Data-driven subgroups of prediabetes and the associations with outcomes in Chinese adults

Identification of distinct trajectories in preclinical type 2 diabetes and their associations with outcomes

“…2e-j, and Supplementary Figs. [5][6][7][8][9][10][11][12][13][14][15][16]. For instance, Glucose and HbA1c levels escalate more rapidly for S2 than for S1 over time (Figs.…”

“…Progressions of other biomarkers were presented in Supplementary Figs. [5][6][7][8][9][10][11][12][13][14][15][16]. R is the Pearson's correlation between biomarkers and SuStaIn stages for each subtype.…”

“…Third, the magnitude and extent of interindividual differences in progression trajectories of preclinical-T2DM and their alterations on clinical outcomes, such as cardiovascular, renal, ophthalmological, neurological, and cognitive functions, remain undetermined. Existing studies have clustered prediabetes into subtypes regardless the progressive stage of prediabetes and explored their associations with T2DM as well as its complications 6 , but such approaches fail to capture the longitudinal heterogeneity of preclinical-T2DM. Conversely, the progressive heterogeneity is a hallmark of T2DM, which could be sourced from the preclinical stage of the T2DM continuum.…”

Deconvoluting prediabetes: A path to understanding the origins of complications

“…In a large population-based study of people above 40 years of age that was conducted in multiple communities across China, one-third of all recruited people had prediabetes according to the broader criteria of the American Diabetes Association. 1 The researchers set out to investigate this prediabetic population of ∼56,000 people to identify clusters of similarities based on 12 variables reflecting anthropometric, glycemic, and other metabolic traits. By using multiple runs of k-means clustering with varying cluster counts on randomly selected subsets of the population and identifying pairs of individuals repeatedly being grouped together, they derived a consensus pairwise similarity matrix for the whole study population, which led to the identification of an optimum segregation of 6 clusters ( Figure 1 ).…”

“… A study of 56,000 people with prediabetes in China stratified the population into 6 clusters, among which were two clusters of impaired insulin secretion and a cluster with hyperinsulinemic insulin resistance. 1 A follow-up showed different risk of complications, demonstrating prediabetes heterogeneity. …”