An independent validation of a clinical prediction rule for the diagnosis of cervical radiculopathy with radicular pain.
Francis Grondin, Chad Cook, Toby Hall and 3 others
The 2003 four-test rule for cervical radiculopathy is confirmed in 85 new patients.
All four negative rules out the diagnosis. A new three-test cluster (shoulder abduction, Spurling's arm, two of four neurodynamic tests) catches more true cases at the same confidence.
Key findings
1When all four tests in the 2003 rule (cervical distraction, Spurling's, ULNT1, cervical rotation <60°) are negative, no patient with cervical radiculopathy was missed (post-test probability 0%). When all four are positive, the post-test probability is 100%, but this combination identifies only 17.9% of the CR patients in the sample.
2A novel three-test cluster (modified shoulder abduction, Spurling's arm pain, at least 2 of 4 ULNT positive) also rules out CR when all three are negative (post-test probability 0%), and when all three are positive the post-test probability is 100% while capturing 37% of CR patients, compared with 17.9% for the 2003 rule at the same confidence level.
3Spurling's test with reproduction of familiar arm symptoms had the highest specificity of any single test, with a positive likelihood ratio of 34.37 and a post-test probability of 94.10%.
Still to come
How it was doneWhat they foundWhat it means for PTs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
Only 27 patients had cervical radiculopathy, so confidence intervals around the key numbers are wide and the 100% post-test probability rests on a small denominator. One physical therapist performed all index tests and no interrater reliability was measured, so you cannot tell how reproducible the findings are across examiners. The reference standard (clinical exam plus MRI) was made by a single clinician. The novel cluster was derived from the same 85 patients it was tested on, and the authors note that backward stepwise regression can overfit. Independent replication in a larger, different sample is needed before this cluster can be recommended. The sample was recruited from a single Neurosurgery Department in a French-speaking setting, which may limit generalisability to other referral patterns or populations.
Declared interests
The authors declare no competing interest.
The easy way to misread this
Do not adopt the new three-test cluster as a replacement for the 2003 rule. It was derived from the same 85 patients it was tested on, and the authors flag the regression method as prone to overfitting. It needs independent replication in a larger, different sample before you change your assessment protocol.
Summarised by AI from the full paper, without a clinician reviewing it. Check it against the source before it changes what you do. Read it on PubMed →