Functional Limb Weakness: Would Treating Her Anxiety Actually Help Her Arm?
A single patient with functional right-arm weakness and longstanding, recently worsened generalized anxiety. The disagreement isn't whether to treat her anxiety — everyone agrees she should — it's whether doing so is honestly expected to improve her arm, or only sold to her that way.
K.M., a 46-year-old woman, has sung alto in her community choir for two decades and can still name the exact rehearsal, eight months ago, when she first noticed she couldn't hold her folder steady with her right hand. The weakness spread over the following weeks until she couldn't reliably lift a coffee cup or turn a doorknob with that arm, though nothing about it followed a pattern any single nerve or nerve root would produce. An MRI of her brain and cervical spine came back normal. EMG and nerve conduction studies came back normal. A neurologist eventually gave her the diagnosis her workup had been circling toward the whole time: functional neurological disorder, functional limb weakness, a real neurological condition without a structural lesion to explain it. She stopped singing lead parts months ago and now mostly mouths along, embarrassed by how little control she has over an arm every scan says should work fine.
Her anxiety predates the arm by years — generalized, low-grade, manageable enough that she never sought treatment for it — but it has climbed sharply over the past year while she's managed her mother's declining health alone, and her current GAD-7 score reflects real, clinically significant anxiety rather than ordinary stress. Whether treating that anxiety would also improve her arm is a genuinely open question the literature doesn't cleanly answer: the nearest RCT-level evidence, LaFrance and colleagues' trial in psychogenic nonepileptic seizures, found sertraline alone did not significantly reduce seizure frequency against treatment as usual — a different functional symptom, but the closest test available of whether treating mood symptoms reliably moves a functional neurological symptom on its own. Her team has to decide whether to frame today's medication as a plausible route to a stronger arm, or as a treatment for her anxiety that happens to run alongside, but isn't shown to cause, whatever happens to her weakness next. The comparison is not as lopsided as it first looks: specialist physiotherapy, the intervention built for her symptom, has its own equivocal record — Physio4FMD, the largest randomized trial of it, missed its primary physical-function endpoint at twelve months while favoring specialist treatment on patients' own rated improvement.
Deciding how to frame today's prescription
Her anxiety and her functional weakness aren't two unrelated problems that happen to coexist — FND is understood to involve limbic and autonomic circuitry that anxiety drives and sustains. Treating the anxiety is a real, mechanistic intervention on one of the maintaining factors for her weakness, not just a parallel treatment we're offering out of general good practice.
The mechanism is plausible — I want to be clear I'm not disputing that circuitry exists. But the nearest real trial evidence argues against expecting it to translate into symptom improvement: LaFrance and colleagues' 2014 trial randomized 38 patients with psychogenic nonepileptic seizures, a closely related functional diagnosis, across four arms — sertraline alone, CBT-informed psychotherapy alone, both together, or treatment as usual. The psychotherapy arms cut seizure frequency by 51 and 59 percent. The sertraline-only arm did not significantly reduce seizures at all, though it did improve depression scores.
A different functional symptom, granted — but look at what the four-arm design lets us separate, because it's exactly your claim. That trial's sertraline arm did move mood and did not move the functional symptom. If anxiety were driving the motor circuitry the way you're describing, that is the one result we should not have seen. If we tell her today that treating her anxiety is likely to help her arm, we're promising something this literature specifically tested and did not find.
I'd rather we stop arguing about which of you is right on the mechanism question, but I'm not going to overstate my own side to do it. The largest trial of specialist physiotherapy for functional motor disorder — Physio4FMD, Nielsen and colleagues, 355 patients across England and Scotland — missed its primary endpoint. Self-reported physical function at twelve months was no better than usual community physiotherapy. What it did show was a significant advantage on patients' own rated impression of change in their motor symptoms. So the honest summary is that this is the intervention built specifically for her symptom, with mixed rather than settled evidence behind it.
That still makes it the better bet for the arm than an SSRI, because mixed evidence aimed directly at functional motor symptoms beats a mechanistic story with a negative trial attached to it. Treat the anxiety — it's real, it's disabling on its own, and it doesn't need her arm's permission to deserve treatment. But tell her plainly that today's prescription is for how she's been feeling, and the physiotherapy program is the primary plan for her arm — two separate goals, not one treatment doing double duty because the story is more satisfying that way.
Agreed: start escitalopram today for her anxiety, refer to specialized FND physiotherapy today as the primary plan for her arm, and tell her honestly that these are two separate treatments for two separate problems rather than one drug doing double duty.
Not agreed, and left as a genuinely open mechanistic question rather than resolved:
The psychiatrist's mechanistic read gains real support in her specific case, even though LaFrance's trial data couldn't establish it as a general effect.
The psychologist's caution is vindicated for her specifically — and the team agreed in advance not to treat that outcome as a treatment failure, since physiotherapy, not the SSRI, was always the primary plan for the arm.