Recent reports indicate that SARS-CoV-2 triggers the secretion of immunoglobulin A, the most abundant antibody isotype produced in humans, and a major component of the mucosal immune response. However, conventional IgA serology does not discriminate monomeric IgA from dimeric IgA (dIgA). Dimeric IgA is the major polymeric form of IgA produced by IgA secreting plasma cells found mainly in the mucosal sites and is the precursor of mucosal secretory IgA. Dimeric IgA and secretory IgA are among the first lines of defence on the mucosal lining, neutralising toxins and pathogens, preventing adhesion of pathogens, and facilitating clearance of foreign antigens. We describe the first discriminatory ELISA for SARS-CoV-2 dIgA based on the highly selective binding of dIgA to a chimeric form of the secretory component (CSC). SARS-CoV-2 dIgA was measured in a longitudinal panel of 168 plasmas (n=30) that included samples sourced from hospitalised cases of COVID-19 and a commercial panel with mild infection, a cross-sectional panel with PCR-confirmed SARS-CoV-2 infection (n=199) with mild COVID-19, and pre-COVID-19 samples (n=200). The assay demonstrated specificity of 99%, 98.5% and 97.5% for dIgA, IgG and IgA, respectively. Analysis of dIgA in the longitudinal panel revealed that 86.6% of subjects tested positive for dIgA by day 13 post PCR diagnosis. In the individuals with mild COVID-19, plasma dIgA levels increased within the first 2 weeks post diagnosis, but were short lived and declined rapidly, while IgG levels were sustained. By contrast, the plasma of hospitalised COVID-19 patients possessed higher and sustained levels of dIgA, up to 245 days post PCR diagnosis. The use of CSC allows the specific measurement of dIgA in the immune response to SARS-CoV-2 and our results suggests that severe COVID-19 is associated with more sustained and higher-levels dIgA in plasma than mild COVID-19.