The benefits of physical activity have been scientifically confirmed in thousands of studies. However, for athletes who engage in extensive high-intensity physical activity for professional sports, exercise does not protect against sudden cardiac death.
As the name suggests, this death is defined as an unexpected death within an hour due to severe, immediate heart disease rather than due to external factors. Sudden cardiac death without an obvious cause remains a problem for athletes and a major problem for physical health scientists to address. Recently, researchers at Italy’s University of Foggia used machine learning to analyze data from more than 700 athletes to solidify and confirm previously identified potential risks for sudden cardiac death.
Several factors are associated with sudden cardiac death, including age. The victim, who was under 35 years old, had a genetic heart disease that caused his death. For athletes under 35, hypertrophic cardiomyopathy, in which the heart muscle enlarges and prevents proper blood flow, is associated with 36% of sudden cardiac deaths, and the risk increases with age. For non-athletes over the age of 35, atherosclerotic coronary artery disease, a buildup of cholesterol and fat in the blood vessels, is commonly associated with 73% of sudden cardiac deaths.
The second common factor in sudden cardiac death is gender. This is because men are at higher risk than women. This difference may be due to differences in exercise-induced cardiac adaptations and ventricular remodeling between men and women, as found in previous studies from the University of London and the University of Oulu. These university studies also found differences in the prevalence of cardiac scar tissue, with men having more of it, suggesting that estrogen plays a protective role in women.
Finally, demographics are associated with sudden cardiac death. According to the Minneapolis Heart Institute Foundation, African American athletes are three times more likely to experience sudden cardiac death. The common theme that young cardiac death victims are athletes has led medical experts and researchers to believe that strenuous exercise can exacerbate existing heart abnormalities through dehydration and electrolyte imbalances, or may cause them in similar ways.
Early detection of sudden cardiac death is difficult to non-existent due to the randomness of the condition and variation in risk. Victims show no symptoms or signs of danger, and autopsies often reveal that young people’s hearts are structurally normal. A possible cause of sudden cardiac death in these “normal” hearts is electrical abnormalities within the heart, which could explain 41% of cases, but no studies have confirmed this link.
So far, the only structural abnormalities found in victims of cardiac death are hypertrophy or scar tissue in the muscles of the heart chambers, although the clinical significance of confirming that these conditions caused sudden death is debated. There is so much ambiguity surrounding this condition that further research is needed to identify other risks and confirm its association with sudden cardiac death. The rise of AI and machine learning using large datasets has enabled researchers to more effectively identify and confirm these associations.
At the University of Foggia, researchers used machine learning to analyze data from 711 athletes. First, the researchers examined factors identified in previous studies. But after refining the data and AI model, the researchers narrowed the total of variables down to eight key factors.
To make the analysis more thorough, the researchers created two data subsets. One was analyzed using eight core elements and the other using elements identified in other studies. For this analysis, the researchers used hierarchical clustering, which groups data points into “family trees” to show their relationships. The relationship between each data point and the risk factors for sudden cardiac death narrows the family tree into a single cluster representing the common factors that connect them.
After applying this clustering method, the researchers identified 21 clusters between the two data subsets. Although this method was efficient for analyzing large amounts of data simultaneously, researchers lacked clinical relevance and interpretation, and the analysis yielded no new observations.
In the absence of clinical formulas or metrics used by medical professionals treating athletes, the researchers’ clustering confirmed that the risks identified in previous studies were still risks, but did not provide further specific confirmation. Despite this lack of corroboration, the researchers suggest that the next step in the analysis focuses on each cluster individually, with an emphasis on the medical significance of the clusters.
From there, the researchers will utilize advanced mathematics and statistics using machine learning to analyze large amounts of clinical data from victims of sudden cardiac death. The ultimate goal is to better understand this data analysis, interpret its significance, and begin to apply it to athlete health. With this understanding, we can move from a mathematical expression to a practical tool for identifying warning signs of sudden cardiac death, thereby removing “sudden” from the title.
