A randomized pilot found lower reported cocaine craving with lisdexamfetamine without a significant reduction in cocaine use in the randomized sample. Craving and a urine-based use endpoint should not be substituted for each other.
The trial assigned 43 cocaine-dependent participants to placebo or lisdexamfetamine, 70 mg, for 14 weeks. There were 21 participants in the placebo group and 22 in the active group.
Cocaine use was assessed through urine testing for benzoylecgonine, a cocaine metabolite. This was not urine amphetamine measurement, a blood cocaine concentration, or an individual impairment test.
The randomized sample showed no significant difference in cocaine use between medication conditions. Participants receiving lisdexamfetamine nevertheless reported significantly less craving than those receiving placebo.
These findings can coexist. A subjective urge may change without a detectable change in the study’s use endpoint, and a urine metabolite measure has a different time window and meaning from a current craving rating.
The authors discussed a benefit among people who completed the full lisdexamfetamine course. Completion, however, is a post-randomization characteristic. A completer comparison does not carry the same protection against selection bias as the original randomized analysis.
The study did not achieve its planned sample size, and treatment completion was limited. These features reduce precision and complicate interpretation of subgroup observations.
Diarrhea, headache and anxiety were reported more frequently with lisdexamfetamine. Blood pressure, heart rate and body weight did not differ significantly between conditions. Those nonsignificant averages should not erase the symptom differences or establish universal safety.
The trial supports further investigation while preserving its unsuccessful randomized cocaine-use comparison. It does not establish lisdexamfetamine as an effective cocaine-use treatment on this evidence alone or provide a basis for interpreting a particular person’s functional state from a specimen.
Related reading: Self-report and urine outcomes can disagree.