Demographic matching in organ impairment studies

Organ impairment (OI) studies, covering hepatic impairment (HI) and renal impairment (RI), are a critical part of clinical drug development for both small molecules and certain classes of large molecules. These studies help characterise pharmacokinetics (PK) in populations with impaired organ function, which is essential for safe and effective dosing.

Demographic matching plays a key role in ensuring proper characterisation of PK by minimising or excluding confounding factors such as age, weight, sex, race and other variables. Yet, despite its importance, there is little guidance on how to implement matching in practice or on the strategies available to select appropriate control groups.

To address this gap, ICON’s Consulting & Advisory experts conducted a systematic literature review of 170 renal (RI) and 173 hepatic impairment (HI) studies spanning 25 years to assess the use and implications of different matching strategies and to propose a framework for improved selection of matching methods.

Demographic matching matters—but which method?

Regulators have long recognised the need for demographic matching. The first FDA guidance on clinical evaluation of PK in participants with organ impairment was published in 2003. However, regulators have not specified a preferred methodology, leaving sponsors and researchers to make their own decisions. Over time, the number of studies performed without demographic matching has decreased, reflecting growing scientific rigor.

Why does demographic matching matter? When designing organ impairment studies, demographic matching is essential to ensure that differences in PK are attributable to organ function rather than unrelated demographic factors. Factors such as age, weight and sex can confound PK results, making it harder to isolate the effect and therefore impact safety, efficacy and labeling considerations.

Our review of 25 years of literature shows two dominant strategies: individual matching (IM) and group matching (GM), with several variations in GM.

Individual vs group matching explained

Individual matching (IM) pairs each participant with organ impairment to a demographically similar control. This approach offers robust comparability but requires a larger sample size. IM offers the highest level of comparability because it ensures that each impaired participant has a closely matched counterpart. This reduces variability and strengthens the scientific robustness of the study. However, IM comes with practical challenges: it requires a larger control group, often equal to the combined number of participants across all impairment severity groups—which is also why it is sometimes referred to as a 1:1 method.

Group matching GM takes a different approach by matching the control group to the overall demographic profile of the organ impairment group rather than matching individuals 1:1. This method typically requires fewer control participants, making it attractive when sample size or recruitment feasibility is a concern. GM variants include:

GM1: Matching the distribution – The control group is selected to mirror the full range and distribution of demographic characteristics (age, weight/BMI, sex) found in the organ impairment groups.
GM2: Matching the mean – The control group is matched to the mean demographic values of the organ impairment groups within predefined acceptance limits. This is simpler and faster than GM1 but does not capture demographic extremes, which may limit comparability in heterogeneous populations.
GM3: Partial matching – Only one impairment group (e.g., moderate impairment) is matched to controls, or matching is applied selectively across groups. This approach is sometimes used when full matching is impractical, but it introduces variability and potential bias.

Of the two main methods for matching, they each became a preference for the two indications considered. IM is more common in hepatic impairment studies (55%), reflecting the higher variability and complexity of hepatic PK. GM dominates renal impairment studies (70%), likely due to the relative predictability of renal clearance and the feasibility advantages of GM. Within GM, GM2 (matching the mean) is the most frequently used variant.

Figure 1: Schematic representation of common matching strategies (*partial matching example depicts scenario where only a single group is matched)

What criteria should be used in demographic matching?

The choice of matching criteria should be driven by factors most likely to influence pharmacokinetics (PK). Our review shows that age, weight or body mass index (BMI), and sex are the most commonly used criteria, and they remain the foundation of demographic matching. Typical acceptance margins are ±10 years for age and ±10–20% for weight or BMI, which balance scientific rigor with recruitment feasibility.

While BMI is frequently used, its relevance to PK is debated. BMI categorises obesity but does not always reflect physiological changes that affect drug absorption, distribution or clearance. In many cases, body weight or an adjusted weight descriptor provides a more reliable basis for matching, particularly for drugs where distribution or clearance correlates with body size.

Additional factors can be considered when supported by pharmacology or population PK data:

Smoking status: relevant for drugs metabolised by CYP1A2
Pharmacogenomics: when genetic polymorphisms influence metabolism or transport
Race or ethnicity: when ethnic differences in PK are suspected or documented

However, including more than three criteria significantly complicates recruitment and may delay study timelines. For rare factors such as specific genotypes, feasibility often dictates whether matching is practical. In such cases, capturing the data and addressing variability through population PK analysis may be a better alternative.

In short, start with age, weight and sex, apply acceptance margins that reflect regulatory norms, and add other criteria only when justified by drug-specific and data-informed pharmacokinetics considerations.

Recruitment, sample size and study duration

Typical organ impairment studies recruit 6–10 participants per group. IM generally requires more control participants than GM, but it allows parallel recruitment, which can offset complexity. GM may reduce sample size but often involves staggered recruitment, adding logistical challenges. Recruitment and study duration are also influenced by site-level factors such as access to participant databases and operational priorities.

Our review found no statistically significant difference in overall study duration between IM and GM strategies, although GM1 tends to take longer due to the difficulty of matching across the full demographic distribution. Practical challenges such as staggered recruitment for GM can add complexity.

Factoring a standard framework: ICON’s recommended strategy

Based on the evidence, ICON recommends:

IM when PK variability is high or demographic influence is uncertain, as it ensures the most comparable populations and covers demographic extremes.
GM when PK is well characterised and variability is low, as it reduces sample size without compromising scientific integrity.
Use population PK data to guide matching criteria and acceptance margins where available. In the absence of such data, apply common thresholds: ±10 years for age and ±10–20% for weight or BMI.
Unless otherwise justified, matching should also be implemented when enrolling the targeted patient population, as dictated by recruitment feasibility of the specific patient population.

This framework aligns with current EMA and FDA recommendations and provides practical thresholds for age, weight and BMI margins.

Matching for stronger research and regulatory confidence

Demographic matching improves scientific robustness and regulatory confidence. Both FDA and EMA expect demographic matching but do not mandate a specific method. Studies without matching can still gain regulatory approval if population PK analyses show that demographic factors have no clinically meaningful impact on PK. For tailored matching strategies and expert support, connect with us and out consulting and asset development solutions experts.

While this review focuses on organ impairment studies, the principles may also apply to other clinical pharmacology studies, such as ethnic bridging or genotype–PK investigations, where demographic matching can help exclude confounding effects.

For a deeper dive into the insights and survey methodology, read the full published research in the AAPS Journal here.