Bioequivalence for Narrow Therapeutic Index Drugs: How Tight Is Tight Enough

Bioequivalence is not a single fixed standard, and grasping that is the key to judging when a generic substitution is scientifically supported. For most oral drugs, a generic must show that the 90 percent confidence interval of the test-to-reference ratio of exposure falls entirely within 80.00 to 125.00 percent. For narrow therapeutic index drugs, that window is deliberately tightened through reference-scaled average bioequivalence, which also adds a check on variability. For highly variable drugs, the same scaling logic runs the other way and can widen the window. The FDA finalized this framework in its May 2026 guidance, which sets out how each case is handled.

What the standard limits actually mean

The familiar 80.00 to 125.00 percent range is often misread as a claim that a generic can differ from the brand by up to 20 or 25 percent in any given patient. That is not what it says. The criterion applies to the 90 percent confidence interval around the ratio of population average exposures, typically peak concentration and area under the concentration-time curve. The interval, not a single point estimate, must sit inside the limits. In practice, approved generics cluster much closer to the reference than the boundaries imply, because a product whose true average drifts toward an edge produces a confidence interval that spills past it and fails.

The 80 to 125 window itself encodes a judgment: that a roughly 20 percent difference in average exposure is not clinically meaningful for most drugs. The apparent asymmetry (80 on one side, 125 on the other) disappears on a logarithmic scale, where the two limits are mirror images. Exposure data are analyzed after log transformation precisely so the comparison is symmetric and the statistics behave.

Why narrow therapeutic index drugs need something tighter

A narrow therapeutic index drug is one where small differences in dose or blood concentration can produce serious therapeutic failure or life-threatening harm. Warfarin, levothyroxine, several antiepileptics, tacrolimus, and lithium are the classic teaching examples. For these agents, the premise behind the standard window breaks down: a 20 percent shift in exposure may be exactly the difference between a therapeutic and a toxic or subtherapeutic state.

The FDA's response, described in detail by Donnelly and colleagues in Clinical Pharmacology and Therapeutics in 2025, is reference-scaled average bioequivalence built on a four-way, fully replicated crossover design. In that design, each healthy volunteer receives the reference product twice and the test product twice. Replication does two things a conventional two-period study cannot. It measures the reference product's own within-subject variability directly, and it lets the same data compare the variability of the test product against the reference.

How scaling tightens or loosens the window

Reference-scaled average bioequivalence ties the acceptance limits to the within-subject standard deviation of the reference product, often written as s-sub-WR. When the reference is very consistent from dose to dose, the limits contract inward from 80 to 125, demanding a closer match. As reference variability rises, the limits relax.

For narrow therapeutic index drugs the FDA applies a firm ceiling. The scaled limits are capped at the conventional 80.00 to 125.00 percent, and that cap takes over once the reference within-subject standard deviation reaches roughly 0.21, the point at which the scaled window would otherwise exceed the standard one. Below that threshold, a low-variability narrow index drug faces a genuinely narrower target than an ordinary generic does. The analysis in Jiang and colleagues, published in the AAPS Journal, illustrates how the scaled criterion pulls the effective limits toward a tighter band such as roughly 90 to 111 percent when the reference is highly reproducible.

Scaling alone is not enough for these drugs, because a test product could match the average yet swing more widely within patients. So the FDA adds a variability comparison criterion: an explicit statistical test that the test product is not meaningfully more variable than the reference. The recommended rule requires the upper bound of the 90 percent confidence interval for the ratio of test to reference within-subject standard deviations to stay below a set constant. A generic must therefore clear two hurdles at once, matching both the average exposure and the consistency of that exposure.

Highly variable drugs and the same idea in reverse

Highly variable drugs, defined by a within-subject variability of about 30 percent or greater, create the opposite problem. Their exposure bounces so much between doses that even the reference product compared against itself can fail the standard window by chance, which would make some genuinely equivalent generics unapprovable. Here reference-scaled average bioequivalence widens the limits in proportion to the reference variability, up to a capped maximum, so the criterion reflects the drug's intrinsic noise rather than penalizing it. The 2026 final guidance consolidates the statistical appendices for both the narrow index and the highly variable cases into one document, replacing the 2001 guidance and finalizing the December 2022 draft.

What this means for reading a substitution

For a scientifically literate reader, the practical takeaway is that not all generics are held to the same numerical standard, and the standard is stricter, not looser, exactly where the stakes are highest. A generic narrow therapeutic index drug that reaches the market has been tested in a replicated design and has met both a scaled average limit and a variability limit that an ordinary generic never faces. That is a stronger evidentiary bar than the headline 80 to 125 figure suggests. It does not eliminate every clinical consideration, since factors like formulation excipients and individual response still matter, but it does mean the regulatory framework already recognizes that tight drugs deserve tight limits. This article is educational and not medical advice; decisions about switching a specific medication belong with the prescriber and pharmacist who know the individual case.