A nice surprise article from JBJS on low back pain. The image is CC by Wanderer99.

Low back pain has a history of failed trials supporting surgery (see Brox et al. and Fairbank et al.) and I was therefore thrilled when I encountered Dengler et al.’s randomized controlled trial (RCT) on sacroiliac joint dysfunction. They showed that arthrodesis outperformed conservative management both early and up to two years after surgery.

While I’m familiar with the concept of sacroiliac joint dysfunction it is not something I regularly diagnose. Thankfully it is pretty clearly defined in the ClinicalTrials registration (all had to be fulfilled):

• Patient has pain at or close to the posterior superior iliac spine (PSIS) 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 3 of 5 physical examination maneuvers specific for SI joint pain (i guess they mean these).
• Patient has improvement in lower back pain NRS of at least 50% of the pre injection NRS score after fluoroscopic controlled injection of local anesthetic into affected SI joint(s) (including previous documented test <6 months ago).

What I like

The fact that it is about 100 patients is critical. Surgery is a high-risk task and if you need lots of patients, e.g. as seen for clavicles, you are most likely finding abnormalities in the scores and not treatments actually worth writing about. You should be able to see the effect in 100-200 patients when it comes to treating diagnoses that involve pain and loss of function. This concept is closely related to the minimal clinically important difference (MCID) where we need to show that our treatment reaches a change that actually matters to the patient.

The clinical trial registration was clean and without any significant changes. The requirement of registering your trial is a major improvement as this reduces the ability to make changes to fit the results. I’ve seen examples where authors motivate a treatment option based on a single item in a score of 10-20 questions. In this study we find that all scores follow the expected path and it seems sound.

Concerns

It is important to remember that the surgery has associated complications that can be severe. Complications are frequently not that well handled by scores – the failure signal from a complication can drown in a score’s general fluctuations. I.e. a patient with a score of 0 may barely be noticed if the other patients fluctuate between 20 and 80. It is therefore important to properly inform all patients of the risks involved even if the surgery appears successful.

The average age of the patients was 49 years in the arthodesis group (the youngest was 27 years). The life expectancy hints that many of these patients will have their implant for another 25+ years. 2 years is nothing at this scale and I hope that the authors intend to keep publishing updates, figure 7 in the paper show how the bone has retracted from the implant and is certainly cause for concern. Hopefully there will be a proper post-market surveillance, I strongly recommend seeing the documentary the bleeding edge.

Summary

Dengler et al. have performed an excellent study showing the benefit in a patient group that is often difficult to treat. The effect seems clear and robust although safety issues are still a concern and post-market surveillance is a must.

In case you’ve come this far I strongly recommend a laugh with a solid Reddit post showing an inversion table fail.