Over the past years, neutrophil extracellular traps (NETs) were shown to contribute to states of acute and chronic inflammatory disease. They are composed of expelled chromatin and decorated by neutrophil-derived proteins. Therefore, the analysis of DNA complexes with myeloperoxidase (MPO) by ELISA has become an attractive tool to measure NET formation in in vitro and in vivo samples. When we used a published MPO-DNA ELISA protocol and included an isotype control for the anti-MPO coating antibody, we observed high assay specificity for in vitro prepared NET samples, whereas the specificity for in vivo plasma samples was low. In addition, the assay failed to detect in vitro generated MPO-DNA complexes when spiked into plasma. Therefore, we set out to improve the specificity of the MPO-DNA ELISA for plasma samples. We found that the use of Fab fragments or immunoglobulins from different species or reversal of the antibody pair led to either a high background or a low dynamic range of detection that did not improve the specificity for plasma samples. Also, the use of higher plasma dilutions or pre-clearing of plasma immunoglobulins were ineffective. Finally, we found that a commercial reagent designed to block human anti-mouse antibodies and multivalent substances increased the detection window between the MPO antibody and isotype control for highly diluted plasma. We applied this modified ELISA protocol to analyze MPO-DNA complexes in human blood samples of acute and chronic inflammatory conditions. While markers of neutrophil activation and NET formation such as MPO, elastase and citrullinated histone H3 correlated significantly, we observed no correlation with the levels of MPO-DNA complexes. Therefore, we conclude that ELISA measurements of MPO-DNA complexes in human plasma are highly questionable regarding specificity of NET detection. In general, plasma analyses by ELISA should more frequently include isotype controls for antibodies to demonstrate target specificity.