What is IVIVC?

An in vitro in vivo correlation (IVIVC) is a predictive mathematical model that describes the relationship between an in vitro property of a dosage form and a relevant in vivo response. When performing IVIVC for formulation development, the in vitro property is primarily dissolution or drug release and the in vivo response is primarily a drug’s plasma concentration or the amount/rate of drug absorbed.1. In other words, IVIVC expresses the relationship between drug release in a dissolution apparatus and how that translates to the amount of drug that enters the bloodstream following administration. This type of relationship is likely to exist when a drug has high solubility and dissolution is the rate limiting factor in the process of drug absorption. IVIVC is important for many different scenarios but is especially important for extended release oral formulations.

Why Conduct IVIVC?

An IVIVC model is recommended by regulatory authorities for most modified release dosage forms. The main advantage of IVIVC is that it provides a mechanism for evaluating the change in in vivo absorption based on in vitro dissolution changes when there are small changes in a formulation. Once a validated IVIVC model has been established, it can be used to predict bioavailability and bioequivalence (BA/BE) based on in vitro data that are already available. In such cases, dissolution test results can be used to provide the desired information without the need for any clinical BE studies with human subjects. Another advantage of IVIVC is that it conveys a better understanding of the drug product itself. This can help establish a wider drug product acceptance criteria and formulation stability. IVIVC can also be especially useful for predicting the in vivo effects of changes to the formulation components, manufacturing site, or process. This is extremely important during initial product development however, the value of IVIVC does not end there. Establishing an IVIVC model can be even more valuable after the product has been approved by determining the impact of post-approval manufacturing changes, site of manufacture changes, and any issues with individual lots of manufactured products. All of this can be determined without having to repeat costly in vivo BE studies.

Benefits of IVIVC

IVIVC can benefit programs in many ways and for a variety of submission types. IVIVC analyses can be used to support:

  • Abbreviated New Drug Applications (ANDA)
  • New Drug Applications (NDA) for oral drugs with extended release characteristics
  • Abbreviated Antibiotic Drug Applications (AADA) as a surrogate for in vivo BE determinations

IVIVC can also be used to support biowaivers, which allows Sponsors to waive in vivo BA and/or BE study requirements. When requesting biowaivers for drug manufacturing changes, IVIVC can be used in lieu of certain otherwise required in vivo studies if sufficient safety and efficacy have been established. Recently, IVIVC has been used in the Quality by Design (QbD) framework to establish clinically meaningful drug product specifications using dissolution as the endpoint.IVIVC can be used for setting dissolution specifications such as:

  • supporting strength change justification
  • small changes in the formulation
  • changes to the site of manufacture
  • batch-to-batch quality control

FDA Guidance for IVIVC

The FDA Guidance, “Extended Release Oral Dosage Forms: Development, Evaluation, and Application of In Vitro/In Vivo Correlations,” is more than 20 years old but remains the definitive source of regulatory thinking on IVIVC. At the time of its release, the ability to accurately and precisely predict expected BA characteristics for an extended release product from its dissolution profile had been a long sought-after goal. The recommendations within the guidance cover IVIVC for oral, extended release drug products that are being developed for regulatory review as part of an NDA, ANDA, or AADA. The guidance outlines:

  • how to develop an IVIVC model
  • how to evaluate predictability
  • how to use IVIVC to establish specifications for dissolution
  • how to apply IVIVC as a surrogate for in vivo BE studies

Levels of IVIVC

There are three primary IVIVC categories, known as Levels A, B, and C. There is also a subcategory known as multiple Level C correlation. Level A is the most common type of IVIVC and historically used primarily for NDAs and investigational new drug (IND) applications.2 Level C can be useful in the early stages of development and is the second most common. Level B and Multiple Level C correlations are comparatively rare.

Level A

Level A correlation is generally linear and represents a point-to-point relationship between in vitro dissolution rate and in vivo input rate, although nonlinear correlations can be considered if found appropriate. IVIVC correlation is commonly estimated by a two-stage procedure that includes deconvolution followed by comparing the amount of drug absorbed to the drug dissolved. Level A should be used when demonstrating an IVIVC relationship for two or more formulations with different release rates. When feasible, Level A is the approach most highly recommended by the FDA and is considered the most informative. The FDA guidance suggests that two or more formulations with different release rates (e.g., slow, medium, and fast release formulations differing by 10%) should be used in demonstrating an IVIVC relationship. The use of only one formulation may be considered for formulations when in vitro dissolution is independent of the dissolution test conditions (e.g., medium, agitation, and pH). According to the guidance, “The model should predict the entire in vivo time course from the in vitro data.”

