Trigeminal nerve stimulation (TNS), which has been cleared in the US for treating attention-deficit/hyperactivity disorder (ADHD), is not effective in children with the condition.

Results of a large randomized, sham-controlled study showed TNS did not improve core symptoms or related clinical and cognitive features of ADHD.

“This treatment is very safe, but it’s not working for ADHD. My message to psychiatrists is not to use this treatment and to families not to buy the device. The results suggest the clearance of the device by the Food and Drug Administration (FDA) should be revisited, said the study investigator,” Katya Rubia, PhD, professor of cognitive neuroscience at the Institute of Psychiatry, Psychology & Neuroscience, King’s College London, London, England, told Medscape Medical News. 

The findings were published online on January 16 in Nature Medicine.

The Path to Approval

ADHD affects approximately 5% of school-aged children and is characterized by symptoms of inattention and/or hyperactivity-impulsivity that interfere with daily functioning, along with impaired executive functioning.

First-line treatments include stimulants such as methylphenidate and amphetamine, but these can cause adverse effects, and adherence is poor. Second-line nonstimulants, including atomoxetine, guanfacine, and clonidine, are typically less effective than stimulants and can also lead to challenging side effects.

Noninvasive TNS delivers an electric current through battery-powered electrodes placed on the forehead. The stimulation targets the supratrochlear and supraorbital branches of the ophthalmic division (V1) of the trigeminal nerve.

The effects of TNS on the brain are believed to enhance attention and arousal mechanisms. TNS is also believed to stimulate the release of neurotransmitters — particularly noradrenaline, as well as dopamine, glutamate, gamma-aminobutyric acid, and serotonin — that are important for arousal, attention, and emotion regulation.

Based on a double-blind, randomized controlled trial in 62 unmedicated children, the FDA granted clearance in 2019 for external TNS as the first nonpharmacologic treatment for ADHD.

The pilot study showed that 4 weeks of nightly real vs sham TNS significantly decreased parent-rated symptoms on the ADHD Rating Scale (ADHD-RS) with a medium effect size (Cohen’s d = 0.5).

A Second Look

The new multicenter, double-blind phase 2b trial included 150 participants with ADHD (mean age, 12.6 years), who were mainly White individuals (79.3%), male (64.7%), and unmedicated (60.7%). The children were randomly allocated to real TNS or sham TNS and were asked to use the stimulator each night for about 8 hours during sleep.

Active TNS was delivered at a frequency of 120 Hz with a 250-μs pulse width, administered in a constant duty cycle of 30 seconds on and 30 seconds off. In the sham condition, electrical stimulation was delivered for 30 seconds once per hour (totaling 4 minutes over the night) at a lower frequency and pulse width designed to be perceptible but not therapeutically effective.

Adherence to the intervention (93.3%) was extremely high. This, the investigators noted, likely reflects the preference of parents and children for nonpharmacologic treatments.

The primary outcome was improvement in core symptoms, as measured by the parent-rated ADHD-RS score. In both groups, ADHD symptoms improved; at week 4, there was no significant difference between groups (estimated adjusted mean difference (aMD), 0.83; 95% CI, -2.47 to 4.13; P = .622; Cohen’s d = 0.09).

Rubia speculated that improvements in the sham group may be partially attributable to parents who, eager to secure their child’s participation in the trial, overstated baseline symptom severity and subsequently offered more realistic evaluations.

As for secondary outcomes, there was no significant between-group difference in the ADHD-RS total score at the 6-month follow-up (aMD, -0.29; 95% CI, -3.17 to 2.59; P = .845; Cohen’s d = -0.03). There were also no significant differences in symptoms of depression and anxiety, emotional dysregulation, or sleep.

Results also showed no effect of the intervention on an objective measure of vigilance/sustained attention, wrist-held measures of hyperactivity, or pupil diameter, a key physiological measure of arousal and autonomic nervous system activity.

Impact on Mind-Wandering?

However, there was a difference favoring the intervention for mind-wandering, a core feature of ADHD. The total score at week 4 on the Mind Excessively Wandering Scale (MEWS) showed a statistically significant group difference (aMD, -2.17; 95% CI, -4.33 to -0.01; P = .049; Cohen’s d = -0.27) for real vs sham TNS.

