Author’s response to reviews Title: Heavy Metal Exposure and Nasal Staphylococcus aureus Colonization: Analysis of the National Health and Nutrition Examination Survey (NHANES) Authors:

-R1 Comment: Sample size is missing. Response: Sample size was added to the methods section of the abstract. R1 Comment: The wording suggests cause and effect, e.g. 'both metals were protective against MSSA'. No causal relationships are proven, and the associations were often weak or notsignificant. The wording around the strength of the associations in the whole manuscript could be toned down. Response: Agreed, language throughout the manuscript was changed to highlight associations instead of causality. R1 Comment: Line 26: Can the authors explain what 'co-select' means in this application? Response. We agree this is a difficult term not often used in EH research, but known and used regularly in infectious disease epidemiology and in previous work describing how metals may confer antibiotic resistance. Text within the abstract was edited to highlight the main point that heavy metal exposure may be promulgating antibiotic resistance. We added details to the main body of the text to describe the many potential mechanisms of co-selection but hesitate to add this level of detail in the abstract in order to ensure readability. We would be happy to add text now in the main body regarding co-selection into the abstract if necessary. R1 Comment: Resultsadd sample size. Some sweeping statements are given regarding metal concentrations found and MRSA. On further examination these differences between the GMs are not statistically different, so either the language needs to reflect this or p values given. Response: We agree with the reviewer this was missing, sample size has been added to the methods. Language has been altered to emphasize that GMs are not statistically different despite overall significance in p for trend analyses. R1 Comments: 'both metals were protective against MSSA': This assumes causation. A dose response relationship was shown, but with adjustment for confounders, there is little difference in the ORs of the quartiles (CIs overlap). Response: Agreed, language changed to highlight associations instead of causality R1 Comment: Conclusions: the authors refer to 'current' population exposure when in fact the samples were taken 13-16 years ago and environmental exposures and indeed blood Pb/Cd levels may have changed during this time. How do the authors know that the metal exposures and colonisations was concurrent? Response: The word current was changed to general, to imply that study participants were not necessarily occupationally exposed to either metal. The metal exposure and colonization were concurrent because samples used to assess both were collected at the same time. Further explanation of this has been added to the methods section of the paper.