Different diabetes phenotypes associated with varying prevalence of complications

Patients were categorized into five groups: mild age-related diabetes, mild obesity-related diabetes, severe autoimmune diabetes, severe insulin-resistant diabetes, and severe insulin-deficient diabetes.


A patient's diabetes phenotype may be associated with risk for comorbidities and complications, such as nonalcoholic fatty liver disease (NAFLD) and diabetic neuropathy, a recent study found.

Researchers conducted comprehensive phenotyping and assessed laboratory variables of 1,105 patients with newly diagnosed type 1 or type 2 diabetes in the German Diabetes Study. Patients were ages 18 to 69 years and had a known disease duration of fewer than 12 months. Exclusion criteria included diabetes of other causes (e.g., monogenic diabetes syndromes, diseases of the exocrine pancreas, and gestational diabetes); pregnancy; and acute or severe chronic heart, hepatic, renal, or psychiatric diseases. At baseline, patients were classified into five clusters: mild age-related diabetes (MARD), mild obesity-related diabetes (MOD), severe autoimmune diabetes (SAID), severe insulin-resistant diabetes (SIRD), and severe insulin-deficient diabetes (SIDD). Results were published online on July 22 by The Lancet Diabetes & Endocrinology.

Overall, 386 (35%) patients were classified as having MARD, and they were generally older than those in other clusters, with only minor metabolic abnormalities. The 323 (29%) patients with MOD were characterized by obesity, substantial adipose tissue insulin resistance, and moderate whole-body insulin resistance. The 121 (11%) patients with SIRD were characterized by high body mass index (BMI) and both whole-body and adipose-tissue insulin resistance. The 247 (22%) patients with SAID had positive glutamic acid decarboxylase antibodies and were more likely to be younger and have a relatively low BMI, poor glycemic control, and overt insulin deficiency. Finally, the 28 (3%) patients with SIDD showed similarities with those in the SAID group, but none had glutamic acid decarboxylase antibodies.

The groups had different rates of progression of diabetes-related complications within five years of diagnosis. Patients with SIRD had the highest prevalence of hepatic fibrosis (NAFLD fibrosis score >0.6) at baseline and at five years compared to those in the other clusters. Patients with SIRD also had the lowest estimated glomerular filtration rate and highest prevalence of stage 2 and stage 3 nephropathy at baseline, with similar results at five years. At baseline and at five years, patients with SIDD had the highest prevalence of confirmed diabetic sensorimotor polyneuropathy. Patients with SIDD also had the highest prevalence of confirmed cardiac autonomic neuropathy at baseline, but not at five years. For some patients, there was change in categorization over time. At five years, 367 patients were reassessed, at which point 128 (35%) had MARD, 106 (29%) had MOD, 88 (24%) had SAID, 35 (10%) had SIRD, and 10 (3%) had SIDD.

Limitations of the study include the German Diabetes Study's inclusion and exclusion criteria, which likely affected the number of patients allocated to specific clusters, the authors noted. They added that the study is ongoing (resulting in follow-up of relatively few patients) and that it is not representative of the general population; therefore, the results cannot be generalized to community-based practice.

The results highlight the heterogeneity of type 2 diabetes and may help explain the differences in disease progression and response to glucose-lowering treatment seen in clinical practice, an accompanying comment said. However, there may not yet be a reliable way to move toward a precision medicine approach for treatment of type 2 diabetes, the comment said.

“[The study] reported that patients can shift from one cluster to another over time. Such a shift could be affected by several environmental factors and pharmacological treatment,” the editorialist wrote. “Precision medicine for type 2 diabetes, therefore, could require even more accurate profiling of individuals belonging to a given cluster, by integrating genetic and omics data, digital phenotyping, sensor-based behavioural monitoring, and pharmacogenomics.”