CLN - Feature

Redefining Diabetes

Will New Classifications Lead to Better Treatment Strategies?

Heather Lindsey

For many of the 425 million people worldwide with diabetes, managing their disease and forestalling its serious complications remain frustratingly out of reach. Patients with type 2 diabetes in particular present with a wide range of disease and respond quite differently to treatments. Researchers and clinicians have long recognized that the current labels of type 1 and type 2 diabetes don’t accurately capture this heterogeneity.

Ongoing research aims to better elucidate the underlying mechanisms behind hyperglycemia from either too little insulin or insufficient insulin effectiveness. Along this line of investigation, researchers in Sweden have proposed a new classification system for diabetes consisting of five subtypes rather than the current two (Lancet Diabetes Endocrinol 2018;6:361-9).

 “Treatment of type 2 diabetes has not been very successful,” said Leif Groop, MD, PhD, senior study author and a physician and professor of diabetes and endocrinology at Lund University Diabetes Centre in Malmö. “Lumping all patients together makes it very difficult to individualize the right treatments.”

If verified, the research group’s proposed classifications could enable physicians to offer better tailored therapies, he suggested, something the entire field is striving for. “The real question we need to address is how do we figure out what drug to give to what person versus another,” stressed Zachary T. Bloomgarden, MD, MACE, clinical professor of medicine at the Icahn School of Medicine at Mount Sinai in New York City and editor of the Journal of Diabetes.

Beyond Types 1 and 2

Diabetologists readily acknowledge that the type 1/2 dichotomy is not sophisticated enough, but whether Groop’s and his colleagues’ proposed classification will ever make its way into practice remains to be seen, said Clare J. Lee, MD, MHS, assistant professor of medicine in the division of endocrinology, diabetes, and metabolism at Johns Hopkins Medicine in Baltimore. “I don’t think this one article is going to change the way we classify things tomorrow, but I think it certainly lays the foundation for looking deeper into a better understanding of the heterogeneity within diabetes that’s not quite captured by just differentiating as type 1 versus type 2.”

Groop and his colleagues conducted a cluster analysis of 8,980 patients with newly diagnosed diabetes from the All New Diabetics in Scania cohort. To create their clusters, the scientists assessed glutamic acid decarboxylase antibodies (GADA), age at diagnosis, body mass index, hemoglobin A1c, and homeostasis model assessments of ß-cell function (HOMA 2-B) and insulin resistance (HOMA-IR) based on C-peptide concentrations. They also used patient records to evaluate prospective data on complications patients developed and on their prescriptions.

The investigators replicated this process in nearly 6,000 patients from three other Scandinavian-based cohorts. Their analysis revealed three subgroups representing relatively more severe disease, and two large subgroups in which most patients have mild disease that does not substantively progress, said Groop (see Box, below). The researchers did not seek to replace classifications of existing, specific diagnoses, such as maturity onset diabetes of the young or diabetes secondary to pancreatitis, and they excluded such cases from the analysis.

The Complications Equation

This research provides some insight into which patients are more likely to develop complications and could help physicians better determine who needs closer screening for retinal versus kidney damage, said Lee. The article also notes that people with age-related diabetes generally don’t develop sequelae, “so perhaps we could reduce the diabetes-related complications screening frequency for this group,” she said.

Potentially, doctors could redirect resources in diabetes care from individuals with mild forms of disease to those in more severe subgroups who need it, said Groop.

In a commentary accompanying the study, Robert Sladek, MD, proclaimed it “compelling” that simple parameters assessed at the time of diagnosis might reliably stratify patients according to prognosis, but he cautioned in an interview that potentially relaxing therapy in a group of patients categorized as having a less aggressive clinical course based on one study, “is probably not a good thing to be doing.” Sladek, an associate professor of medicine and human genetics at McGill University and the Genome Quebec Innovation Centre in Montreal, also expressed hope that Groop’s and other investigators’ research might offer solutions for one of the most vexing clinical challenges today, that current diabetes therapies may reduce disease symptoms but not the risk of developing complications.

This is why researchers and clinicians are interested in looking for diabetes treatments that reduce blood sugar and also prevent the development of cardiovascular disease or eye and kidney complications, he said.

Genetic Analysis and Clustering

A genetic association analysis Groop and his colleagues conducted that supported the proposed five subtypes could help with the latter. “I think the paper supports the idea that genetic variances may be helpful in understanding one’s risk for diabetes-related complications,” said Lee. “There’s some justification for looking in future studies at genetic differences that put some patients at a higher risk for, say, eye versus kidney complications.”

If a genetic classification system could be developed, patients would not change subgroups as they aged, said Sladek. Genetic data from randomized controlled trials for diabetes treatments also could be analyzed to identify subgroups of patients who respond and who don’t respond to therapy. He pointed to another recent analysis that found five novel clusters of type 2 diabetes genetic loci: two associated with insulin production and processing in pancreatic beta cells and three related to mechanisms of insulin response (PLoS Med 2018;15:e1002654).

Genetic analysis will help scientists find genes, proteins, or cellular processes to target, said Sladek. Researchers also need to find a genetic or blood test that determines whether an individual patient will respond to that therapy or not, he added.

While further classifying genetic features of diabetes could be helpful in categorizing patients, “it’s only part of the answer,” stressed Loren Wissner Greene, MD, MA, an endocrinologist and clinical professor of medicine at NYU Langone Health in New York City. Autoantibodies, particularly GADA status, are also important in understanding patients’ treatment needs, she noted.

The Journey to Personalized Care

As intriguing as Groop’s and other researchers’ proposed classifications might be, they will require considerably more vetting to advance into use clinically, cautioned Andrew Ahmann, MD, a professor of medicine and director of the Harold Schnitzer Diabetes Health Center at Oregon Health and Science University in Portland. One challenge with reclassifying diabetes is that none of the new systems are anywhere near universally accepted, he said. They also don’t yet meet the test of clinical practicality for the physicians caring for these millions of patients.

Although the Swedish researchers evaluated six relatively simple variables, using them in clinical practice would be much more complicated and require a new laboratory approach and clinical interpretation for diagnosis, added Ahmann. “We do not routinely obtain all of the tests used to characterize the proposed clusters,” he said.

Because Groop and his colleagues evaluated a mostly white Scandinavian group of patients, research data need to be replicated in other populations, said Lee. If validated, then clinicians would need to consider whether baseline fasting glucose and fasting insulin, as well as other study measures, need to be adopted, she said. Groop agreed with the need to validate his team’s model in more diverse populations. He and his colleagues are planning to launch more extensive diabetes classification research in populations in India, China, and Mexico.

Ultimately, the Swedish analysis and others to come will, according to Ahmann, “eventually lead to precision medicine approaches that will make a difference, particularly as technology matures.”

Heather Lindsey is a freelance medical writer in Maplewood, New Jersey. +Email: [email protected]