A double-sandwich enzyme immunoassay method was developed for determination of serum immunoglobulin A (S-IgA) and mucosal secretory immunoglobulin A (sIgA) in duodenal brush samples obtained via endoscopy and the relationship between enteric mucosal sIgA, salivary sIgA and S-IgA in dogs was examined. Twenty healthy dogs underwent routine endoscopy. A brush sample from the duodenal mucosa was obtained and washed in PBS, with a serum sample being taken concurrently. A saliva sample was collected from twelve of these dogs. S-IgA and sIgA with total protein concentrations in the duodenal washings and saliva samples were determined. A significant negative correlation (r = -0.64, P = 0.0059) was found between duodenal sIgA/protein ratios and S-IgA concentrations. Saliva sIgA/protein ratios did not correlate with sIgA/protein ratios of duodenal samples. The method described here allows for direct assessment of duodenal IgA; therefore indirect measures based on serum IgA or salivary IgA can be avoided. In addition, these indirect measures appear to be poor indicators of duodenal sIgA competence in dogs.