Central or peripheral? This week the answer came back both ways.
This week the literature handed me a lesson I teach every course but rarely get to show with numbers: the site is not the source, and the source is not always the spine either. Sixty-three people with lateral hip pain, forty percent of whom had already been told to have back surgery. Four thousand hip replacements that went worse when nobody dealt with the back. Same question, opposite answers, same fix. In between, a needling trial that measured the muscle instead of asking the patient how they felt. And before all that, a safety paper I want every needler to read before Monday, about one sensation that should end a needle advance on the spot. Let's go.
I'll start with the hard one. A spine group in Korea reviewed four patients who developed an acute spinal cord injury immediately after a needle went into the neck. Two kinds of needle, cervical nerve blocks and what they call targeted acupuncture, and the same story every time. The patient felt an instantaneous, excruciating electric-shock sensation as the needle went in. Then weakness. Then paresthesia. Then an MRI that showed a bright spot sitting dead center in the cord, right on the path the needle took, not in the territory of an artery. The authors' point is that we blame particulate steroid and blood vessels for these injuries, and sometimes it is simply the needle. Their rule is one sentence long: when the patient reports a shock, you stop advancing. Now.
Four cases. That's the evidence level, and I'm not going to dress it up. It also doesn't matter. You don't need a randomized trial to tell you not to put a needle into a spinal cord.
Here is why I'm putting it in a dry needling newsletter when none of the four were dry needled by a physio. We teach the twitch as the goal. The local twitch response is the muscle answering the needle, a quick contraction, sometimes a jump, often a dull ache the patient recognizes. An electric shock is not that. It is sharp, it shoots, it goes somewhere, and the patient's face changes. Those are two different signals from two different tissues, and if you can't tell them apart in the moment, you are not S.A.F.E. The F is for Feel. Feel means the needle in your hand and the patient in front of you, and the patient just told you that you are in the wrong tissue.
A second paper the same week makes the bigger case. A Danish group pulled every reported adverse event from dry needling they could find across six databases, 130 of them, and sorted them into 31 categories. Bilateral pneumothorax was the only one they called potentially life-threatening. Epidural hematoma, dysautonomia, nerve injury, and pneumothorax made the severe list. Serious events are rare, and I believe that. The sentence I'd frame is this one: pneumothorax and peripheral nerve injury “remain of greatest concern as they are, at least in part, ascribable to clinician error.” Somebody else wrote our safety lecture.
One anatomy note before I move on, because the upper trap is the muscle most of us needle most. A cadaver and surface-mapping study found that the spot where patients localize their tender point in the upper trap, roughly the upper-middle interface along a line from C7 to the acromion, sits right over a consistently placed superficial neurovascular corridor, with nerve fascicles clustered against the deep fascia. The author makes no claims about what that means. I'll make one. Palpate first, find the structure, and angle away from it.
Add one line to your consent script and mean it: “If you feel an electric shock, say so.” Then do what you said. In the neck in particular, dose the depth, know the lamina, and treat a shock-like report as the end of that needle, not a reason to redirect it. And if you teach, show the difference between a twitch and a shock out loud, with the patient's words, so the next needler knows what they're listening for.
Central before peripheral. Most weeks I argue it from a neuro exam and a referral pattern. This week someone put electrodes on it, and someone else argued it from a shoulder blade.
The electrodes first. Sixty-eight outpatients with a lumbar disc herniation, leg pain and back pain, were randomized to five daily sessions of either conventional acupuncture or a needling technique the authors call Qihuang needling, which they describe as micro-invasive needling aimed at myofascial dysfunction, with the targets picked by the Selective Functional Movement Assessment. Movement screen first, then the needle goes where the screen points. The primary outcome was not pain. It was multifidus activity on surface EMG at rest. Multifidus activity went up more in the movement-guided group, by about seven and a half microvolts, and all seven of the secondary EMG measures went the same way. Disability on the Oswestry came in seventeen points better, and three out of four patients in the needling-by-screen group hit the responder bar against one in three in the comparison arm. No serious adverse events.
