Does Stopping d-Amphetamine Increase Methamphetamine Craving? What a 34-Person Residential Trial Actually Tested

Short answer: In a small randomized residential experiment, people with a history of frequent methamphetamine use first received supervised oral d-amphetamine. When some participants were then switched to placebo, their reported methamphetamine craving increased relative to participants who continued d-amphetamine. The primary group-by-time craving interaction was reported as p = 0.0354. Yet the experiment did not establish a blood concentration associated with craving, a reliable clock for a person’s last methamphetamine use, or a test of legal intoxication or driving impairment.

The distinction matters in toxicology and litigation. A person can report craving or withdrawal without a contemporaneous measured methamphetamine concentration, and a physiological or psychological response to stopping a prescribed stimulant is not a chemical identification of that drug in an evidentiary specimen. The underlying source is Thompson and colleagues’ 2021 pilot clinical trial, Utility of a controlled amphetamine withdrawal paradigm among adults who use methamphetamine. Its measurements, limitations, and negative findings must be read together.

What scientific question did the experiment answer?

Investigators wanted to know whether abruptly ending a standardized period of d-amphetamine administration could produce measurable withdrawal-related responses in residential participants with substantial histories of methamphetamine use. This is narrower than demonstrating that d-amphetamine is an effective maintenance treatment for methamphetamine use disorder. It is also narrower than identifying which drug, if any, a particular person consumed immediately before a crash, arrest, workplace event, or medical presentation.

The experiment was designed as a possible model for later withdrawal-treatment research. Standardizing administered medication before randomization reduces one source of uncertainty in exposure history. It does not remove differences in prior methamphetamine use, age, health, metabolism, drug tolerance, mood, or how participants use symptom-rating scales. Researchers therefore collected multiple symptom and physiological outcomes rather than treating a single measure as a comprehensive withdrawal diagnosis.

Who participated, and how many independent people were studied?

Of 40 eligible adults who consented, 34 entered the study. Their demographics included approximately 47% women, and the enrolled group was reported as entirely White. Participants had recent frequent methamphetamine use and entered a supervised residential treatment setting. Study eligibility required self-report of use on at least 15 of the prior 30 days and at least half a gram per week, together with a positive urine screen for amphetamines in the month before enrollment. That screen did not supply a quantitative contemporaneous methamphetamine value.

The randomization compared 17 participants assigned to continued d-amphetamine with 17 assigned to discontinue it and receive placebo. Three weekly assessments, repeated visits, scores, and pulse measurements are observations on those people—not additional independent participants. Some comparisons included fewer people because of attrition: a source table of intake and post-stabilization measures displayed 12 participants per group, and only eight in each arm completed the full four-week course.

Crucially, the researchers did not administer a formal standardized diagnostic assessment to establish that every participant met a defined methamphetamine-use-disorder diagnosis. These were selected adults with an extensive methamphetamine-use history. The results should not be generalized to every patient, all racial or ethnic groups, adolescents, healthy drivers, or forensic examinees without additional evidence.

How was d-amphetamine administered and discontinued?

Every participant first received observed, immediate-release oral d-amphetamine during week one. The starting schedule was 10 mg twice daily; the daily amount rose by 20 mg until 60 mg/day, given as 30 mg twice daily, was reached on day three. This was a controlled clinical exposure, not a verified continuation of a participant’s previous illicit methamphetamine product. d-Amphetamine and methamphetamine are related stimulants but are not interchangeable names for the same molecule.

At the beginning of week two, a blinded randomization assigned one group to continue d-amphetamine while the other received matching placebo capsules. That separation continued through weeks two and three. At the start of week four, both groups received placebo so that all participants ended on a washout phase. The study therefore compared withdrawal from supervised d-amphetamine against continued supervised d-amphetamine in people who previously used methamphetamine. It did not randomize people to new illicit methamphetamine exposure.

Both groups remained in residential treatment with staff-observed medication administration and clinical monitoring. The investigators excluded several potentially interacting medications and important medical conditions. Such intensive supervision differs substantially from an unobserved episode of street-drug exposure or an ordinary roadside encounter.

What did investigators actually measure?

