An IQ Industry Perspective on Informing Dosing Recommendations in Patients With Renal Impairment

Abstract

The IQ CPLG Organ Impairment WG conducted a survey to understand how various IQ member companies develop dosing recommendations for patients with renal impairment. Eighteen IQ member companies participated in the survey. The survey results were summarized by the working group in light of the regulatory renal impairment guidance documents as well as recent publications from the pharmaceutical industry and nephrology community. There were two important learnings from the survey: (i) pharmaceutical companies do not use a consistent methodology to assess renal function in their drug development programs. (ii) there has been significant improvement in predicting how kidney function affects drug pharmacokinetics (PK) and thus dose recommendations. Applying model‐based methods such as population PK, physiologically‐based PK (PBPK), and virtual controls has enabled earlier prediction of how kidney function influences PK, leading to the participation of more patients with impaired kidney function in Phase 2 and 3 trials.

The International Consortium for Innovation and Quality in Pharmaceutical Development (IQ) Organ Impairment Working Group (WG) conducted a survey in 2020 among the member pharmaceutical companies to collect information on the current approaches and experiences for evaluating the impact of kidney function impairment on drugs under development and informing dosing recommendations.

IQ is a not‐for‐profit organization of pharmaceutical and biotechnology companies with a mission of advancing science and technology to augment the capability of member companies to develop transformational solutions that benefit patients, regulators, and the broader research and development community. The mission of the IQ Organ Impairment working group is to enhance the current paradigm for informing dosing recommendations for patients with organ impairment. Among others, the objectives of the working group are to map out multiple pathways to facilitate the inclusion of participants with organ impairment in Phase 2/3 studies and to enhance current practices for conducting Phase 1 clinical pharmacology studies in participants with organ impairment.

Aligned with the WG mission and prompted by the draft FDA guidance on renal impairment (RI) studies issued in 2020, ¹ the WG conducted a survey to understand how various member companies develop dosing recommendations for patients with renal impairment. The survey questions were related to whether and how dedicated renal impairment studies are conducted, the timing of the study in relation to the overall drug development program, formulae utilized for kidney function assessment, kidney impairment severities included in the study, and differences in approach for new chemical entities vs. new biologic entities. These results are summarized and discussed in this paper, and selected results are discussed in the context of the recently issued final FDA renal impairment guidance. ² This paper is intended to spark further conversations among industry and regulatory experts on preferred and contemporary methods for evaluating the impact of renal impairment on PK and provide recommendations on how to improve the current paradigm to enroll patients with renal impairment in Phase 2/Phase 3 studies.

IQ SURVEY

The survey questions were drafted by the IQ Organ Impairment working group members and were reviewed and endorsed by the IQ Board and the IQ CPLG. The survey consisted of 32 questions and was divided into three parts. Part 1 (Question 1–Question 12) focused on the conduct of dedicated pharmacokinetic studies in participants with renal impairment. Part 2 (Question 13–Question 26) focused on enrolling participants with renal impairment in Phase 2 and Phase 3 studies. Part 3 (Question 27–Question 32) had general questions on participating companies, regulatory interactions, and future needs in the area of informing dosing recommendations in participants with renal impairment. The survey questions are listed in Supplementary Information. The survey was conducted according to IQ's Bylaws and Antitrust Guidelines between October and December 2020. Each IQ member company was requested to provide a single response to the survey questions to represent company's position. Survey responses were received from 18 of a total of 37 IQ member companies in 2020. Responses from individual companies were kept confidential and were anonymized by the IQ independent third‐party secretariat (Faegre Drinker Biddle & Reath LLP). Survey results were reviewed and analyzed by the IQ WG team members and relevant responses to Parts 1 and 2 are summarized in this publication. Half (9) of the biopharmaceutical companies that responded to the survey were large companies with > 10,000 employees, while seven were medium‐sized (1000–10,000 employees) and two were small (< 1000 employees) [IQ Survey Part 3].

