Detection of recurrent venom post-antivenom in snake envenoming is commonly reported and thought to be due to insufficient antivenom. However, relatively few reports of recurrence have venom measurement, and in most cases patients clinically improve, despite venom detected post-antivenom. We hypothesized that persistent or recurrent venom detection post-antivenom is due to detecting bound venom. Multiple (>4) serum samples were available from 255 Russell's viper (Daboia russelii) envenomed patients. Enzyme-linked immunosorbent assay was used to measure venom, antivenom and venom–antivenom (VAV) complexes. In 79/255 (31%) there was persistent/recurrent venom detected despite antivenom being present. In these post-antivenom samples, VAV was also detected at the same time as venom was detected. Anti-horse (aH) antiserum was bound to UltraLink (UL) resin and added to in vitro venom–antivenom mixtures, and 15 pre- and post-antivenom samples from patients. There was significantly less free venom detected in in vitro venom–antivenom mixtures to which ULaH had been added compared to those without ULaH added. In 9 post-antivenom patient samples the addition of ULaH reduced venom detected by a median of 80% (69%–88%) compared to only 20% in four pre-antivenom samples. This suggests that the detection of persistent/recurrent venom post-antivenom is due to bound and not free venom.