The role of minimally invasive sacroiliac (SI) joint fusion surgery in the treatment of low back pain recently came under severe criticism in an editorial commentary in the influential Journal of Bone and Joint Surgery (American edition). (See Lee, 2019.)
This criticism is jarring because minimally invasive SI-joint fusion surgery is becoming increasingly popular after a wave of industry-sponsored RCTs showing positive results. Yet the criticism appears difficult to refute. And it calls into question recent favorable reviews of this form of surgery—from prominent insurers, payers, and policy-making groups.
Whether the architects of those reviews chose to ignore the issues brought up in this critical editorial or simply conducted shallow, incomplete investigations is not clear.
The editorial commentary focused on an RCT comparing minimally invasive SI joint fusion to nonoperative care in 103 patients by Julius Dengler, MD, and colleagues. (See Dengler et al., 2019.) The study came to positive conclusions.
“For patients with chronic sacroiliac joint pain due to joint degeneration or disruption, minimally invasive sacroiliac joint arthrodesis with triangular titanium implants was safe and more effective throughout 2 years in improving pain, disability, and quality of life compared with conservative management,” according to Dengler et al. (See study details below.)
However, the editorial commentary by orthopedic surgeon Mark C. Lee, MD, identified what he termed “several weaknesses” in the design and conduct of the commercially sponsored study—and also questioned the validity of the underlying diagnostic method that attributes back pain to the SI joint. Lee offered the reminder that there is no fully validated, widely accepted method of determining that back pain stems from the SI joint, which basically calls into question all the results of this study.
Uncertainty Regarding the Diagnosis of SI Joint Pain
“The study's central limitation, and the limitation of the plethora of recent industry-supported investigations into percutaneous sacroiliac joint arthrodesis, is that there remains no gold standard for the diagnosis of sacroiliac joint pain. The myriad pain generators around the pelvis, hip, and lumbar spine, inclusive of disc disease, facet disease, degenerative labral tears, and simple musculoligamentous discomfort, make it challenging to attribute pain symptoms directly to the sacroiliac joint,” according to Lee. So attributing back pain to the SI joint involves an uncomfortable level of guesswork.
Given the uncertainty surrounding this diagnosis, it is not clear whether minimally invasive SI joint fusion is addressing pain from the SI joint or other structures around the spine and pelvis. In a recent e-mail, Lee said he worries that, without better research, minimally invasive SI joint surgery will end up being used inappropriately for a variety of spinal conditions.
“I worry about this condition becoming mainstream as I anticipate it will be used to treat all sorts of low back pain,” Lee commented. To prevent that from happening, he would like to see a variety of further studies before minimally invasive SI joint fusion surgery is widely utilized outside of research settings.
“I would like to see more studies establishing the diagnosis, offering reproducible physical exam, radiographic and diagnostic/therapeutic injection approaches. Once diagnostic criteria are established, then well-performed randomized trials should be done,” Lee suggested.
“Magic Bullet Destined for Disrepute?”
Lee's commentary follows another, even more critical, review of SI joint fusion from neurosurgeons Robert W. Bina, MD, and R. John Hurlbert, MD, PhD, in 2017. In an article entitled “Sacroiliac Fusion: Another ‘Magic Bullet’ Destined for Disrepute,” they also questioned the evidence regarding SI joint involvement in low back pain. (See Bina and Hurlbert, 2017).
They pointed out that widely cited estimates of the contribution of the SI joint to back pain in studies and systematic reviews are not based on adequate scientific evidence. Those estimates—culled from multiple studies—suggest that the SI joint is responsible for some 10% to 30% of back pain in clinical settings.
“However, calculating the true prevalence of SI joint dysfunction as the cause of low back pain is rife with difficulty, as there are no ‘gold standard’ criteria by which to make the diagnosis,” according to Bina and Hurlbert. And they suggest that existing prevalence studies are incapable of documenting the true prevalence of the disease “because of inconsistent inclusion criteria, loose radiographic definitions, nonspecific clinical findings, and varied techniques. Consequently, the true prevalence of SI-related low back pain is unknown.”
Bina and Hurlbert also asserted that the evidence on interventions for SI joint-related pain, including SI joint fusion, is significantly flawed.
