Clinical input has supported the use of controlled diagnostic blocks with at least 75% pain reduction for diagnosis of sacroiliac pain.  Clinical input supported the use of corticosteriods for the treatment of SIJ pain.  Based on clinical input and the established use of injections to diagnose and treat pain in other joints, controlled diagnostic (2 blocks with anesthetics of different duration) and therapeutic (corticosteroid) injections may be considered medically necessary for the diagnosis and treatment of SIJ pain.

Treatment of SIJ pain with corticosteriods, radiofrequency ablation (RFA), stabilization or minimally invasive SIJ fusion has also been explored.  Sacroiliac joint (SIJ) arthrography using fluoroscopic guidance with injection of an anesthetic has been explored as a diagnostic test for SIJ pain.  Duplication of the patient's pain pattern with the injection of contrast medium suggests a sacroiliac etiology, as does relief of chronic back pain iwth injection of local anesthetic.

For individuals who have SIJ pain who receive RFA, the evidence includes 4 small RCT's using different techniques of applying radiofrequency and case series.  Relevant outcomes are symptoms, functional outcomes, quality of life, medication use, and treatment related morbidity.  For RFA with a cooled probe, the two small RCT's reported short-term benefits, but these are insufficient to determine the overall effect on health outcomes.  The RCT on palisade RFA of the sacroiliac joing did not include a sham control.  Another sham-controlled RCT showed no benefit of RFA.  Further high-quality controlled trials are needed that compare this procedure in defined populations with sham control and with alternative treatments.  The evidence is insufficient to determine the effects of the technology on health outcomes.

A number of radiofrequency generators and probes have been cleared for marketing through the U.S. Food and Drug Administration’s (FDA) 510(k) process. One device, the SInergy® by Halyard (formerly Kimberly-Clark), is a water-cooled single-use probe that received FDA clearance in 2005, listing the Baylis Pain Management Probe as a predicate device. The intended use is in conjunction with a radiofrequency generator to create radiofrequency lesions in nervous tissue. FDA product code: GXD.

Several percutaneous or minimally invasive fixation/fusion devices have received marketing clearance by FDA. These include the SI-FIX Sacroiliac Joint Fusion System (Medtronic), the IFUSE® Implant System (SI Bone), the SImmetry® Sacroiliac Joint Fusion System (Zyga Technologies), Silex™ Sacroiliac Joint Fusion System (X-spine Systems) and the SI-LOK® Sacroiliac Joint Fixation System (Globus Medical). FDA Product Code: OUR.

This policy was created in 2000 and has since been periodically updated with searches of the MEDLINE database. The most recent literature review was performed through March 25, 2015. Following is a summary of key references to date.

Diagnosis

The use of diagnostic blocks to evaluate sacroiliac joint pain builds on the experience of use of diagnostic blocks in other joints to evaluate pain. Blinded studies with placebo controls (although difficult to conduct when dealing with invasive procedures) are ideally required for scientific validation of sacroiliac joint blocks, particularly when dealing with pain relief well-known to respond to placebo controls. In the typical evaluation of a diagnostic test, the results of sacroiliac diagnostic block would then be compared with a criterion standard. However, no current criterion standard for sacroiliac joint injection exists. In fact, some authors have positioned sacroiliac joint injection as the criterion standard against which other diagnostic tests and physical exam may be measured.1 Finally, one would like to know how the results of a diagnostic test will be used in the management of the patient and whether the subsequent treatment plan results in beneficial health outcomes.

At the time this policy was created, there was minimal literature regarding sacroiliac joint blocks. For example, Schwarzer et al reported on a case series of 43 patients with unexplained low back pain below L5-S1.2 These 43 patients were chosen opportunistically from a larger group of patients referred for discography or zygapophyseal joint blocks. Maigne et al reported on a series of 54 patients with low back pain who received double sacroiliac joint block.3 Similar to the Schwarzer et al study, this study was primarily designed to demonstrate that sacroiliac pain exists and to assess its prevalence. No studies were identified that described how the results of sacroiliac joint arthrography might be used in the management of the patient.

The 2009 practice guidelines from the American Pain Society (APS) were based on a systematic review that was commissioned by APS and conducted at the Oregon Evidence-based Practice Center.4,5 The systematic review concluded that no reliable evidence existed to evaluate validity or utility of diagnostic sacroiliac joint block as a diagnostic procedure for low back pain with or without radiculopathy, with a resulting guideline recommendation of insufficient evidence. Data on sacroiliac joint steroid injection were limited to 1 small controlled trial, resulting in a recommendation of insufficient evidence for therapeutic injection of this joint. In 2010, Manchikanti et al published critical reviews of the APS guidelines for interventional techniques, including sacroiliac injections.6,7 Evidence for diagnostic sacroiliac injections was considered to be fair to poor, and no additional literature was identified since a 2009 systematic review by Rupert et al.8

In 2013, the American Society of Interventional Pain Physicians (ASIPP) published an updated evidence review and guidelines.8-12 Seven studies met the inclusion criteria of 75% to 100% relief with dual blocks. The prevalence of sacroiliac joint pain ranged from 10% to 44.4% with false-positive rates ranging from 20% to 26%. The evidence for diagnostic sacroiliac intra-articular injections was considered to be good, with 75% to 100% pain relief as criterion standard with controlled local anesthetic or placebo blocks.

