Susan L. Samson, MD, PhD, FRCPC, FACE 

From Label to Etiology: Applying the 2026 AACE Framework

Experts conclude by emphasizing that the 2026 AACE classification algorithm is a call for clinicians to pause, reassess presumed type 2 diabetes, and recognize that getting the etiology right is foundational to optimizing downstream treatment.

In the closing segment, Samson reiterates that the differential diagnosis for adult-onset diabetes is far broader than the traditional assumption that most patients have type 2 diabetes. She emphasizes that clinicians should not simply accept a preexisting type 2 diabetes label when patients establish care, particularly if the disease course does not align with the expected phenotype. Instead, the 2026 AACE algorithm encourages a systematic reassessment of diagnosis when glycemic patterns, insulin requirements, weight trajectory, or associated comorbidities raise concern.

Samson notes that individuals who clearly do not fit a typical type 2 phenotype—whether because of autoimmunity, pancreatic disease, endocrinopathy, or genetic features—should be referred for endocrinology evaluation and more detailed classification.

Umpierrez reflects on his decades of practice in diabetes and endocrinology and candidly acknowledges how often he now questions whether earlier patients may have been misclassified as having type 2 diabetes. He underscores that misclassification of adult-onset type 1 diabetes as type 2, pancreatic diabetes, MODY, or cortisol-mediated diabetes is not merely academic; it can profoundly alter downstream treatment decisions, risk for acute complications such as diabetic ketoacidosis, and the trajectory of chronic complications.

In his view, the 2026 AACE classification algorithm provides a practical, clinically oriented framework that can help busy clinicians recognize when additional testing, imaging, or specialist referral is warranted.

Together, Samson and Umpierrez characterize the new algorithm as an “invitation to pause”—a structured prompt to ask whether the current diagnosis truly explains the patient’s presentation and treatment response. By embedding classification logic alongside therapeutic guidance, the 2026 AACE algorithm aims to ensure that the right diagnosis underpins decisions about pharmacologic therapy, technology use, and long-term risk reduction.

The experts conclude that widespread adoption of this approach has the potential to improve diagnostic accuracy, individualize care, and ultimately enhance outcomes for adults living with diabetes who might otherwise be managed under an oversimplified label of type 2 diabetes.

Videos

Learn how clinicians spot uncommon diabetes causes—from pancreatic disease and MODY to steroids, transplants, and checkpoint inhibitors—before treatment.

How the 2026 AACE Algorithm Broadens the Lens on Diabetes

Experts outline the broader differential diagnosis embedded in the new algorithm, spanning pancreatic, monogenic, treatment-related, and rare genetic forms of diabetes that may masquerade as type 2 in adults.

In this section, Samson invites Umpierrez to walk through the remaining categories included in the 2026 AACE diabetes classification algorithm beyond autoimmune diabetes and hypercortisolism.

Umpierrez first discusses pancreatic disorders that can lead to diabetes, including pancreatic cancer, chronic pancreatitis, cystic fibrosis–related diabetes, and hemochromatosis. He notes that in subspecialty centers with large cystic fibrosis populations, diabetes is a frequent complication and carries distinct management considerations. Weight loss and underweight status in a patient labeled as having type 2 diabetes, particularly when combined with severe insulin requirements or recurrent hypoglycemia, should prompt clinicians to consider pancreatic causes or other atypical forms of diabetes rather than assuming typical type 2 disease.

The conversation then turns to monogenic diabetes, particularly maturity-onset diabetes of the young (MODY), which Samson and Umpierrez emphasize has been recognized for decades but remains underdiagnosed. Umpierrez suggests clinicians should suspect monogenic diabetes when there is an autosomal dominant pattern of diabetes across 3 generations, diagnosis at a young age, absence of type 1 autoantibodies, and preserved or only mildly reduced C‑peptide levels. He notes that calculators and clinical prediction tools are available to estimate the likelihood of MODY and guide genetic testing. T

he algorithm also highlights glucocorticoid-induced diabetes and posttransplant diabetes, both increasingly seen in practice, as well as checkpoint inhibitor–associated autoimmune diabetes, which is emerging with broader use of immune checkpoint inhibitors in oncology. In this latter group, onset can be abrupt and reminiscent of type 1 diabetes, often with positive autoantibodies and low C‑peptide, necessitating rapid initiation of insulin therapy.

Finally, Samson emphasizes that the 2026 AACE algorithm also acknowledges rare inherited forms of diabetes, including mitochondrial diabetes syndromes that may present with diabetes and sensorineural deafness or lipodystrophic phenotypes characterized by loss of subcutaneous fat, hypertriglyceridemia, fatty liver disease, and marked insulin resistance.

Umpierrez notes that, although infrequent, these conditions are important to recognize because they carry distinct systemic risks and may respond differently to standard therapies.

Together, Samson and Umpierrez argue that the value of the new algorithm lies not only in cataloging these entities but also in providing a structured pathway that prompts clinicians to reconsider a default type 2 label when the clinical picture, family history, or treatment response is atypical.

Explore signs of uncommon diabetes—from pancreatic disease and MODY to steroid, transplant, and checkpoint-inhibitor autoimmune causes.

Hypercortisolism in Refractory Diabetes, CATALYST Data

Experts examine the role of hypercortisolism in difficult-to-control diabetes, highlighting findings from the CATALYST trial and the practical implications of dexamethasone suppression testing in this population.