PART 1

Conduct of dedicated renal impairment pharmacokinetic studies

Based on survey responses, for both small and large molecules, the decision to conduct a dedicated (reduced or full design) renal impairment study is usually made in Phase 1 for oncology drugs, and during Phase 2 or Phase 3 of development for other therapeutic areas. Factors driving the timing of this decision include the extent of renal excretion of unchanged drug or its active metabolite, modality (small molecule or large molecule), therapeutic area/ indication/ target population, and duration of treatment.

The decision to conduct renal impairment studies is mainly driven by animal and human (single and/multiple ascending dose) data on urinary excretion of the drug, human mass balance study data, intended target population (including the impact of disease on renal function), and extent of metabolism. Other factors that could play a role include preclinical renal toxicity profile of the drug, drug site of action, data from other drugs in the class, and involvement of transporters in the disposition of the molecule. The considerations are consistent with the FDA guidance which indicates that a renal impairment PK study is recommended for small molecules and therapeutic proteins and peptides with a molecular weight less than 69 kDa when impaired renal function is likely to alter the PK of the unchanged drug or its active metabolites because they could be substantially eliminated (> 30%) by the kidneys, and a full pharmacokinetic study design is recommended for such molecules. For drugs that are predominantly eliminated via nonrenal routes and likely to be used in patients with impaired renal function, a reduced pharmacokinetic study design can be used to determine the impact of renal impairment on PK and provides a worst‐case scenario of the impact. In fact, survey responders indicated that on average ~60% of renal impairment studies conducted for small molecules were full design studies, while ~40% were reduced design studies suggesting that a good proportion of NCEs in development at this surveyed cohort of companies comprise of molecules with substantial renal excretion.

With regards to enrollment of severe renal impairment vs. ND‐ESRD (non‐dialysis dependent—end‐stage renal disease) participants in dedicated renal impairment studies, most survey participants recommended enrollment of severe renal impairment participants over ND‐ESRD participants with the justification that ND‐ESRD participants constitute a vulnerable population which is difficult to enroll. Additionally, ND‐ESRD participants are at risk for progressing into dialysis dependent—end‐stage renal disease (DD‐ESRD) very quickly (sometimes during the trial). The participants with severe renal impairment are considered more stable in terms of renal function compared to ND‐ESRD participants. Per the recently issued FDA guidance and EMA guidance, PK data in participants with DD‐ESRD are recommended to be collected for drugs that are likely to be administered to these patients to determine the contribution of the specific replacement modality (e.g., intermittent hemodialysis) to the elimination of the drug and its potentially active metabolites.

Survey responders indicated that challenges with enrolling ND‐ESRD or severe RI subjects in dedicated RI study included participant availability, safety of the drug in this population, and impact of concomitant medications. Dialysis schedule and fluctuating renal function in the subjects were the other challenges in conducting a study in DD‐ESRD subjects. These challenges could be mitigated by using multiple, experienced and/or dedicated sites, increased advertising, staged enrollment, dose reduction/up‐titration.

Formulae for estimation of renal function

Adults

When the survey was originally conducted, 61% of the responders indicated a preference for the Cockcroft‐Gault (C‐G) formula for enrolling participants in dedicated renal impairment studies. No specific preference for formula was indicated by the remainder of the responders. Most responders indicated that they do not measure GFR in renal impairment studies. Interestingly, survey responses indicated an almost equal split in the use of C‐G, the modification of diet in renal disease (MDRD) equation, or Chronic Kidney Disease Epidemiology Collaboration (CKD‐EPI) equation to assess/monitor renal function in adults enrolled in Phase 2/3 trials. Since then, several publications have proposed a move to more contemporary equations for estimating GFR. ³ , ⁴ , ⁵ Additionally, the National Kidney Foundation and American Society of Nephrology recommended a refit of the CKD‐EPI_Scr and CKD‐EPI_Scr‐Scys equation without the race variable for adults. This prompted the IQ Organ Impairment working group to publish a manuscript ³ on the appropriateness of these formulae. More recently, the working group re‐surveyed the IQ member companies for formula preference to evaluate whether the new race‐independent CKD‐EPI formula ⁶ was being adopted by the pharmaceutical industry. Fourteen IQ member companies responded to the new survey and the use of C‐G equation has declined to 23%. Seventy‐seven % of the responders preferred using eGFR (41%: CKD‐EPI. 27% MDRD; 9% race‐independent CKD‐EPI).

