A wearable, artificial intelligence (AI)-driven, continuous-use transcutaneous peripheral nerve stimulation (TPNS) device was linked to significant improvement in upper limb tremor in adult patients with essential tremor (ET), results from the TRANQUIL trial showed.

The Felix NeuroAI Wristband (FaskikI Inc) stimulates the radian, median, and ulnar nerves. It received FDA clearance in July — largely on the basis of the TRANQUIL findings.

In the multicenter, randomized clinical trial, participants who received continuous daily TPNS with the Felix wrist device had significantly greater improvement on the modified Activities of Daily Living (mADL) subscale of the Essential Tremor Rating Assessment Scale (TETRAS) at 90 days than their counterparts who received sham stimulation, meeting its primary endpoint.

The most common device-related adverse event (AE) was mild skin irritation, occurring in 29% of the active group vs none of the sham group. Other AEs in the TPNS group included nausea, arthralgia, and worsening of existing arthritis in the thumb — but only in one participant for each AE. No serious AEs occurred.

“This was a positive study when there haven’t been that many positive multicenter trials of a nonsurgical treatment for essential tremor,” lead author William Ondo, MD, director of the Movement Disorders Clinic at Houston Methodist Neurological Institute, Houston, told Medscape Medical News.

“This fills a niche for a certain type of patient. And for them, I think this will provide a nice benefit,” said Ondo, who is also a professor of neurology at Weill Cornell Medicine, New York City.

As reported at the time by Medscape Medical News, the researchers presented preliminary results from TRANQUIL at the American Academy of Neurology (AAN) 2025 Annual Meeting. More detailed findings were published online on October 20 in JAMA Neurology.

Large Treatment Gap

Patients with ET experience involuntary movements of the upper limbs that can interfere with daily activities, such as eating, drinking, and writing. Approximately 1%-2% of individuals worldwide have ET.

Current treatment usually consists of medication, with the most common being propranolol and primidone, or surgery such as deep brain stimulation (DBS) or magnetic resonance-guided focused ultrasound.

However, 30%-50% of patients with ET don’t respond to these medications — with many experiencing AEs. Response to second-line meds are also “unpredictable and rarely adequate,” the investigators wrote.

They added that a substantial number of individuals with ET “do not want, or may not be candidates for, surgical procedures.” 

“So there’s still a pretty large gap in our treatment armamentarium that new medications and devices will hopefully fill,” Ondo said.

The FDA cleared noninvasive transcutaneous afferent patterned stimulation (TAPS) in 2018 for the treatment of hand tremor symptoms in ET. The agency cleared the Cala kIQ system, which includes a wrist-worn device, in 2023 for tremor relief.

That device calibrates a patient’s tremor frequency at initiation and then alternately stimulates the median and radial nerves in the forearm.

TPNS differs from TAPS in that it “delivers stimulation driven by cloud-based [AI] that continuously adjusts stimulation based on patient activity and predictive analytics,” the investigators wrote. It provides stimulation to the radial, median, and ulnar nerves. The device can be worn on either hand — or on both hands if using two devices.

Clinically Meaningful?

The investigators enrolled 125 adults with upper-extremity tremor and a clinical diagnosis of ET (mean age, 65 years; 50.4% men; mean tremor duration, 11.4 years) from 12 outpatient neurology clinics in the US and China in 2024. All had a tremor severity score of at least 2 on TETRAS performance subscale tasks and a score of at least 7 on the total performance subscale.

In a 2:1 design, 83 of the participants were randomly assigned to receive TPNS and 42 received sham treatment, which looked identical to the active device but generated brief vibrations instead of electrical stimulation. All participants used their assigned devices during waking hours for 90 days.

The primary outcome was change in daily activities at 90 days, as measured by the mADL subscale of TETRAS. Secondary outcomes included Clinical Global Impression of Improvement (CGI-I) and CGI-Severity (CGI-S) scores and Patient Global Impression of Improvement (PGI-I) and PGI-Severity (PGI-S) scores.