Rubia said the team did not want to overstate the finding, noting that it was only one of 16 outcomes, the difference was only marginally significant, and the researchers did not adjust for multiple testing.

She said the finding may have been spurious: A post hoc analysis of older participants (aged 13.6-19 years) showed no effect on mind wandering, while younger children “clicked on everything,” possibly because they did not understand the MEWS questions.

Another analysis, which was restricted to participants aged 8-13.5 years (roughly matching the age range of the pilot study), showed no significant between-group difference in ADHD-RS total scores at week 4 (aMD, 0.55; 95% CI, -3.73 to 4.83; P = .80). 

A subgroup analysis of the 91 unmedicated children and adolescents — compared with 62 unmedicated participants in the pilot study — showed no significant between-group difference in ADHD-RS total scores at week 4. 

Blinding in the study worked very well, Rubia said, adding that the success of the blinding was “quite unique.” She noted that, unlike in her study, the control condition in the pilot study did not involve any stimulation.

“Our findings are more conclusive than those of the pilot study because it was a larger group and had a better sham,” said Rubia. “The results highlight the importance of having very large trials with very tight control conditions.”

The intervention was safe, with no serious adverse events and similar side-effect profiles across groups. The most common side effects included sleep disturbances, drowsiness, headaches, and feelings of nervousness or hyperactivity, most of which participants rated as mild.

The most commonly reported adverse device effects were headache (real TNS, 21.3%; sham TNS, 17.3%) and difficulties falling asleep or sleep disturbances (real TNS, 20%; sham TNS, 9.3%).

The study had several limitations, including a high rate of missing teacher-rated data (80%) due to low teacher participation and the potential for bias in parent-reported ratings. In addition, although adherence was reported to be very high, it was self-reported and may have been overestimated.

The inclusion of medicated participants was another potential confounder; however, results were unchanged when analyses were restricted to nonmedicated participants. 

Rubia and her team are currently completing a functional MRI analysis to see how TNS affects brain activation.

In the US, TNS devices are available by prescription, and families are buying and using them, said Rubia.

In light of the new findings, Rubia said she intends to write to the FDA, although she does not anticipate there will be any significant changes.

Not the Final Word?

Reached for a comment, David W. Goodman, MD, assistant professor of psychiatry and behavioral sciences, Johns Hopkins University School of Medicine, Baltimore, said the new study provides more evidence that TNS doesn’t work well for ADHD.

“I think this fits within the body of literature that says that while we’d like to hope it’s promising, it’s not terribly convincing that this treatment works,” he told Medscape Medical News.

However, he noted that the trial’s size — 150 participants — is large for studies of this kind. He also said that the negative findings in the subgroup analysis of unmedicated children are an important point.

Also commenting for Medscape Medical News, James J. McGough, MD, professor of clinical psychiatry at the University of California, Los Angeles, and an author of the pilot study, said the new results are interesting and warrant careful consideration but emphasized that this is not the final word and that more evidence is needed.

McGough said the new study differs from the pilot study in several ways that may help explain its negative results. For example, medicated children were allowed to participate, and although the subgroup analysis of unmedicated participants showed no effect of TNS, McGough questioned how many of those participants were adolescents.

He added that relying on parents’ assessments of their teen’s behavior — particularly when they may not spend extensive time with them — can introduce substantial variability into the data.

He also highlighted differences in how the primary outcome measure, the ADHD-RS, was assessed. In the pilot study, investigators completed the scale using a fixed, established scoring system, whereas in the new trial it appears that parents completed it themselves, a shift that may have introduced greater variability in how children’s symptoms were rated, McGough said.

McGough also questioned the remarkably high placebo response rate observed in the current study, suggesting it may reflect parents guessing when rating their children’s symptoms and providing responses that were closer to random.

He and his colleagues are conducting a larger study involving 225 unmedicated children with ADHD, aged 8-12 years. One of the study’s goals is to determine whether brain changes in children who respond to TNS are associated with improvements in symptoms.

“Our current working idea is that there are some kids in whom this treatment works. It may be that we can identify them on a biological basis before prescribing treatment,” he said.

The study received funding from the Efficacy and Mechanism Evaluation Programme, a partnership between the National Institute for Health and Care Research (NIHR) and the UK Research and Innovation Medical Research Council, and the NIHR Maudsley Biomedical Research Center.

Rubia, Goodman, and McGough reported having no relevant disclosures.