Does Lisdexamfetamine Produce a Lower Plasma Amphetamine Peak? A 24-Person Direct Comparison

Short answer: No lower peak was demonstrated. In a randomized, blinded, three-session crossover experiment, equimolar oral lisdexamfetamine and immediate-release d-amphetamine produced similarly high modeled peak concentrations and overall exposure to active d-amphetamine in physically separated human plasma. Lisdexamfetamine delayed the start of the plasma increase and moved the estimated peak later, but modeled peak concentrations were 118 versus 120 ng/mL and exposure to infinity was 1,817 versus 1,727 ng·h/mL; investigators did not find a significant difference in either peak or exposure. Peak self-reported drug liking was also similar. These findings do not establish a comparable concentration in unseparated forensic whole blood, a particular person’s time of last ingestion, or the safety or impairment of an individual driver.

The source is the original 2017 clinical investigation by Dolder, Strajhar, Vizeli, Hammann, Odermatt, and Liechti, Pharmacokinetics and Pharmacodynamics of Lisdexamfetamine Compared with D-Amphetamine in Healthy Subjects, published in Frontiers in Pharmacology: DOI 10.3389/fphar.2017.00617. The paper’s Methods, pharmacokinetic table, physiological outcome table, concentration-time plots, and Discussion permit a narrower conclusion than either “the prodrug is weaker” or “the two medications are indistinguishable.”

What did the investigators actually test?

Twenty-four healthy adults, twelve women and twelve men, took part in a randomized, double-blind, placebo-controlled crossover trial. Each participant attended three test days and received a single oral dose of lisdexamfetamine, a single oral dose of d-amphetamine, and a placebo in a balanced order. The sessions were separated by at least seven days. This means the biological sample size was twenty-four people, not seventy-two separate people. Each person contributed many serial concentration samples and repeated ratings; those observations were correlated within that person.

The study compared pharmacokinetics—what happened to circulating active amphetamine over time—with pharmacodynamics—what participants reported and how cardiovascular and pupil measurements changed. It was not a trial of sustained ADHD treatment, a controlled simulated-driving challenge, an assessment of an alleged criminal event, or a study enrolling chronic stimulant users. Its strengths lie in the within-person placebo-controlled design, supervised timed dosing, serial specimens, and simultaneous measurement of subjective and physiological outcomes.

Why does equimolar dosing matter?

The active treatments were 100 mg lisdexamfetamine dimesylate salt and approximately 40.3 mg d-amphetamine sulfate salt. Despite the striking difference between the two labeled masses, each treatment corresponded to about 29.6 mg of d-amphetamine base. Lisdexamfetamine is a covalently linked prodrug that must be converted into pharmacologically active d-amphetamine; an oral capsule’s salt weight is not the same measurement as its delivered active moiety.

Simply comparing “100 mg” against “40 mg” as though these values were equal chemical entities would materially misstate the design. Investigators selected relatively high single doses; the paper specifically acknowledged that 100 mg lisdexamfetamine was above the cited 70 mg maximum recommended daily dose. Because only one equimolar high-dose pair was studied in healthy adults after fasting, this experiment cannot prove identical behavior at lower therapeutic doses, after food, during repeated dosing, or after nonoral administration.

Which blood fraction was chemically measured?

Study personnel placed venous catheters and collected lithium-heparin blood samples before dosing and repeatedly for up to twenty-four hours after administration. Importantly, the collected blood was immediately centrifuged. Investigators isolated plasma and froze that fraction for subsequent quantitative analysis. The amphetamine results in the article therefore describe separated venous plasma, not unseparated venous whole blood. Both originate from venous collection, but their reported concentration matrices are not interchangeable.

Active d-amphetamine in plasma was assayed using ultra-high-pressure liquid chromatography coupled to tandem mass spectrometry. The study gave a lower detection limit of 0.26 ng/mL and a lower quantification limit of 0.78 ng/mL; detailed validation was placed in supplementary methods. Plasma profiles were available for 24 lisdexamfetamine sessions and 23 d-amphetamine sessions because a venous catheter could not be placed in one participant in the latter condition. There were still twenty-four enrolled biological people, and twenty-three—not twenty-four—contributed the comparison’s direct immediate-release pharmacokinetic profile.

What did the investigators measure, and what did they model?