Treatment

Hansen et al published an updated systematic review of sacroiliac joint interventions in 2012.13 The primary outcome was short-term (≤6 months) or long-term (>6 months) pain relief. Evidence was classified as good, fair, or limited/poor based on the quality of evidence. A total of 11 studies (6 randomized, 5 nonrandomized trials) met inclusion criteria. Review found that evidence for intra-articular steroid injections is limited/poor, as is the evidence for periarticular injections (local anesthetic and steroid or botulinum toxin). For radiofrequency neurotomy, the evidence for cooled radiofrequency was found to be fair (2 RCTs), while evidence for conventional radiofrequency or pulsed radiofrequency was limited/poor. The 2013 ASIPP evidence review found no additional studies on intra-articular or periarticular injections besides those identified by Hansen et al in 2012.9

Therapeutic Sacroiliac Injections

The available literature on therapeutic sacroiliac injections is limited, consisting of 1 small RCT that compared intra-articular injection with physical therapy or manual therapy, 1 small RCT that compared steroid injections with prolotherapy, and case series.13 There are no RCTs that compare therapeutic sacroiliac injections with placebo to determine the efficacy above placebo.

A 2013 study randomized 51 patients with sacroiliac joint and leg pain to physiotherapy, manual therapy, or intra-articular injection of corticosteroid.14 Diagnosis of sacroiliac joint pain was based on provocation tests and not sacroiliac joint injections. In a blinded assessment, 25 patients (56%) were considered to be successfully treated at the 12-week follow-up visit based on complete relief of pain and improvement in the visual analog score (VAS) for pain. Physical therapy was successful in 20%, manual therapy in 72%, and intra-articular injection in 50%.

Kim et al reported a randomized double-blind, controlled trial of intra-articular prolotherapy (see policy No. 2.01.26) compared with steroid injection for sacroiliac joint pain in 2010.15 The study included 48 patients with sacroiliac joint pain, confirmed by 50% or greater improvement in response to a single local anesthetic block, who had failed medical treatment. Intra-articular dextrose water prolotherapy or steroid injections were administered under fluoroscopic guidance on a biweekly schedule, with a maximum of 3 injections. Injections were stopped when pain relief was 90% or greater, which required a mean of 2.7 prolotherapy injections and 1.5 steroid injections. Pain (numeric rating scale) and disability scores (Oswestry Disability Index [ODI]) were assessed at baseline, 2 weeks, and monthly after completion of treatment. At 2-week follow-up, pain and disability scores were significantly improved in both groups, with no significant difference between the groups. Pain on the numerical rating scale improved from 6.3 to 1.4 in the prolotherapy group and from 6.7 to 1.9 in the steroid group. At 6 months after treatment, 63.6% of patients in the prolotherapy group remained improved from baseline (≥50%), compared with 27.2% in the steroid group. At 15-month follow-up, the cumulative incidence of sustained pain relief was 58.7% in the prolotherapy group compared with 10.2% in the steroid group. The median duration of survival (recurrence of severe sacroiliac joint pain) was 3 months for the steroid group.

Results from these small trials are insufficient to permit conclusions on the effect of this procedure on health outcomes. Comparisons with placebo, ideally using sham injections, are needed to determine the degree of benefit over placebo.

In 2007 Weksler et al reported results of diagnostic/therapeutic blocks in a series of patients who were referred for low back pain and disc herniation without claudication or neurologic abnormalities.16 Fifty patients who had at least 3 positive pain provocation tests for sacroiliac joint dysfunction received sacroiliac injection of bupivacaine and betamethasone. Pain, assessed by VAS, improved from 7.8 to 1.3 at 30 minutes after the injection. At a 12-week follow-up, 46 patients (92%) reported VAS scores of 3 or less. Four patients required hospitalization for an unanticipated motor block.

Questions also remain about intra-articular versus periarticular sources of sacroiliac pain. For example, 1 prospective comparison found that periarticular lidocaine injections (25/25 patients) were more effective than intra-articular injection (9/25 patients).17

Radiofrequency Denervation

The literature on radiofrequency denervation of the sacroiliac joint is limited. Two small RCTs using a cooled radiofrequency probe were identified. A third RCT used palisade sacroiliac joint radiofrequency neurotomy.

