Clinical utility of ADHD symptom thresholds to assess normalization of executive function with lisdexamfetamine dimesylate treatment in adults

Abstract

Objectives: This analysis assessed the relationship of various cutoff scores of the ADHD Rating Scale IV (ADHD-RS-IV) to levels of improvement in ADHD-related executive function (EF), measured by the Brown ADD Scale for Adults (BADDS), which may provide a measure of clinically meaningful EF improvement after ADHD treatment.

Methods: Post hoc analysis of a 4-week, open-label, dose-optimization phase in a double-blind, placebo-controlled study of lisdexamfetamine dimesylate (LDX) in adults with ADHD. The BADDS for Adults, a validated, normed, self-report measure of EF in ADHD, provides a qualitative measure to rate treatment progress. The ADHD-RS-IV assesses current symptom status based on DSM-IV criteria. Postbaseline ADHD-RS-IV scores were categorized according to four cutoff criteria of symptom remission: (1) ADHD-RS-IV total score ≤ 18; (2) ADHD-RS-IV total score ≤ 10; (3) no ADHD-RS-IV item scored >1; and (4) ADHD-RS-IV total score ≤ 18 and ≤ 2 items per subscale with a score of 2. Sensitivity and specificity of criteria for identifying participants with optimal BADDS scores were assessed using receiver operating characteristics (ROC). Safety evaluation included treatment-emergent adverse events (TEAEs).

Results: At endpoint, 85/127 participants had optimal BADDS scores. Linear ANOVA indicated limited overlap between BADDS and ADHD-RS-IV scores (r (2) = 0.20; P < 0.0001). Specificity was similar for criteria 1-4 (0.46, 0.39, 0.39, and 0.42), as were ROC (0.699, 0.776, 0.732, and 0.668). Sensitivity was high for criteria 2 and 3 (0.96, 0.92), lower for criteria 1 and 4 (0.72, 0.75). TEAEs were consistent with those of stimulants.

Conclusion: Criteria 2 and 3 had satisfactorily high sensitivity, but no criteria had adequate specificity. AUC comparison indicated that criteria 2 and 3 ADHD-RS-IV thresholds may be more accurate assessments of EF normalization as measured by the BADDS. The open-label design, small sample size, and selection criteria limit the applicability of these results to a larger treatment population.