“At the present time, evidence establishing short-term and long-term outcomes is misleading because of vague diagnostic criteria, flawed methodology, bias, and limited follow-up.” And they believe that both the indications for SI joint fusion surgery and current surgical methods “are unlikely to stand up to close scrutiny or weather the test of time.”
RCTs with Inherent Limitations
Unfortunately, RCTs comparing minimally invasive SI joint fusion to nonoperative care face a variety of methodological challenges. A group of Norwegian researchers recently pointed out at ClinicalTrials.gov that the identification of a viable control group (against which to compare minimally invasive SI-joint fusion) is a difficult issue.
“Newer mini-invasive surgical approaches have shown promising results in scientific studies. [However,] it is difficult to find an adequate control group for surgery as most patients already have tried conservative and alternative treatments without effect,” according to Thomas Johan Kibsgård, PhD, MD of Oslo University Hospital and colleagues. (See Kibsgård et al., 2019).
And members of this type of control group will naturally be discontented with continuing unproductive treatment—and inclined to cross over to surgery. And when they do, this complicates statistical analysis of the RCT, leading to questionable results.
So, in many respects, the single best study design to evaluate the benefits and risks of minimally invasive SI-joint fusion surgery would be a placebo-controlled RCT. And that is exactly the type of study that Kibsgård et al. have recently undertaken in Norway.
“The investigators want to examine whether there is a difference in SI joint pain in patients operated with minimally invasive arthrodesis of the SI joint compared to a sham-operated control group. Patients with SI joint pain are included. They will be randomized to either surgery with arthrodesis or sham surgery. Neither patient nor health personnel who work with the patient after the surgery will know what has been done. The primary end point for the study is sacroiliac joint pain on the operated side after 6 months,” according to the research team. They expect to complete the study in 2023.
Support From Multiple Payers and Policy-Makers
As mentioned above, the shortcomings in the body of evidence on SI-joint fusion surgery have not prevented several major review organizations and insurers from giving minimally invasive SI joint surgery positive reviews.
UnitedHealthcare, the largest private insurer in the United States, recently concluded that minimally invasive SI joint fusion “is indicated” for patients who meet seven clinical and imaging criteria. (See https://www.uhcprovider.com/content/dam/provider/docs/public/policies/medadv-guidelines/p/percutaneous-minimally-invasive-fusion.pdf
The UnitedHealthcare policy did not mention the lack of a validated definition of SI-joint pain nor the corresponding controversy over the prevalence of SI-joint symptoms.
NICE (the National Institute for Clinical Excellence), a policy-making arm of the National Health Service in the UK, recently concluded that “Current evidence on the safety and efficacy of minimally invasive sacroiliac (SI) joint fusion surgery for chronic SI pain is adequate to support the use of this procedure provided that standard arrangements are in place for clinical governance, consent and audit.”
The NICE guidance also failed to mention the controversy over the definition of SI-joint pain. And it appeared to accept without comment the past estimates of SI joint pain.
“Chronic pain in the lower back triggered from the sacroiliac (SI) joint occurs in 15% to 30% of patients with low back pain. The causes of SI joint pain include degenerative sacroiliitis, osteoarthritis, SI joint disruptions from trauma or pregnancy, problems after lumbar spinal fixation techniques, anatomical abnormalities such as scoliosis, infection, gout, tumour or idiopathic causes,” according to NICE.
The NICE claim that SI-joint pain accounts for 10-30% of low back pain (as opposed to 10-30% of chronic low back pain) would suggest that the SI-joint contributes to millions of cases of low back pain.
This is quite a leap from the existing evidence. This is a clearly a shallow and less than accurate review. NICE has an excellent reputation for evidence-based reviews. This guidance will clearly undermine that perception.
So who is right regarding the risks and benefits of minimally invasive SI-joint fusion surgery? And the answer is “No one knows.” It will take an extensive research effort to resolve these issues. And discussion of the uncertainties in this area should be front and center in shared decision-making and informed-consent discussions among patients and providers.
SI Joint a Controversial Explanation for Back Pain for More Than a Century
By way of background, the SI joint has been a controversial explanation for low back pain, and a speculative surgical indication, for more than a century.
