New Way to Classify Adult-Onset Diabetes, Explained

A novel classification system for adult-onset diabetes, eschewing the usual type 1 and type 2 designation in favor of five subgroups, distinguishes unique disease progression and risk for related complications, researchers reported.

The five unique subgroups based on severity and underlying disease mechanism analysis -- reported by Emma Ahlqvist, PhD, of Lund University in Sweden, and colleagues in the Lancet Diabetes & Endocrinology -- were:

  • Cluster 1: Severe autoimmune diabetes (SAID)
  • Cluster 2: Severe insulin-deficient diabetes (SIDD)
  • Cluster 3: Severe insulin-resistant diabetes (SIRD)
  • Cluster 4: Mild obesity-related diabetes (MOD)
  • Cluster 5: Mild age-related diabetes (MARD)

The first cluster -- defined as severe autoimmune diabetes -- was the least common, with only 6.4% of the cohort fitting this type. Clinical characteristics included early-onset of the disease and being GADA positive, along with a lower BMI, poor metabolic control, and insulin deficiency. This group had the largest proportion of patients prescribed insulin (42%), as well as the shortest time to insulin use (HR 26.87 versus cluster 5, 95% CI 21.17-34.11).

Cluster 2 was mainly defined as severely insulin-deficient as well as GADA negative. Other clinical characteristics were similar to cluster 1, although this group had the highest proportion of patients on metformin.

The third cluster, accounting for 15.3% of the cohort, was marked by both a high degrees of insulin resistance -- specifically a high HOMA2-IR index -- paired with a high BMI. This group had relatively low use of metformin, despite the authors noting that this group would likely see the most benefit from such treatment.

Around one-fifth of patients belonged to cluster 4, defined by a presence of obesity without insulin resistance.

The fifth and most common cluster, which comprised nearly 40% of patients with adult-onset diabetes, was marked by an older age of diabetes onset, without as poor a metabolic profile as other clusters.