The re‐survey indicated that the use of eGFR has gone up and the use of Cockroft‐Gault has gone down among the responding companies which is a change in the right direction. ³ The EMA guidance on the evaluation of the pharmacokinetics of medicinal products in patients with decreased renal function does not recommend a specific formulae for renal function assessment. ⁷ The recently released FDA guidance on the pharmacokinetics in patients with impaired renal function alludes to measured GFR, measured creatinine clearance, estimated GFR, and estimated creatinine clearance (using alternative body weight metrics) as some of the acceptable means to estimate renal function. ² Considering the pros and cons of different methods/equations, the FDA notes in the guidance that any contemporary, widely clinically accepted estimating equation could be used for the population being studied. The inclusion of a recommended method for assessing renal function in the guidance documents will facilitate the allocation of subjects to appropriate renal function category without concern of reallocation due to a different method during regulatory review. Given the lack of recommendation for a particular method for estimation of renal function in the draft and final FDA and the final EMA guidance documents, it is not surprising that the use of Cockcroft‐Gault and MDRD equations has persisted in the pharmaceutical industry as seen in responses to the original and new survey. Unless dealing with in‐clinic patients, measuring GFR or creatinine clearance is rarely used due to the inconvenience as well as potential inaccuracies (unless patient is catheterized). In terms of estimated renal function, labs have largely switched to IDMS‐standardized serum creatinine measurements, which were used to design the CKD‐EPI equation, while the older non‐IDMS standardized methods were used for Cockroft‐Gault equation. CKD‐EPI equation (race‐independent equation in the United States since 2020) is the most used and accepted equation in clinical practice in the United States currently and is supported by the KDIGO guidelines. ⁸ Hence, CKD‐EPI equation would fit the description of a contemporary equation described by the FDA guidelines and its use in clinical programs would make the results more translatable to and consistent with clinical practice.

Pediatric

For pediatric participants, most (62%) survey responders indicated the use of the Original Schwartz, followed by limited usage of Bedside/Modified/Interim Schwartz, CKiD Schwartz, and Cockcroft‐Gault (for adolescents only) formulae.

In response to the re‐survey, most (64%) participants reported using Bedside/Modified/Interim Schwartz followed by Original Schwartz (28%) and limited usage of CKiD and Cockcroft‐Gault formulae suggesting that more companies have switched to using the Bedside Schwartz formula. As with the recommendations related to the estimation of renal function in adults, the EMA and FDA guidance documents do not provide specific recommendations for formulae to be used/ avoided for the pediatric population. Most laboratories have moved to IDMS‐traceable serum creatinine measurements which would preclude the use of the Original Schwartz equation. While equations like the CKiD‐Schwartz equation are available, they require the measurement of biomarkers beyond serum creatinine, which makes them less attractive for routine implementation in clinical trials. As such, the Bedside/Modified/Interim Schwartz equation or the newly available U25 equation could be considered more contemporary for use in pediatric patients. ⁹ However, the U25 equation needs further validation for routine implementation in clinical practice or trials.

Recently, a full age spectrum equation (EKFC equation) was introduced by a group in Europe. ¹⁰ This equation would have the advantage of being applicable across various age groups but is not frequently used in clinical routine and more validation would be needed. In a small validation vs. Asian‐modified CKD‐EPI equation, no clinically meaningful difference was observed in Chinese CKD patients. ¹¹

Markers of renal function assessment and bioanalytical methodology

Most companies did not measure any marker of renal function other than serum creatinine in late‐phase studies on a regular basis. The most commonly measured marker of renal function other than serum creatinine was cystatin C, particularly with drugs that inhibited renal transporters or conditions that altered creatinine metabolism. Several bioanalytical methods are available to analyze creatinine concentrations and the choice of bioanalytical methods impacts the GFR formulae that can be employed. ³ Some participating companies indicated the use of enzymatic method in late‐phase trials, whereas others reported using the Jaffe method or did not provide a response. Some responders commented that the method has not been standardized across their company and both methods have been used in different programs. The majority of the participating companies indicated using IDMS‐standardized bioanalytical method. The aspects of bioanalytical method for measuring serum creatinine and its impact on the choice of method have not been discussed in any of the regulatory guidance documents.

