C-Peptide Testing for Diabetes Type Differentiation
Measure C-peptide when the glucose is <220 mg/dL (<12.5 mmol/L) in insulin-treated patients with ambiguous diabetes classification, using cut-offs of <200 pmol/L (<0.6 ng/mL) for type 1 diabetes and >600 pmol/L (>1.8 ng/mL) for type 2 diabetes. 1, 2
When to Measure C-Peptide
C-peptide testing is primarily indicated in insulin-treated patients when diabetes classification is unclear 1. The most common scenario is distinguishing type 1 from type 2 diabetes in adults who present with atypical features—for example, someone with a type 2 phenotype (older age, higher BMI) who presents in ketoacidosis 1.
Timing Considerations:
- Glucose requirement: Measure fasting C-peptide when simultaneous fasting plasma glucose is <220 mg/dL (<12.5 mmol/L) 1
- Random sampling acceptable: A random C-peptide sample (with concurrent glucose) within 5 hours of eating can replace formal stimulation testing 2
- Avoid testing: Do not measure C-peptide within 2 weeks of a hyperglycemic emergency 2
- Optimal timing: C-peptide is most useful after 3-5 years from diagnosis when persistent insulin secretion suggests type 2 or monogenic diabetes rather than type 1 3
Critical Glucose Context
If C-peptide is <600 pmol/L (<1.8 ng/mL) and the concurrent glucose is <70 mg/dL (<4 mmol/L) or the person may have been fasting, repeat the test 2. This prevents false interpretation from physiologically suppressed C-peptide during hypoglycemia. Very low levels (<80 pmol/L [<0.24 ng/mL]) do not need repeating 2.
C-Peptide Cut-Off Values
Clear-Cut Ranges:
- <200 pmol/L (<0.6 ng/mL): Type 1 diabetes 2
- >600 pmol/L (>1.8 ng/mL): Type 2 diabetes—circumstances of testing don't matter at this level 2
- 200-600 pmol/L (0.6-1.8 ng/mL): Indeterminate zone—usually consistent with type 1 diabetes or MODY, but can occur in insulin-treated type 2 diabetes, particularly in people with normal/low BMI or long disease duration 2
Supporting Research Evidence:
Recent meta-analysis supports a fasting/random C-peptide cut-off of ≤0.20 nmol/L (≤0.6 ng/mL) for type 1 diabetes and ≥0.30 nmol/L (≥0.9 ng/mL) for type 2 diabetes 4. A 2024 study found ≤0.16 nmol/L (≤0.48 ng/mL) showed 92.9% sensitivity and 97.6% specificity for distinguishing type 1 from type 2 diabetes 5. These align well with the guideline-recommended 200 pmol/L threshold.
Clinical Algorithm
Step 1: Determine if C-peptide testing is needed
- Patient on insulin with unclear diabetes type
- Clinical features don't clearly fit type 1 or type 2
- Age >35 years at diagnosis with autoantibody-negative status
- Type 2 phenotype but presented in ketoacidosis
Step 2: Ensure appropriate testing conditions
- Glucose <220 mg/dL (<12.5 mmol/L)
- Not within 2 weeks of hyperglycemic emergency
- Random sample within 5 hours of eating OR fasting sample
- Record concurrent glucose level
Step 3: Interpret results
- <200 pmol/L: Diagnose type 1 diabetes—absolute insulin requirement confirmed
- >600 pmol/L: Diagnose type 2 diabetes—consider non-insulin therapies
- 200-600 pmol/L: Indeterminate—consider islet autoantibodies, genetic testing for MODY, clinical features, and duration of diabetes
Important Caveats
Assay variability remains a significant limitation. Despite calls for standardization, C-peptide measurements still show significant variation between manufacturers 6. The ADA/EASD cut-offs assume standardized assays that don't yet exist universally in practice.
Long-standing type 2 diabetes can have very low C-peptide levels due to β-cell exhaustion, potentially mimicking type 1 diabetes 5. Duration of diabetes and clinical trajectory matter.
Insulin antibodies can interfere with C-peptide assays in some methods, though this is less common with modern assays 1.
The indeterminate zone (200-600 pmol/L) requires clinical judgment incorporating BMI, age, autoantibody status, family history, and response to therapy 2. Don't rely on C-peptide alone in this range.