Is there utility in testing IgA-endomysial antibodies in patients with weak-positive or equivocal IgA-tissue transglutaminase antibodies in the diagnosis of coeliac disease? A critique of current NICE guidance (NG20).

BackgroundCurrent coeliac disease (CD) NICE guidelines recommend testing IgA-endomysial antibodies (EMA) following a weak-positive IgA-tissue transglutaminase antibody (tTGA). Outside of patients with very high IgA-tTGA results, a positive IgA-EMA necessitates duodenal biopsy to confirm CD diagnosis, meaning a positive IgA-EMA does not alter the diagnostic pathway. Therefore, to be helpful, a negative IgA-EMA needs to reliably exclude CD.ObjectivesWe aimed to evaluate the negative predictive value (NPV) of IgA-EMA, following a weak-positive/positive IgA-tTGA, and to evaluate whether IgA-EMA result (positive or negative) affects duodenal biopsy rates.MethodsRetrospective patient cohort (n = 963) study of patients with IgA-EMA and IgA-tTGA testing, with or without evidence of duodenal biopsy. The NPV of IgA-EMA was assessed by comparison to duodenal biopsy. Duodenal biopsy rates were compared between patients with a positive/negative IgA-EMA (after positive/weak-positive IgA-tTGA).ResultsThe NPVs for CD of a negative IgA-EMA, in the context of a weak-positive or positive IgA-tTGA, were 41% and 0%, respectively (n = 45). There was a significant reduction in the proportion of patients who had a duodenal biopsy with a negative IgA-EMA (9.4%) compared to patients with a positive IgA-EMA (28.5%), following a positive/weak-positive IgA-tTGA (n = 963).ConclusionIgA-EMA does not reliably exclude CD following a positive/weak-positive IgA-tTGA result. Our data indicates that clinicians are utilizing a negative IgA-EMA, following a positive/weak-positive IgA-tTGA result, to inappropriately exclude CD. We recommend IgA-EMA be exclusively used in the context of a 'non-biopsy' approach to CD diagnosis, following a high positive IgA-tTGA, and that a negative IgA-EMA result should not be used to exclude CD in the context of a weak-positive/positive IgA-tTGA.