Diabetes in the young
Log in to starLast updated 2mo ago
Diabetes mellitus presenting in childhood, adolescence, or young adulthood that doesn't fit the classic type 1 (autoimmune, ketosis-prone, antibody-positive) or type 2 (obese, insulin-resistant, acanthosis nigricans) phenotype. The clinical posture is to not anchor on type 1 by default, because monogenic diabetes is undiagnosed in an estimated 1-2% of young diabetes cases and the management changes substantially: many MODY subtypes respond to sulfonylureas instead of insulin, and most patients with neonatal diabetes can switch off insulin entirely once a K-ATP channel cause is confirmed.
Four data points narrow the differential:
- Age of onset: under 6 months almost never type 1 (autoimmunity doesn't develop that fast); 6 months to puberty is the diagnostic gray zone where MODY and atypical T1D overlap; puberty onward is mostly autoimmune unless other clues intervene.
- Insulin requirement: persistently insulin-independent or sulfonylurea-responsive after the honeymoon suggests monogenic.
- Autoantibodies: GAD, IA-2, insulin, ZnT8 negative across multiple titers is suspicious for non-type 1.
- Family history: strong dominant pedigree across multiple generations, especially with mild non-insulin-requiring diabetes, points to MODY.
A useful screening pattern: age of onset < 25 + family history of diabetes in successive generations + preserved C-peptide + autoantibody-negative = consider MODY genetic testing.
Neonatal diabetes (< 6 months of age)
Diabetes diagnosed under 6 months is almost always monogenic and is the highest-yield "do not call this type 1" scenario.
- KCNJ11 and ABCC8 (encode the K-ATP channel Kir6.2 and SUR1): the most common causes of permanent neonatal diabetes. Sulfonylureas (glyburide) bind SUR1 and close the channel, restoring insulin secretion: most patients can transition off insulin entirely, often within days. Some KCNJ11 variants cause DEND syndrome (Developmental delay, Epilepsy, Neonatal Diabetes) where high-dose sulfonylurea may also improve neurologic features.
- INS (insulin gene): permanent neonatal diabetes from misfolded proinsulin and beta-cell ER stress; requires lifelong insulin (no sulfonylurea benefit).
- 6q24 (paternal PLAGL1/HYMAI overexpression): the most common cause of transient neonatal diabetes mellitus: hyperglycemia in the first weeks, remits by 6 months, often recurs at adolescence.
- GATA6, GATA4: neonatal diabetes + pancreatic agenesis or hypoplasia + congenital heart disease + biliary tract anomalies (syndromic).
- EIF2AK3 (Wolcott-Rallison syndrome): neonatal diabetes + skeletal dysplasia + episodic liver dysfunction.
MODY (Maturity-Onset Diabetes of the Young)
MODY is the autosomal-dominant, monogenic, non-autoimmune diabetes category. Diagnosed typically before age 25, mild at onset, not insulin-requiring at diagnosis, strong family history across generations.
- HNF1A-MODY (MODY3): the most common form. Glycosuria at relatively normal blood glucose (low renal threshold) is a classic clue. Marked sulfonylurea sensitivity: switching from insulin to a low-dose sulfonylurea often achieves better control. Progressive over decades.
- GCK-MODY (MODY2): mild fasting hyperglycemia from birth, stable over a lifetime. Patients often have a fasting glucose of 100-150 mg/dL with a normal HbA1c around 6-7%. Usually does not need treatment and does not progress to microvascular complications. Often discovered incidentally in pregnancy.
- HNF4A-MODY (MODY1): macrosomia + transient neonatal hyperinsulinemic hypoglycemia + later-onset sulfonylurea-responsive diabetes. The infant phenotype precedes the adolescent/young-adult diabetes by years.
- HNF1B-MODY (MODY5) / RCAD syndrome: Renal Cysts And Diabetes. Diabetes + renal structural anomalies (cysts, dysplasia, hypoplasia) + genital tract anomalies + hyperuricemia/gout + abnormal LFTs + pancreatic atrophy. Often arises from 17q12 deletions, with associated neurodevelopmental features.
- Other MODY subtypes (PDX1, NEUROD1, KLF11, CEL, PAX4, INS, BLK, APPL1): individually rare.
