Background: No head-to-head clinical trials have been published comparing guanfacine extended release (GXR) and atomoxetine (ATX): two nonstimulants approved for the treatment of attention-deficit/hyperactivity disorder (ADHD). However, other study designs or methods could be used to indirectly compare these two medications. Matching-adjusted indirect comparison (MAIC) is a recent methodology that utilizes individual patient data (IPD) from clinical trials for one treatment and published aggregate data from another treatment to estimate the relative efficacy of both, providing rapid, reliable comparative efficacy results.
Objective: The aim of this study was to compare the efficacy of GXR and ATX for the treatment of ADHD using MAIC.
Study design: A systematic literature search was conducted to identify ATX and GXR trials published through December 2012. Studies were selected for MAIC analyses on the basis of having comparable trial characteristics and study designs. Summary data from selected ATX trials and IPD from selected GXR trials were used. MAIC methodology ensured comparable populations: target doses for the 'base case' comparison were selected on the basis of maximum effective dosage ranges from the US FDA-approved product labels (GXR 0.09-0.12 mg/kg/day, ATX 1.2 mg/kg/day for children and adolescents weighing ≤70 kg). Individuals from GXR trials were selected if they matched inclusion/exclusion criteria from selected ATX trials; selected GXR IPD were then re-weighted to match the published ATX trial mean baseline characteristics and placebo outcomes. Sensitivity analyses were conducted, examining different dosage ranges and repeating the analysis in a larger number of trials, allowing for larger and more heterogeneous trial populations.
Main outcome measure: The primary outcome measure was change in ADHD Rating Scale IV (ADHD-RS-IV) total score.
Results: Using MAIC in the base case comparison, significantly greater reductions in mean (standard error; SE) ADHD-RS-IV total scores from baseline to end of study were observed in patients treated with GXR relative to ATX [-7.0 (2.2); p < 0.01]. Significantly greater reductions for GXR over ATX were also demonstrated for hyperactivity/impulsivity [-3.8 (1.2); p < 0.01] and inattention [-3.2 (1.3); p < 0.05] subscales of the ADHD-RS-IV. Similar results were observed in MAIC sensitivity analyses evaluating other dosage ranges and using more heterogeneous trial populations (e.g., larger randomized sample, broader subject weight range, additional trials). Mean (SE) decreases in ADHD-RS-IV total scores were greater for GXR relative to ATX when including IPD for those administered GXR at lower than target dosage (0.075-0.090 mg/kg/day) compared with ATX at target dosage (1.2 mg/kg/day), with a relative improvement of -6.0 (2.7) (p < 0.05). Reductions in ADHD-RS-IV total scores were also greater for GXR in another MAIC examining GXR at target dosage (0.09-0.12 mg/kg/day) and a broader range of ATX dosages (including three additional trials evaluating ATX ≥1.2 mg/kg/day); relative improvement for GXR versus ATX administered at target dosage or higher was -7.6 (1.4) (p < 0.01).
Conclusion: After adjusting for difference in baseline trial characteristics using MAIC, GXR appears to be more efficacious than ATX for the treatment of ADHD. Results were consistent in a variety of dosage range comparisons and within increasingly heterogeneous trial populations.