A small pilot study found that oral THC:CBD extract reaches higher peak blood concentrations faster than nabiximols oromucosal spray in healthy fasting adults, but total drug absorption was similar between the two formulations. No therapeutic outcomes were measured, meaning these pharmacokinetic differences cannot yet inform clinical decisions about which product works better for patients.

Clinicians prescribing cannabis-based medicines face a marketplace with dozens of formulations and almost no head-to-head comparative data to guide selection. Nabiximols, the most widely studied oromucosal THC:CBD product, has a regulatory track record in multiple countries, but oral cannabis extracts are increasingly common and often preferred by patients. Understanding whether these products deliver cannabinoids to the bloodstream differently is a prerequisite for designing the clinical outcome studies that would actually tell us which works better for a given patient. This pilot represents one of the first controlled pharmacokinetic comparisons between these two commercially available product types.

Cannabinoid-based medicines are available in multiple delivery formats, but clinicians have had remarkably little comparative pharmacokinetic data to guide product selection. Nabiximols, an oromucosal spray containing roughly equal parts THC and CBD, has accumulated the most extensive clinical trial data of any cannabis-derived medicine. Oral extracts of similar cannabinoid ratios are increasingly prescribed, particularly in jurisdictions where nabiximols is unavailable or cost-prohibitive. Differences in route of administration are expected to alter how quickly and how extensively cannabinoids reach systemic circulation: oromucosal delivery theoretically permits partial sublingual absorption that bypasses hepatic first-pass metabolism, while oral ingestion routes cannabinoids through the gastrointestinal tract and liver. This crossover pilot study was designed to characterize those pharmacokinetic differences in a controlled setting.

In 12 healthy fasting adults, a single dose of oral THC:CBD extract produced significantly higher peak plasma concentrations of both THC (p = 0.007) and CBD (p < 0.035) compared to nabiximols. Elimination half-lives were significantly shorter for the oral extract for both THC (p < 0.0001) and CBD (p < 0.001). However, total drug exposure as measured by AUCinf did not differ significantly between formulations for either cannabinoid, and time to peak concentration (Tmax) was also statistically similar. A sex difference was noted in nabiximols CBD Cmax (higher in males, p < 0.05) but with only six participants per sex, this finding remains exploratory. No serious adverse events were reported. The authors appropriately characterize these results as preliminary and call for larger studies incorporating clinical outcome measures, fed-state conditions, and multiple-dose regimens.

Same Total Exposure, Different Peak: What a 12-Person Pharmacokinetic Pilot Tells Us, and Doesn’t, About THC:CBD Formulation Choice

Two products. One cannabinoid ratio. Completely different peak blood levels. A new pilot study puts oral THC:CBD oil head-to-head against the oromucosal nabiximols spray, and the results are intriguing enough to matter, but preliminary enough to demand caution before they reach a prescription pad. What the study actually tested is narrow and specific: whether 12 healthy adults, fasting and closely monitored in a clinical pharmacology unit, showed different plasma cannabinoid concentration curves after a single dose of each product. They did. The oral extract produced a higher and sharper peak for both THC and CBD, with a faster decline. But the total amount of drug that made it into the bloodstream was statistically indistinguishable between the two formulations. This is the paper’s most underappreciated finding. It suggests that the theoretical oromucosal absorption advantage of nabiximols, the idea that spraying the drug inside the cheek partly bypasses the liver, may not operate as cleanly as the marketing implies. Much of that spray appears to be swallowed and processed through the gut just like any oral dose.

I want to be fair about what this paper does well before discussing its limitations. The crossover design is smart for a pharmacokinetic study: each participant serves as their own control, reducing the noise of inter-individual metabolic differences. The bioanalytical methods are validated with tight precision values, and the dense 16-timepoint sampling over 24 hours gives a rich concentration curve. These are not trivial design choices, and they make the pharmacokinetic characterization itself internally credible. The central methodological problem, however, is one of statistical architecture. The investigators used independent-sample t-tests rather than paired tests to analyze their crossover data. In precise terms, this ignores the within-subject correlation that is the entire statistical rationale for a crossover design, reducing power and potentially altering which comparisons reach significance. Think of it this way: if you weigh the same person before and after a meal, you compare each person’s two readings. If instead you average the “before” group and the “after” group separately and compare those averages, you lose the precision of the pairing. With only 12 people, that lost precision is not a minor bookkeeping issue; it could meaningfully change results. Compounding this, the study was unblinded, doses were not perfectly matched (10 versus 10.8 mg THC), and fasting conditions may exaggerate Cmax differences that would be blunted in the real world, where patients rarely take cannabinoid medicines on an empty stomach.

What would I tell a patient asking about this study? That a faster-absorbing formulation reaching higher blood levels sounds compelling, but a faster-absorbing aspirin reaching higher blood levels does not mean it treats a headache better. Until we have outcome data, Cmax differences remain pharmacologically interesting but clinically hypothetical. To a colleague, I would say this is a well-executed starting point that deserves independent replication in an adequately powered sample under realistic conditions, including fed states and multiple-dose regimens. To a policymaker, I would say this pilot illustrates exactly the kind of comparative product research the field desperately needs, but it would be premature to use a 12-person study with industry co-authors to drive formulary or reimbursement decisions. The durable lesson from this paper is one that recurs throughout cannabinoid medicine: pharmacokinetic differences between formulations are necessary but not sufficient evidence for therapeutic differentiation. A higher peak blood level tells you the drug got there faster and in greater concentration. It does not tell you whether it worked better, lasted longer in its effects, or was safer. That bridge between pharmacokinetics and clinical outcomes must be built with clinical trial data, not assumed.