Aydin et al published a meta-analysis of radiofrequency ablation (RFA) for sacroiliac pain in 2010.18 Nine studies were included that reported the primary outcome measure of a reduction of pain of 50% or greater, including 1 randomized placebo controlled study, 3 prospective observational studies, and 5 retrospective studies. All of the studies used injection of local anesthetic to determine if RFA was indicated for the patient. Seven studies reported follow-up to 3 months, and 6 studies reported follow-up to 6 months. Meta-analysis indicated that half or greater of the patients who received RFA to the sacroiliac joint showed a reduction in their pain of 50% or more at 3 and 6 months. Analysis found no evidence of publication bias, but heterogeneity in studies was observed for the 6-month follow-up. This systematic review is limited by the low quality of included studies and lack of RCTs. In addition, as noted by the authors, no standards have been established for the specific nerves to ablate or type of technique.

The single RCT included in the systematic review was published in 2008.19 This study examined the effect of lateral branch radiofrequency denervation with a cooled probe in 28 patients with injectiondiagnosed sacroiliac joint pain. Two of 14 patients (14%) in the placebo-control group reported pain relief at 1-month follow-up. None reported benefit at 3-month follow-up. Of 14 patients treated with radiofrequency denervation, 11 (79%) reported pain relief at 1 month, 9 (64%) at 3 months, and 8 (57%) at 6 months.

In 2012, Patel et al reported a randomized double-blind placebo-controlled trial of lateral branch neurotomy with a cooled radiofrequency probe.20 Twelve-month follow-up was reported in 2015.21 Fiftyone patients who had a positive response to 2 lateral branch blocks were randomized in a 2:1 ratio to lateral branch radiofrequency or sham. At 3-month follow-up, significant improvements in pain (-2.4 vs - 0.8), physical function (14 vs 3), disability (-11 vs 2), and quality of life (0.09 vs 0.02) were observed for radiofrequency treatment compared with controls (all respectively). With treatment success defined as a 50% or greater reduction in the numeric rating scale (NRS), 47% of radiofrequency-treated patients and 12% of sham patients achieved treatment success. The treatment response was durable out to 12 months in the 25 of 34 patients who completed all follow-up visits.21 Of the 9 patients who terminated study participation, 4 were considered treatment failures (12% of 34).

No additional studies were identified in the 2013 ASIPP evidence review, which concluded that evidence is limited for conventional radiofrequency neurotomy, limited for pulsed radiofrequency neurotomy, and fair for cooled radiofrequency neurotomy.9

In 2014, Zheng et al reported an RCT of palisade sacroiliac RFA in 155 patients with ankylosing spondylitis (AS).22 Palisade RFA uses a row of radiofrequency cannulae perpendicular to the dorsal sacrum. Inclusion criteria were age 18 to 75 years; diagnosis of AS; chronic low back pain for at least 3 months; axial pain below L5; no peripheral involvement; pain aggravation on manual pressing of the sacroiliac joint area; and at least 50% pain relief following fluoroscopically guided anesthetic injection into the joint. Patients who met the inclusion/exclusion criteria were randomized to either palisade RFA or celecoxib. Blinded evaluation found that RFA resulted in lower global VAS scores compared with celecoxib (2.8 vs 5.0, respectively, p<0.001), as well improved scores for secondary outcome measures. This study is limited by the lack of a sham control.

Arthrodesis

The literature on arthrodesis (open or minimally invasive) for sacroiliac joint pain includes 1 RCT on minimally invasive fusion, 1 cohort study comparing open and minimally invasive sacroiliac fusion, and a number of case series.

In 2015, Whang et al reported an industry-sponsored non-blinded RCT (NCT01681004) of the iFuse Implant System in 148 patients.23 Inclusion in the study was based on the determination of the sacroiliac joint as a pain generator from a combination of a history of sacroiliac joint?localized pain, positive provocative testing on at least 3 of 5 established physical tests, and at least a 50% decrease in sacroiliac point pain after image-guided local anesthetic injection into the joint. The duration of pain before enrollment averaged 6.4 years (range, 0.47-40.7 years). Prior treatments in the control group included physical therapy (78.3% of subjects), intra-articular steroid injections (91.3%), and RFA of the sacroiliac nerve roots (8.7%).

Patients were assigned in a 2:1 ratio to minimally invasive sacroiliac joint fusion (n=102) or to nonsurgical management (n=46). Nonsurgical management included a stepwise progression, depending on individual patient needs, of pain medications, physical therapy (98%), intra-articular steroid injections (73.9%), and RFA of sacral nerve roots (45.7%). The primary outcome measure was 6-month success rates, defined as the proportion of treated subjects with a 20-mm improvement in sacroiliac joint pain in the absence of severe device-related or neurologic adverse events or surgical revision. Missing values were considered to be treatment failures, and the study was considered to meet its end point if there was a posterior probability for superiority of fusion of at least 0.975 by Bayesian analysis. Patients in the control arm could crossover to surgery after 6 months. Baseline scores indicated that the patients were severely disabled, with VAS pain scores averaging 82.3 of 100 and ODI scores averaging 61.9.