The study authors used six parameters including age at diagnosis, BMI, glutamate decarboxylase antibodies (GADA; to identify patients with autoimmune diabetes), HbA1c, homoeostatic model assessment 2 (HOMA2)-B (to assess β-cell function based on C-peptide concentration), and HOMA2-IR (to assess insulin sensitivity) to create a data-driven classification.
The following clusters were identified:
Cluster 1: (n= 577; 6·4%) characterised by early-onset disease, relatively low BMI, poor metabolic control, insulin deficiency, and presence of GADAappendix, and was labelled as severe autoimmune diabetes (SAID).
Cluster 2: (n=1575; 17·5%) labelled as severe insulin-deficient diabetes (SIDD), was GADA negative but otherwise similar to cluster 1: low age at onset, relatively low BMI, low insulin secretion (low HOMA2-B index), and poor metabolic control.
Cluster 3: (n= 1373; 15·3%) labelled as severe insulin-resistant diabetes (SIRD) and characterised by insulin resistance (high HOMA2-IR index) and high BMI.
Cluster 4: (n= 1942; 21·6%) characterised by obesity but not by insulin resistance, and labelled as mild obesity-related diabetes (MOD).
Cluster 5: (n= 3513; 39·1%) labelled as mild age-related diabetes [MARD]) were older than patients in other clusters, but showed, similar to cluster 4, only modest metabolic derangements.
The study authors note that the new clusters do not represent different aetiologies of diabetes, nor that this clustering is the optimal classification of diabetes subtypes.
A related editorial notes that whilst this classification is compelling, the study was conducted in reasonably homogeneous Scandinavian population therefore the generalisability of the results might be limited by genetic or environmental risk factors.