Results showed that the TPNS group had a greater reduction in mADL score at 90 days than the sham stimulation group (-6.9 points vs -2.7 points; P < .001).

Asked whether this reduction difference was perhaps too small to be considered clinically meaningful, Ondo said the investigators felt having at least a three-point improvement would be clinically meaningful. So the study showing a more than four-point greater improvement for TPNS over sham cleared that hurdle, he said.

He added that a one-point improvement could move an individual to being able to eat soup with a spoon when before they could not. “One-point increments are actually fairly large,” Ondo said.

Additionally, 61.3% of the TPNS group vs 25% of the sham group had a reduction of 20% or greater on the mADL at 90 days (P < .001). There were also greater improvement at 90 days for the active vs sham groups in total ADL score (P < .0001) and in item scores for hygiene (P < .0001), writing (P = .03), pouring (P = .0003), using keys (P = .03), and carrying food trays (P = .04).

In addition, 69% vs 44% showed improvement on the CGI-I (P = .02) and 68% vs 50% had improvement on the PGI-I (P = .04). There were no significant between-group differences on the CGI-S or PGI-S.

Coming Soon

Among the 29 mild skin irritation events that occurred among 24 members of the TPNS group, only five required topical treatment. There were three cases of moderate skin irritation in the TPNS group vs two cases in the sham group.

There were also reports of one each of the following in the active treatment group: nausea, arthralgia, limb discomfort, worsening of arthritis in the thumb, muscular weakness, and involuntary muscle contractions.

Ondo reported that the device used in the study “is almost available commercially.”

A company release from July noted its plan to have the Felix wristband available by prescription “in select US” regions by the end of 2025, with nationwide availability expected in the first half of 2026.

Ondo said this device should fill an unmet need for patients who are not satisfied with other ET treatments.

Although there have been no head-to-head trials, he said he thinks peripheral nerve stimulators aren’t as effective as DBS or thalamotomy. However, “they are obviously much easier to try, with less risk. I think there’s a pretty significant range of people who have had unsatisfactory response to medications but don’t want to go to brain surgery yet,” he said.

‘Exciting Therapy’

In an accompanying editorial, Theresa A Zesiewicz, MD, Department of Neurology at the University of Florida in Tampa, Florida, and colleagues noted that the clinical significance of the study meeting its primary endpoint “remains uncertain.”

In addition, they noted that improvement on several of the individual ADL item scores were “changed by fractional amounts” and that several of the secondary outcomes showed no benefit for TPNS.

“Clinicians should be aware of the variety of peripheral devices available to treat action tremor and that few have provided the kind of clinically meaningful benefit that patients with ET seek,” the editorialists wrote. Future advances and studies may “change that scenario,” but that remains to be seen, they added.

Commenting for Medscape Medical News, Jamie Adams, MD, chief of the Movement Disorders Division and associate director of the Center for Health + Technology at the University of Rochester, Rochester, New York, noted that there is a real need for additional ET therapies.

“The field and patients are desperately in need of better treatments. Our oral medications don’t work a good portion of the time and I appreciate that there are people working on noninvasive technologies,” said Adams, who was not involved with the research.

She applauded the use of longitudinal data in the current study and that the patient population came from two countries. She also said the device’s use of AI to tailor treatment was “very interesting,” but would have liked to have seen whether it provides long-term benefit.

“Still, I think it’s another exciting therapy for people with tremor,” Adams said.

Ondo reported having received grants from Houston Methodist Research Institute during the conduct of the study; personal fees from ACADIA, Neurocrine, TEVA, Supernus, AbbVie, Merz, and Revance; and grants from Biogen, AskBio, Harmony, Cerevance, and Dystonia Coalition outside the submitted work. Financial relationships for the other investigators are fully listed in the original article. Zesiewicz reported having conducted studies with Sage, Praxis, Jazz, Cala, and Boston Scientific. One of the other two editorialists reported having received research support from the National Institute of Neurological Disorders and Stroke and publishing royalties for Merritt’s Neurology. Adams reported having no relevant financial relationships.