A medication trial can show both improved mood scores and reduced drug use without proving that the mood improvement caused the drug-use change. A secondary ADAPT-2 analysis examined that possibility using mediation models, producing a promising hypothesis with substantial uncertainty.
Was this an independent treatment trial?
No. Jha and colleagues reanalyzed the existing ADAPT-2 cohort, using only the first six-week stage and only participants with at least mild depressive symptoms at baseline: a Patient Health Questionnaire-9 score of five or higher. The analytic sample was 326, including 84 assigned to active medication and 242 to placebo.
The active regimen combined 380 mg extended-release naltrexone by intramuscular injection every three weeks with oral extended-release bupropion targeted to 450 mg daily. This selected, opioid-free-at-randomization population had moderate or severe methamphetamine use disorder. The analysis does not establish effects in all people with depression or all people using methamphetamine.
What improved, and how was it measured?
Depressive symptoms were measured weekly by self-reported PHQ-9, using a one-week recall period rather than the usual two-week period. The estimated medication-versus-placebo difference at week four was −2.52 PHQ-9 points, standard error 0.81. The study also examined depression response, defined as at least a 50% score reduction, and remission, defined as a score below five.
These are related but different outcomes. A continuous trajectory of improvement can carry information that a threshold-based response label discards. PHQ-9 symptom categories are also not an independent diagnostic interview establishing major depressive disorder in each participant.
How was mood related to later methamphetamine response?
Methamphetamine treatment response required at least three negative urine screens among four possible screens during weeks five and six. Twenty-three of 326 participants met that endpoint: 14/84 with active medication and 9/242 with placebo. It was not a definition of continuous abstinence throughout treatment.
Greater early PHQ-9 improvement, represented by a modeled slope, was associated with later urine response: odds ratio 3.74, 95% confidence limits 1.28–10.93. By contrast, the separate week-four depression-response and remission categories were not significantly associated with methamphetamine response: their confidence intervals included one. Both the positive slope result and those null categorical findings belong in the account.
What does “24.8% mediated” mean?
The mediation model estimated that early depression change accounted for 24.8% of the medication effect on the defined urine endpoint. Its bootstrapped 95% confidence interval was 6.7%–60.3%. That wide interval signals considerable uncertainty; it is not a precise decomposition of how the medicines work.
Treatment assignment was randomized, but the amount of each person’s mood improvement was not. Mood change may share causes with reduced use, better engagement or other changes during treatment. The temporal ordering helps frame a mechanism hypothesis, but does not by itself eliminate mediator–outcome confounding or establish that directly treating depression would reproduce the same drug-use benefit.
What additional checks temper the interpretation?
The exploratory cross-lagged analysis found one significant mood-response-to-next-visit urine association, from week three to week four; other cross-domain estimates were nonsignificant. The paper also reported missing PHQ-9 values, increasing to 69 at week four. Missing mood outcomes and the selected baseline symptom threshold constrain inference.
This analysis adds a mechanism question to the original ADAPT-2 efficacy result; it is not an independent replication. It does not validate choosing an antidepressant solely to reduce methamphetamine use or using mood improvement as proof of abstinence or functional recovery.
Source
Jha MK et al. Early Change in Depressive Symptom Severity With Naltrexone-Bupropion Combination and Its Association With Reduction in Methamphetamine Use in ADAPT-2 Trial. Journal of Clinical Psychiatry. 2025;86(3). doi:10.4088/JCP.25m15825. Secondary analysis of historical randomized human-trial data.
Leave a Reply