The central outcome was a 0–20 visual-analogue rating of methamphetamine craving over the preceding 24 hours. Two other symptom tools were the Methamphetamine Withdrawal Assessment (MAWA) and the Methamphetamine Selective Severity Assessment (MSSA). Researchers also measured Hamilton anxiety and depression scores, performance on a Stroop color-word task, pulse, postural blood pressure, retention, and spontaneously reported adverse events.

These were symptom, cognitive, and physiological measurements. A reported craving score of 10, for example, is not 10 ng/mL of methamphetamine, d-amphetamine, or any metabolite. An improvement in a questionnaire is not evidence that urine or blood drug concentration decreased by a corresponding amount. Heart rate is a direct physiological measurement, but it is not a drug-specific analytical result: age, hydration, medical conditions, arousal, withdrawal, and other factors may affect it.

The study explicitly stated that urine specimens were not collected to test for illicit substances during the trial because the administered d-amphetamine would confound interpretation of amphetamine testing. This is among the most important forensic limitations. There was no measured time series of parent methamphetamine in properly collected venous whole blood, separated plasma, or serum and no method-validated comparison of methamphetamine to d-amphetamine in those matrices.

The design and the changing denominators

Failure to track denominators is an avoidable expert-review error. A graph with observations on numerous days does not turn 16 final completers into a large independent sample. Nor should 34 enrolled participants be described as 34 successfully completed month-long medication courses. Repeated-measures statistical methods can retain observed data from partial completers, but cannot automatically replace data that were never collected.

What happened to methamphetamine craving after the placebo switch?

During week one, before the groups received different study medications, craving and several symptom scores fell over time. This preliminary improvement occurred in both eventual assignments, consistent with effects of the residential environment, supervised treatment, standardized stimulant exposure, and natural change over time. It cannot be credited solely to an active-versus-placebo comparison because everyone was given active d-amphetamine in that week.

After randomization, the researchers reported a craving group-by-time interaction of p = 0.0354 in the immediate discontinuation comparison. Their overall interpretation was that craving rose or remained higher after the switch to placebo compared with continued d-amphetamine. A later reported comparison at week-two visit three also favored continued medication, with p = 0.0251 for craving. That is group-level evidence consistent with an effect of stopping a standardized stimulant regimen on self-reported craving in this particular setting.

It is not proof that every participant abruptly developed intense withdrawal symptoms. The statistical model concerns an average trajectory in a small and shrinking sample. An individual’s rating can remain low, fall, or move differently from the group pattern. Moreover, craving is a subjective construct with clinical value but is not a chemical fingerprint; it cannot distinguish recent methamphetamine ingestion from its absence without independent toxicological and case evidence.

Which outcomes were positive, and which remained nonsignificant?

A statistically significant result on one outcome does not erase genuine null results on other prespecified measures. MAWA scores declined over time in both groups; MSSA scores also fell within both groups but were not significantly different between arms at the sampled time points. Anxiety improved over time within arms without a demonstrated between-treatment advantage. The Stroop cognitive test showed no significant study effect. These separate domains should not be collapsed into a blanket conclusion that medication either prevented or caused all features of a methamphetamine withdrawal syndrome.

What did the trial show about heart rate and tolerability?

The trial’s cardiovascular observations deserve as much attention as craving. Even during the initial week, when everyone was receiving d-amphetamine, the group that would later receive placebo had higher average heart rates than the continuation group. In a post-stabilization source table, standing heart rate averaged 107.5 versus 93.4 beats/minute in eventual placebo and continuation groups, respectively (p = 0.0048). Baseline age also differed: mean age was 34.18 versus 40.25 years (placebo versus continuation; p = 0.0450).

Those differences caution against attributing every pulse contrast after randomization to the medication switch. Physiological differences were present even before different treatments began. The investigators reported that most heart-rate measures diminished over time, and most blood-pressure contrasts were not significant. A single study participant was removed for a clinically concerning elevated pulse despite being asymptomatic; the investigators described reported adverse events as mild.

Increased heart rate was recorded six times in the eventual placebo arm and four times in the continuation arm during the initial week. Counts of reported events are not necessarily counts of distinct individuals, and the first-week labels refer to future assignment, not two different drugs then being received. This small pilot cannot rule out uncommon cardiovascular complications or establish that 60 mg/day is appropriate for unsupervised treatment.