This study sits at the very beginning of the research arc for comparative cannabis product pharmacokinetics. While nabiximols has decades of clinical trial data supporting its use in multiple sclerosis spasticity and other conditions, almost none of that evidence was generated alongside head-to-head pharmacokinetic comparisons with oral extracts. This pilot provides the first controlled data point for designing such comparisons, but it does not advance our understanding of whether the observed Cmax differences translate to meaningful differences in symptom relief, psychoactive experience, onset of action, or adverse events in actual patients.

From a pharmacological standpoint, the shorter half-life observed with the oral extract may have implications for dosing intervals in chronic therapeutic use: faster clearance could necessitate more frequent dosing to maintain steady-state concentrations. The higher 11-OH-THC peak (which narrowly missed significance at p = 0.054) is consistent with greater hepatic first-pass metabolism and warrants attention in future studies, as 11-OH-THC is pharmacologically active and may contribute to both therapeutic and adverse effects. Clinicians should note that all data were collected under fasting conditions, and food is known to substantially alter oral cannabinoid absorption. The single most actionable recommendation from this study is not a change in prescribing practice, but rather a call for clinicians to maintain awareness that pharmacokinetic profiles may differ meaningfully between formulations and to counsel patients that switching between products may alter their experience even at nominally equivalent doses.

This is an original pilot pharmacokinetic study using a randomized counterbalanced crossover design in healthy volunteers. It sits in the lower tiers of the evidence hierarchy because it is preliminary, lacks clinical outcome measures, and was designed to generate descriptive pharmacokinetic data rather than test therapeutic hypotheses. The single most important inference constraint is that pharmacokinetic parameters, however precisely measured, cannot be used to infer clinical superiority or inferiority of either formulation without corresponding efficacy and safety data.

Prior pharmacokinetic work, including studies by Karschner and colleagues, has characterized nabiximols absorption in isolation, and Itin and colleagues have explored how formulation variables affect oral cannabinoid bioavailability. What has been conspicuously absent is a direct, controlled comparison of two commercially available THC:CBD products with different routes of administration in the same participants. This study begins to fill that gap but does not yet challenge or confirm any prior therapeutic claims about either product. The finding that AUCinf was equivalent despite different Cmax profiles is broadly consistent with the hypothesis that a substantial portion of oromucosal nabiximols is swallowed rather than absorbed transmucosally, a possibility suggested in earlier phase I work. Independent replication with fed-state conditions and clinical endpoints would meaningfully extend these preliminary observations.

The most consequential analytic choice was the use of independent-sample t-tests rather than paired t-tests for a crossover design. In a crossover study, each participant receives both treatments, and the statistical analysis should leverage that pairing to reduce variance. Using unpaired tests ignores within-subject correlation and reduces statistical power, which in a study of 12 people is not a trivial loss. It is plausible that some findings that narrowly missed significance, most notably the 11-OH-THC Cmax comparison at p = 0.054, could have reached significance with paired analysis. Conversely, some significant findings might have been strengthened or weakened. A mixed-effects model incorporating period and sequence effects, standard for crossover PK studies, would have been more informative and more appropriate for this design.

The most likely overinterpretation is concluding that the oral extract is therapeutically superior to nabiximols because it achieved higher peak blood concentrations. This exceeds the evidence because Cmax is a pharmacokinetic parameter, not a clinical outcome. A higher peak does not necessarily mean stronger symptom relief, faster onset of therapeutic effect, or a better safety profile. It may simply mean the drug reached and left the bloodstream faster. Similarly, the observation that AUCinf was equivalent should not be interpreted as evidence of therapeutic equivalence, because equivalent total exposure can produce different pharmacodynamic effects if the concentration-time curve has a different shape. Drawing conclusions about sex-based differences from six individuals per group is also premature and should not influence prescribing.

This carefully designed but very small pilot study demonstrates that oral THC:CBD extract and nabiximols oromucosal spray produce measurably different plasma concentration profiles despite similar total drug exposure in 12 healthy fasting adults. It contributes a needed descriptive pharmacokinetic comparison to a data-poor field. It does not establish therapeutic superiority of either formulation, nor does it support changes to prescribing practice. Its most important contribution may be clarifying the research agenda: adequately powered, independent studies comparing these products on clinical outcomes are urgently needed.

Does this study prove that oral cannabis oil works better than nabiximols spray?

No. The study measured blood levels only, not symptoms, pain relief, spasticity, or any other health outcome. Higher peak blood levels do not automatically mean better therapeutic effects. Until clinical outcome studies are conducted, neither formulation can be called superior based on this data.

Should I ask my doctor to switch me from nabiximols to an oral extract based on this study?

This study does not provide a basis for switching products. It involved only 12 healthy people who were not being treated for any medical condition, and the participants took only a single dose while fasting. The findings do not apply to patients taking these medicines regularly with food for chronic conditions.

Why does it matter that the total drug exposure was the same even though the peaks were different?

Total drug exposure (AUC) reflects the overall amount of cannabinoid that reaches the bloodstream. The fact that both formulations delivered similar total amounts despite different peak patterns means the body ultimately absorbed comparable quantities of THC and CBD. Whether a faster, higher peak or a slower, more sustained profile is better for a particular condition is a clinical question this study was not designed to answer.

Were there any safety concerns with either product?

No serious adverse events were reported during this study. However, the study was far too small and too brief to serve as a safety assessment for either product. Safety profiles should be evaluated through larger and longer studies conducted in patients with actual medical conditions.

Have thoughts on this? Share it: