The “Clinical Data Interchange Standards Consortium” (CDISC) is a not-for-profit organization that develops data standards for drug development. CDISC works with global regulatory agencies to develop guidelines and requirements that influence the standards for both clinical and nonclinical study data. Standard CDISC formatting guidelines help enable a more efficient submission and review process by providing regulatory agencies with clean, consistent, and standardized data. Within CDISC, an important subset of clinical trial data is pharmacokinetic data. Pharmacokinetics (PK) is the study of a drug’s time course profile within the human body via the processes of absorption, distribution, metabolism, and excretion (ADME). In simpler terms, it is the study of how a drug moves throughout the body. The analysis of PK data is an important aspect of clinical trials that helps define a drug’s concentration, exposure, and other PK parameters.
Sources of PK CDISC Data
It is critical to understand the differences between the sources and types of PK data collected during a trial. Equally as important is understanding how to combine those different data sources in the correct format in order to develop a dataset that can be used in a noncompartmental PK analysis (NCA). An NCA provides the most elementary PK information of a drug, such as peak concentration, elimination half-life, etc. The two main sources of data necessary for an NCA are:
- Electronic data capture
- Bioanalytical lab data
Electronic Data Capture
Data collected directly at the clinical study site are stored in an electronic data capture (EDC) system. Information about all aspects of a clinical trial is recorded in the EDC system. This includes the dates and times of each sample collected which is necessary for PK analysis. Data collected in the EDC range from:
- subject enrollment and randomization
- demographics
- dosing amounts and drug accountability
- vital signs and medical history
- and more
Bioanalytical Lab Data
Unlike the study data collected and stored solely in the EDC system (such as demographics), subjects’ PK samples, such as blood, urine, or plasma, require additional processing by a bioanalytical (BA) lab for PK analysis. PK samples collected from the clinical study site are transported to a BA lab to be processed. The BA lab analyzes the PK samples and provides drug concentration values that are not captured in the EDC system which are needed for PK analysis. Data generated by the BA lab tends to contain minimal study information, limited to only what is necessary to uniquely identify the PK samples along with the drug concentration results.
SDTM and ADaM Datasets
Data from the EDC system and data from external vendors (such as ECG results and the study deviations log) are then converted into CDISC datasets, including SDTM (study data tabulation model) and ADaM (analysis data model). SDTM datasets are a direct reformatting of the raw data available from the source, while ADaM datasets build on the foundation of available SDTM datasets to provide data in an analysis-ready format.
PC and ADPC Domains
The pharmacokinetic concentrations (PC) domain is created by combining information from the BA lab data, raw EDC data, and other SDTM domains. The PC domain often makes use of data from the exposure (EX) domain which contains drug dosing information as well as the trial visits (TV) domain which contains information on how study visits are formatted for a given CDISC package. The BA lab data and raw sampling time data from the EDC system are merged together based on the uniquely identifying information for each sample available in both the BA lab and EDC datasets. This often includes variables such as subject, matrix, sampling day, and nominal sampling times. Once these two sources are combined, additional information from other sources (such as additional SDTM domains) can be added. The ADaM dataset used to perform an NCA is the “analysis dataset of pharmacokinetic concentrations” (ADPC) domain and is a direct translation of the information in the SDTM PC domain. The ADPC domain is generated following the completion of the PC domain. The ADaM dataset format is less strict than that of SDTM and allows additional information to be added that is necessary to perform analyses, such as NCA. Generation of an ADPC domain includes the addition of:
- subject demographics
- treatment and dosing information
- calculated elapsed time following dosing
- imputation of concentration values that were below the limit of quantification
- flagging for both the analysis and associated tables listings and figures (TLFs)
PP and ADPP Domains
Upon completion of the NCA, the results are then compiled to generate the pharmacokinetic parameters (PP) domain. Similar to the PC domain, the PP domain is a reorganization of the raw data into a standard format, and it describes the PK parameters calculated from time-concentration profiles in the PC. This process makes use of the raw parameter export(s) as well as other SDTM domains (typically just the PC domain, though other domains can sometimes be used). An analysis counterpart to the PP domain is the “analysis dataset of pharmacokinetic parameters” (ADPP) domain which follows the completion of the PP domain. This dataset includes similar, additional columns to the ADPC such as subject demographics and flagging for any additional analysis, such as statistical analysis and TLFs.
Relating Records Domain
The related records (RELREC) domain is often overlooked but is an important part of a PK CDISC package. The PK-specific RELREC domain is used to relate the PC domain to the PP domain in an effort to highlight the PK concentrations used in calculating the PK parameters. The identifying variable used to relate the records between datasets is normally the unique subject identifier (USUBJID) and the sequence number (–SEQ) in the PC and PP domains. The RELREC domain aids in making the final CDISC package cohesive and traceable as it specifies cross-domain relationships.
Conclusions
In order to create a clean and comprehensible clinical trial submission that includes PK data and analysis, it is important for the analyst to understand both the sources of data received that will be used in the analysis as well as the specific formatting and dataset requirements that are part of a CDISC submission package. Effective communication between analysts and data managers helps ensure that the submitted package provides a clear representation of the data that was collected and analyzed. Allucent is an industry leader in PK CDISC standards and has extensive experience with generating PK datasets for legacy, planned, and ongoing studies. Allucent can help guarantee your program’s datasets are in compliance with the FDA’s required CDISC standards. Contact us to learn more about our expert PK CDISC team and services.