Why do the negative retention results matter?

Only 16 of the 34 enrolled participants finished the study, eight in each assignment. That is roughly 53% attrition. Reasons reported included voluntary withdrawal, facility-rule violations, family emergencies, refusal of medication, seeking outside medical care, and one investigator-directed discharge for elevated pulse. The study did not detect a retention difference between assignments during week one (log-rank p = 0.8921) or after the week-two transition (p = 0.7975).

Absence of statistical evidence for a retention difference is not evidence that dropout was identical under all plausible study conditions. The trial was small, and people who leave may differ systematically from those who remain. Repeated-measures analysis assumed missingness at random. If symptoms or treatment tolerance contributed to leaving the residential program, that assumption could matter. Group averages near the end are therefore less reliable as a complete description of all enrolled patients’ courses.

Does this establish a treatment for methamphetamine use disorder?

No. The objective was primarily feasibility and the production of a controlled discontinuation response for future medication research. The authors interpreted the regimen as relatively well tolerated under their selection criteria and residential supervision, and suggested that a controlled withdrawal model could be informative. They did not demonstrate durable abstinence, relapse prevention, recovery of occupational performance, or safety of routine off-label maintenance outside the trial.

It is especially important not to treat the administered dose as a dosing recommendation. Participants had substantial stimulant-use histories, monitoring, standardized capsules, and exclusion criteria. The pilot was not designed to evaluate individualized prescribing, was not powered for rare adverse reactions, and did not determine the best dose for treating a withdrawal syndrome. Its positive craving observation motivates further research rather than establishing clinical practice for a different patient.

Were methamphetamine and d-amphetamine analytically distinguished?

The distinction exists in the experimental design: the histories concerned methamphetamine, while the supervised medication was pharmaceutical d-amphetamine. The study did not publish serial quantitative confirmatory assays distinguishing methamphetamine from d-amphetamine during the blinded medication periods. Investigators withheld repeated illicit-substance urine testing because administered study medication would complicate amphetamine urine-screen interpretation.

In a forensic case, a urine class screen may react to multiple compounds and does not necessarily distinguish a parent chemical, metabolite, isomer, or specific medication. Proper chemical identification would require the actual method, chromatographic and mass-spectral identification criteria when applicable, relevant reference standards, and the correct specimen and collection time. For a legal whole-blood claim, a validated method for that actual whole-blood matrix is needed; plasma, serum, and urine data cannot be silently substituted.

Neither the prescribed 60 mg daily dose nor the self-reported methamphetamine history supplies a case-specific blood value. Converting the dose to a hypothetical ng/mL concentration would require pharmacokinetic assumptions and still would not be a direct measurement of a legal specimen. The design contains no such validated person-specific bridge.

Could craving reveal a person’s last use or current impairment?

No. Craving is a response that can vary over hours and days, even in an observed abstinence program. This study assessed desire for methamphetamine over the prior 24 hours, not the concentration of methamphetamine circulating at the time of the rating. A person may report withdrawal symptoms after earlier exposure or experience craving for reasons not resolved by the present data. Conversely, lack of reported craving cannot establish that no stimulant was taken.

Likewise, the study supplied no impairment calibration. It did not test standardized roadside driving, a validated simulator with individual classification thresholds, reaction-to-hazard during operation of a vehicle, or criminal-event behavior. One nonsignificant Stroop result does not prove all cognitive skills were normal. One significant craving contrast does not prove a particular person’s judgment or driving was impaired. Those are different constructs with different evidentiary requirements.

How should a forensic expert handle a similar real-world case?

Start with the exact scientific question. If the dispute is whether a drug was chemically present, the core evidence is the properly collected specimen, laboratory method and chromatographic confirmation, calibration, analyte identity, numerical result and uncertainty. If the dispute is whether a person had withdrawal symptoms, review history, examination, timing, medication records and alternative explanations. If the dispute is the cause of a crash, examine the event-specific evidence rather than promoting a small treatment trial into a causal reconstruction.

A clinical history of prescribed d-amphetamine and suspected illicit methamphetamine exposure raises questions about molecular identity and laboratory interpretive scope. The same numerical concentration in two different physical matrices is not necessarily equivalent. The site guides on biological specimens and matrix distinctions, confirmatory drug testing, and therapeutic drug monitoring explain why method, timing and collection context need to travel with the reported finding.

