Adjunctive cariprazine for the treatment of major depressive disorder: Number needed to treat, number needed to harm, and likelihood to be helped or harmed

L Citrome,I Reda,M Kerolous
DOI: https://doi.org/10.1016/j.jad.2024.10.040
2025-01-15
Abstract:Background: The number needed to treat (NNT) for efficacy and number needed to harm (NNH) for tolerability/safety were evaluated for adjunctive cariprazine in major depressive disorder (MDD). Methods: Data were extracted from five randomized, double-blind, placebo-controlled trials of adjunctive cariprazine in MDD. NNTs (response, remission, severity shift) and NNHs (discontinuations due to adverse events [AEs], AEs, laboratory shifts) were determined in dose groupings; likelihood to be helped/harmed (LHH) was calculated. Results: NNTs (95 % CI) for adjunctive cariprazine versus placebo were statistically significant at week 6/early termination for response on the Montgomery-Åsberg Depression Rating Scale (MADRS), as defined by a decrease in total score ≥ 50 % (doses ≥ 1 mg/d = 12 [9-21]; 1-2 mg/d = 12 [8-25]; 2-4.5 mg/d = 14 [9-43]) and other response/remission outcomes. NNHs for cariprazine versus placebo were generally ≥ 10 for AEs that were statistically significant; an apparent dose-response was seen for akathisia (lower dose = 24 [17-43]; higher dose = 9 [7-11]). LHHs were ≥ 1 (acceptable benefit/harm ratio) for MADRS total score response versus most important cariprazine AEs in most dose groupings. For response versus discontinuation because of an AE, adjunctive cariprazine 1-2 mg/d had a more favorable response/tolerability profile in indirect comparison with other approved atypical antipsychotics. Limitations: Post hoc analysis; indirect comparisons. Conclusions: Patients receiving adjunctive cariprazine encountered benefits more often than harms; NNT values at week 6/early termination were statistically significant versus placebo on response/remission outcomes across dose groupings from the five pooled studies. Adjunctive cariprazine was well tolerated; NNH values versus placebo were generally > 10, with better akathisia tolerability in the lower-dose range.
What problem does this paper attempt to address?