Choice of reference group and analyses to guide dosing recommendations

The survey results indicated that in dedicated renal impairment studies, the preferred reference group was demographically‐matched participants with normal renal function as it is the typical regulatory requirement and due to ease of recruitment and interpretation of results. This is consistent with the FDA and EMA guidance documents that indicate participants with impaired renal function should be similar to the control group with respect to factors known to affect the drug's PK, for example, demographics like age, sex, race, ethnicity, pharmacogenetic factors, or weight. Both the FDA and EMA guidance documents suggest that the reference group should best represent renal function in the efficacy/safety study population and, thus, may not necessarily be the group with normal renal function. As such, some respondents preferred the use of participants with normal renal function or mild renal impairment in the case of drugs with a wide therapeutic index as these were more representative of the target population that may be older and also allowed for a smaller study size.

With regard to data analysis methods, most responders indicated that they used both pairwise comparison (GMR (Renal Impairment/Control) and 90% CI) and regression analysis for analyzing dedicated renal impairment study data. Pairwise comparison was preferred by some as it is a simple method and commonly used, particularly for reduced design studies while regression analysis was preferred by others because it uses totality of data, minimizes the impact of outliers, is amenable to extrapolation, and can also be used to compare data across studies and is therefore more informative. Responses also indicated that both methods (pairwise comparison and regression analysis) were presented in the clinical study report. However, regulators now clearly recommend the use of regression analysis (or mathematical models) as the preferred method due to the potential for identifying more clinically meaningful thresholds for dose modification.

For deriving dosing recommendations in renal impairment participants, most responders preferred matching based on point estimate (GMR RI to Healthy in dedicated RI study) because this method is simple, easy to implement, and commonly used, particularly given the limited sample size in these studies which makes a CI‐based approach impractical. Some responders preferred matching based on confidence intervals (CI of the mean in RI to predefined “no‐effect boundary”) as it was considered practical and based on clinical relevance. A small number of responders also preferred matching to the exposure range (10^th–90^th percentile) observed in clinical trials as a more dynamic method. The FDA guidance proposes that dosing recommendations should be based on exposure‐matching to a reference group with an acceptable benefit–risk profile for the drug (not necessarily the normal renal function group). The FDA and EMA guidance documents also mention the use of other approaches for dosing recommendation, for example, pharmacokinetic simulations that project systemic exposures within the 5^th and 95^th percentiles or a specific target range of the reference group or establishing no‐effect boundaries. Regardless of the matching approach, the use of modeling and simulation to derive dosing recommendations is preferred.

PARTS 2 AND 3

Inclusion of participants with organ impairment in phase 2/phase 3 clinical studies

In general, the majority of participating companies (89%) attempt enrolling participants with renal impairment in Phase 2/3 studies. Totality of available knowledge on the drug is used to inform the decision. Factors driving the decision include type of the molecule (small or large), availability of data from a dedicated renal impairment study, urinary excretion of the drug or human mass balance data, or modeling and simulation (PBPK or population PK) predictions of exposure change in participants with renal impairment, renal toxicity profile of the drug, data from other drugs in the class, involvement of transporters in the disposition of the molecule, and intended target population. The typical approach is to exclude participants with severe renal impairment unless information is available from a dedicated renal impairment study or modeling and simulation. Few cases (< 3) have been reported where participants with severe renal impairment have been enrolled in Phase 2/3 studies. Notably, participants with moderate and severe renal impairment are more commonly included in Phase 2/3 studies for drugs developed to treat rare diseases or cancer. About 55% of survey responders indicated that model‐based approaches (such as population PK or PBPK models) are being regularly used to predict the impact of renal impairment on exposure. These early predictions facilitate step‐wise enrollment of more participants with increasing severity of renal impairment in Phase 2/3 studies. However, there are instances where the enrollment of these participants in Phase 2/3 studies is limited due to a lack of availability of predictive models in early clinical development.

