What are the diagnostic criteria for celiac disease, including serologic screening, duodenal biopsy findings, HLA-DQ2/DQ8 typing, and special considerations for children?

Celiac Disease Diagnostic Criteria

In adults, celiac disease diagnosis requires both positive IgA anti-tissue transglutaminase (IgA-TG2) serology AND duodenal biopsy showing villous atrophy while the patient is consuming a gluten-containing diet. 1

Diagnostic Algorithm for Adults

Step 1: Serologic Screening (Patient Must Be on Gluten-Containing Diet)

First-line test:

• IgA-TG2 (sensitivity 93.4%, specificity 95.8%) 2, 3
• Total IgA level (to identify IgA deficiency) 1

If IgA deficient:

• Use IgG-TG2 or IgG-DGP (deamidated gliadin peptide) 1

Second-line confirmatory test (if needed):

• IgA-EMA (endomysial antibodies) - highest specificity at 99% 2, 3
• Note: Routine EMA confirmation is no longer recommended with high-performance validated ELISA assays 4

Step 2: Duodenal Biopsy

Biopsy is mandatory for adult diagnosis with rare exceptions (coagulation disorders, pregnancy) 1

Biopsy protocol:

• Minimum 4 biopsies from second part of duodenum 4
• Consider bulb biopsies (increases diagnostic yield by doubling detection rate) 5

Histologic criteria for definite celiac disease:

• Villous atrophy (Marsh 3) is required for definitive diagnosis 1
• Increased intraepithelial lymphocytes (IELs) ≥25/100 enterocytes 1
• Crypt hyperplasia 6

"Probable" celiac disease:

• Lesser degrees of damage (≥25 IELs but no villous atrophy) may indicate celiac disease 1
• Trial of gluten-free diet can support diagnosis 1

Step 3: HLA-DQ2/DQ8 Typing (Selective Use)

When to order HLA testing:

• Equivocal or discrepant serology and histology 1, 5
• Seronegative villous atrophy 5
• Patient already on gluten-free diet 5
• Diagnostic uncertainty requiring clarification 4

Interpretation:

• ≥90% of celiac patients are HLA-DQ2.5 positive 1
• Majority of remaining patients are HLA-DQ8 positive 1
• Negative HLA-DQ2/DQ8 effectively rules out celiac disease (>99% of patients carry these alleles) 5, 7
• Positive HLA does NOT confirm diagnosis (present in ~30% of general population) 5

Special Considerations for Children

Children can be diagnosed WITHOUT biopsy if ALL criteria met: 8, 9

• IgA-TG2 ≥10 times upper limit of normal (ULN)
• Positive IgA-EMA in second blood sample
• At least one symptom present (any symptom, or specifically malabsorption symptoms for 100% PPV)
• HLA-DQ2/DQ8 positive (per 2012 ESPGHAN guidelines, though newer data suggests this may not be necessary) 9, 10

Validation data: This no-biopsy approach has PPV of 99.75% (any symptom) to 100% (malabsorption symptoms only) 9

Recent evidence suggests further simplification: Biopsy and HLA testing may be unnecessary in children with IgA-TG2 ≥10x ULN, regardless of symptoms 10

Diagnostic Scenarios and Interpretation

Definite Celiac Disease:

• Positive IgA-TG2 (or IgG-DGP/IgG-TG2 if IgA deficient)
• Villous atrophy on biopsy
• Patient on gluten-containing diet 1

Potential Celiac Disease:

• Positive serology
• Positive HLA-DQ2/DQ8
• No villous atrophy on biopsy 1

Seronegative Celiac Disease:

• Can occur but is rare 11
• Diagnosis established by clinical and histologic response to gluten-free diet
• Compatible HLA haplotypes required 11

Critical Pitfalls to Avoid

1. Never start gluten-free diet before testing - will invalidate both serology and histology 12

2. Patient must consume gluten for ≥6 weeks before testing if already avoiding gluten 2

3. Check total IgA level - IgA deficiency occurs in ~2-3% of celiac patients and causes false-negative IgA-based tests 1

4. Biopsy quality matters - inadequate sampling or poor orientation causes misdiagnosis 6, 5

5. Consider differential diagnoses for villous atrophy if seronegative: common variable immunodeficiency, autoimmune enteropathy, tropical sprue, medication-induced enteropathy (olmesartan, NSAIDs, mycophenolate), Crohn's disease, giardiasis 11, 1

6. Elevated antibodies do NOT always mean ongoing gluten ingestion - can occur in refractory celiac disease 11

Genetic Background

• HLA-DQ2.5 accounts for 40% of genetic variance (most patients DR3, DQ2 positive) 1
• 39 non-HLA loci account for only 14% of genetic variance 1
• 10% prevalence among first-degree relatives 1
• 75% concordance in monozygotic twins 1