The fine print, and it matters. Five days, immediate effects only, and the authors say plainly they cannot tell you whether any of it lasts or which part of the package did the work. And I don't yet know what Qihuang needling is in the hand: how deep, what tissue, whether it is what you and I would call deep dry needling of the paraspinals or something else wearing a similar name. I'm reading the full text before this goes anywhere near a course slide. Hold it loosely. What I'll take from it today is the design. A radicular population. A movement screen choosing the target. A needle. And the verify measure is the segment itself, the multifidus, the little brains muscle that goes quiet when the root is unhappy. That's assess, treat, verify, written out by a group that has never heard of our model. The segment got treated, and the segment answered.
Now the shoulder blade, because it is the same story one region up. A group in Thailand took 144 people, 36 in each of four groups: chronic neck pain whose scapula rests in downward rotation, chronic neck pain whose scapula rests protracted, chronic neck pain with a normal resting scapula, and controls. They measured cervical range, the flexion-rotation test, which is the upper-neck rotation test you use for cervicogenic headache, joint position error at the neck and the shoulder, and pressure pain thresholds at the C5-C6 facets, the upper trap, the levator, and the tibialis anterior. The shin. Nowhere near the neck. The group whose scapula sat low was worse than everybody: more pain, more headache, more dizziness, a worse flexion-rotation test, bigger position errors, and the lowest pressure thresholds, including at the shin. The protracted group, the rounded-forward posture everybody worries about, looked like the group with no scapular problem at all.
Why does a low shoulder blade come with a tender shin? Because the shin is not the problem. A threshold dropping at a site the neck doesn't refer to is the nervous system turning the whole volume up, and that is a central story. And the muscles holding that scapula down, the levator and the upper trap, are the guarding muscles of the upper cervical segments. The muscle is the fuse. It is doing a job for a segment that is not happy, and the headache, the dizziness, and the restricted flexion-rotation test are all telling you which segment. Cross-sectional, so association only, and headache was a side symptom in a neck-pain cohort. The direction is the one we teach.
Two more bricks. A scoping review mapped fifty-one studies where somebody applied manual therapy to the lumbopelvic spine to treat an isolated lower-limb problem; nearly two thirds were knees, and strength was among the commonest outcomes. Treating the spine for the periphery is not a fringe idea, it's in the literature, even if almost no one describes the technique well enough to repeat it. And from the surgical side, one in four of 173 people who had the offending disc removed still had the leg pain six weeks later, and nothing on the MRI predicted who. The scan doesn't tell you who gets better. The nerve does.
Treat the segment, then measure the segment. In the low back, that's the prone multifidus lift, the segmental palpation, the myotome you tested before, the single-leg stance; needle the level, re-test, and if it moved, load it the same visit. In the neck, look at the shoulder blade before you touch anything. If it sits low, run the flexion-rotation test and the neuro exam, press on the shin, and if the shin is tender you have a turned-up system and a gentler dose. Needle the segment and the guarding muscles, levator and upper trap, before you treat the shoulder, and re-test the flexion-rotation on the table. If better, make sure to load it. Also remember last week's Briefing where we discussed the window for reinforcement after the posterior tibialis was needled. Seventy-two-hour window; love that we have support for that.
Here is how it walks in. Lateral hip pain, worse lying on that side, worse on stairs, a limp by the end of the day. There's a lumbar MRI in the chart, and it's ugly, because the patient is sixty and everybody's lumbar MRI is ugly at sixty. Someone has already said the word surgery.
The RMO question, out loud: what is driving this, and is the driver central or peripheral? I teach that every patient has a spinal segmental issue until you prove otherwise. People hear the first half of that sentence and forget the second half. This week the second half showed up in print, twice, from opposite directions.
Direction one. A neurosurgery group in Turkey looked at 153 consecutive patients with MRI-documented degenerative lumbar disease and clinically matching symptoms. Sixty-three of them had a hip that looked like greater trochanteric pain syndrome on exam, and all sixty-three had hip MRI findings to match. They treated the hip, an ultrasound-guided bursa injection and two weeks of anti-inflammatories, and the lateral hip pain fell from seven and a half out of ten to two at a month and under two at three months. Here's the number I keep coming back to. Twenty-five of those sixty-three, forty percent, had already been given a lumbar surgical recommendation somewhere else before anyone recognized the hip. And the more outpatient visits a patient had logged before the diagnosis, the more likely someone had told them to have their spine operated on. Uncontrolled, retrospective, an injection rather than a needle, and the authors are careful to say so. The direction is the point. The spine took the blame for a hip.