The experience with regulatory agencies worldwide has been mixed with regard to enrolling participants with renal impairment in Phase 2/3 studies ranging from encouraging enrollment to requiring a dedicated renal impairment study prior to enrollment. In some instances, regulatory agencies allowed the enrollment of participants with mild or moderate renal impairment without the need for a dedicated renal impairment study. Hurdles to enrolling participants with renal impairment in Phase 2/3 studies identified by the survey participants could be clustered into three categories:

1. Safety: Safety concerns for participants with severe renal impairment or ND‐ESRD participants in addition to the potential to confound the overall safety database.

2. Study protocol planning: Ethical and safety concerns raised by IRBs or by regulatory agencies in some cases. Participants may get excluded due to the use of concomitant medications excluded in the protocol. Clinical trial protocol planning may also run into difficulties in implementing a different dose regimen for participants with renal impairment due to the lack of pharmacokinetic data in participants with renal impairment at the time of planning Phase 2/3 studies.

3. Efficacy: Potential confounding on efficacy such as the impact of concomitant medications for participants with renal impairment, in addition to not having a good understanding of the impact of renal impairment on PK/PD at the time of Phase 2 study.

Majority of participating companies reported a successful initial NDA/BLA filing without conducting a dedicated renal impairment study where specific dosing recommendations (e.g., contraindication, administration is not recommended) for drug use in patients with renal impairment were included in labeling. These cases included drugs with low urinary excretion/primarily cleared hepatically, large molecules (e.g., monoclonal antibodies) or topicals for which modeling data indicated that dose adjustment would not be warranted in patients with renal impairment. The possibility of doing the initial submission without renal impairment data for oncology/rare diseases was indicated. On the other hand, cases were reported where regulatory agencies did not agree on including dosing recommendations for patients with renal impairment in labeling without conducting a dedicated renal impairment study. In other instances, regulatory agencies have requested a dedicated renal impairment study as a post‐marketing commitment or requirement.

This practice among the surveyed pharmaceutical companies is consistent with the FDA and EMA guidance documents that provide the possibility of evaluating the impact of renal function on PK as a part of Phase 2/3 trials using population PK. The guidance suggests that if the number of patients with severe or ND‐ESRD patients is insufficient in these trials, a reduced design renal impairment PK study may be needed.

In general, the majority (> 80%) of participating companies preferred to conduct reduced renal impairment studies with utilization of modeling and simulation approaches to predict exposure in patients with renal impairment of lesser severity and enrolling participants with renal impairment in phase 2/3 trials.

SUMMARY AND RECOMMENDATIONS

The survey focused on understanding current practices in the pharmaceutical industry for informing dosing recommendations in patients with renal impairment. There were two important learnings from the survey:

1. Pharmaceutical companies use different methods for assessment of kidney function in their drug development programs. Cockcroft‐Gault, original MDRD, and original Schwartz equations are still utilized by many companies for this purpose even though it is inconsistent with guidance from the nephrology community (e.g., KDIGO (Kidney Disease Improving Global Outcomes) or ASN (American Society of Nephrology)), and better alternatives (e.g., CKD‐EPI equation) are available. In fact, of the 17 NCEs approved by the FDA in 2023, 13 included labeling information on the impact of kidney function and of these six used Cockroft‐Gault formula for adults (data on file compiled from Drugs@FDA). In contrast, clinical labs and clinical practices in the United States and elsewhere have switched to more contemporary equations, like CKD‐EPI, that are based on IDMS‐standardized serum creatinine bioanalytical methods. As a result, labeling recommendations for dose adjustments for patients with renal impairment may not be accurate, are not in line with clinical practice, and therefore not optimal. A clearer recommendation from regulatory agencies toward using equations that are based on IDMS‐standardized bioanalytical methods can help mobilize the pharmaceutical industry to adopt the more contemporary equation(s) recommended by nephrology community. In fact, there is also a clear need from regulatory agencies to align on recommended method(s) to assess renal function and guide dose adjustment in partients with renal impairment to be inline with those used in clinical practice by nephrologists. Inclusion of a recommendation for a specific contemporary method for the estimation of eGFR in adults and pediatric participants would go a long way in enabling consistent application across the industry for dose recommendations in patients with renal impairment. One consideration is for regulatory agencies to recommend the use of the CKD‐EPI formula for adults. Since the FDA has recently engaged in a discussion with the pediatric nephrology community and pharmaceutical industry about dose adjustments for renal impairment in pediatric patients, they could also consider providing recommendations for the use of an appropriate method for pediatric participants.

