Interventional procedure overview of transcutaneous electrical stimulation of the trigeminal nerve for ADHD

Analysis

Follow up issues: Outcomes were assessed weekly or at baseline and week 4 and 5. One person randomised to the sham group withdrew after 3 weeks. One additional participant in each group withdrew between weeks 4 and 5. qEEG data for 3 participants were excluded due to excessive movement artifact.

Study design issues: This RCT assessed the outcomes of TNS for children with ADHD. People were recruited through community advertisements and internet postings. Randomisation was 1:1, using random block lengths of 4 and 6, to active or sham TNS. Participants, parents, and staff were blinded (except for 1 staff member who managed the study devices). To maintain the blind, active and sham systems were identical in appearance and operation, and participants were informed via a scripted presentation that 'pulses may come so fast or so slowly that the nerves in the forehead might or might not detect a sensation'. Outcomes included (see Outcome measures for full descriptions):

For statistical analysis, a general linear mixed model was used with treatment group (active vs. sham), time (in weeks), and group-by-time interactions to test for differential treatment effects as primary predictors, along with subject level random intercepts. This accounts for repeated measures and for missing data. Categorical outcomes were assessed using chi square. Effect size differences between groups were estimated using Cohen's d and NNT. For Cohen's d, cut-off values for small, medium, and large effects were defined as 0.2, 0.5, and 0.8, respectively. For NNT, small, medium, and large effects were defined as 9, 4, and 2, respectively. ADHD-RS was identified as the primary outcome a priori. However, there was no hierarchical testing of endpoints or adjustment for multiple comparisons. p<0.05 was considered statistically significant.

Study population issues: There were no statistically significant differences between the groups for age, sex, race/ethnicity, height, weight, vital signs, IQ, ADHD subtype, or baseline behavioural ratings.

Key efficacy findings

Key safety findings

Number of people analysed: 32 active treatment; 30 sham

There were no serious adverse events in either group and no participant withdrew for adverse events.

Analysis

Follow up issues: Data was available for 51 total people (31 active treatment; 20 sham crossover).

Study design issues: This secondary analysis of McGough (2019) evaluated whether baseline cognitive or EEG characteristics were predictors of positive response and associated with ADHD symptom reduction within the original RCT sample. The sample was divided into responders and non-responders. Responders were those with a 25% or greater reduction in ADHD-RS score from baseline and the end of active treatment (week 4 for people randomised to active treatment and week 9 for the people randomised to sham who chose to crossover). Non-responders were those with a less than 25% reduction in ADHD-RS score. Outcome measures (in addition to those described in McGough, 2019) included (see Outcome measures for full descriptions):

Analysis of variance (ANOVA) was used to identify baseline differences between responders and non-responders. Two analyses were used to then predict response status: linear regression analyses for prediction of post-treatment ADHD-RS Total scores and receiver operating characteristics curve analysis to determine area under the curve for prediction of responder status. Significant baseline predictors of ADHD symptoms were then tested for TNS treatment-related change by responder status and time (pre-/post-TNS) using repeated measures ANOVAs.

Pearson correlations between baseline predictors and ADHD symptoms were used to characterize degree of change occurring in both variables with TNS treatment. Partial eta squared was used as the measure of effect size and was interpreted as follows: small: 0.01, medium: 0.06, large: 0.14. Two procedures were used to control for multiple comparisons: only variables that were significant at p<0.05 in the baseline profile were further tested for prediction of treatment outcomes and treatment related change; and a p-value of p<0.01 was used as the threshold for statistical significance in subsequent analyses.

Study population issues: Responders and non-responders did not differ on any demographic or baseline clinical variables, including age, gender, IQ, or socioeconomic status.

Key efficacy findings

Key safety findings

Safety findings were reported in McGough, 2019 (Study 1).

Analysis

Follow up issues: Two participants were lost to follow up prior to Visit 4 outcome assessments, one each at Visits 2 and 3. One further participant was lost to follow up after Visit 6.

Study design issues: This single arm, open-label trial was a pilot study of the use of TNS for ADHD. Children were recruited and followed prospectively. Outcomes included (see Outcome measures for full descriptions):

The safety population included all participants with at least 1 night's exposure to TNS. The treatment population included all participants with outcomes data at week 4, the first post-baseline point at which primary behavioural and cognitive outcomes were obtained. Behavioural and cognitive measures were assessed for change over time with the general linear mixed model. p<0.05 was considered statistically significant. No adjustments for multiple comparisons were made.

Key efficacy findings

Key safety findings

Number of people analysed: 24

*ADHD symptom.