Does a Lisdexamfetamine Oral Solution Match Capsules in Plasma? A 29-Volunteer Study

Short answer: In a randomized two-period crossover study, a 70 mg lisdexamfetamine dimesylate oral solution and a 70 mg reference capsule produced bioequivalent exposure to the inactive prodrug and the released active dexamfetamine in the 29 healthy volunteers who completed both treatments. The test/reference geometric-mean ratio for dexamfetamine plasma peak concentration was 99.07% (90% confidence interval 96.93%–101.25%), and the ratio for dexamfetamine exposure through the last quantifiable sample was 98.05% (94.30%–101.95%). Those intervals fell inside the prespecified 80%–125% bioequivalence boundaries. The finding concerns controlled drug pharmacokinetics in separated plasma; it does not validate a blood-amphetamine impairment threshold, predict a particular person’s driving, or establish when an unsupervised dose was taken.

The original controlled trial by Rizea-Savu, Duna, Panagiotopoulos, and Sandulovici was published in Frontiers in Pharmacology in 2022: Single Dose Comparative Bioavailability Study of Lisdexamfetamine Dimesylate as Oral Solution Versus Reference Hard Capsules in Healthy Volunteers. Reading its complete analytical validation and clinical results reveals why the measured prodrug and active drug must remain separate in any defensible scientific or forensic interpretation.

What did the investigators actually compare?

The investigators tested two completed pharmaceutical dosage forms containing the same nominal amount of lisdexamfetamine dimesylate: 7 mL of an oral solution formulated at 10 mg/mL, and a commercially available 70 mg Elvanse hard capsule. A volunteer received one formulation in each of two study periods, with treatment sequence randomized and a 14-day washout. All dosing occurred after at least ten hours of fasting. This was not a comparison of 70 mg active dexamfetamine with 70 mg lisdexamfetamine; the administered material was the lisdexamfetamine prodrug in both arms.

An oral liquid differs from a hard capsule at the point of administration because the capsule must release and dissolve its contents. The central question was whether that physical difference would change the amount or rate of prodrug and active-moiety exposure. It was not a trial of attention, safe driving, intoxication, symptom control, or optimal dosing for an individual patient.

How many people were studied, and which denominator is appropriate?

Thirty-two distinct healthy adults enrolled and received the first study treatment. They comprised 13 men and 19 women, were 18–45 years old, and met the investigators’ health and body-mass screening criteria. Three participants left voluntarily before their crossover treatment. Thus 29 people, not 32, formed the paired pharmacokinetic bioequivalence population. Those same 29 people contributed many timed samples across the two treatments; a sample count, a 72-hour time series, or the two periods cannot legitimately be converted into a larger independent human sample size.

Investigators separately assessed taste and usability in 28 people considered sensitive to taste and smell. That subset should not be treated as the 29-person paired kinetic analysis, nor as a separate trial of medication efficacy. All participants in the trial were described as healthy Caucasian adults; generalizing without qualification to children, medically complex patients, altered renal function, repeated dosing, or people taking interacting medications would exceed this experiment.

Why does lisdexamfetamine produce dexamfetamine?

Lisdexamfetamine is a prodrug, which means that the administered chemical is converted to another pharmacologically active compound. After absorption, erythrocyte-associated hydrolysis releases dexamfetamine. That biological conversion helps explain why different initial dissolution conditions can leave the active-moiety exposure more similar than the intact prodrug’s exposure. The study’s actual evidence consists of separately measured plasma concentration profiles for lisdexamfetamine and dexamfetamine; its interpretation of erythrocyte-limited conversion draws on established prodrug pharmacology rather than a direct measurement of red-cell hydrolysis rate in each participant.

Do not interchange “lisdexamfetamine,” “dexamfetamine,” “amphetamine,” and “methamphetamine.” The article quantified the specific lisdexamfetamine prodrug and its dexamfetamine product. It did not administer methamphetamine, measure an illicit methamphetamine exposure, compare mixed amphetamine salts, or prove anything about enantiomer composition in an unrelated forensic specimen. A correctly named analyte is essential before any concentration, dose, or impairment discussion begins.

Which biological material did the laboratory actually test?

Investigators collected roughly 5 mL of venous blood at each scheduled point into K2EDTA tubes. They refrigerated and centrifuged the specimens, separated plasma, divided it into aliquots, and froze it within the stated handling window. Reported concentrations in ng/mL therefore describe plasma. The fact that the sample originated as blood does not make the reported number a validated concentration in unseparated venous whole blood.

