Condition: Trouble du déficit de l'attention avec ou sans hyperactivité

In a network meta-analysis of 133 double-blind randomized trials (10,068 children and adolescents; 8,131 adults), all licensed medicines beat placebo on clinician-rated core symptoms at around 12 weeks. In children and adolescents, methylphenidate SMD was -0.78 (95% CI -0.93 to -0.62) and amphetamines -1.02 (-1.19 to -0.85); in adults, amphetamines -0.79 (-0.99 to -0.58) and methylphenidate -0.49 (-0.64 to -0.35). Weighing efficacy and tolerability together, the authors named methylphenidate the first-choice medicine for children and adolescents and amphetamines for adults. Measured in: 133 double-blind randomized controlled trials in children, adolescents and adults with ADHD, analyzed as a network meta-analysis. The trials mostly ran about 12 weeks, so this measures short-term symptom control, not the long-term effects the authors said still need study. The effect sizes are averages, and the choice of medicine is prescribed and monitored by a clinician.

Across 32 randomized trials, behavioral interventions improved positive parenting (SMD 0.63) and reduced conduct problems (SMD 0.31) on probably blinded ratings. The reduction in core ADHD symptoms seen on unblinded ratings (SMD 0.35) did not reach significance once assessors were probably blind to who had been treated. Measured in: 32 randomized controlled trials of behavioral interventions in children and adolescents with ADHD. The core-symptom benefit rested on ratings by people who knew the child was being treated; the durable, blinded gains are in parenting quality and conduct, not in the ADHD symptoms themselves, and the interventions did not improve parent mental well-being.

Across 32 studies (up to 896 adults), cognitive behavioral therapy beat control on self-reported ADHD symptoms (Hedges g 0.65, 95% CI 0.44 to 0.86) and functioning (g 0.51, 0.23 to 0.79). Pre-to-post effects were larger (symptoms g 1.00), studies with active control groups showed smaller effects, and medication status did not moderate the benefit. Measured in: 32 published and unpublished studies of cognitive behavioral treatment in adults meeting diagnostic criteria for ADHD. Much of the evidence is self-report, and effects shrank against active comparison groups, so part of the pre-to-post change reflects attention and expectation, not the therapy content alone. Longer treatments were not associated with better outcomes.

Across ten randomized placebo-controlled trials (699 children), omega-3 fatty acid supplementation produced a small but significant reduction in ADHD symptoms, and the eicosapentaenoic acid (EPA) dose within supplements correlated with efficacy. The authors judged the effect modest relative to stimulant medication and showed no sign of publication bias or heterogeneity. Measured in: Ten randomized placebo-controlled trials of omega-3 supplementation in 699 children with ADHD symptomatology. The effect is small, and the authors framed omega-3 as an add-on to medication or an option for families declining other treatment, not a replacement for it. Participants were children.

A meta-analysis found restriction diets reduced ADHD symptoms with a small effect (g 0.29, 95% CI 0.07 to 0.53). For synthetic food colors, the parent-rated effect was g 0.18 (0.08 to 0.24), falling to 0.12 after adjustment for possible publication bias; teacher and observer ratings were not significant. The authors estimated about 8% of children with ADHD may have symptoms related to synthetic colors. Measured in: 24 publications on synthetic food colors and 10 additional studies on dietary restriction in children with ADHD or ADHD symptoms. The color effect was small, strongest on parent ratings, and vulnerable to publication bias, while teacher and observer ratings were not significant. It points to a sensitive subgroup, not a general dietary cause of ADHD.

Across 16 randomized trials (759 children), cognitive training improved working-memory test scores (verbal SMD 0.52; visual 0.47) but had limited effect on core ADHD symptoms once probably blinded raters were used (total ADHD SMD 0.20, 95% CI 0.01 to 0.40; hyperactivity/impulsivity not significant), and no significant effect on academic performance. Working-memory training on its own did not move ADHD symptoms. Measured in: 16 randomized controlled trials of cognitive training in 759 children with ADHD. Gains showed up on the trained tasks and on unblinded ratings but largely faded on blinded measures of real-world symptoms. Commercial working-memory programs in particular did not reduce ADHD symptoms.

The MADDY trial gave medication-free children with ADHD and irritability a broad-spectrum micronutrient formula or placebo for eight weeks under double-blind conditions. On the pre-registered primary outcome, the clinician global impression of improvement, the supplement clearly beat placebo. The parent-rated symptom checklist (CASI-5 composite) did not separate, so the finding reads as a broad global improvement more than a big drop in symptom counts. It sits alongside omega-3 as a supplement adjunct with modest support, not a replacement for assessment and first-line treatment.

This is distinct from the working-memory and brain-training apps already covered: neurofeedback trains brainwave patterns directly. The same blinding pattern holds, and it is the heart of the finding. When only the least-blinded assessors rate improvement, neurofeedback separates from control by a small-to-moderate amount; under blinded ratings or against sham neurofeedback, it does not. Neurofeedback is time-intensive and often costly, so this pattern matters for families weighing it.

Pooling eight randomized trials (n=249), short-term aerobic exercise had a moderate-to-large effect on attention (SMD 0.84), hyperactivity (0.56) and impulsivity (0.56) in children with ADHD, with smaller effects on executive function (0.58) and anxiety (0.66). Yoga showed a suggestion of benefit on core symptoms. Measured in: Eight randomized controlled trials of exercise interventions in children and adolescents with ADHD. The trials were small, short-term and mostly unblinded, so the pooled effect is likely inflated and durability is unknown. This supports exercise as an adjunct to established treatment, not a substitute for it.

In the INCA trial, 100 children aged 4 to 8 were randomized to a five-week few-foods elimination diet or a healthy-diet control. On masked pediatrician ratings, the between-group difference on the ADHD rating scale was 23.7 points (95% CI 18.6 to 28.8). In the double-blind challenge phase, ADHD symptoms relapsed in 19 of 30 responders (63%) after reintroducing foods. Measured in: 100 children aged 4 to 8 with ADHD in the Netherlands and Belgium (INCA randomized controlled trial). Only the assessing pediatrician was masked in the main open-label phase, the diet is highly restrictive and hard to sustain, and later reviews could not confirm effects of this magnitude. It is a demanding elimination approach best done with dietitian supervision, not a first move.

In a randomized trial of 244 children aged 5 to 12 with ADHD, a brief behavioral sleep program reduced ADHD symptom severity versus usual care at six months (adjusted mean difference -3.7, 95% CI -6.1 to -1.2; effect size -0.4) and cut moderate-to-severe sleep problems (46% vs 34% at six months). About a third to a half of the symptom benefit was mediated by improved sleep. Measured in: 244 children aged 5 to 12 with ADHD across 21 general pediatric practices in Australia. The effect on ADHD symptoms was small and part of it ran through the sleep improvement itself. It applies to children who had a sleep problem to fix, which was an entry criterion, not to every child with ADHD.