Aortic Stiffness Case Study

Introduction

The stiffness of the aorta has been found to be a predictor of cardiovascular risk, making it potentially useful for stratifying patients and as a target for therapy. However, the gold standard method for assessing aortic stiffness (measurement of carotid-femoral pulse wave velocity) is time consuming and requires a trained operator. Alternative approaches have been proposed, which assess aortic stiffness from the characteristics of pulse waves. These could potentially be incorporated into widely used medical devices, such as pulse oximeters. However, these arterial stiffness indices may not be as precise as the gold standard, and may be adversely influenced by other cardiovascular properties besides aortic stiffness. Therefore, the aim of this case study was to assess the performance of three arterial stiffness indices derived from the finger PPG pulse wave, and to investigate how they are affected by changes in other cardiovascular properties besides aortic stiffness.

The case study is described in this article.

Reproducing this case study

Requirements

You will need the following to reproduce the case study:

• The pwdb_data.mat file, which contains all of the data required for this case study. This file is available for download, and was produced when exporting the database.
• The Matlab ® script for this case study: asi_case_study.m
• The Matlab ® script for analysing the pulse wave database: analyse_pwdb.m
• The Matlab ® script for setting the paths for post-processing: setup_paths_for_post_processing.m
• An installation of Matlab ®

Steps

To reproduce the case study:

1. Make a directory in which to store the data and analysis, which will be known as the storage folder. For instance, you might choose: C:/Desktop/PW_studies/
2. Obtain the pwdb_data.mat file by either: (i) downloading it (see links here); or (ii) generating it yourself by reproducing the database (NB: this will take a long time).
3. Store pwdb_data.mat in a subdirectory of the storage folder named pwdb_#/exported_data/, where # is the number of the database. For instance, this directory might be: C:/Desktop/PW_studies/pwdb_1/exported_data/
4. (This can be skipped if already done previously) Open the setup_paths_for_post_processing.m, and adjust PATHS.storage_folder to the path of the storage folder (such as C:/Desktop/PW_studies/).
5. Open the co_case_study.m script in Matlab ®.
6. Run the script, specifying the number of the PWDB to be analysed (e.g. number 1). The script can be run using a command such as:

asi_case_study(1);

7. The plots showing the results of the case study should be saved in C:/Desktop/PW_studies/pwdb_1/case_studies/ . The following plots are relevant to this case study: (i) ASI_correlation_plot_hl_AGI_mod, (ii) ASI_correlation_plot_hl_RI, and (iii) ASI_correlation_plot_hl_SI.

In the article describing the results of the case study, we also presented results on the determinants of the three ASIs. The plots showing these results can be generated by: 8. Open the analyse_pwdb.m script in Matlab ®. 9. Run the script, specifying the number of the PWDB to be analysed (e.g. number 1). The script can be run using a command such as:

analyse_pwdb(1);

10. The plots showing these results should be saved in C:/Desktop/PW_studies/pwdb_1/analysis/figures . The following plots are relevant to this case study: (i) sens_AGI_mod, (ii) sens_RI, and (iii) sens_SI.