Richard A. Deyo, MD, wrote a seminal commentary on low back pain in the New England Journal of Medicine in 1991. He pointed out that the history of medical care for low back pain has featured a series of unvalidated fads, including diagnostic and treatment approaches that have done serious harm to patients. He opened up the NEJM commentary with mention of open fusion surgery for pain attributed to the SI joint. (See Deyo, 1991.)
“Earlier in this century, sacroiliac joint disease was thought to account for many cases of back pain, and this led to many fusions of the sacroiliac joint. Coccydynia was a popular diagnosis that led to a wave of coccygectomies, a procedure now almost completely abandoned. The use of chymopapain injections for herniated lumbar intervertebral disks enjoyed explosive growth in the early 1980s, but has declined greatly. Recent clinical trials have challenged the efficacy of many popular treatments, including lengthy bed rest, traction, and transcutaneous electrical nerve stimulation. Each of these fashions arose from a seductive pathoanatomical theory, leading to the popularity of a diagnosis that justified the corresponding treatment. Such fads are not innocuous; they may lead to unnecessary morbidity and costs, as well as to embarrassment for professionals,” according to Deyo.
And, as mentioned previously, some have suggested that minimally invasive surgery for sacroiliac joint pain might arrive at a similar fate. But that remains to be seen, pending better quality evidence.
RCT of 103 Patients
In the new study, Julius Dengler, MD, and colleagues from nine European surgery centers performed a trial of 103 patients randomly allocated to minimally invasive SI joint surgery or nonoperative care. Study subjects in the nonoperative care group were allowed to cross over to surgery after six months.
Dengler et al. investigated an implant system developed by SI-Bone of Santa Clara, California. One author of the study was an employee of SI-Bone and four other authors were consultants to the company. “The study sponsors participated in study design, data collection, data analysis, data interpretation, and writing of the report. One author of this study (J.D.) had full access to all study data and had final responsibility for the decision to submit for publication,” according to Dengler et al.
To enter the study, all patients had to meet the study's definition of having pain stemming from the SI joint. These included meeting the following criteria:
- Patient has pain at or close to the posterior superior iliac spine with possible radiation into buttocks, posterior thigh, or groin and can point with a single finger to the location of pain (Fortin Finger Test);
- Patient has at least three of five physical examination maneuvers specific for SI joint pain (see study for details); and
- Patient has improvement in lower back pain numeric rating scale score of at least 50% of the preinjection score after fluoroscopic controlled injection of local anesthetic into affected SI joint(s) (including previously documented test less than six months ago).
The study design excluded patients with (1) severe low back pain due to other causes, such as lumbar disc degeneration, lumbar disc herniation, lumbar spondylolisthesis, lumbar spinal stenosis, lumbar facet degeneration, and lumbar vertebral body fracture; (2) SI pathology caused by autoimmune disease (e.g. ankylosing spondylitis) and/or neoplasia (e.g. benign or malignant tumor) and/or crystal arthropathy; (3) history of recent (less than one year) fracture of the pelvis with documented malunion; (4) nonunion of sacrum or ilium or any type of internal fixation of the pelvic ring; and (5) spine surgery during the past 12 months, and other conditions that make treatment with the IFuse Implant System infeasible.
However, the study did allow patients who had previously undergone spinal fusion surgery (more than 12 months previously) to enter the study. More than one-third of the patients in both groups had previously undergone lumbar spinal fusion.
Extensive Crossover Complicated Study Interpretation
At the six-month follow-up mark, 43% of the nonsurgical group (21 patients) crossed over to surgery, leaving only 28 to continue nonoperative care.
This level of crossover, of course, poses major challenges in an RCT. The authors came up with a less-than-satisfying solution. “The impact of cross-over from conservative management to sacroiliac joint arthrodesis was investigated using a last-observation-carried forward approach, substituting the last observation prior to cross-over for subsequent values.”
The primary outcome measures were self-reported pain at two years (except among the crossover group), disability as measured by the Oswestry Disability Index (ODI) and quality of life as assessed by EuroQol.
Considerable Improvement in Pain and Disability in the Surgery Group
The mean baseline pain scores in the surgery and nonoperative groups were 77.7 and 73 points, respectively.