Some sample-handling validation experiments used prepared or fortified whole-blood specimens to evaluate possible interconversion during handling. Those were methodological controls. They do not convert the participants’ plasma pharmacokinetic results into legal whole-blood amphetamine results. The trial did not report concurrent measured parent drug in oral fluid, urine, hair, sweat, vitreous humor, actual exhaled breath, or muscle tissue, and it provided no plasma-to-whole-blood conversion factor for application to a case.

How was the parent drug distinguished from the prodrug analytically?

The scientists developed and validated an HPLC–tandem mass-spectrometry assay with positive-ion multiple-reaction monitoring for both compounds. Dexamfetamine was tracked through the 136.040 → 119.000 m/z transition, while lisdexamfetamine used 264.170 → 83.900. Separate isotope-labeled internal standards—amphetamine-d8 and lisdexamfetamine-d4—were used to control analytical variability. Chromatographic separation employed an RP-amide column and a water/acetonitrile gradient; instrument performance was evaluated rather than assumed from a brand name.

Eight-point calibration covered 2–200 ng/mL dexamfetamine and 1–100 ng/mL lisdexamfetamine in fortified blank plasma, with weighted 1/x² fitting. Accordingly, the lower ends of those ranges were method-specific quantification limits, not universal human impairment cutoffs. The investigators tested selectivity against blank plasma including lipemic and hemolytic specimens and a range of possible co-medications. A calibration correlation near 0.99960 for either analyte describes an analytical fit; it is not 99.96% diagnostic accuracy in drivers or criminal-case samples.

What were the directly measured peak and exposure results?

The key descriptive results came from the 29 complete paired participants. The solution and capsule generated fairly similar mean plasma dexamfetamine peaks, while their intact lisdexamfetamine values differed somewhat more. Table 1 distinguishes the inactive prodrug from the active moiety and labels the statistics rather than mixing pharmacokinetic measures with behavior.

Measurement note: Means are descriptive rather than participant-level thresholds, AUC describes integrated concentration over time, and Tmax values are medians. Peak-to-peak numerical similarity alone is not the formal statistical definition of bioequivalence; the paired geometric-mean ratios and confidence intervals are needed.

How strong was the formal bioequivalence finding?

Bioequivalence was evaluated on logarithmically transformed pharmacokinetic endpoints using a prespecified 80%–125% interval for the test/reference ratio. The appropriate question is whether each 90% confidence interval falls entirely within those boundaries. For the released dexamfetamine, both peak and AUC estimates were very close to 100%. The prodrug showed slightly greater differences between formulations, yet its intervals also met the stated criterion.

The table is strong evidence of matched systemic availability under the specified 70 mg, fasting, single-dose conditions. It is not evidence that any two drug products are clinically interchangeable under every circumstance, that individual concentration curves are identical, or that every person’s subjective or psychomotor response must be the same. The numerical 90% intervals are not probabilities that a particular person had a particular blood level.

Did the solution reach the circulation sooner?

Mean reported lag time for the measured intact prodrug was 21 ± 10 minutes with solution and 29 ± 14 minutes with capsule. The corresponding mean lag times for released dexamfetamine were 35 ± 14 and 47 ± 14 minutes. These are descriptive time-to-first-measurable patterns, consistent with different preparation and dissolution steps before conversion. They should not be mistaken for a fixed “start of effect” time in a particular patient, or a clock that can reliably reconstruct when an unidentified person swallowed the medication.

The investigators explicitly found no statistically significant solution–capsule difference in Tmax: for lisdexamfetamine p = 0.06 and for dexamfetamine p = 0.21. It would be scientifically incorrect to label the observed median time differences proven in this sample. Failure to cross a significance threshold also does not prove exact equality of every concentration-time point or of clinical onset.

What happened to the two analytes during elimination?

The parent prodrug declined quickly. In the studied plasma profiles, lisdexamfetamine had an estimated mean terminal half-life of about 0.44 hours for the solution and 0.56 hours for the capsule, and its mean concentrations were below the assay’s quantification limit within about four hours. Dexamfetamine persisted much longer, with an estimated half-life of 11.42 versus 11.21 hours. Measurable active-drug plasma exposure remained above the assay’s quantification limit as far as about 48 hours in the reported profiles.

Half-life is a population model summary under a specified sampling and analytical framework. It is not a universal duration of cognitive change, legal impairment, clinical benefit, or detectability by every laboratory. The paper noted a statistically significant formulation difference for apparent prodrug half-life (p = 0.02), although the mean difference was only roughly seven minutes; the active dexamfetamine half-life difference was not significant (p = 0.80). That study-specific positive-and-null pairing is more informative than a simplified slogan about identical kinetics.

Did the trial measure red-blood-cell conversion?