Level B

Level B correlation uses the same data used in Level A, but is based on the principles of statistical moment analysis. The mean in vitro dissolution time of the drug is compared to either:

  1. the mean in vivo residence time or
  2. the mean in vivo dissolution time

Level B is the least useful for regulatory purposes because it does not reflect the actual in vivo plasma level curves. Also, in vitro data from a Level B correlation cannot be used to justify the extremes of quality control.

Level C

Level C correlation involves determining the relationship between in vivo pharmacokinetic (PK) parameters (e.g. Cmax, AUC, Ka) and in vitro dissolution data at a single point. Level C does not reflect the complete shape of the plasma concentration time curve, which is a critical factor for determining the performance of extended release products. It is also not sufficient for obtaining a bioequivalence waiver. Level C can predict Cmax and AUC, which can help you to establish BA and BE.

Multiple Level C

Multiple Level C correlation relates one or several PK parameters of interest to the amount of drug dissolved at several time points and can be as beneficial as Level A correlation. However, if multiple Level C correlation is possible, the existence of Level A correlation is also highly likely and often preferable. Sample sizes for these types of studies should be between 6 and 36, and only human data are permitted. Crossover studies are preferred, but parallel studies may be acceptable. The reference product to develop an IVIVC can be intravenous, immediate release, or an aqueous oral solution. Although a fasting state is preferred, a fed state may be acceptable for safety reasons.

5 General IVIVC Considerations

  1. Successful IVIVC relationships demonstrate that different release rates of two or more formulations result in corresponding differences in absorption profiles via the same absorption mechanism.
  2. For each of the formulations studied, the release rates, as measured by percent dissolved, should differ by at least 10%. This should also result in in vivo profiles that show comparable differences in Cmax and AUC between each formulation (i.e., formulations that differ by a given percentage in vitro should show a corresponding difference in vivo).
  3. The predictive performance of an IVIVC model is estimated as prediction error. The evaluation of this prediction error is based on either internal datasets and/or additional external datasets and depends upon the intended application of an IVIVC analysis and the therapeutic area of the drug.
  4. IVIVC development is much more likely to successful for drugs with high solubility, when in vitro and in vivo data are available from a variety of formulations including a solution formulation as a reference.
  5. Historically, the overall FDA acceptance rate of IVIVC submissions has been less than 50%. Factors that may contribute to low IVIVC success rates include:
  • Not selecting appropriate formulation amounts and types for IVIVC development and validation
  • Not reviewing exploratory plots to help guide model building and selection
  • Not investigating the reasons behind inconclusive predictability, quality, and richness of input data
  • Not choosing enough parameters for parameterization


IVIVC has many advantages. Not only does IVIVC provide a better understanding of the dosage form, but it provides a predictive tool that can eliminate the need for certain clinical BE studies, help in interpreting batch-to-batch variability, and help to optimize formulation development, thus streamlining product development and manufacturing. However, developing an appropriate IVIVC model is not trivial. If you would like to discuss whether an IVIVC approach would be appropriate for your development program, contact Alucent today. Our consultants have many years of experience developing effective IVIVC models for our clients.


  1. Guideline on the pharmacokinetic and clinical evaluation of modified release dosage forms (EMA/CPMP/EWP/280/96 Corr1), European Medicinal Agency, Nov 2014.
  2. Suarez-Sharp S, 2012, IVIVC workshop
  3. Suarez-Sharp S, Li M, Duan J, Shah H, Seo P. Regulatory Experience with In Vivo In Vitro Correlations (IVIVC) in New Drug Applications. AAPS J. 2016 Nov;18(6):1379-1390.
  4. FDA Guidance for Industry. Extended Release Oral Dosage Forms: Development, Evaluation, and Application of In Vitro/In Vivo Correlations. Sep 1997.
Share this: