116.7A). The electrode is repositioned along the dorsal sacral ala as described, and stimulation repeated. The electrode is manipulated along the dorsal sacral ala until the lateral branch is identified. The lateral branch of the dorsal ramus of L5 does not have a known motor component, and localization depends solely on sensory findings.

Fig. 116.7 (A) Typical L5 lateral branch electrode placement. Left SIJC denervation. (B) Typical S1 lateral branch electrode placement. Left SIJC denervation. (C) Typical S2 lateral branch electrode placement. Right SIJC denervation. (D) Typical S3 lateral branch electrode placement. Left SIJC denervation

Once a potential symptomatic branch has been identified, patient-blinded stimulation (with faux stimulation) is performed at the minimum stimulating voltage to determine whether patient responses are consistent. If consistent reproduction of pain is clearly distinguishable from nonpainful stimulation, a radiofrequency thermal lesion is created. Whereas maximum electrical stimulation occurs just proximal to the exposed end of the electrode, the maximum concentration of heat generated with ionic monopolar RF lesioning occurs further proximally along the exposed electrode tip. For this reason, a single thermal lesion at the site of maximum stimulation may not adequately coagulate the target nerve, especially if high-impedance surrounding tissues are present. For this reason, several sequential lesions overlapping the 45° isotherm of the previous lesion are recommended.

Prior to lesion generation, 0.5–1.0 mL of local anesthetic (e.g. 0.5–0.75% bupivacaine) is injected. As the lateral branch nerves are closely invested within a labyrinthine network of ligaments surrounding the dorsal sacral plexus, injected local anesthetic may not be directed in the area desired; the electrode may have to be rotated in order to achieve diffusion of sufficient local analgesia sufficient to permit painless coagulation. Care should be taken to return the electrode to the exact position where stimulation resulted in pain reproduction prior to lesion generation. The author creates three overlapping 90° C lesions, each of 60 seconds duration, on each symptomatic branch identified. Lesion parameters, including initial temperature and impedance, volts, watts, and milliamperes, should be recorded. Any complaints of pain, especially that which radiates down the lower extremity, should result in the immediate cessation of lesion generation and confirmation of appropriate electrode placement well dorsal to the segmental spinal nerves.

Once the L5 lateral branch has been located and mapped, attention is directed towards mapping of the lateral branch nerves arising from the dorsal ramus of S1. As the S1 dorsal ramus supplies the most lateral branch nerves going to the SIJC, care should be taken to map this level carefully. More than one symptomatic branch may be encountered. Because the lateral branch nerves arising from the dorsal sacral plexus are invariably located lateral the cephalocaudad meridian of the foramen, mapping efforts are systematically dedicated to this area (i.e. if the dorsal foramina was viewed as a clock face, between the 12 o’clock and 6 o’clock position) (see Fig. 116.7B). The lateral margin of the sacral foramina is frequently difficult to visualize on fluoroscopy, but can be easily identified by feel as the electrode tip ‘slips over’ the edge of the foramen. Stimulation of the lateral branch nerves or communicating arcade can result in referral of pain to the lower extremity, groin, or abdomen. It is imperative that such sensation be differentiated from stimulation of the sacral spinal nerves themselves, which may occur if the electrode is inadvertently advanced through the dorsal foramina. If the depth of electrode placement is a concern, nonionic radiopaque contrast may be injected through the electrode and oblique and lateral fluoroscopic imaging used to verify placement of the electrode along the dorsal sacral plate. Any symptomatic branches identified at S1 are lesioned as previously described for the dorsal ramus of L5.

In a sequential fashion, the lateral branch nerves of S2 and S3 are mapped. Symptomatic branches are coagulated as described (see Fig. 116.7C,D). At the conclusion of the procedure, the electrodes and introducer catheter are removed.

Anecdotally, the author has found it useful to administer a dose of high-potency antiinflammatory corticosteroid at the conclusion of the procedure, which appears to decrease the frequency of early postoperative dysesthesiae. The author follows the practice of Dr. M. T. Kline, who uses a single 10 mg intravenous dose of dexamethasone.

The patient’s back and buttocks are cleansed of antiseptic and antibiotic ointment dabbed over the skin puncture site(s). A small adhesive dressing is applied, and the patient discharged to the recovery room. Prior to discharge, most patients experience minimal to no discomfort. A neurological examination is documented prior to discharge.

