Psychiatry · Medical Literature
[Corrections] Correction to Lancet Psychiatry 2026; 13: 485–95
Abstract
Nourredine M, Jurek L, Hamza T, et al. Pharmacological interventions for ADHD: a systematic review and dose–effect network meta-analysis. Lancet Psychiatry 2026; 13: 485–95—In this Article, the modafinil data in figure 5 has been incorrectly presented as data for guanfacine. The graph caption has been updated to illustrate that the data on drug effect on symptom severity are for modafinil and not for guanfacine. This correction has been made to the online version as of June 12, 2026.
Clinical relevance
Across ADHD medications (stimulants and non-stimulants) and age groups (children/adolescents and adults), what are the dose–effect relationships for both efficacy (symptom severity reduction) and tolerability, and do benefits continue to accrue above licensed maximum doses? __SNAPSHOT__ {"studyDesign":"Systematic review and dose–effect network meta-analysis of double-blind randomized controlled trials (RCTs); analyzed using a three-level hierarchical Bayesian framework with restricted cubic splines","setting":"Multi-national; international collaborative study (University of Southampton, University of Lyon, University of Bern, and others); included RCTs identified via the MED-ADHD database","sampleSize":"113 double-blind RCTs included in the dose–effect network meta-analysis (drawn from a broader systematic review of 164 studies); total participants: approximately 14,138 children/adolescents and 11,016 adults (~25,154 total)","population":"Individuals aged ≥5 years with ADHD; analyzed separately as children/adolescents (<18 years) and adults; note: children (5–<12 years) and adolescents (12–<18 years) could not be analyzed separately due to sample size constraints","intervention":"Oral monotherapy with stimulants (methylphenidate, amphetamines, lisdexamfetamine) and non-stimulants (atomoxetine, guanfacine, viloxazine, modafinil) across a range of doses","comparator":"Placebo (within the network meta-analysis framework, allowing both direct and indirect comparisons across doses)","primaryOutcome":"Dose–effect curves for ADHD symptom severity (efficacy); tolerability outcomes also assessed across the dose range","followUp":"Not reported (duration of individual included RCTs not specified in available abstract/summary)","fundingSource":"Not reported in available sources; notable conflicts of interest: one co-author (TH) was employed by F. Hoffmann-La Roche at time of writing and holds company shares (declared independent conduct); other co-authors received honoraria or travel support from various pharmaceutical companies for unrelated activities"}
Key findings
⚠️ CORRECTION NOTICE — PRIMARY DATA INTEGRITY ALERT: This summary pertains to a published correction (Lancet Psychiatry, June 12, 2026) to the original article (Lancet Psychiatry 2026; 13: 485–95). The correction discloses that modafinil data in Figure 5 of the original publication was incorrectly labeled and presented as data for guanfacine. The graph caption has been updated in the online version to correctly attribute these data to modafinil. All quantitative findings reported below derive from supplementary sources describing the original article and should be interpreted with caution pending review of the corrected figure. From the original article (pre-correction findings, as described in secondary sources): The study identified dose–effect plateaus for several ADHD medications. In children/adolescents, methylphenidate reached peak efficacy at approximately 45 mg/day, amphetamines at approximately 25 mg/day, and lisdexamfetamine at approximately 55 mg/day; guanfacine reached peak efficacy at approximately 4 mg/day. In adults, amphetamines reached peak efficacy at approximately 50 mg/day, while methylphenidate efficacy increased without a clearly identified plateau. No evidence was found that exceeding FDA-licensed maximum doses improved average efficacy for any drug or age group; higher doses were associated with increased side effects. Specific p-values, confidence intervals, and standardized mean differences for individual agents are not reported in available sources. NNT/NNH: not reported. Because the correction specifically involves the modafinil/guanfacine figure, any previously reported or interpreted dose–effect data attributed to guanfacine in Figure 5 must be treated as unreliable until readers consult the corrected online version.
Practice implications
CRITICAL CAVEAT: This is a correction notice, not a primary research article. The correction identifies a material data labeling error in Figure 5, in which modafinil dose–effect data were incorrectly displayed under the guanfacine label. Clinicians and researchers who read, cited, or applied the original article's Figure 5 — particularly any conclusions drawn specifically about guanfacine's dose–effect curve — should revisit the corrected online version before using those data in clinical decision-making, guideline development, or further research. The remainder of the study's figures and findings are not stated to be affected by this correction, though only the online version has been updated. The study's broader conclusions — that subtherapeutic dosing is a clinical concern and that exceeding licensed maximum doses does not improve average outcomes — were derived from a large, methodologically rigorous network meta-analysis and appear robust to this isolated figure correction. However, as only the abstract and secondary sources are available for this summary, the full scope of the correction's impact cannot be independently verified. Findings apply to adults and children/adolescents (≥5 years) treated with oral ADHD monotherapy; children <5 years, those on polypharmacy, and highly treatment-resistant patients were not adequately represented. __LIMITATIONS__ CORRECTION-SPECIFIC: A confirmed data labeling error in Figure 5 (modafinil data attributed to guanfacine) undermines confidence in at least one drug's dose–effect presentation in the original publication; the full downstream impact on reader interpretation is unknown.,Correction applied only to the online version as of June 12, 2026; print versions and any citations or derivative works based on the original figure remain uncorrected.,Children aged 5–<12 years and adolescents aged 12–<18 years could not be analyzed separately due to insufficient sample sizes, limiting age-stratified precision.,Separate analyses by rater type (clinician, parent, teacher, self-report) were not feasible; a hierarchical rater prioritization approach was used instead, which may introduce rating-source bias.,Network meta-analysis relies on the transitivity assumption; unverified effect modifier imbalance across the network could bias indirect comparisons.,Follow-up duration of included RCTs is not reported, limiting conclusions about long-term dose–effect relationships.,RCT populations may not represent real-world ADHD patients; trial eligibility criteria can systematically exclude patients with comorbidities or those at extremes of the dose range.,Conflict of interest: one co-author employed by and holding equity in a pharmaceutical company (Roche) at time of writing, though the study was declared independent of that company.,Only the abstract and secondary sources were available for this teaching summary; full methodological details, effect sizes, confidence intervals, and the corrected figure itself could not be directly reviewed. __BOTTOM_LINE_CATEGORY__ low_certainty __BOTTOM_LINE__ This correction notice alerts clinicians that Figure 5 of Nourredine et al. (Lancet Psychiatry 2026; 13: 485–95) contained a labeling error in which modafinil dose–effect data were incorrectly presented as guanfacine data; the online version has been corrected as of June 12, 2026. Any clinical or research conclusions drawn specifically from the original Figure 5 guanfacine data should be revisited using the corrected version. The study's broader, methodologically robust finding — that exceeding licensed maximum doses of ADHD medications does not improve average efficacy and is associated with increased adverse effects — is not reported to be affected by this correction.