The fluorescent probe 2′,7′-dichlorofluorescein diacetate (DCFH-DA) together with the enzyme horseradish peroxidase (HRP) is widely used in nanotoxicology to study acellular reactive oxygen species (ROS) production from nanoparticles (NPs). This study examined whether HRP adsorbs onto NPs of Mn, Ni, and Cu and if this surface process influences the extent of metal release and hence the ROS production measurements using the DCFH assay in phosphate buffered saline (PBS), saline, or Dulbecco's modified Eagle's medium (DMEM). Adsorption of HRP was evident onto all NPs and conditions, except for Mn NPs in PBS. The presence of HRP resulted in an increased release of copper from the Cu NPs in PBS and reduced levels of nickel from the Ni NPs in saline. Both metal ions in solution and the adsorption of HRP onto the NPs can change the activity of HRP and thus influence the ROS results. The effect of HRP on the NP reactivity was shown to be solution chemistry dependent. Most notable was the evident affinity/adsorption of phosphate toward the metal NPs, followed by a reduced adsorption of HRP, the concomitant reduction in released manganese from the Mn NPs, and increased levels of released metals from the Cu NPs in PBS. Minor effects were observed for the Ni NPs. The solution pH should be monitored since the release of metals can change the solution pH and the activity of HRP is known to be pH-dependent. It is furthermore essential that solution pH adjustments are made following the addition of NaOH during diacetyl removal of DCFH-DA. Even though not observed for the given exposure conditions of this study, released metal ions could possibly induce agglomeration or partial denaturation of HRP, which in turn could result in steric hindrance for H2O2 to reach the active site of HRP. This study further emphasizes the influence of HRP on the background kinetics, its solution dependence, and effects on measured ROS signals. Different ways of correcting for the background are highlighted, as this can result in different interpretations of generated results. The results show that adsorption of HRP onto the metal NPs influenced the extent of metal release and may, depending on the investigated system, result in either under- or overestimated ROS signals if used together with the DCFH assay. HRP should hence be used with caution when measuring ROS in the presence of reactive metallic NPs. © 2021 The Authors.