Comparative efficacy and acceptability of pharmacological, psychological, and neurostimulatory interventions for ADHD in adults: a systematic review and component network meta-analysis

Abstract

Background: The comparative benefits and harms of available interventions for ADHD in adults remain unclear. We aimed to address these important knowledge gaps.

Methods: In this systematic review and component network meta-analysis (NMA), we searched multiple databases for published and unpublished randomised controlled trials (RCTs) investigating pharmacological and non-pharmacological interventions for ADHD in adults from database inception to Sept 6, 2023. We included aggregate data from RCTs comparing interventions against controls or any other eligible active intervention for the treatment of symptoms in adults (ages ≥18 years) with a formal diagnosis of ADHD. Pharmacological therapies were included only if their maximum planned doses were considered eligible according to international guidelines. We included RCTs of at least 1-week duration for medications, of at least four sessions for psychological therapies, and of any length deemed appropriate for neurostimulation. For RCTs of medications, cognitive training, or neurostimulation alone, we included only double-blind RCTs. At least two authors independently screened the identified records and extracted data from eligible RCTs. Our primary outcomes were efficacy (change in ADHD core symptom severity on self-rated and clinician-rated scales at timepoints closest to 12 weeks) and acceptability (all-cause discontinuation). We estimated standardised mean differences (SMDs) and odds ratios (ORs) using random effects pairwise and component NMA, dismantling interventions into specific therapeutic components. This study was registered with PROSPERO (CRD42021265576). People with relevant lived experience were involved in the conduct of the research and writing process.

Findings: Of 32 416 records, 113 unique RCTs encompassing 14 887 participants were eligible for analysis (6787 [45·6%] females, 7638 [51·3%] males, 462 [3·1%] sex not reported). The RCTs encompassed pharmacological therapies (63 [55·8%] of 113 RCTs; 6875 participants), psychological therapies (28 [24·8%] of 113 RCTs; 1116 participants), neurostimulatory therapy and neurofeedback (ten [8·8%] of 113 RCTs; 194 participants), and control conditions (97 [85·8%] of 113 RCTs; 5770 participants). For reduction of ADHD core symptoms at 12 weeks on both self-reported and clinician-reported rating scales, atomoxetine (self-reported scale SMD -0·38, 95% CI -0·56 to -0·21; clinician-reported scale -0·51, -0·64 to -0·37) and stimulants (0·39, -0·52 to -0·26; -0·61, -0·71 to -0·51) had higher efficacy than placebo (Confidence in Network Meta-Analysis [CINeMA] ranging between very low and moderate). Cognitive behavioural therapy (-0·76, -1·26 to -0·26), cognitive remediation (-1·35, -2·42 to -0·27), mindfulness (-0·79, -1·29 to -0·29), psychoeducation (-0·77, -1·35 to -0·18), and transcranial direct current stimulation (-0·78; -1·13 to -0·43) were better than placebo only on clinician-reported measures. Regarding acceptability, all therapeutic components were similar to placebo other than atomoxetine (OR 1·43, 95% CI 1·14 to 1·80; CINeMA moderate) and guanfacine (3·70, 1·22 to 11·19; high), which had lower acceptability compared with placebo. Baseline severity of self-reported ADHD core symptoms, year of publication, percentage of male individuals, and percentage of individuals with ADHD and another mental health condition did not explain the heterogeneity observed in unadjusted non-component models of self-reported ADHD core symptoms. Treatment length had little effect on heterogeneity.

Interpretation: Stimulants and atomoxetine were the only interventions with evidence of beneficial effects in terms of reducing ADHD core symptoms in the short term, supported by both self-reported and clinician-reported ratings. However, atomoxetine was less acceptable than placebo. Medications for ADHD were not efficacious on additional relevant outcomes, such as quality of life, and evidence in the longer term is underinvestigated. The effects of non-pharmacological strategies were inconsistent across different raters. Our network meta-analysis represents the most comprehensive synthesis of available evidence to inform future guidelines in the field.

Funding: UK National Institute for Health and Care Research.