A Course on Bioequivalence: Unit 11 - Introduction to Narrow Therapeutic Drug Index (NTID) Products

Authors

Yoni Nazarathy

Andreas Noack

This is Unit 11 of a complete bioequivalence analysis course using Pumas. There are 15 units in the course in total.

1 Unit overview

In this short unit we broadly define Narrow Therapeutic Index Drugs (NTIDs) and explain their clinical characteristics. The standard bioequivalence acceptance criteria are often not sufficient for NTIDs. We touch on the simple approach of narrowing the limits to a fixed level, and defer the reference scaled approach to the sequel

We use the following packages.

using Bioequivalence  # available with Pumas products
using PharmaDatasets # available with Pumas products
using SummaryTables

We have used all of these packages in previous units.

2 What is a Narrow Therapeutic Index Drug?

A Narrow Therapeutic Index Drug (NTID), sometimes called a Narrow Therapeutic Ratio (NTR) drug, is a medication for which a small difference in dose or blood concentration can lead to serious therapeutic failures or adverse drug reactions.

In simpler terms, the therapeutic window — the range between the minimum effective concentration and the minimum toxic concentration — is very small.

  • Below the window: The drug is ineffective, potentially leading to the progression of a serious disease.
  • Above the window: The drug can be toxic, causing life-threatening side effects.

Classic examples of NTIDs include:

  • Warfarin (anticoagulant)
  • Lithium (for bipolar disorder)
  • Digoxin (for heart failure)
  • Tacrolimus (immunosuppressant for organ transplant)
  • Theophylline (for respiratory diseases)

For these drugs, precision is paramount.

3 Beyond the 20% Standard Bioequivalence Criteria

The standard 80-125% limits for a bioequivalence trial which implement at TOST (see Unit 3) are based on the clinical judgment that for most drugs, a \(20\%\) difference in plasma concentration will not cause a clinically significant difference in safety or efficacy.

As an example consider this analysis of the AUC point with this example dataset arising from a \(2 \times 2\) crossover trial with many subjects:

data = dataset(joinpath("bioequivalence", "RT_TR", "SLF2014_8"));

Let’s analyze this data with pumas_be:

pumas_be(data)
Observation Counts
Sequence ╲ Period 1 2
RT 288 288
TR 429 429
Paradigm: Non replicated crossover design
Model: Linear model
Criteria: Standard ABE
Endpoint: AUC
Formulations: Reference(R), Test(T)
Results(AUC) Assessment Criteria
R Geometric Marginal Mean 132.3
Geometric Naive Mean 138.5
T Geometric Marginal Mean 123.6
Geometric Naive Mean 149.4
Geometric Mean T/R Ratio (%) 93.42
Degrees of Freedom 715
90% Confidence Interval (%) [86.81, 100.6] Pass CI ⊆ [80, 125]
Variability CV (%) | σ̂ 99.27 | 0.8281
ANOVA Formulation (p-value) 0.1277
Sequence (p-value) 0
Period (p-value) 0

The GMR point estimate of 93.42 in percentage is surrounded by a confidence interval of [86.81, 100.6] which passes the standard criteria and hence we see Pass.

Yet the lower limit of 86.81 in percentage indicates that there is a non-negligible chance of the test product to be quite significantly below the window.

A simple approach: narrowing the limits

A simple approach for handling NTID drugs is specified in Section 4.19 of EMA (2010). That section reads:

In specific cases of products with a narrow therapeutic index, the acceptance interval for AUC should be tightened to 90.00-111.11%. Where Cmax is of particular importance for safety, efficacy or drug level monitoring the 90.00-111.11% acceptance interval should also be applied for this parameter. It is not possible to define a set of criteria to categorise drugs as narrow therapeutic index drugs (NTIDs) and it must be decided case by case if an active substance is an NTID based on clinical considerations.

