Experiencing a positive effect from amphetamine did not reliably identify people who also liked alcohol or THC in a small controlled study. The investigators expected responses across drug classes to correlate. Most did not, despite clear average effects from the administered drugs.
The result distinguishes a group-average drug effect from the ability to predict an individual’s response to another drug. It does not establish that the drugs lack shared mechanisms, and it was not a test of taking them together.
What was the trial design?
Wardle, Marcus and de Wit studied 24 healthy social drinkers, 12 men and 12 women, aged 21–31 years. Each completed six 5.5-hour laboratory sessions in a randomized, counterbalanced, double-blind, double-dummy design: one session with each active drug and three corresponding placebo sessions. Sessions were separated by at least 48 hours.
The oral doses were 20 mg d-amphetamine in a capsule, 7.5 mg delta-9-tetrahydrocannabinol in a capsule, and alcohol in a beverage at 0.8 g/kg for men and 0.7 g/kg for women. The sex-specific alcohol doses are specified in the methods; compressing the protocol to “0.8 g/kg for everyone” would be inaccurate.
Capsules were given at 11:30 a.m. and beverages consumed over 15 minutes beginning at 12:30 p.m. This staggered schedule aimed to align peak subjective effects around 1 p.m., approximately 90 minutes after the capsule and 30 minutes after drinking began. Active drugs were administered on separate sessions, rather than in combination.
Who does the sample represent?
Participants were selected for light drug use and social drinking, with no lifetime dependence on alcohol, cannabis or stimulants. Eighty-three percent reported lifetime cannabis use; 33% reported lifetime nonmedical stimulant use. They were not a sample of people with stimulant use disorder.
Twenty-eight participants were randomized, and four discontinued: three because of scheduling conflicts and one because of alcohol-related nausea and vomiting. Replacements yielded the final 24-person analyzable sample. The participant count, rather than the number of sessions or questionnaire ratings, is the relevant denominator for the individual-difference analyses.
The trial was registered after enrollment. The authors explained that they initially had not regarded this acute laboratory protocol as a clinical trial. That timing is part of the study’s methodological record.
What did the researchers measure?
Participants rated feeling a drug effect, liking, disliking, feeling high and wanting more on visual analog scales. Investigators compared changes around the planned peak with changes in the nearest paired placebo session. A separate questionnaire assessed drug-related response patterns, including amphetamine-like stimulation.
Using three placebo sessions helped avoid creating artificial correlations by subtracting the same unusual placebo response from every active-drug response. The analysis concerned placebo-adjusted subjective ratings, not blood amphetamine concentrations or an objective measure of impairment.
What happened after d-amphetamine?
Compared with placebo, d-amphetamine increased ratings of feeling a drug effect, liking the effect and wanting more. The reported probabilities were below .001, below .001 and .001, respectively. Its increase in feeling high had p=.004 and was treated as marginal under the investigators’ stricter correction for multiple group-level comparisons.
Those findings describe average responses within this selected sample at one dose. They do not imply that every participant liked amphetamine or wanted another dose, nor do they establish later dependence.
Did the same people respond positively across drugs?
The primary correlations generally failed to support that prediction. For example, placebo-adjusted liking after amphetamine correlated −.131 with liking after alcohol and −.091 with liking after THC. Neither provided evidence of the hypothesized positive cross-drug relationship.
The paper also explored whether people who experienced stimulant-like alcohol effects showed a closer relationship between alcohol and amphetamine stimulation. In the 12-person stimulant-responder subgroup, the correlation was .601, with p=.05; in the other subgroup it was .008. However, the direct comparison between the two correlation coefficients was not significant, p=.14.
That last test matters. A relationship appearing nominally significant in one small subgroup and not another does not establish that the subgroups differ. The authors regarded the finding as preliminary and in need of replication.
Did personality explain a general susceptibility?
No measured personality trait predicted responses across multiple drug types. Exploratory analyses suggested an alcohol-specific relationship involving constraint, a trait related to behavioral control, but did not establish a personality profile that predicted broadly positive responses to all three substances.
The study used multiple comparisons and distinguished corrected significance from marginal findings. Treating every p-value below .05 as equally confirmatory would overstate its exploratory results.
What limits the conclusions?
The sample was small, young and selected. Only one dose of each drug was tested. Alcohol produced stronger overall “feel drug” ratings than the other drugs despite attempts to match intensity. Oral THC responses were especially variable, and oral isolated THC does not reproduce every feature of inhaled cannabis.
Blinding was imperfect: 70.8% identified a stimulant somewhere in the amphetamine session, and 91.7% identified alcohol in its active session. Expectancies and pre-existing preferences could therefore influence subjective ratings. Nonsignificant correlations in 24 people also do not prove that the true relationships are exactly zero.
The study did not assess combined-use safety, driving, later substance use or dependence outcomes. Its useful contribution is narrower: positive average amphetamine effects did not translate into a reliable general predictor of positive responses to other drug classes in this experiment.
Scientific source
Wardle MC, Marcus BA, de Wit H. A Preliminary Investigation of Individual Differences in Subjective Responses to D-Amphetamine, Alcohol, and Delta-9-Tetrahydrocannabinol Using a Within-Subjects Randomized Trial. PLoS ONE. 2015;10(10):e0140501. doi:10.1371/journal.pone.0140501. Full text. Historical exploratory controlled human experiment.
Leave a Reply