Remember, too, that symptoms may improve during structured residential care for several reasons, while age, hydration, preexisting cardiovascular differences and selective attrition may affect comparisons. A scientifically credible report should describe the original outcome as reported, preserve the actual group denominators, and distinguish inferential limits from directly observed data.

What are the strongest and weakest aspects of this evidence?

Strengths include a genuinely randomized, blinded medication switch after a common supervised stabilization week; direct observation of dosing; a residential environment; serial measurement of multiple clinically relevant symptom domains; and explicit reporting of nonsignificant results. Those design elements make the source more informative than a simple unsupervised recollection of methamphetamine discontinuation.

Limitations include just 34 enrolled people, 16 full completers, a demographically narrow group, age and pulse imbalances, the absence of a formal use-disorder diagnostic assessment for everyone, and incomplete observation of why some participants dropped out. The study also lacked ongoing chemical differentiation of illicit methamphetamine and administered d-amphetamine and did not provide measured blood/plasma/serum drug pharmacokinetics. Its primary craving effect was detected in a small pilot with multiple outcomes and time points.

Accordingly, the defensible inference is that controlled cessation of oral d-amphetamine can alter reported methamphetamine craving trajectories in selected residential patients. The evidence does not prove universal withdrawal, durable treatment efficacy, a reproducible impairment threshold, or a person’s drug concentration and responsibility for a specific event.

Frequently asked forensic questions

1. Did the study actually administer methamphetamine to the participants?

No. The adults reported frequent prior methamphetamine use, but the randomized medication was immediate-release oral d-amphetamine. The compounds should not be relabeled interchangeably in a report or deposition.

2. Was the randomized sample 34 people or 24 people?

The study enrolled and randomized 34 people, 17 per arm. A later source table contained data from 12 per arm, and only eight per arm finished the four-week course. Each denominator describes a different analysis stage.

3. Did craving rise when the administered d-amphetamine was stopped?

The reported group-by-time craving interaction during the immediate discontinuation comparison was significant (p = 0.0354), with the overall trajectory indicating higher craving after switching to placebo. This is a group result, not a deterministic prediction for each participant.

4. Were all withdrawal and cognitive measures significantly worse after the switch?

No. There were significant findings for craving and a selected MAWA visit, but MSSA treatment-group comparisons, anxiety treatment-group contrasts, Stroop cognition and most systolic blood-pressure comparisons did not show significant differences. The data do not support a blanket effect on every function.

5. Was urine or blood used to measure methamphetamine during the trial?

Eligibility included an earlier positive amphetamine-class urine screen. The researchers did not collect ongoing illicit-drug urine tests during the blinded treatment because the study medication would complicate interpretation. They did not report quantitative contemporaneous forensic whole-blood methamphetamine or separate plasma/serum values.

6. Can an elevated heart rate identify whether someone used methamphetamine or stopped d-amphetamine?

No. Even during the shared medication-stabilization week, future randomized arms differed in age and heart rate. Pulse is not drug-specific. Medication records, clinical context, properly identified chemical findings and alternative medical causes require separate evaluation.

7. Can this trial establish impaired driving, relapse, or the precise last drug dose?

No. It evaluated supervised medication discontinuation, symptom ratings, mood, basic cognitive tasks and vital signs. It did not determine individual forensic blood concentration, actual driving, relapse over a validated follow-up period, or causation of a case-specific event.

Bottom line for scientific and legal interpretation

This controlled experiment supports a narrow finding: in selected residential adults with frequent methamphetamine-use histories, continuing d-amphetamine after common stabilization was associated with less reported craving than abruptly switching to placebo at relevant points in week two. The study’s genuine null outcomes and major attrition limit broader claims.

The most important boundary is analytical. Symptoms, prescribed dosage and routine vital signs are not measurements of methamphetamine in an identified specimen. An expert should not invent a drug concentration, assume urine or plasma stands in for legal whole blood, or turn a group craving effect into individualized impaired-driving or event-causation testimony. Consult the complete original Thompson et al. primary study for the methods, source tables, statistical analyses and limitations.