Mitochondrial diabetes
- MIDD (Maternally Inherited Diabetes and Deafness): m.3243A>G in MT-TL1 (the same variant that causes MELAS at higher heteroplasmy). Diabetes + sensorineural hearing loss + maternal inheritance pattern (mother and maternal aunts/grandmother often affected, with no male-to-male transmission). Macular pattern dystrophy on fundus is a supporting finding. Avoid metformin (lactic acidosis risk in mitochondrial disease).
Other monogenic syndromic diabetes
- Wolfram syndrome (DIDMOAD) (WFS1, AR): Diabetes Insipidus + Diabetes Mellitus + Optic Atrophy + Deafness, in roughly that temporal order during the first 2-3 decades. Diabetes is insulin-requiring (beta-cell loss, not autoimmune); optic atrophy is the high-yield discriminator.
- Lipodystrophy syndromes (LMNA, AGPAT2, PPARG, BSCL2): severe insulin resistance + characteristic fat-distribution phenotype + hypertriglyceridemia + hepatic steatosis.
- Rabson-Mendenhall, Donohue (leprechaunism) (INSR): extreme insulin resistance from insulin receptor defects; distinct phenotypes.
- Cystic fibrosis-related diabetes: emerges in adolescence in patients with established cystic fibrosis.
- Diabetes diagnosed before 6 months of age is almost never type 1. Test KCNJ11, ABCC8, INS, 6q24 within weeks of diagnosis. A positive K-ATP channel result lets the patient transition from insulin to sulfonylurea.
- Three-generation pedigree of mild non-insulin-requiring diabetes, autoantibodies negative, C-peptide preserved → MODY; test HNF1A, HNF4A, GCK, HNF1B.
- Mild fasting hyperglycemia (100-150 mg/dL) from birth or childhood, stable HbA1c around 6-7%, no microvascular complications in older relatives → GCK-MODY; usually no treatment needed.
- Diabetes + sensorineural hearing loss + maternal inheritance → MIDD (m.3243A>G); avoid metformin.
- Diabetes + central diabetes insipidus + optic atrophy in the first two decades → Wolfram.
- Diabetes + renal cysts + genital tract anomaly → HNF1B-MODY / RCAD (often a 17q12 deletion).
- Macrosomic newborn + neonatal hyperinsulinemic hypoglycemia + family history of diabetes → HNF4A-MODY.
- Pancreatic autoantibodies (GAD-65, IA-2, insulin, ZnT8): negative across multiple titers raises suspicion for non-type 1.
- C-peptide at diagnosis and after the honeymoon: preserved C-peptide off insulin (or on minimal insulin) more than 3-5 years after diagnosis is atypical for type 1.
- HbA1c, fasting glucose, OGTT: characterize severity and progression.
- Three-generation family history: age of onset, treatment, complications, hearing loss, renal anomalies, optic atrophy.
- Targeted MODY panel (HNF1A, HNF4A, GCK, HNF1B, ABCC8, KCNJ11, INS, others) if the autoantibody-negative, family-history-positive criteria are met.
- Neonatal diabetes panel (K-ATP channel genes, INS, 6q24 methylation, EIF2AK3, GATA6) urgently for any patient diagnosed under 6 months.
- Mitochondrial m.3243A>G testing if there is sensorineural hearing loss or maternal inheritance.
- Renal ultrasound + UA + uric acid if HNF1B or RCAD is on the differential.
- Diabetes MODY probability calculator (Exeter): a clinical pre-test probability tool that helps decide which patients warrant the genetic panel.
- Any diabetes diagnosed before 6 months gets a neonatal diabetes panel today, not in 6 months. A positive KCNJ11 or ABCC8 result can pull the patient off insulin and onto sulfonylurea with dramatic improvement in glycemic control and quality of life.
- GCK-MODY is the diagnosis you make so that you can stop treating. Patients are often referred with "treatment-resistant" mild hyperglycemia; once the diagnosis is made, treatment is discontinued and the patient is reassured.
- Pregnancy is a clarifying moment for MODY. A pregnant patient with mild fasting hyperglycemia who has had it for years is often GCK-MODY; treatment depends on whether the fetus inherited the variant.
- Avoid metformin in m.3243A>G carriers. Lactic acidosis risk is amplified.
- The high-yield missed diagnosis in MODY is HNF1A-MODY mistaken for type 1; the patient is on insulin, develops hypoglycemia easily, and a sulfonylurea trial yields much better control than insulin did.