Ejection Fraction Endpoints

This example illustrates example analysis datasets for the following endpoints:

• From CV domain
  • Percent change in ejection fraction over time
  • Decline in ejection fraction over a time period (yearly) grouped by a decline greater than 10.0%.
• From the LB domain
  • Percent change in NTproBNP over a period of time (yearly).

Source Data

The SDTM examples used as the source data are from Section 2.1, Basic CMR Tests - Systolic Function. The CV dataset contains 16 rows, 1-8 for visit 1 and 9-16 for visit 6. Of these rows, the CVTESTCD = "LVEF_C" and CVTESTCD =  "RVEF_C" representing the Ventricular Ejection Fraction, Calculated (%) for left and right were selected for the analysis. The LB dataset test where LBTESTCD = "BNPPRONT" was used to add the value of BNPPRONT to compute the percent change over time, and then it was added as a ptential covariate in each row for the last analysis.  As in all ADaM datasets, the Subject Level Analysis Dataset (ADSL) was merged in to the SDTM datasets to capture all the necessary demographics, treatments and other required variables for analysis of the data.

Analysis Datasets

The example analysis dataset includes treatment and demographic information drawn from the ADSL, defined in the Metadata Table below. In this example, only some of the required variables, plus important stratification variables from ADSL, are shown for illustrative purposes. For example, one variable added; body surface area at baseline (BSABL)  was computed for baseline records in ADSL. However, since the subjects are children, the body surface area changes over time. Therefore, it would probably be necessary to have a separate dataset to represent this over time.

The table below shows the example analysis datasets and unique parameters for this example. The Subject Level Analysis Dataset (ADSL)  is a necessary component in the creation process of subsequent analysis datasets for human trials. The ADSL incorporates demographics, treatment groups, study dates, and stratification variables. In this example, the ADSL is not illustrated, but serves as a source for standard ADaM variables. Additional information can be found in Section 2.3.1 of the ADaM Implementation Guide (ADaMIG) v1.3.

The ADCVEF dataset includes tests pertinent to this analysis, those for which CVTESTCD is equal to "LVEF_C" or "RVEF_C". The relationship dataset RELREC illustrates the relationship between datasets. 

The ADCVNTRP dataset is a subset of the Laboratory Results (LB) SDTM dataset, for which LBTESTCD is equal to "BNPPRONT". In this case, many lab draws may occur throughout the year, but only those linked to a specific visit are included. By subsetting the LB dataset with LBLNKID not missing, laboratory data that aligns with CV data can be filtered out and sorted using USUBJID and VISIT. If LBLNKID is not available, a windowing strategy could be used to select the laboratory draw closest to that visit for merging with CV, computed in AVISIT.

ADCMRI is an efficacy analysis dataset with only a few records for more complex modeling of changes in ejection fraction, plus the ProB-type test results. It is an ADaM Basic Data Structure (BDS) dataset, with additional variables added from ADCVEF.

Analysis Datasets

The subsequent section provides the data structures for the example ADaM datasets. In this example, columns for derived variables BASE, CHG, PCHG and CHGCAT1 have been added to facilitate analyses of the efficacy endpoints. Specifically, example demographic data for ADSL were created for an 8 year old male. Also, a custom variable for BSA at baseline (BSABASE), was added from ADSL.

Example 1 

This example dataset shows the findings and additional analysis variables associated with:

• Left Ventricular Ejection Fraction, Calculated (%)
• Right Ventricular Ejection Fraction, Calculated (%)
• N-Terminal ProB-type Natriuretic Peptide (pg/mL)

The dataset above can also be set up to provide multivariate analyses by taking the lab tests of interest, transposing and merging with the ADCVNTPP dataset. In that case, the records from LB are not needed as a row.

The variables transposed and merged to the rows created from CV are:

Example Analysis Results Metadata (ARM) Tables

This table may be included in the submission package for regulatory agencies. It describes the analyses conducted for the study (usually limited to primary and secondary  endpoints) with details on data and analysis procedures used. The first example uses the dataset ADCVNTPR, selecting for the parameter "N-Terminal ProB-type Natriuretic Peptide (IU/L)".