The original authors interpreted similar active-dexamfetamine exposure in light of known conversion of lisdexamfetamine by red-cell-associated enzymes. But the participant endpoint was not an individual erythrocyte conversion experiment. Plasma samples were removed from blood after refrigerated centrifugation, then subjected to LC–MS/MS. The study’s controls also tested whether the two analytes interconverted spuriously in prepared samples during storage and handling, with results inside the stated ±15% acceptance band across investigated conditions.

Those controls matter because sample degradation or ex vivo conversion can distort apparent prodrug/active-metabolite relationships. A carefully validated assay reduces that concern in this experimental protocol. It does not follow that an unrelated forensic collection, a different anticoagulant, delayed separation, room-temperature storage, or a different detection method has automatically passed the same validation. Specimen history is an integral part of interpretation.

Could plasma matrix effects have distorted quantification?

Plasma includes endogenous molecules capable of suppressing or enhancing electrospray ionization. In validation, the investigators observed unnormalized lisdexamfetamine matrix factors of 1.458 and 1.384 at two evaluated concentrations, suggesting ion enhancement. They corrected with a matched labeled internal standard, reporting normalized factors near 0.994 and 0.992. For dexamfetamine, corresponding unnormalized factors were about 0.986 and 0.948, with normalized factors 1.049 and 1.055. These are instrument-and-matrix-specific analytical observations, not evidence of increased or decreased drug exposure in the volunteers.

The team also tested accuracy and precision, recovery, short- and long-term storage, carryover, dilution, hemolysis, and incurred-sample reanalysis. The reported reanalysis accuracies were 98.71% for lisdexamfetamine and 97.85% for dexamfetamine under the method’s procedure. Those findings support analytical reliability in the study’s setting, but clinical causation and person-specific toxicological interpretation remain separate propositions.

Was oral-solution acceptability investigated separately?

Yes. Among 28 taste- and smell-sensitive participants, 19 initially described daily use of the oral solution as easy and nine as manageable. Ten minutes later, 17 selected easy and 11 manageable. The percentages concern those 28 respondents, not the entire 32 enrolled participants and not the 29-completer PK group. A favorable response to an administration question does not prove actual long-term adherence, effectiveness in a child, or safety of an unauthorized dosing device.

The participants were supervised, and the liquid was measured by syringe under the study protocol. Outside a controlled setting, formulation strength, proper measuring equipment, caregiver practices, swallowing problems, and medical supervision may all matter. This paper tested a particular development formulation and reference capsule; it should not be used as permission to substitute arbitrary compounded or unapproved solutions.

What adverse events and negative safety comparisons were reported?

The study recorded 21 non-serious adverse-event episodes occurring among 14 individuals. Of those episodes, 17 were categorized as treatment-emergent and four as not related to study medication. Headache appeared in seven reported cases, tachycardia in five, and nausea in four; an isolated orthostatic-hypotension episode was also noted. Do not report 21 adverse events as 21 affected participants or infer a causal incidence rate from the raw episode count alone.

Researchers did not identify a statistically significant difference between formulations in adverse-event incidence or in the number of participants affected. The small, healthy, single-dose trial was not designed to exclude rare complications, establish long-term cardiovascular safety, or predict risk with interacting drugs. Likewise, an absence of a between-formulation safety difference here does not imply that a person can safely exceed a prescribed dose or drive while experiencing symptoms.

Why does bioequivalence not establish equivalent driving performance?

Plasma concentrations establish measured exposure in a chemical compartment; driving performance depends on many additional factors. These participants did not perform a simulator drive, standardized roadside evaluation, vigilance test, lane-position test, hazard-response task, or witness-memory assessment as part of the reported bioequivalence outcomes. The study therefore contains no treatment-by-driving endpoint and cannot prove either improved or impaired driving after either formulation.

Even perfect agreement in mean plasma AUC would not establish that every participant had the same task performance. Individual factors such as dose timing, tolerance, other medications, health, fatigue, sleep deprivation, and baseline ability were not systematically tested as a legal impairment classifier here. The paper does not report a validated threshold for saying that a driver with some reported plasma or whole-blood concentration was unsafe, safe, inattentive, or capable of a particular reaction time.

How should a forensic toxicologist use the study in a real case?

First, identify what was actually reported: venous whole-blood result, separated plasma result, serum, urine, or a qualitative screen. Second, confirm whether the laboratory named dexamfetamine, total amphetamine, lisdexamfetamine, methamphetamine, or another analyte, and whether the method resolved relevant stereochemistry. Third, document dose history, specimen collection time, matrix handling, analytical validation, and potential co-exposures before interpreting the number. This sequence avoids creating a drug-time or behavior conclusion that the assay never supported.

