Low Endotoxin Recovery (LER) is the loss of ability of the industry-standard lipopolysaccharide (LPS)-detection method, the Limulus Amebocyte Lysate (LAL) assay, to detect LPS in formulations containing chelating agents and surfactants. LER may potentially interfere with the testing of biopharmaceutical formulations for the absence of endotoxins. The mechanism of LER remains unresolved, with some suggesting that it may be caused by changes in LPS aggregate size. Here, we offer a comprehensive analysis of the effects of chelating agents, surfactants and multivalent cations on both LPS masking and LPS aggregate sizes. In the presence of certain divalent or trivalent cations, surfactants and chelating agents an apparent correlation between LER effects and a reduction in the size of LPS aggregates is observed. However, for other examples of the same type of excipient no such correlation exists. This indicates that any previously proposed correlation between aggregate size and LER is merely coincidental and does not contribute to the LER mechanism. Additionally, adding Mg2+ cations fully prevented LER over 7 days, while trivalent cations only temporarily delayed it, raising a question about the role of electrostatic interactions and aggregate rearrangements in the LER effect. We propose that, despite the apparent general reduction in aggregate sizes in some LER conditions, the mechanism of LER should be attributed to modification of the surface of LPS aggregates, rather than their size or LPS disaggregation. These findings can be used to direct further research into LER mechanisms and to reconsider the strategies to mitigate LER in LAL assays.