2. The assessment of the impact of kidney function on PK has reached a new frontier in drug development. Majority of survey participants use model‐based approaches such as population PK and PBPK modeling to predict the impact of kidney function on PK, which can enable the inclusion of patients with impaired kidney function in Phase 2 and 3 trials. This is in line with the FDA guidance on renal impairment and the draft guidance on diversity and inclusion recommending the inclusion of participants with renal impairment early in the program to allow a more robust and clinically meaningful assessment of efficacy and safety in these participants in Phase 2/3 studies. ¹² , ¹³ However, greater progress is required to enroll them earlier in development and facilitate inclusion of participants with more severe renal impairment into clinical trials.

Traditionally, PBPK or population PK modeling has been used only retrospectively to assess the impact of kidney function impairment on exposure. ¹⁴ , ¹⁵ Recently, PBPK modeling has also been used to prospectively evaluate the impact of renal impairment on exposure ¹⁶ and inform inclusion of participants with mild renal impairment in early (e.g., first in human or proof of mechanism patient study) clinical studies. Results from these participants with mild renal impairment could then be used to validate or update the model as needed and predict the impact for participants with a greater degree of renal impairment and facilitate their inclusion in future clinical trials. This step‐wise approach facilitates mitigation of risks associated with inclusion of participants with renal impairment during development and is in line with recommendations from the FDA to better represent real‐world patient populations in clinical trials. Moreover, for drugs not expected to be impacted significantly by renal impairment, this approach provides a more robust dataset for population PK modeling in support of label recommendations.

More recently, efforts have been made to replace the healthy control group in dedicated renal impairment studies with a PopPK‐based characterization of drug PK in healthy participants without renal impairment. ¹⁷ , ¹⁸ , ¹⁹ Benefits of this approach include assessment of the impact of renal impairment on exposure using all available data from healthy participants instead of a small cohort of participants that is more susceptible to error due to interindividual variability. This approach also enables shorter timelines and lower costs for the dedicated renal impairment study which in turn may facilitate earlier inclusion of renally impaired participants in Phase 2/3 trials.

In summary, 18 IQ member companies participated in the survey on current approaches and experiences for conducting renal impairment studies and informing dosing recommendations in participants with kidney function impairment. The IQ organ impairment working group summarized the survey results in light of the regulatory renal impairment guidance as well as recent publications from industry and nephrology community. In general, industry practice for renal impairment participants is in line with the regulatory expectations at the time. However, there were also important learnings that could be gleaned from the survey. One is that there is no consistency in the choice of method for renal function assessment used by pharmaceutical companies during drug development. This is not in line with the current clinical practice and recommendations from the nephrology community in the United States and other parts of the world.

Finally, significant progress has been made in the prediction of the impact of kidney function on drug pharmacokinetics and thus dose recommendations. Implementation of model‐based approaches such as population PK, PBPK, and virtual controls has allowed much earlier prediction of the impact of kidney function on PK, resulting in the inclusion of more participants with impaired kidney function in Phase 2 and 3 trials as well as reduced study designs. This allows the collection of efficacy and safety data in these participants in drug development programs consistent with the FDA's diversity and inclusion plan.

FUNDING

No funding was received for this work.

CONFLICT OF INTEREST

The authors declare no competing interests in this work.

ETHICS STATEMENT

None.

DECLARATION

AS, VS, JK, and IY are employees of the pharmaceutical companies listed under their affiliation and may own stocks in these companies.

Supporting information

Data S1