Diagnostic accuracy of clusters of pain provocation tests for detecting sacroiliac joint pain: systematic review with meta-analysis

Key points

1. 01A cluster means several provocation tests (distraction, compression, thigh thrust, Gaenslen, and sacral thrust) are performed together and called positive when more than 3 are positive.
2. 02The pooled positive likelihood ratio was 2.13 and the negative likelihood ratio was 0.33, both fairly weak shifts in probability.
3. 03At a 20% pre-test prevalence, a positive cluster moved the probability of SIJ pain only to about 35%, not enough to confidently rule it in.
4. 04A negative cluster dropped the probability to about 8%, so it is more helpful for ruling the joint out than ruling it in.
5. 05GRADE certainty was very low, so these estimates could shift as better studies appear.

How it was conducted

Design

Diagnostic test accuracy systematic review with bivariate random-effects meta-analysis (PRISMA-DTA, QUADAS-2, and GRADE; prospectively registered, PROSPERO CRD42020180268)

Search

Seven databases searched to September or October 2020; 2195 records screened and 5 studies included

Participants

Adults with low back pain, with or without leg pain, across the 5 studies; per-study sample sizes 34 to 150 and mean ages 42 to 58 years

Index test

A cluster of pain provocation tests, scored positive when more than 3 of 5 to 7 tests were positive

Reference standard

Image-guided intra-articular local anaesthetic blocks of the sacroiliac joint (IASP criteria), with double blocks preferred over single blocks

What they found

• Pooled positive likelihood ratio 2.13 (95% CI 1.2 to 3.9) and negative likelihood ratio 0.33 (95% CI 0.11 to 0.72).
• Pooled diagnostic odds ratio 9.01 (95% CI 1.72 to 28.4).
• Assuming a 20% pre-test prevalence: positive post-test probability 35% (95% CI 32 to 37) and negative post-test probability 8% (95% CI 6 to 10).
• GRADE certainty of evidence was very low for all outcomes.

Limitations

• Only 5 studies met inclusion and GRADE certainty was very low, so the pooled estimates are unstable.
• Reference standards varied, and only one included study used the more accurate double anaesthetic block.
• Differences in which tests made up the cluster and the positivity threshold limit how cleanly the results pool.
• The accuracy applies to ruling the joint in or out as a pain source, not to identifying a specific structural cause.

Why it matters

For patients

A positive set of sacroiliac joint tests does not confirm the joint is causing your pain, but a negative set makes the joint an unlikely source.

For clinicians

Use sacroiliac provocation clusters mainly to help rule the joint out; a positive cluster alone is not strong enough to rule it in, and a guided anaesthetic block remains the reference standard.

For readers

Clinical tests often carry this nuance: better at excluding a diagnosis than confirming it, and only as trustworthy as the very-low-certainty evidence behind them.