No-Biopsy Approach to Celiac Disease Diagnosis and Positive Predictive Value (Based on Population)

In medical school, I distinctly remember my anatomy professor discussing how a test’s accuracy was affected by the underlying population tested. He said, ‘Well if you do an RPR test for syphilis among prostitutes and it is abnormal, then it is likely accurate as the test performs better in a population with a higher risk of the condition. Whereas if you run the same test on the faculty, the test would perform with much less accuracy. Well, actually, may be the faculty here are not the best example…’

This discussion comes to mind with the recent publication regarding the No-Biopsy Approach for Celiac disease in adults:

MG Shiha et al. Gastroenterol 2024; 166: 620-630. Open Access! Accuracy of the No-Biopsy Approach for the Diagnosis of Celiac Disease in Adults: A Systematic Review and Meta-Analysis

In this systematic review and meta-analysis with 12,103 adult patients from 18 studies, the key findings:

• The pooled prevalence of biopsy-proven celiac disease in the included studies was 62%.The proportion of patients with IgA-tTG ≥10×ULN was 32%. 
• The summary sensitivity of IgA-tTG ≥10×ULN was 51%, and the summary specificity was 100%
• However, the positive predictive value of the no-biopsy approach to identify patients with celiac disease was 65%, 88%, 95%, and 99% if celiac disease prevalence was 1%, 4%, 10%, and 40%, respectively in hypothetical cohorts.

The associated editorial (B Lebwhol, pg 557: Moving Away From Biopsy Confirmation of Celiac Disease) notes the following:

The duodenal biopsy, long the cornerstone of a celiac disease diagnosis, has lost some of its luster in recent years. As data emerged that the specificity of tissue transglutaminase (TTG) IgA for duodenal villus atrophy increases in accordance with the degree of antibody elevation in children, in 2012 the European Society for Pediatric Gastroenterology and Nutrition adopted a biopsy-free pathway for the diagnosis of celiac disease in symptomatic children who met stringent criteria, including a ≥10-fold TTG IgA elevation and an elevated endomysial antibody on a separate blood draw.¹ With further data supporting the positive predictive value of a highly elevated TTG IgA² these guidelines were modified in 2020 to now include asymptomatic children.³

..the specificity of an elevated TTG IgA seems to be lower in people with type 1 diabetes.⁸

Studies employing a serial serology strategy, such as a ≥10-fold TTG IgA elevation that persists over a number of months, might yield specificities that are so high that that they are impervious to decreases in positive predictive values wrought by low underlying prevalences…

Most patients with celiac disease do not mount a sufficiently high TTG IgA to rely exclusively on a serological diagnosis. And the minority who do meet the stringent criteria of a biopsy-free diagnosis might wisely choose to confirm their life-long diagnosis with a biopsy.

My take: Ironically, this study, which showed that a TTG IgA ≥10 x ULN was highly indicative of celiac disease in adult patients with a 100% specificity and a positive predictive value of 98%, makes an argument for ongoing duodenal biopsies in those at lower risk for celiac disease. However, the lower positive predictive values (in some groups) are based on hypothetical cohorts. Shared decision-making is important and relying on a single elevated lab test is particularly problematic.

Related blog posts: