Assessing the Risk of Legionella Infection through Showering with Untreated Rain Cistern Water in a Tropical Environment

Round 1

Reviewer 1 Report

This is a review of manuscript water-1109670 titled "Assessing the risk of Legionella infection through showering with untreated rain cistern water in St. Thomas, Virgin Islands" that calculates bacterial risks of using cistern water for showering, with particular focus on an environment recently devastated by hurricanes. The manuscript is very well written and with a clear and effective narrative. I recommend publication after addressing just a few minor points in the text.  

Lines 138-157: the baseline distribution of total bacteria from the Crabtree study was for untreated cistern water. But what do we know about the treated or untreated status of the cistern water in the samples used from Jiang et al., (2020) which is used in this study? 

Figure 2: similar vein to above, was there any indication that higher income water users used treatment systems more prevalently? and were the samples randomly selected across income level?

Zhou et al., study showed in cold water usage, fewer aerosols but smaller aerosols were generated. In the discussion of the present study, this difference in size is not really noted; in fact, the opposite is pointed out due to the chimney effect during convective flow more effectively allowing smaller aerosols to rise. Nevertheless, these distinctions should be indicated particularly because of the greater health risk from smaller aerosols  

L313-316: please comment in the text why the Mab (mass deposited in the A-B region) was not a significant model parameter. Is this due to a dilution effect? Although more mass delivered would indicate higher absolute bacteria counts, it would be also represent more water mass, thus diluting the bacteria concentration and not consistently impacting the model output?

other comments:

Title: please consider to change from stating the specific location to "in a tropical environment" to make it more widely applicable

Lines 60-70: it is not clear what you mean by connecting the overflow from the sewers to the contamination of RHRW systems. I assume you mean more opportunity for water contaminants to become airborne and contaminate a RWRW system? Add a few words to clarify this in the text

Author Response

See the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Overall, this is a very well written manuscript and very clearly explained for some, rather complex topics, such as QMRA. The introduction provides a very good explanation of the previous research/problem/research question. I have some questions, however, about some of the assumptions used to put the model together and how applicable they really are.

Major comments:

Why was P<0.10 chosen as significant? Typically 0.05 used; if standard value is used water safety among income is nonsignificant (0.058)

Why no culturing? Why no use of a Legionella specific gene (like mip) as opposed to 16S rRNA next gen sequencing which can cross-react/detect other non-Legionella bacteria due to gene conservation. It is a big assumption that proportion of Legionella is the same genome vs. culture. What is the evidence for this claim?

Is it appropriate to use HPC counts to estimate Legionella when Legionella cannot grow on HPC media and HPC is not representative in previous studies comparing WBPs and HPC- how was HPC done there is no method in the manuscript? was it consistent in different cisterns? The method chosen to do HPC could lead to large differences between different systems and lead to consistency issues in being able to compare results.

Not considering growth in indoor plumbing/showerheads could be a big limitation of this study that should have more discussion. If the systems were abandoned or disused due to natural disaster the pipes in the buildings will be stagnant and may allow for more Legionella growth even than usual in pipes.

Why was concentration in water aerosols assumed to be equal to the water? This is unlikely to be the case, unless there's literature evidence that points to this. Previous studies (Feazel et al.) note preferential aerosolization of NTM and other WBPs from water so there may be even more in the aerosols than the liquid water.

Minor comments:

• 160- considerably higher
• 172- were considered
• Figure 2- "Low" and "High" should say income on the figure; the way this is presented is somewhat confusing. Should be 4 independent questions parts A-D not one line up. The %s on the Y-axis seem to merge together and the Safe/Somewhat safe/Unsafe vs. Yes/No should have different color patterns for the figure
• 381- agreed
• 398- yes this is a problem; there will likely be Legionella growing in plumbing especially when these systems are stagnant and not being used after a natural disaster. See major point above
• pppy is not defined in the manuscript but used as an acronym
• 4.2 section- ensure Legionella is italicized throughout
• 490- Development? Developing?

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

The authors present an interesting QMRA analysis of roof harvested rain water. The paper is generally well written and provides additional information to an emerging field. 

General remarks:

1. The paper is currently too long and lacks brevity. The authors should consider reducing unecessary detail in places.
2. The risk perception elements of this work feel a little superfluous and not related to the QMRA aspects. I would strongly encourage the authors to consider removing these elements and perhaps publishing a shorter report to detail these. The questions about the goverments response to the hurricane are completely irrelevant and should be removed to add clarity. 
3. The major limitation of the study is the use of WGS data for Legionella estimation. Although the authors acknolwedge this limitation i think the limitation needs to be made more explicit. The model currently assumes that all detected Legionella by WGS is a) viable b) infectious and c) released in aerosols. Caution is advised. Pathogenic Legionella were never detected in the samples.
4. I question the significance of this being a post-hurricane study. I can't quite see how a hurricane would influence Legionella contamination of roof water tanks. The authors do not postulate how this might arise. Detail regarding this should be included or the strong references to the hurricane removed. 
5. Introduction line 58-59 - please reword. LD and Pontiac Fever are not symptoms of Legionella infection, they are collective terms for disease manifestations. 
6. I have a major issue with the survey. At no point do the authors explictly ask the respondents if they use roof harvested rain water for showering. They only ask about tap water. Can the authors please clarify? Also, showering is mixed with bathing - how did the authors differentiate showering from bathing. The former has much greater aerosol inhalation risk. 

Author Response

Please see the attachment

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

I appreciate the authors' attention to responding to my initial comments there are just two items that I don't believe were adequately addressed/add some confusion for me.

The authors note "To our knowledge, there has not been a reliable and efficient method to estimate the concentration of culturable Legionella in drinking water" (Lines 360-362). While I understand the interest in the use of qPCR as described in the response and that this study is being generated from data that was not necessarily collected for this purpose, I do not understand the sentence as written. I agree that Legionella is fastidious to culture but that does not mean there are not methods that are well accepted and reliable used in the field to culture Legionella from drinking water including ISO 11731 and the CDC method. This sentence and any other references to there not being methods for Legionella culture should be removed. The authors may benefit from review of these culture methods and publications utilizing them.

In the response, the authors note "The standard HPC culture media and methods were used in the study report" however these are not defined or described in the document or methods section. Even within the SMEWW Method 9215 (if that's what's being used? it's unclear/not mentioned explicitly) there are 4 different acceptable HPC methods and 5 different culture media types. There can be differences between the results on these methods/media types even while still using SMEWW 9215. This should be addressed within the manuscript and whether or not all the HPC cultures were done with the same method/media conditions, otherwise, there could be substantial differences in the HPC results based on the method selected in addition to the noted differences in cistern conditions. This should be clarified in the methods and the discussion (especially if methods differed) as a limitation of the study as being a survey.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

The authors have addressed all my concerns. Happy to recommend for publication. 

Author Response

Thank you again for the time for helping us to improve this manuscript.

Round 3

Reviewer 2 Report

I thank the authors for providing additional clarification to my comments within the manuscript and am happy for this manuscript to move on to acceptance.