“At 2 years, the mean low back pain improved by 45 points (95% confidence interval [CI], 37 to 54 points) after sacroiliac joint arthrodesis and 11 points (95% CI, 2 to 20 points) after conservative management, with a mean difference between groups of 34 points (p < 0.0001),” according to Dengler et al.
Oswestry Disability Index scores approached 60 in both groups at baseline. “The mean ODI improved by 26 points (95% CI, 21 to 32 points) after sacroiliac joint arthrodesis and 8 points (95% CI, 2 to 14 points) after conservative management, with a mean difference between groups of 18 points (p < 0.0001). Parallel improvements were seen in quality of life,” according to Dengler et al.
However, there was no blinding in this study. Neither the surgeons nor patients were blinded. And outcome assessors and study personnel weren't blinded either. As many studies have pointed out over the years, lack of blinding can lead to inflated outcome scores, especially for subjective outcomes such as pain. For example, if surgeons, patients, and study personnel all believe that one treatment in an RCT is superior to the other, these expectations can certainly influence outcomes.
“Blinding patients to the treatment they have received in a controlled trial is particularly important when the response criteria are subjective, such as alleviation of pain, but less important for objective criteria, such as death. Similarly, medical staff caring for patients in a randomised trial should be blinded to treatment allocation to minimise possible bias in patient management and in assessing disease status,” noted Simon Day and the late Douglas Altman in a commentary in BMJ. (See Day and Altman, 2000).
Dengler et al. acknowledged that lack of blinding was a short-coming in this RCT. “As the intervention was not blinded, it is possible that extraneous factors associated with sacroiliac joint fusion [SIJF] may have contributed to greater responses in the SIJF group. The purpose of our study was not to estimate the proportion of the surgery effect attributable to the device (as would a sham study) but rather to estimate the increment in relief of pain, disability, and poor quality of life potentially observable in standard practice.”
Continuing Opioid Use
So the study results by Dengler et al. were positive. However, it was clear that neither approach was a panacea for low back pain. The prevalence of opioid use was 56% and 47.1% in the surgical and nonsurgical groups at baseline. There was an improvement in opioid use in the surgical group. But at two-year follow-up, a full third of the surgical patients were still taking opioids. Patients in the nonoperative group made no significant progress in terms of opioid consumption, with 45.7% still reporting taking opioids at long-term follow-up. Many would regard these opioid outcomes as ominous results. And there is clearly a need for long-term follow-up of these and similar patients—to see what level of medical treatment these patients require over the long run.
Other Issues in the RCT by Dengler et al.
In his editorial commentary in JBJS, orthopedic surgeon Lee took issue with several design aspects of this study. “The study should be evaluated in light of several weaknesses,” he asserted. “First 35% to 37% of enrolled subjects had prior lumbar spine surgery [i.e., fusion surgery]. Previous spine surgery may confound the reported effectiveness of this surgical intervention, either enhancing or diminishing the improvement in pain scores.”
For example, given the diagnostic uncertainty surrounding SI-joint pain, some would argue that patients who previously had fusion surgery might be seeking treatment for failed surgery syndrome rather than SI-joint pathology.
As mentioned above, Lee pointed out that physical tests for SI-joint pain, and their combination, can stress multiple structures in the spine, pelvis, and hip area. Imaging tests, he noted, are generally “unhelpful” in the absence of clear inflammatory or structural disease.
“Further, although sacroiliac joint diagnostic injections are thought to be reliable for identifying sacroiliac joint pain, the false-positive rate is reported to be up to 20% and there exists wide practice variability in terms of injection technique, medication type and dosage, and threshold used for a positive response,” according to Lee. (See Szadek et al., 2009.)
Overall, according to Lee, the profile of a patient who would benefit from minimally invasive SI-joint fusion “remains murky.”
Readers who like to see a more positive review of the diagnosis and treatment of sacroiliac joint pain can find one by David W. Polly, Jr., MD at Neurosurgery Clinics of North America. (See Polly, 2017) Polly was the lead author of one of the most widely cited RCTs comparing minimally invasive SI-joint fusion to nonoperative care. (See Polly et al., 2016.) Additionally, readers can access the study by Dengler et al., and the commentary by Lee, for free at the website of JBJS. (See references below).
Disclosures: None declared.