ANDS/ANDIS Classification of Adult-Onset Diabetes
The ANDIS (All New Diabetics in Scania) classification stratifies adult-onset diabetes into five distinct subgroups based on six clinical and biochemical variables: GAD antibodies, age at diagnosis, BMI, HbA1c, HOMA2-β (beta-cell function), and HOMA2-IR (insulin resistance). 1
The Five ANDIS Subgroups
1. SAID (Severe Autoimmune Diabetes)
- GAD antibody-positive
- Similar to classical type 1 diabetes/LADA
- Characterized by autoimmune β-cell destruction
- Represents approximately 2-3% of adult-onset diabetes 2
- Highest risk for retinopathy and neuropathy (adjusted HR 1.35 and 2.58 respectively) 3
2. SIDD (Severe Insulin-Deficient Diabetes)
- GAD antibody-negative
- Young age at diagnosis
- Low BMI
- Severe insulin deficiency (low HOMA2-β)
- High HbA1c at diagnosis
- Represents approximately 18-31% of cases 3, 2
- Highest risk for retinopathy (HR 2.11), neuropathy (HR 2.13), kidney disease including kidney failure (HR 2.94), stroke (HR 1.32), and increased total mortality (HR 1.44-1.52) 3
3. SIRD (Severe Insulin-Resistant Diabetes)
- Severe insulin resistance (high HOMA2-IR)
- High BMI
- GAD antibody-negative
- Represents approximately 15-32% of cases 3, 2
- Most dangerous subgroup for metabolic complications despite lower HbA1c
- Highest prevalence at diagnosis of: hypertension (69.8%), dyslipidemia (44.4%), cardiovascular disease (30.6%), and kidney disease (10.9%) 3
- Highest risk for kidney failure (HR 3.41), steatotic liver disease (HR 3.29), heart failure (HR 1.55), atrial fibrillation (HR 1.32), myocardial infarction (HR 1.51), and increased total mortality (HR 1.44-1.52) 3
- Critical finding: SIRD develops end-organ damage early and represents a high-risk subgroup not identified by conventional glycemia-based risk factors 3
4. MOD (Mild Obesity-Related Diabetes)
- High BMI
- Young age at diagnosis
- Relatively preserved insulin secretion
- Moderate insulin resistance
- Represents approximately 17-18% of cases 2
- Increased risk for atrial fibrillation (HR 1.58) and total mortality (HR 1.44-1.52) 3
5. MARD (Mild Age-Related Diabetes)
- Older age at diagnosis
- Modest metabolic derangements
- Most benign course
- Represents approximately 17-40% of cases 2
- Used as the reference comparator group for risk assessment 3
Clinical Implications and Limitations
Key Findings from Validation Studies
- Inflammatory markers (CRP, IL-6) and GLP-1 levels are similarly elevated across all subtypes, indicating common pathological pathways despite different clinical phenotypes 2
- Lifestyle factors, education level, and psychosocial parameters differ more significantly between subtypes than metabolic or inflammatory markers 2
- SIDD-like and SIRD-like groups show lower educational status, requiring tailored patient education programs 2
- Gut microbiota and dietary patterns show only minor differences between subtypes 2
Critical Limitations for Clinical Implementation
The ANDIS framework faces major implementation barriers 4:
- Diagnostic complexity requiring GAD antibody testing and HOMA2 indices
- High costs
- Limited access to required diagnostics even in high-income countries
- Limited therapeutic differentiation over existing guidelines 4
- No prospective intervention studies demonstrating improved outcomes with subgroup-specific treatment 5
Practical Adaptation
The framework can be simplified using accessible markers: HbA1c, BMI, and abdominal circumference 4, though this reduces precision.
Relationship to Current Guidelines
The 2025 ADA Standards of Care acknowledge that diabetes classification is being reconsidered based on genetic, metabolomic, and other characteristics 6, but continue to use the conventional four-category system (Type 1, Type 2, specific types, and gestational diabetes) 6. The guidelines note that "many groups are working on combining clinical, pathophysiological, and genetic characteristics to more precisely define the subsets of diabetes" 6 but have not yet incorporated ANDIS into standard practice.
Important Caveat
ANDIS classification is a research framework, not yet a clinical standard. While it provides valuable prognostic information about complication risk, particularly identifying SIRD as a high-risk subgroup requiring aggressive early intervention 3, it has not replaced conventional classification in clinical practice guidelines 6.