Using technology to predict risk of heart disease for people with lupus

 

Scientific Study title:

SLE-CALCULATOR: Systemic Lupus Erythematosus CArdiovascuLar Disease Event Risk PrediCtion Using Machine LeArning Techniques and NOvel ThRombotic Autoantibodies

Study Start Date:

2021

Study Start Date:

2023

 

Why Did We Do This Research?

Cardiovascular disease (CVD), such as heart disease and stroke, is the most common cause of death in people with systemic lupus erythematosus (SLE). However, the reason for a higher risk of CVD in people with lupus is unknown. Researchers think it could be due to inflammation from SLE and/or the medications used to treat it. An accurate way to measure the risk of CVD is needed to identify lupus patients who may benefit from CVD prevention interventions. Therefore, the goal of this study was to create a SLE CVD prediction tool, that could be used in clinical settings to identify risks for heart disease and stroke early in the life of people with lupus.

 

What Did We Do?

First, we developed a new tool, called SLECRISK, specifically designed to predict CVD risk in SLE patients. This tool was created using machine learning. Machine learning teaches computers to learn information from data using statistics. Variables specific to SLE like disease activity, duration, and certain blood markers were included in the SLECRISK tool.

We studied patients who were 18 years of age or older, enrolled in the Brigham and Women’s Hospital SLE cohort. We collected data at the beginning of the study, including information such as traditional CVD risk factors (e.g., blood pressure), patient demographics, and SLE-specific characteristics (e.g., serum creatinine). We followed patients for ten years to track the occurrence of major adverse cardiovascular events, such as heart attacks, strokes, or cardiac deaths. We used medical codes and had cardiologists confirm each event. We compared SLECRISK with existing tools like the American College of Cardiology/American Heart Association (ACC/AHA) Pooled Cohort Risk Equations, Framingham Risk Score (FRS), and modified FRS (mFRS).

What Did We Find?

A total of 1,243 SLE patients over 10 years were included in this study, during which there were 90 major adverse cardiovascular events. SLECRISK was found to be more sensitive in identifying patients at moderate to high risk (≥7.5%) of major adverse cardiovascular events compared to the ACC/AHA model. SLECRISK also identified significantly more high-risk patients than ACC/AHA, particularly younger women with severe SLE who didn’t show many traditional CVD risk factors. This research is important because it shows that SLECRISK is a more sensitive and slightly more accurate CVD prediction tool.

The Research Team

Principal Investigator:

May Y. Choi, MD, FRCPC, Research Scientist, Arthritis Research Canada (University of Calgary)

Co-Investigators:

Karen H. Costenbader, MD, MPH (Brigham and Women’s Hospital & Harvard Medical School)

Brittany Weber, MD, PhD (Brigham and Women’s Hospital & Harvard Medical School)

 

Who Funded This Research?

Canadian Institutes of Health Research and the Lupus Foundation of America

 

Related Publications:

  • Choi MY, Guan H, Yoshida K, Paudel M, Kargere BA, Li D, Ellrodt J, Stevens E, Cai T, Weber BN, Everett BM, Costenbader KH. Personalizing cardiovascular risk prediction for patients with systemic lupus erythematosus. Semin Arthritis Rheum. 2024 Aug;67:152468. doi: 10.1016/j.semarthrit.2024.152468. Epub 2024 May 17. PMID: 38788567; PMCID: PMC11214838.
  • Shantz E, Elliott SJ, Sperling C, Buhler K, Costenbader KH, Choi MY. Towards an understanding of the biopsychosocial determinants of CVD in SLE: a scoping review. Lupus Sci Med. 2024 Jul 24;11(2):e001155. doi: 10.1136/lupus-2024-001155. PMID: 39053931; PMCID: PMC11284934.