With pumas_be, we can use the EMA_NarrowTherapeuticIndex indication as a second argument to implement such tighter criteria:

pumas_be(data, EMA_NarrowTherapeuticIndex)
Observation Counts
Sequence ╲ Period 1 2
RT 288 288
TR 429 429
Paradigm: Non replicated crossover design
Model: Linear model
Criteria: EMA Narrow ABE
Endpoint: AUC
Formulations: Reference(R), Test(T)
Results(AUC) Assessment Criteria
R Geometric Marginal Mean 132.3
Geometric Naive Mean 138.5
T Geometric Marginal Mean 123.6
Geometric Naive Mean 149.4
Geometric Mean T/R Ratio (%) 93.42
Degrees of Freedom 715
90% Confidence Interval (%) [86.81, 100.6] Fail CI ⊆ [90, 111]
Variability CV (%) | σ̂ 99.27 | 0.8281
ANOVA Formulation (p-value) 0.1277
Sequence (p-value) 0
Period (p-value) 0

Notice in the heading list that we see Criteria: EMA Narrow ABE. Further, as you can see, the confidence interval limits under Criteria are at [90, 111] and in this case, the output presents Fail as expected.

Towards a more complex approach: reference scaled average bioequivalence

The FDA approach for NTID drugs is more complex and relies on reference scaled average bioequivalence (RSABE). The key document is FDA (2012). We discuss this approach in the sequel where we see it is also applicable to highly variable drugs. At this point, let us just present output from this approach.

One requirement is to have fully replicate designs so that the within subject variability can be estimated both for the reference and test products. Assume we have this dataset:

data = dataset(joinpath("bioequivalence", "RTTR_TRRT", "PJ2017_4_3"));

We now use the FDA_NarrowTherapeuticIndex indication as a second argument for pumas_be. Here we carry out the analysis both for AUC and Cmax:

pumas_be(data, FDA_NarrowTherapeuticIndex)
Observation Counts
Sequence ╲ Period 1 2 3 4
RTTR 8 8 8 8
TRRT 9 9 9 8
Paradigm: Replicated crossover that supports reference scaling
Model: Mixed model (unequal variance)
Criteria: FDA RSABE for NTI
Endpoint: AUC
Formulations: Reference(R), Test(T)
Results(AUC) Assessment Criteria
R Geometric Marginal Mean 7152
Geometric Naive Mean 7131
T Geometric Marginal Mean 7412
Geometric Naive Mean 7454
Geometric Mean T/R Ratio (%) 103.6
Degrees of Freedom 15.24
90% Confidence Interval (%) [99.31, 108.2] Pass CI ⊆ [80, 125]
Variability CVᵣ (%) | σ̂ᵣ 8.02 | 0.0801
CVₜ (%) | σ̂ₜ 10.84 | 0.1081
Variability Ratio (%) 135
ANOVA Formulation (p-value) 0.1624
Sequence (p-value) 0.3184
Period (p-value) 0.665
Reference Scaling Params Reference Scaling Constant 1.11
Reference Scaling Analysis Geometric Mean T/R Ratio (%) 103.7
Standard Error (Log Scale) 0.0253
90% Confidence Interval (%) [99.2, 108.4]
Degrees of Freedom 15
Howe's Approx RSABE Stat (95%) 0.0001105 Fail ≤ 0
Variability Ratio Quantile (95%) 2.118 Pass ≤ 2.5 
pumas_be(data, FDA_NarrowTherapeuticIndex, endpoint = :Cmax)
Observation Counts
Sequence ╲ Period 1 2 3 4
RTTR 8 8 8 8
TRRT 9 9 9 8
Paradigm: Replicated crossover that supports reference scaling
Model: Mixed model (unequal variance)
Criteria: FDA RSABE for NTI
Endpoint: Cmax
Formulations: Reference(R), Test(T)
Results(Cmax) Assessment Criteria
R Geometric Marginal Mean 1150
Geometric Naive Mean 1150
T Geometric Marginal Mean 1049
Geometric Naive Mean 1048
Geometric Mean T/R Ratio (%) 91.23
Degrees of Freedom 40.82
90% Confidence Interval (%) [83.18, 100.1] Pass CI ⊆ [80, 125]
Variability CVᵣ (%) | σ̂ᵣ 21.17 | 0.2094
CVₜ (%) | σ̂ₜ 26.87 | 0.264
Variability Ratio (%) 126.2
ANOVA Formulation (p-value) 0.1019
Sequence (p-value) 0.5596
Period (p-value) 0.4916
Reference Scaling Params Reference Scaling Constant 1.11
Reference Scaling Analysis Geometric Mean T/R Ratio (%) 91.63
Standard Error (Log Scale) 0.0518
90% Confidence Interval (%) [83.65, 100.4]
Degrees of Freedom 15
Howe's Approx RSABE Stat (95%) -0.01049 Pass ≤ 0
Variability Ratio Quantile (95%) 1.98 Pass ≤ 2.5

Observe that the criteria is FDA RSABE for NTI.

