Secondary treatment failure (STF) of botulinum toxin A (BoNT/A) therapy in cosmetic indication has been postulated as production of antibody against active sites of BoNT/A in unresponsive patients. To prove of concept, detection of anti-BoNT/A antibody is required, however, current enzyme-linked immunosorbent assay (ELISA) detects human IgGs against whole BoNT/A molecule. We developed an inhibition ELISA to quantify antibodies bound to the active sites of BoNT/A using three mouse monoclonal antibodies targeting translocation domain, receptor binding site and catalytic domain of BoNT/A prior to processing ELISA to detect human IgG (hIgG) against BoNT/A. Adults naïve to BoNT/A, or treated and responsive (toxin-response), or treated but unresponsive (toxin-tolerance) were recruited. Detection of hIgG revealed that naïve volunteers had basal level of hIgG against whole BoNT/A, whereas its level was significantly lower than those hIgG in BoNT/A-exposed cohorts. Higher anti-BoNT/A levels in sera from volunteers ever-exposed to BoNT/A indicates that BoNT/A may provoke immune responses in BoNT/A-treated cohorts. Inhibition ELISA demonstrated that levels of BoNT/A-specific hIgG in tolerance patients had a dramatic decrease in mouse monoclonal antibody blockage, suggesting presence of hIgG specific to BoNT/A's three active sites in STF patients. Therefore, our ELISA detected hIgG against whole BoNT/A protein and BoNT/A active sites suggesting that human antibodies may cause STF. To compare with frontalis test, our inhibition ELISA provided good accuracy at 83.1% (50% sensitivity and 89.9% specificity). Our test may help clinicians to diagnose possibility of STF and also to monitor immune status against BoNT/A.