BACKGROUND: The revised paediatric criteria for coeliac disease allow omission of duodenal biopsies in symptomatic children who have specific serology and coeliac disease-associated genetics. It remains unclear whether this approach is also applicable for adults with various clinical presentations. AIM: To evaluate the accuracy of serology-based criteria in adults with variable pre-test probabilities for coeliac disease. METHODS: Three study cohorts comprised adults with high-risk clinical coeliac disease suspicion (n = 421), moderate-risk family members of coeliac disease patients (n = 2357), and low-risk subjects from the general population (n = 2722). Serological and clinical data were collected, and "triple criteria" for coeliac disease comprised transglutaminase 2 antibodies >10× the upper limit of normal, positive endomysium antibodies, and appropriate genetics without requirement of symptoms. The diagnosis was based on intestinal biopsy. RESULTS: The diagnosis of coeliac disease was established in 274 subjects. Of these, 59 high-risk subjects, 17 moderate-risk subjects, and 14 low-risk subjects fulfilled the "triple criteria". All had histologically proven coeliac disease, giving the criteria a positive predictive value of 100%. Altogether, 90 (33%) of all 274 newly diagnosed patients could have avoided biopsy, including 37% among high-risk, 20% among moderate-risk, and 48% among low-risk patients. No histological findings other than coeliac disease were found in the biopsies of "triple positive" subjects. CONCLUSIONS: Coeliac disease can reliably and safely be diagnosed without biopsy in adults fulfilling the "triple criteria" regardless of the pre-test probability. Revised criteria would enable the number of endoscopies to be reduced by one-third.