Regulation of cocaine self-administration in humans: lack of evidence for loading and maintenance phases

Abstract

Background: In rodents, cocaine self-administration under a fixed-ratio schedule and with timeout intervals limited to the duration of the infusions is characterized by an initial burst of drug intake (loading) followed by more stable infusion rates (maintenance). We sought to examine whether similar phases might characterize self-regulated cocaine use in humans.

Methods: 31 Non-treatment seeking, cocaine dependent subjects participated in three (8, 16, and 32 mg/70 kg/infusion), self-regulated, 2-h cocaine self-administration sessions under a fixed-ratio 1, 5-min timeout schedule. Data were assessed for visual (e.g., by graphs of cumulative numbers of infusions) and statistical evidence of change in phase (by step-function analyses of individual infusion rates).

Results: Graphs of cumulative infusions over time suggested a single, linear rate of self-administration over 2h at each cocaine dose. Statistical analyses of infusion data by generalized estimating equation (GEE) models also failed to support a loading/maintenance pattern (suggesting, if anything, the possibility of increasing infusion rates over time).

Conclusions: Our findings fail to support the existence of distinct loading and maintenance phases of self-regulated cocaine administration in humans at behaviorally relevant doses. Several factors may account for these observations including differences between humans and rodents in self-regulated drug intake.

Fig. 1a

Average number of cumulative cocaine infusions for 8 mg (blue line), 16 mg (pink line) and 32 mg (orange line) depicted over time (i.e., 120 min self-administration period). Cumulative infusion curves appear linear and without evidence of discrete loading/maintenance phases.

Fig. 1b

Cumulative cocaine infusions for each of 10 representative subjects depicted over time (i.e., 120 min self-administration period) for 8 mg (top), 16 mg (middle), and 32 mg (bottom) doses. Individual cumulative infusion curves are without evidence of distinct loading/maintenance phases.

Fig. 2a

Graph depicting hypothesized probabilities of cocaine infusions over time for 8 mg (blue dashed line), 16 mg (pink dashed line), and 32 mg (orange dashed line) doses as fit by a step-function model. Specifically, loading to maintenance phases are predicted to be associated with descending “steps” for each cocaine dose (i.e., with the probability of infusions being higher for lower as compared to higher doses, with the probability of infusions being higher for loading as compared to maintenance periods, and with the transition being later for lower as compared to higher doses).

Fig. 2b

Graph depicting actual probabilities of cocaine infusions over time (minutes) for 8 mg (blue line), 16 mg (pink line), and 32 mg (orange line) doses as best-fit by “step-point” analyses. In contrast to predictions, step-function modeling revealed statistically significant ascending “steps” for each cocaine dose.

Fig. 2b

Graph depicting actual probabilities of cocaine infusions over time (minutes) for 8 mg (blue line), 16 mg (pink line), and 32 mg (orange line) doses as best-fit by “step-point” analyses. In contrast to predictions, step-function modeling revealed statistically significant ascending “steps” for each cocaine dose.