Here, Samson broadens the discussion from autoimmune etiologies to secondary endocrine causes of diabetes, focusing on hypercortisolism. She notes that diabetes can occur in the context of overt Cushing syndrome and acromegaly, in which treatment of the underlying endocrinopathy often leads to improved glycemic control and, in some cases, remission of diabetes.

Samson then turns to milder forms of cortisol excess, including so-called autonomous cortisol secretion from adrenal nodules, in which patients may lack classic Cushingoid features but still exhibit significant metabolic disturbances. She explains that the new 2026 AACE algorithm explicitly lists hypercortisolism among the conditions that warrant further evaluation when patients exhibit unexplained hyperglycemia or require unusually high insulin doses.

Umpierrez, who served as a principal investigator for the CATALYST trial, describes the study’s design and key findings. The trial enrolled more than 1000 adults with difficult-to-control type 2 diabetes, many of whom also had hypertension or required multiple glucose-lowering agents. Participants underwent an overnight 1‑mg dexamethasone suppression test, with morning cortisol levels measured the following day. A cortisol concentration greater than 1.8 μg/dL was considered indicative of nonsuppressible cortisol or hypercortisolism.

According to Umpierrez, approximately 23.8% of these patients met criteria for hypercortisolism, and the prevalence was even higher (around 36.6%) among those taking multiple antihypertensive medications. Subsequent adrenal imaging revealed that about one-third of patients with abnormal suppression tests had adrenal abnormalities, most commonly unilateral adrenal nodules.

Building on these data, Samson and Umpierrez discuss the therapeutic and practical implications of identifying hypercortisolism in patients with “type 2” diabetes that is refractory to standard management. Samson notes that in CATALYST, patients with hypercortisolism who entered a treatment phase with the glucocorticoid receptor antagonist mifepristone experienced improvements in weight, waist circumference, and hemoglobin A1c—on the order of a 1.5% reduction over 24 weeks. She cautions that while more research is needed to define screening strategies and long-term outcomes, the findings support a more active search for cortisol excess in patients with unexplained insulin resistance and metabolic instability.

Practically, Samson emphasizes that a positive dexamethasone suppression screen should prompt referral to an endocrinologist for comprehensive evaluation and management of hypercortisolism. Within the 2026 AACE algorithm, this approach reinforces the broader message that clinicians should consider endocrine comorbidities when glycemic control is unexpectedly difficult.

Learn key clues that diabetes isn’t type 2, and how autoantibody tests plus C‑peptide guide accurate adult type 1/LADA diagnosis.

Red Flags for Misdiagnosed Type 2, When to Order Antibodies and C‑Peptide

Experts describe clinical warning signs that should prompt reconsideration of a type 2 diabetes label and outline how autoantibody and C‑peptide testing can help differentiate type 1 and autoimmune diabetes in adults.

In this segment, Samson shifts the focus from baseline phenotype to dynamic clinical clues that may signal a misdiagnosis of type 2 diabetes. She discusses scenarios in which patients with presumed type 2 diabetes exhibit poor response to noninsulin therapies despite strong adherence, rapidly progress to insulin dependence within a few years of diagnosis, or require very high insulin doses (eg, >2 units/kg) suggestive of extreme insulin resistance or another underlying process.

Additional red flags include the development of ketosis, particularly in the setting of sodium-glucose cotransporter 2 (SGLT2) inhibitor use, and lack of expected response to oral hypoglycemic agents such as metformin or insulin secretagogues. Samson notes that these patterns should trigger clinicians to question whether the original diagnosis of type 2 diabetes is correct.

Umpierrez then details how autoantibody testing and C‑peptide measurement are integrated into the 2026 AACE classification algorithm when type 1 diabetes or latent autoimmune diabetes in adults (LADA) is suspected. He explains that glutamic acid decarboxylase 65 (GAD65) antibodies are the most commonly used and often remain detectable in adults, including those with LADA. Additional autoantibodies include those against IA‑2, zinc transporter 8, and insulin. He cautions that up to 10% of patients with type 1 diabetes may be antibody-negative and that immunosuppressive therapies, including corticosteroids, can mask antibody positivity. When testing for insulin autoantibodies, he emphasizes that patients must be insulin-naive; exogenous insulin exposure can itself induce autoantibody formation.

Regarding C‑peptide, Samson and Umpierrez explain that its interpretation must be contextualized with simultaneous glucose levels and recent glycemic control. Umpierrez notes that severe hyperglycemia or recent diabetic ketoacidosis can suppress endogenous insulin secretion—a phenomenon of glucose toxicity—leading to transiently low C‑peptide levels.

The 2026 AACE algorithm, aligned with recommendations from major diabetes societies, considers a low C‑peptide (eg, <200 pmol/L) coupled with at least 1 positive autoantibody as supportive of type 1 diabetes in adults, particularly at very low C‑peptide levels (<35 pmol/L) after age 35 years. By contrast, a C‑peptide concentration greater than approximately 600 pmol/L suggests type 2 diabetes, with intermediate levels requiring integration of clinical features, family history, and autoantibody results. Through this framework, Samson and Umpierrez illustrate how the algorithm operationalizes laboratory tools to refine diabetes classification beyond phenotype alone.

Spot hidden type 1/LADA in overweight adults: learn key red flags, and how autoantibodies plus C-peptide guide accurate diagnosis.

Susan L. Samson, MD, PhD, FRCPC, FACE

Articles