The chromatographic instrument directly quantified d-amphetamine in collected plasma specimens at specified times. The familiar single values labeled peak concentration, time to peak, lag time, half-life, and area under the curve were then summarized using a one-compartment pharmacokinetic model with first-order input and elimination. The paper also considered noncompartmental analyses. It would be inaccurate to present the model’s 118 or 120 ng/mL peak as the exact observed maximum from every individual tube—or as a particular participant’s measured concentration.

The investigators did not determine a lisdexamfetamine-prodrug concentration-time profile in this trial. They analyzed the resulting active d-amphetamine. Thus “plasma amphetamine” in this article means measured active-moiety concentrations after either administered product, not measured concentrations of the intact prodrug. It is also distinct from generic urine amphetamine immunoassay positivity, an oral-fluid screening result, a blood-to-plasma conversion factor, or a kinetic reconstruction in an unrelated subject.

Original numeric evidence: plasma kinetics

The following values reproduce the original Table 1 model summaries. They are geometric means with the article’s accompanying 95% confidence intervals, except for the explicitly identified dosing information. Sample availability differs across treatment arms. The table is newly organized for comparison, rather than a reproduction of a publisher figure.

Individual paired comparisons found an average additional 0.6 ± 0.6 hours for lag and 1.1 ± 1.5 hours for peak time after lisdexamfetamine, expressed as mean ± standard deviation. These are differences in the time course, not proof of smaller total exposure. The source reports no significant difference in modeled maximal plasma concentration or total AUC. It also reports a nonsignificant between-drug absorption-rate constant comparison (p = .07). A nonsignificant comparison is not an equivalence trial establishing that every patient’s time course must match.

Is the later peak the same as a weaker drug effect?

No. A delayed rise can coexist with a comparable maximum and comparable integrated exposure. Lisdexamfetamine first requires conversion into d-amphetamine; after the active compound appears in the circulation, the measured disposition need not be dramatically different. This trial illustrates the distinction with a clear separation between lag or peak timing and the magnitude of the modeled peak.

Those patterns nevertheless have practical interpretive limits. At an identical clock time after supervised ingestion, the two formulations need not have identical plasma concentrations, especially during the ascending phase. At a late sampling time, overlap may be more substantial. Neither situation licenses a forensic examiner to work backward from one concentration in an unknown individual to a precise dose, product, or ingestion time. Interperson variability, dosing conditions, drug interactions, and the absence of a validated subject-specific population model prevent that leap.

What happened to maximal drug liking and vital signs?

The authors also followed visual-analog subjective ratings, heart rate, blood pressure, body temperature, pupils, and adverse effects. Many measures were significantly higher after either active dose than after placebo. Crucially, the maximal “drug liking” ratings were not significantly different between the two active oral formulations in the study. Similar peak self-report is a different proposition from proving identical rates of misuse or identical addiction risk in communities and patients.

Original Table 2 reported mean ± standard error of the mean for twenty-four participants. The new evidence table below selects endpoints showing both large placebo-controlled activity and the relatively similar active-condition peaks; the plasma pharmacokinetic sample count exception described above remains unchanged.

For these selected drug-liking, drug-high, heart-rate, systolic and diastolic blood-pressure measures, the table marks both active conditions as significantly different from placebo (generally p < .001). The table’s original notation also indicates a small between-active-drug distinction for peak diastolic pressure. The comparison should therefore be described as broadly similar rather than exactly identical in every pharmacodynamic component. Not all time points were simultaneous maxima, and these maxima cannot be assigned to a particular blood tube or criminal event.

Why is the time course of subjective effects important?

Drug liking, feeling high, stimulation, and concentration were assessed on scales at repeated times. Their onsets followed the slightly later plasma trajectory after lisdexamfetamine, yet the maximal reported effects overlapped materially across active treatments. This matters because it refutes an overly simple story that slower prodrug activation necessarily means much smaller maximal subjective reinforcement after an equimolar oral dose.

Investigators also noted a time-dependent separation between subjective sensations and plasma exposure: many noticeable subjective stimulant effects waned within approximately eight hours while the measured active plasma amphetamine concentrations remained relatively high. Heart-rate effects could persist longer. The source interprets this as acute tolerance or pharmacodynamic adaptation. A person saying they no longer feel strongly stimulated cannot be assumed chemically negative; conversely, a persistent measured concentration alone cannot identify the magnitude of their contemporaneous subjective feeling or objective ability to perform a complex task.

Does this prove equal abuse liability?

No single study in healthy screened volunteers can prove equal risk of compulsive use across different populations, settings, doses, or routes. The researchers argued that at relatively high oral doses, lisdexamfetamine might have an oral-misuse potential closer to immediate-release d-amphetamine than the simple “slow prodrug equals low liking” intuition would suggest. That is an inference from measured time courses and ratings, not a direct measure of future substance-use disorder incidence.

