Cluster analysis based on fasting and postprandial plasma glucose and insulin concentrations

Objective

Plasma glucose and insulin concentrations are clinical markers used to diagnose metabolic diseases, particularly prediabetes and diabetes. In this paper, we conducted a cluster analysis using plasma glucose and insulin data collected during both fasting and 2-h postprandial periods.

Methods

Different clustering experiments were performed by changing the attributes, from one (fasting glucose) to four (fasting and postprandial glucose and insulin) attributes input to a k-means clustering algorithm. Based on the elbow and silhouette methods, three clusters were chosen to perform the clustering experiments. The Pearson correlation coefficient was utilized to evaluate the association between the levels of glucose and insulin within each created cluster.

Results

Results show that one cluster comprised individuals with prediabetes, another cluster consisted of individuals with diabetes, while subjects without prediabetes and diabetes were assigned to a separate cluster. Despite not being used as an attribute, we observed varying age ranges among subjects in the three clusters. Furthermore, significant correlations were found between fasting and postprandial insulin levels, as well as between fasting and postprandial glucose levels, suggesting a consistent relationship between these variables, and highlighting their interdependence in the context of glucose metabolism.

Conclusion

The clustering analysis successfully differentiated individuals into distinct clusters based on their metabolic conditions, confirming that the approach effectively captured the underlying patterns in the plasma glucose and insulin data. Furthermore, despite not being a considered attribute, the varying age ranges observed within the clusters indicate that age may play a role in the development and progression of diabetes. Additionally, the fasting and postprandial associations in insulin and glucose levels exhibited greater strength in the cluster encompassing individuals with diabetes, where insulin production or action is compromised.