ISSN 1662-4009 (online)

ESPE Yearbook of Paediatric Endocrinology (2022) 19 15.4 | DOI: 10.1530/ey.19.15.4


Cell Rep Med. 2022;3(1):100477. doi: 10.1016/j.xcrm.2021.100477. PubMed ID: 35106505

Brief summary: To explore clinical heterogeneity, this study analyzed baseline visit data on 726 adults with newly diagnosed Type 2 diabetes (T2D) adults and identified in 4 distinct profiles (clusters of phenotypes), which predicted differences in subsequent disease progression and anti-diabetic treatments.

It is increasingly recognised that T2D is not a homogenous condition. Separate to this study, recent data show that adolescents with T2D display a more rapidly progressing form of disease associated with need for insulin therapy and high incidence of microvascular complications (see Paper 12.3 in this Yearbook).

This study focussed on T2D in adults, who had at study baseline a mean age of 62 years, mean BMI 30.4 kg/m2, and all were on lifestyle and/or metformin treatment only. They were characterised at baseline in great depth, by a wide range of anthropometry, hormones, lipids, and a frequently-sampled mixed-meal tolerance test. Diabetes progression was assessed by change in HbA1c over the 36 months follow-up. A soft-clustering statistical method was used to sub-group patients based on 32 clinical variables. Four different profiles (‘archetypes’) were identified. Patients with a baseline archetype characterised by obesity, insulin resistance, dyslipidemia, and impaired beta cell function (N=45 individuals, 6% of total) showed the fastest disease progression. Other archetypes included: A) low BMI, older age, high insulin sensitivity, and high cholesterol (N=103; slowest HbA1c progression); B) high BMI, but insulin sensitive and favorable lipid profiles (N=22); C) high BMI, insulin resistance (N=84).

A limitation is that only 35% of patients could be confidently categorised to one archetype, and the majority displayed phenotypes with moderate contributions from two or more archetypes, possibly due to multiple etiological processes. Furthermore, over time there was mixed stability in archetypes. Hence, the study tests a promising idea to map the wide variations in disease presentations to different aetiological processes and inform personalized treatments. However, clustering using even more traits (e.g. omics, genetics, continuous glucose monitoring) may be needed to achieve those aims.

Article tools

My recent searches

No recent searches.

Authors