The participants did not self-administer doses in an unrestricted setting, choose among formulations in an economic choice task, or have clinically monitored years of treatment. Intravenous and intranasal use were not randomized conditions in this experiment. A comparison with published evidence for nonoral administration requires a different primary study and additional questions about absorption and conversion, and should not be silently imported into this article’s results. Risk assessment must separate route-specific pharmacology from measured oral liking after supervised administration.

What do the genuine null results actually mean?

The original source found no significant between-formulation difference in maximal modeled plasma d-amphetamine concentration or AUC, and no significant difference in the main peak subjective liking ratings. The absorption rate-constant comparison reported p = .07, above the study’s conventional significance threshold. Those null findings deserve space beside the significant timing delays and clear active-versus-placebo physiological responses.

A statistical null is not the same as an unmeasured outcome. Researchers did not perform standardized field-sobriety testing, an on-road driving assessment, a crash-risk study, or a blind person-level blood concentration–impairment classifier. There is consequently no nonsignificant p-value for a hypothetical driving result: that outcome was simply absent. The absence of a statistically significant observed difference also should not be misreported as a formally proven narrow equivalence interval. The study was modestly sized and tested only a specific high-dose, fasted condition.

Why does a plasma number not establish a whole-blood cutoff?

Physical specimen identity is essential. The investigators separated plasma from heparinized venous blood before quantitation. Forensic statutory or administrative interpretation may instead concern unseparated whole blood, and some jurisdictions regulate substances or metabolites differently. Concentrations in plasma and whole blood are not interchangeable by replacing a specimen name in a report. Whole-blood/plasma partitioning must be validated for the drug, method, population, time and application if a conversion is contemplated.

The study also did not establish that a specific ng/mL plasma concentration means a person is impaired. An analytical instrument can answer whether a defined analyte was present in a defined sample at a defined concentration. Behavioral performance and psychophysical state depend on additional variables and must be supported by appropriate person-level data. For an actual toxicology case, consult the specimen label, processing records, chain of custody, analytical method, confirmation, collection time, clinical timeline, and legally applicable matrix before applying pharmacokinetic background to the facts.

Can a single concentration reveal the pill and the time of ingestion?

A later average peak after lisdexamfetamine does not provide a unique clock for an unidentified person’s dose. Two participants may absorb and convert at different rates; a person may have used repeated doses or may not have been fasted; the collection interval and laboratory processing may differ. A solitary confirmed d-amphetamine concentration also measures the active compound common to both administered formulations, not the original tablet’s brand identity. This study did not directly assay the intact lisdexamfetamine prodrug, and it did not develop a forensic source-identification test.

Attempting to equate the group geometric mean with an individual’s trajectory additionally ignores modeling assumptions. The estimated time to peak describes a population contrast under controlled conditions; it is not a precision timestamp for a particular driver or decedent. Source attribution or time-since-ingestion questions would require contemporaneous records, independent chemical evidence, appropriate sampling and a validated model with uncertainty quantified for the case at hand.

How should clinicians and forensic experts read the high-dose design?

A 100 mg single lisdexamfetamine dose and equimolar 40.3 mg d-amphetamine sulfate were purposefully chosen to produce noticeable pharmacological effects in screened non-stimulant-using volunteers. That design improves detection of acute subjective and autonomic responses but restricts generalization to titrated chronic treatment. The investigators administered the medications in a monitored hospital setting; the article does not establish the risk of adverse events after unsupervised nonprescribed use, differences during exercise, or a chronic steady-state pattern.

Fasting is another important condition. Food could alter the time profile even if it does not create the same effect for both formulations. Sex-related differences were not evident in the reported pharmacokinetic parameters, but absence of detection in a 12-woman/12-man sample does not guarantee absence in a larger or more diverse population. Clinical interpretation should be cautious when extending these high-dose, single-exposure pharmacodynamic estimates to people with cardiovascular illness, interacting medicines, renal impairment, substance-use history, or variable tolerance.

How does this investigation differ from other lisdexamfetamine research?

A comparative formulation-bioequivalence experiment can ask whether a lisdexamfetamine solution and a lisdexamfetamine capsule have equivalent exposure; this study instead contrasted lisdexamfetamine with the active d-amphetamine formulation. Another investigation can ask whether the same stimulant volunteers performed differently on vigilance, digit-symbol, or stop-signal tasks; that is not this pharmacokinetic paper’s primary endpoint. Distinguishing studies by exact DOI, biological participants, administered substances and dominant question prevents double counting related investigations or inventing a driving outcome from a pharmacokinetic chart.