Without getting into the details of the output yet, observe that in each of the above cases there are multiple criteria (3 in total).

  • The first criteria next to 90% Confidence Interval (%) is simply to pass the standard 80%–125% average bioequivalence.
  • The second criteria next to Howe's Approx RSABE Stat (95%) is FDA’s reference scaled average bioequivalence approach described in the next two units.
  • The final criteria next to Variability Ratio Quantile (95%) is FDA’s approach based on the variability ratio of the test and reference product.

For example for AUC we have a respective Pass, Fail, and Pass under the Assessment column. Hence the analysis fails for AUC as not all three criteria are met.

Similarly, for Cmax we have a respective Pass, Pass, and Pass, and hence the analysis passes for Cmax.

More on the details of these criteria arising from FDA (2012) is in the sequel.

4 Conclusion

This unit introduces Narrow Therapeutic Index Drugs (NTIDs), which are medications where small variations in blood concentration can lead to significant clinical consequences, such as therapeutic failure or severe toxicity. Because of this heightened risk, the standard bioequivalence acceptance range of 80% to 125% is often considered insufficient to ensure switchability between a reference product and a generic. The core challenge with NTIDs is managing this narrow therapeutic window to ensure patient safety and drug efficacy.

To address this challenge, regulatory agencies have developed stricter bioequivalence standards. The European Medicines Agency (EMA) employs a straightforward approach by tightening the acceptance limits for the 90% confidence interval to a more stringent range of 90.00% to 111.11%. In contrast, the U.S. Food and Drug Administration (FDA) uses a more complex method known as reference-scaled average bioequivalence (RSABE). This approach, which requires replicate study designs, evaluates bioequivalence based on three separate criteria: a standard average bioequivalence test, a scaled bioequivalence limit, and a constraint on the variability of the test product relative to the reference.

5 Unit exercises

  1. Defining NTIDs and Their Challenge

    1. In your own words, what is a Narrow Therapeutic Index Drug (NTID)?
    2. Why are the standard 80-125% bioequivalence limits often considered inappropriate for these types of drugs?
    3. Name two examples of NTIDs mentioned in the text.

  2. Applying the EMA’s Tightened Criteria

    Consider the first dataset used in this unit.

    data = dataset(joinpath("bioequivalence", "RT_TR", "SLF2014_8"));

    The analysis under standard criteria yielded a 90% confidence interval for AUC of [86.81, 100.6].

    1. Does this result pass the standard 80-125% bioequivalence criteria?
    2. If this drug were being evaluated as an NTID under EMA guidelines, would it pass the tightened 90.00-111.11% criteria? Explain why or why not.
    3. Write the single pumas_be command that would perform the analysis specifically for an NTID according to the EMA guideline. Run the command to confirm your answer to part (b).

  3. Interpreting FDA RSABE Results

    The unit presents the output of an FDA NTID analysis for Cmax using a replicate design dataset. The code is:

    data_rep = dataset(joinpath("bioequivalence", "RTTR_TRRT", "PJ2017_4_3"));
pumas_be(data_rep, FDA_NarrowTherapeuticIndex, endpoint = :Cmax)

    Look at the output from running this command and answer the following:

    1. What are the three distinct criteria that must all be passed for the overall assessment to be Pass? (Hint: Look at the row labels in the output table).
    2. For each of these three criteria, what was the individual assessment (Pass or Fail)?
    3. What was the final, overall assessment for Cmax, and how is this final conclusion reached based on the individual assessments?

  4. Study Design Considerations

    Based on the information presented in the unit, what is a key difference in the required study design when planning a study for an NTID submission to the FDA versus a submission to the EMA? Why is this design feature necessary for the FDA’s approach?

References

EMA. 2010. “Guideline on the Investigation of Bioequivalence.” 2010. https://www.ema.europa.eu/en/documents/scientific-guideline/guideline-investigation-bioequivalence-rev1_en.pdf.
FDA. 2012. “FDA, Draft Guidance on Warfarin Sodium. Recommended Dec 2012.” 2012. https://www.accessdata.fda.gov/drugsatfda_docs/psg/Warfarin_Sodium_tab_09218_RC12-12.pdf.

Reuse