CBD Formulation Comparisons: Similar Average Exposure Does Not Establish Bioequivalence

A pharmacokinetic comparison must account for uncertainty as well as average exposure. A small trial of a cannabidiol and hydroxychloroquine combination illustrates why similar-looking averages cannot, by themselves, establish bioequivalence.

What the investigators compared

Mbogo and colleagues reported a randomized, open-label phase I trial in 2025. Thirty-six healthy volunteers entered three parallel groups, with 12 participants per group. Each received one oral treatment: the investigational combination IHL-675 A containing 150 mg cannabidiol (CBD) and 200 mg hydroxychloroquine sulfate, 150 mg CBD as Epidiolex, or 200 mg hydroxychloroquine sulfate as Plaquenil.

Treatment followed a high-fat, high-calorie breakfast. The investigators measured drugs and metabolites in plasma obtained from collected blood samples. The comparison therefore describes specific formulations and fed conditions, rather than every product carrying the same labeled dose.

Peak concentration and total exposure answer different questions

C_max is the observed peak concentration. Area under the concentration-time curve, or AUC, summarizes exposure over time. Two formulations can have different peaks while their AUC values are closer together.

For CBD, the combination-to-reference geometric least-squares mean ratio was 152.0% for C_max, with a 90% confidence interval of 101.44% to 227.75%. For AUC to the last measured time, the ratio was 109.9%, with a 90% confidence interval of 82.91% to 145.72%. These are different pharmacokinetic endpoints, not interchangeable measures of clinical benefit.

Why the interval matters

The paper evaluated the ratios against an 80% to 125% acceptance interval. The reported CBD confidence intervals extended outside that interval. The study therefore did not establish bioequivalence using that criterion, even where average exposure appeared close.

The abstract describes pharmacokinetic differences as not statistically significant, while Table 9 gives a CBD peak-ratio confidence interval entirely above 100%. That reporting tension should remain visible. Neither a broad statement of nonsignificance nor a near-unity AUC estimate substitutes for examining the actual endpoint and its interval.

Incomplete terminal estimates create another limit

For hydroxychloroquine, half-life could be estimated in only five of the 12 combination-group participants. The authors regarded the resulting comparison of AUC extrapolated to infinity as unreliable. An extrapolated exposure estimate depends on having an adequately characterized terminal concentration decline.

What this study can support

This was a small parallel-group, single-dose study in healthy volunteers. It does not supply a universal conversion between formulations or show therapeutic superiority. Incannex funded the trial, and four authors disclosed employment and shareholding interests in the company. Those disclosures belong alongside the design and statistical limitations.

The practical lesson is to identify the specimen, formulation, dosing conditions, endpoint and confidence interval before claiming that two products produce equivalent exposure.

Source: Mbogo GW and colleagues. An open-label phase I comparator-controlled clinical trial to assess tolerability and pharmacokinetics of IHL-675 A a fixed dose combination of cannabidiol plus hydroxychloroquine in healthy volunteers. Scientific Reports. 2025;15:19357. Open-access full text.