The Briefing Issue 06

Start with the one that should slow you down. A scoping review in the Journal of Manual and Manipulative Therapy pulled together every reported case of extracranial vertebral artery dissection presenting with neck pain. Here is the picture that came out. Mean age 36.8. Almost two thirds women. Neck pain that was unilateral, severe, and, in their words, "poorly responsive to analgesics," often radiating into the head, shoulder, or arm. In half the cases nobody could identify what started it. Neurological symptoms showed up in two thirds, most often arm weakness and dizziness. Nausea and vomiting in about a quarter.

Read that list again and ask yourself honestly how many of those patients would look like a mechanical neck with a headache on your schedule Monday. Half of them? That is the whole point. Now the caveat, and it is a big one: this is 58 patients pulled from 38 case reports and 8 case series. There is no denominator, so nobody can tell you the sensitivity or specificity of any of these signs. The authors say it themselves, the clinical utility "remains uncertain." This does not give you a screening rule. It gives you a reason to keep asking questions.

It also lines up with the gate in our consensus. Contraindications and precautions get cleared before the needle is on the table, not after, and the anterior neck stays off limits unless you have real training there. A patient who does not walk like a duck does not get treated like one.

This next one is about as clean a demonstration of the model as you will ever get handed, and it came out of a surgical imaging study nobody wrote for us. A hundred patients had anterior cervical decompression and fusion. Ten of them, 10 percent, came out with a C5 palsy, meaning a shoulder that would not go. When the researchers measured how far the cord had drifted backward at C4/5 on the early post-op MRI, the palsy group measured 3.42 mm and the group that was fine measured 2.04 mm. That difference held up after adjustment, odds ratio 2.89, and the cutoff landed at 2.85 mm with an AUC of 0.86.

Nobody touched the deltoid. Nobody injured a muscle. The cord shifted, the C5 root got tethered, and the arm quit working. Roughly a millimeter and a half of tension on a nerve root is the difference between a shoulder that works and one that does not, and I would frame that and hang it in every clinic that still treats the painful shoulder and never looks at the neck. This is silent radiculopathy with a number attached. Ten events in one center, and the authors call the threshold hypothesis-generating, so hold the precision loosely. Hold the principle tightly.

There is a second paper worth stapling to it. A retrospective series of 150 revision lumbar surgeries for failed back surgery syndrome. Pain resolved beautifully, VAS from 7.2 down to about 1, disability from 66 down to 8. And yet sensory improvement showed up in only 15 of 48 patients, 31 percent, and motor improvement in 21 of 39, 53 percent. Oof. So the pain left and roughly half the motor deficits and most of the sensory ones stayed. Pain relief is not the same thing as nerve recovery, and if we use the patient feeling better as our proof the nerve came back, we are going to stop way too early.

Last issue I told you to test your headache patient's capacity instead of their pain, and I said my gut was that you would find a comfortable neck that could not hold a position. This month somebody put numbers on it. A case-control study in Physiotherapy Theory and Practice took 140 young adults, 70 with neurologist-confirmed migraine and 70 without headaches, and measured them.

Deep neck flexor endurance in the migraine group: 22.46 seconds. In the controls: 48.37 seconds. Twenty-two seconds against forty-eight. Extensor endurance was down too, 49.8 against 72.2, and the craniovertebral angle was about 6 degrees smaller, 45.4 against 51.6. Core endurance was lower on every McGill test they ran. When they put it all in a multivariable model, neck flexor endurance and that craniovertebral angle held up on their own.

Now, these were migraine patients. Not cervicogenic headache, not neck pain patients. Migraine, diagnosed by a neurologist, and their necks were still half as durable as everybody else's. That is the argument for examining the neck in every headache patient, not just the ones already labeled as ours. Keep it honest though. Case-control, so this is association, and the authors say so plainly. Young university students, so do not extrapolate it to your 58-year-old. And the paper reports an AUC of 0.99, which is an in-sample number on unvalidated data. Ignore that figure. It is not a diagnostic test and nobody should sell it as one. The endurance numbers are the useful part.

None of this is new ground, even if the imaging is. Gwen Jull ran the trial that put deep neck flexor training on the map more than twenty years ago. Two hundred patients, that time with cervicogenic headache, the kind the neck is already driving, split across four groups. The exercise arm was a low-load program that trained these very muscles, the deep cervical flexors, to hold a position rather than to lift something heavy. Both that exercise and manipulative therapy cut headache frequency and intensity against the controls, the effect sizes were at least moderate and clinically meaningful, and the gains were still there at twelve months. Keep the two papers in their lanes. Jull is cervicogenic headache and a real randomized trial. This month's is migraine and an association. Put them side by side though and the line is hard to miss: the capacity you can now measure is the same capacity Jull showed you can train, and training it changed how often people hurt.

And Jull gave us the standard way to measure it, not just a stopwatch. The craniocervical flexion test is a slow, graded chin nod done against a pressure sensor tucked behind the neck, and it reads whether the deep cervical flexors are doing the holding or whether the big superficial muscles are muscling in to fake it. It is a test of control and low-load endurance, not raw strength, which is exactly the quality that was down in this month's paper. If you want a number you can defend and re-measure in six weeks, that is the tool to learn.