REPRESENTATIVE CASE

RW, a 55-year-old male, presents with long-standing complaints of right low back and posterior hip pain. He is 20 years status post several lumbar surgeries for low back pain, including two laminectomies, and a noninstrumented posterior fusion. None of his surgeries improved his back pain, and 10 years ago he underwent an instrumented pedicle screw fusion at L5–S1, which he reports markedly worsened his symptoms. He occasionally experiences burning pain that radiates over the hamstring region, usually stopping in the popliteal region, with occasional complaints in the right groin without testicular radiation. His symptoms are worsened with static sitting, prolonged ambulation, and lumbar extension. Typically, his symptoms are severe upon first awakening, improve slightly during the late morning hours, and worsen towards the end of the day proportional to activity. His pain awakens him 3–5 times a night. He rates his pain at 85 mm on a 100 mm visual analog scale (VAS).

While living in England over the past 8 years, he underwent a total of 45 caudal epidural steroid injections. Some provided him with incomplete and temporary relief, typically lasting 1–2 weeks. Since relocating to the United States, his family practitioner has prescribed a number of oral neuromodulator agents, muscle relaxants, nonsteroidal antiinflammatory agents, and mild opioid analgesics that have not been helpful. He has undergone numerous therapies over the past decade, including many rounds of physical, massage, acupuncture, and homeopathic therapies, as well as chiropractic and osteopathic manipulations. None of these modalities provided relief. He has been told by a pain psychologist that he must learn to ‘cope with his pain,’ but relaxation, cognitive, biofeedback, and hypnosis techniques have not been successful. Prior urological evaluation of his right groin symptoms demonstrated no abnormalities or hernia. Review of supplied past medical records demonstrates consistent symptom complaints. He works on a full-time basis as a successful oil industry consultant, is married, and has three grown children. He does not smoke cigarettes, but enjoys an occasional cigar. He takes an antihypertensive agent and a cholesterol lowering drug.

On examination, the patient is appropriate, well dressed, and exerts excellent effort on physical examination requests; he demonstrates no symptom exaggeration or magnification. His constitutional examination is unremarkable; he is of average height and build. His neurological examination demonstrates no abnormalities, with the exception of a mild decrease in pinprick sensation in the right L5 distribution, extending to the dorsum of the foot. Motor and reflex examinations are normal. Dural traction signs in the lower extremities are absent. Lower extremity vascular examination is normal. His gait is slightly antalgic to the right, with a minimal amount of pelvic tilt, but tandem heel and toe walking are performed without difficulty. A well-healed midline scar is present between L3 and S1. A well-healed scar consistent with a prior left posterior iliac crest bone graft is present. The patient points to an area just medial to the right PSIS as his primary area of focal discomfort.

His lumbar range of motion is decreased in extension, and limited by pain. Forward flexion and thoracolumbar lateral rotation are performed without discomfort, but lateral flexion, especially to the right is painful. On examination by palpation, increased paravertebral muscle tone is present symmetrically in the lumbar spine. There is mild tenderness to palpation in the dorsal midline and paravertebral regions symmetrically at the lumbosacral junction. There is exquisite tenderness to palpation over the dorsal sacroiliac ligaments on the right. No appreciable tenderness is present over the trochanteric bursa, and there is no intrinsic hip joint pain on examination.

A seven-view plain film radiographic series of the lumbar spine demonstrates the presence of an instrumented fusion at L5–S1 with pedicle screws and bridging instrumentation. The instrumentation appears intact, and there is no evidence of foraminal intrusion of the pedicle screws. A previous intertransverse bony fusion is evident, with more bony incorporation on the right than left. Bilateral L4 and L5 laminectomy defects are present. Intervertebral disc height is reduced at L4–5 (40%) and at L5–S1 (50%), with moderate anterolateral spondylosis at both levels. Flexion and extension images demonstrate no segmental instability. Marked L4–5 hypertrophic zygapophyseal joint (Z-joint) degenerative changes are seen; the L5–S1 Z-joints appear arthrodesed. A bony defect is present, involving the left posterior ilium. Limited imaging of the sacroiliac joints demonstrates mild degenerative changes, predominantly involving the posterior medial poles on both sides.

MRI of the lumbar spine with and without gadolinium enhancement is significant for decreased disc hydration at all lumbar levels, particularly L4–5 and L5–S1. Broad-based disc bulges are present at L3–4, L4–5 and L5–S1. At L4–5, moderate bilateral recess stenosis is present due to a combination of Z-joint hypertrophy and disc bulge, but no clear evidence of central stenosis is present at any level. Sagittal and axial foraminal imaging at L4–5 and L5–S1 is compromised by metallic imaging artifact. Contrast enhanced images demonstrate the present of epidural fibrosis surrounding the thecal sac at L4–5 and L5–S1, circumferentially surrounding the L5 spinal nerves in the lateral recess posterior to the L4–5 disc.