Direction two. Four thousand four hundred Medicare patients had a hip replaced, and the surgeons tracked who hit the threshold for a substantial clinical benefit a year later. If the patient came in with back pain and pain in the other hip, the odds of missing that threshold tripled. Severe back pain on its own nearly tripled it. Those patients were also twice as likely to say they were dissatisfied. The hip got replaced. Whatever the back was contributing did not get treated, and it showed up in the result.
Same question. Opposite answers. Same fix: examine, don't assume. In the first group the segment was innocent and the periphery was guilty, and the proving would have saved twenty-five people a surgical conversation. In the second group the segment was never examined as a driver and the periphery got a new joint it couldn't use. The nerve, by the way, is in the hip's pain either way. A small series this week burned the obturator nerve with pulsed radiofrequency in refractory hip arthritis and watched pain drop nine points on the spot, holding in milder joints and fading by three months in the worst ones. The joint hurts through a nerve. That is not a reason to burn the nerve. It is a reason to ask what else is talking to it.
Why does the mirror case strengthen the model rather than weaken it? Because the RMO is a question, not a verdict. The whole discipline is deciding, in this patient today, how much the segment is driving. If you skip the deciding and just treat the spine, you're doing what the surgeons in the first study did with a bigger needle.
Work the assessment. Pinwheel across L2 to L4 for the front of the hip and thigh, L5 and S1 for the lateral hip and the glutes. Myotomes: hip flexion, quads, glute med in side-lying, and compare them to the back. Reflexes. A slump or a femoral nerve tension test. Then the hip itself: single-leg stance for thirty seconds, resisted abduction, FADIR, and palpation last, over the trochanter and the glute med tendon, asking whether you reproduced the familiar pain. Now you know which end to treat first.
If the neuro exam lights up, needle the segment and its guarding muscles first, re-test glute med strength and the single-leg stance on the table, and only then go to the hip if it still needs you. If the neuro exam is clean and the trochanter reproduces the familiar pain, treat the hip and stop apologizing for it: needle the glute med and minimus, the brainy muscles of the pelvis, load them in side-lying and single-leg, and leave the spine alone unless the re-test says otherwise. Either way, write down the test you used to decide. That's the part the first study's twenty-five patients never got.
The patient felt better and got no stronger. Feeling better is not a verify measure for strength.
Your hands are not the problem. Your twitch is clean, you know your depth, and this issue's safety paper is not about you. This one might be. A trial out of Egypt took forty-five women with chondromalacia confirmed on MRI and split them three ways: dry needling plus exercise, myofascial release plus exercise, exercise alone. Six sessions over three weeks.
The needle did what the needle does. Pain dropped more in both hands-on groups than with exercise alone, and the needle matched the manual work on pain. Then they put everyone on the isokinetic dynamometer. Quadriceps peak torque went up in the myofascial release group. It went up in the exercise-only group. In the needling group it did not move. At all.
Fifteen per arm and three weeks, so hold it as what it is, but I'd rather we talk about it before somebody quotes it at a course. The abstract doesn't say which muscles were needled, how much, or how close to the loading session. My read, and it's a read: a quad that's sore from the needle on the day it's supposed to load may buy you pain relief at the cost of the reload.
The patient felt better and got no stronger, and feeling better is not a verify measure for strength. You already know that. The growth edge is proving it, every time: a strength number before the needle, the same number after the load, and if the number isn't moving, the needle isn't the plan. If you already work that way, teach it. The needle is still the sexy tool, and the dynamometer is where the patient actually gets free.
The site is not the source, and the source is not always the spine. Every patient has a segmental issue until you prove otherwise, and the job is the proving.
A small ask. If you have a hip that turned out to be a back, or a back that turned out to be a hip, send me that. I'd like to run a reader's case through Root Cause Rounds before the year is out. And a heads-up: Thursday you get the first Briefing Special Topics edition, on headaches. The whole five-part argument in one place, the treatment sequence, and a decision tree you can print. It is long on purpose. Make a coffee.
Upcoming courses
Pittsburgh, Pennsylvania
Grand Rapids, Michigan
On the radar
Thanks for reading all the way to the end. Here's a little something for it: use code THEBRIEFING15 for 15% off your next course.