This trial helps explain that a therapeutically administered lisdexamfetamine formulation can yield circulating dexamfetamine and that the released active moiety can be relatively long-lived compared with the intact prodrug. It cannot determine whether a different person’s concentration came from that prescription rather than another amphetamine source without appropriate independent evidence; nor can a single result alone prove the number of prior doses, precise ingestion time, tolerance, or incident causation.

What study-design limitations constrain generalization?

The 29 paired completers support a within-person formulation comparison but provide limited power for subgroup questions and rare safety findings. The trial was open-label as to dosage form, although the pharmacokinetic endpoints came from objective laboratory measurements. All volunteers were healthy adults within a relatively narrow age range, and both treatments were given under fasting, supervised conditions. The trial did not test chronic dosing at steady state, variable meals, drug interactions, or children, and it did not include legally adjudicated impaired drivers.

The sponsor-development context and associated potential interests warrant normal editorial scrutiny, but they do not invalidate the measured concentration tables by themselves. The most defensible reading is narrow: reported 90% bioequivalence intervals met the chosen criterion in this experiment, with observed minor prodrug differences and relatively similar active-moiety kinetics. Questions about individual performance, laboratory cutoffs in another matrix, or an event remain outside the trial’s evidence.

Seven frequently asked forensic questions

1. Did the study enroll 29 or 32 people?

It enrolled and dosed 32 unique healthy adults in the first period. Three did not complete the second treatment, so 29 supplied complete paired crossover concentration profiles for the formal bioequivalence evaluation. The 28-person palatability analysis was a separate subset.

2. Were the stated ng/mL numbers measured in whole blood?

No. Venous blood was drawn, but the laboratory centrifuged the tubes and measured separated plasma. The ng/mL figures are plasma concentrations, not validated unseparated whole-blood levels. Preparative whole-blood stability experiments were method controls rather than the reported clinical PK matrix.

3. Is lisdexamfetamine chemically the same substance as dexamfetamine?

No. Lisdexamfetamine is the administered inactive prodrug, while dexamfetamine is the released active chemical. The trial quantified them separately by LC–MS/MS and found different peak times, half-lives, and exposure values. Neither name should silently be replaced with methamphetamine.

4. Was one treatment proven to have a faster pharmacological onset?

The solution showed shorter descriptive mean lag times, but the formal comparison of plasma Tmax was not statistically significant for prodrug (p = 0.06) or active dexamfetamine (p = 0.21). Clinical attention or perceived onset was not measured as a comparative outcome.

5. Does bioequivalence mean every participant had the same plasma level?

No. Bioequivalence is an inference about test/reference exposure ratios and uncertainty under the study protocol. Individuals’ concentrations varied; the within-subject coefficients of variation differed between the prodrug and the active moiety. The result does not guarantee an identical individual concentration-time curve.

6. Does measurable dexamfetamine 48 hours after dosing prove impairment?

No. The study reported active-moiety plasma concentrations above the method’s lower quantification range in the concentration-time data up to about 48 hours. It did not simultaneously establish observed impairment, driving outcomes, or an individual legal-effect threshold at that time.

7. Can this study prove when someone took a prescription before an incident?

Not by itself. Controlled group PK profiles can inform the range of possible kinetics, but dose history, individual variation, exact specimen type, analytical conditions, and other evidence are required. One case concentration cannot be reverse-engineered into a reliable exact dosing time or cause of an incident merely from this study.

Further reading on specimen interpretation

For the distinction between unseparated blood and plasma, see Biological Matrices. For analytical specificity and verified target identity, see Confirmatory Drug Testing. For the separate clinical meaning of timed drug concentrations, see Therapeutic Drug Monitoring. These background guides provide context; they do not enlarge the empirical outcomes of this trial.

Primary research source and bottom line

Rizea-Savu S, Duna SN, Panagiotopoulos D, Sandulovici RC. Single Dose Comparative Bioavailability Study of Lisdexamfetamine Dimesylate as Oral Solution Versus Reference Hard Capsules in Healthy Volunteers. Frontiers in Pharmacology. 2022;13:881198. https://doi.org/10.3389/fphar.2022.881198.

Bottom line: A supervised lisdexamfetamine oral solution met prespecified plasma-exposure bioequivalence requirements against a reference capsule in 29 healthy paired completers. Dexamfetamine was directly quantified in separated plasma and persisted longer than the intact prodrug. This supports a carefully bounded formulation and laboratory interpretation—not a claim about a particular person’s whole-blood amphetamine concentration, exact last dose, individual functional impairment, or legal incident causation.