Bone scan with SPECT demonstrates no areas of intense focal uptake, although mildly increased uptake is seen in the region of the L4–5 and L5–S1 intervertebral discs, and involving the right acromioclavicular joint.

The patient undergoes intra-articular right-sided L4–5 anesthetic injection (0.75% bupivacaine). Z-joint arthrography demonstrates significant intra-articular sclerotic changes, but the joint is competent. The L5–S1 Z-joints are arthrosed, and cannot be accessed. The patient reports no improvement in pain following L4–5 Z-joint anesthetic injection.

A right L5 selective spinal nerve local anesthetic injection is attempted via a posterolateral approach, but is not technically feasible, due to previous intertransverse bony fusion. Right S1 selective spinal nerve block is performed, with transforaminal epidural steroid injection. Despite the development of excellent S1 sensory hypoesthesia immediately following injection, the patient notes no anesthetic phase improvement in his pain. For 1 week after injection, he notes that his pain is ‘40%’ better, but rapidly returns to baseline.

Provocative rate, volume, and manometrically controlled lumbar discography is then performed at the L2–3 through L5–S1 levels. A posterior transdural midline approach is necessary at L5–S1 due to previous posterolateral fusion. Although significant degenerative changes are present at all levels, including frank annular incompetence at L4–5 and L5–S1, provocative discography is painless at all studied levels.

Right sacroiliac joint arthrography is successfully performed through a trans-DIOL approach. There is moderate joint sclerosis on arthrography, and a ventral capsular defect is present involving the lateral joint pole. No intra-articular anesthetic is injected. DIOL contrast study demonstrates contained contrast spread, and 2.5 mL of local anesthetic (0.75% bupivacaine) is injected. Post procedure, the patient is observed for 2 hours. The patient notes complete (100%) relief of his pain, despite attempts to exacerbate his pain with prolonged walking, static sitting on a hard chair, and lifting. He is enrolled in a saline placebo-controlled study of the right SIJC DIOL.

Following blinded, placebo-controlled injection of the right SIJC DIOL, the patient’s postprocedural analgesic diaries are reviewed. No improvement following saline injection is demonstrated, whereas the patient documents 5 hours of complete relief following bupivacaine injection, with an additional 1.5 hours of near complete (80%) relief, before rapid return to baseline pain. With the diagnosis of right SIJC pain established to the maximum specificity available with validated testing methods, SIJC sensory stimulation-guided RF denervation is recommended. He is told that there is a 65% chance that his symptoms may be reduced by 60%, and an overall 35% chance that his pain may be completely relieved.

Intraoperatively, due to previous fusion, the L5 medial PPR and lateral branch cannot be identified. However, two separate painful lateral branches are identified arising from S1, and a single branch identified at S2. No painful branches are identified arising from S3. Multiple nonpainful branches are found with stimulation at the S1–3 levels. RF thermocoagulation is performed over the identified painful lateral branch nerves. The patient tolerates the procedure well and is discharged with a normal lower extremity neurological examination.

Nursing follow-up telephone calls on postoperative day one and seven are made. The patient complains of mild postprocedural discomfort, but was able to return to work the day following surgery. He notes no buttock dysesthesiae or hypoesthesiae.

At 4-week postoperative clinic follow-up, the patient notes that his procedural discomfort has completely resolved. He complains of no buttock numbness or dysesthesiae. He notes that his posterior hip pain is ‘90% better,’ that he is able to sleep through the night without interruption, and that he no longer has pain with sitting or walking. He notes that his posterior thigh burning pain has completely resolved, as has his groin pain. He is walking 2–3 miles a day, and has started to play golf again.

At 8-week postoperative follow-up, the patient notes that his primary presenting pain is ‘100%’ resolved. His low back VAS pain score is 0/100. His neurological examination is normal, including resolution of the mild pinprick deficit in the right L5 distribution. He now primarily complains of right anterior shoulder pain after playing golf.

This patient is now 2 years status post SIJ RF at the time of this writing, and remains pain free. Intercurrently, he has been referred to an orthopedic sports medicine specialist who is contemplating right shoulder acromioplasty.