Because the same active moiety appears after the two different administrations, chemical identity and pharmacological time course must be analyzed separately. The design provides unusually helpful experimental grounding for oral prodrug kinetics, but it cannot answer every analytical toxicology, clinical treatment, and legal question involving amphetamine. A precise answer is more reliable than a broad assertion about every stimulant medication.

Practical interpretation for an expert report or deposition

A well-supported formulation identifies the experiment: “In a supervised crossover study of 24 healthy adults, equimolar oral lisdexamfetamine produced a later rise and time to peak of plasma d-amphetamine than immediate-release d-amphetamine, without a significant difference in modeled peak concentration or total exposure, and with similar maximal subjective liking.” The phrase “plasma d-amphetamine” should not be shortened to “whole-blood amphetamine” or “amphetamine impairment.” If discussing pharmacokinetic sampling, the d-amphetamine treatment had a measured concentration profile for 23 people, not the full 24.

Then identify the limits: one high-dose oral pair, fasting, selected healthy volunteers, within-person repeated measures, a compartment model for summary PK values, no quantitative prodrug series, and no individual driving or incident assessment. For a case involving actual toxicological evidence, the proper next question is which substance was actually quantified in which physically collected specimen, with which laboratory method and collection chronology. The comparative research is context for a defined hypothesis, not proof of that hypothesis about a person merely because their toxicology report contains the word “amphetamine.”

Frequently asked scientific and forensic questions

1. Did lisdexamfetamine reduce peak plasma d-amphetamine?

Not in this trial’s modeled comparison. The geometric means were 118 ng/mL after lisdexamfetamine and 120 ng/mL after d-amphetamine. The paper reports no significant difference between the active formulations in the peak; it does not formally prove perfect equivalence for all people and doses.

2. Did it delay the time to peak?

Yes. Modeled times to peak were 4.6 versus 3.3 hours, and the mean paired peak-time difference was 1.1 ± 1.5 hours. The absorption lag also rose from approximately 0.8 to 1.5 hours. These estimates describe supervised high-dose oral conditions, not an individualized drug-use clock.

3. Were the reported concentrations from whole blood?

No. Investigators collected venous lithium-heparin blood and physically centrifuged it before quantifying amphetamine in separated plasma. A plasma value should not be asserted as a legal whole-blood result without independent validation.

4. Did the researchers measure intact lisdexamfetamine in the blood samples?

They directly quantified active d-amphetamine in separated plasma. They expressly noted that lisdexamfetamine-prodrug plasma profiles were not determined in this experiment, so the result is not a simultaneous quantified prodrug/active-drug fingerprint.

5. Was the maximum feeling of drug liking substantially lower after the prodrug?

No significant active-formulation difference in the peak liking rating was established. Original means were 48 ± 6.9 versus 51 ± 5.8 percent of scale maximum. Those are group subjective self-ratings, not an objective individual addiction diagnosis or a direct future-misuse outcome.

6. If the subjective high fades, has amphetamine left the plasma?

Not necessarily. The study documents time-dependent divergence: reported subjective effects commonly diminished while plasma d-amphetamine concentrations were still elevated. This is a reason not to treat the time course of a feeling, a heart-rate response, and an analytically confirmed drug level as interchangeable measures.

7. Can this study alone decide whether a person was too impaired to drive?

No. The investigators did not test actual driving, determine person-level impairment classification, or establish a legally applicable whole-blood concentration threshold. A case-specific opinion requires the actual matrix, analytical evidence, time line, behavioral observations, uncertainty and jurisdictional standard rather than extrapolation from group plasma means.

Primary evidence and further reading

Original empirical source: Dolder PC, Strajhar P, Vizeli P, Hammann F, Odermatt A, Liechti ME. Pharmacokinetics and Pharmacodynamics of Lisdexamfetamine Compared with D-Amphetamine in Healthy Subjects. Frontiers in Pharmacology. 2017;8:617. Read the complete primary investigation and its DOI. The interpretation above specifically uses the original Methods and Tables 1–2, rather than extrapolating from an abstract alone.

Related published explanations on this website: why biological specimen matrices cannot be casually interchanged; how confirmatory drug testing differs from screening; and how therapeutic monitoring and clinical interpretation differ from forensic inference.