Long-Term Effect of Oral Exposure to Hexavalent Chromium on Gastrointestinal Cancer Mortality—An Ecological Study in Greece

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

In the ecological study titled “Long-term effect of oral exposure to Hexavalent Chromium on gastrointestinal cancer mortality – An ecological study in Greece” by Katsas et al., the authors examine the long-term impact of Cr(VI) in drinking water on the development and mortality rates of gastrointestinal cancers.

The study is highly compelling and underscores the consequences of inadequate waste management, particularly the failure to ensure proper disposal and the reliance on nature to “take care of it,” which leads not only to major environmental disasters but also to serious and far-reaching public health consequences.

The problem is clearly described, the methodology is appropriate (using SMR with a Poisson distribution, which is well-suited for this type of research), and the results are presented clearly in both tables and figures, thoroughly discussed, and followed by a concise conclusion. The literature cited in the manuscript is appropriate and correctly referenced.

Since most existing literature focuses on occupational exposure to Cr(VI), this study is a valuable step forward in understanding and establishing the presence of environmental exposure to Cr(VI).

The authors acknowledge the inherent limitations of such research, which they have clearly outlined and considered.

Although the manuscript, even in its current form, is entirely satisfactory, I have several recommendations for further improvement. The introduction is highly specific, and it would be beneficial to clarify certain concepts used later in the paper – specifically, the applicable legislation and regulatory framework, as well as the permissible levels of Cr(VI) in water. Additionally, I am interested in understanding why the authors chose the particular time intervals for the second part of the SMR analysis (the data presented in Figure 2).

Author Response

Reviewer: 1

Issues:

1. The introduction is highly specific, and it would be beneficial to clarify certain concepts used later in the paper – specifically, the applicable legislation and regulatory framework, as well as the permissible levels of Cr(VI) in water.

Response

We thank the reviewer for this constructive suggestion. We agree that strongly establishing the unique context of Oinofyta highlights the importance of our study. We have expanded the Introduction to clarify these concepts used later in the paper: “While the global regulatory… industrial pollution in the region” (L71-118). Additionally, the “Study Design and Setting” section in the Methods section was changed accordingly because some information is discussed in detail in the introduction.

2. Additionally, I am interested in understanding why the authors chose the particular time intervals for the second part of the SMR analysis (the data presented in Figure 2).

Response

We thank the reviewer for asking for clarification on this methodological choice. The decision to group the 22-year study period into these specific intervals was driven by the need to balance statistical stability with temporal granularity. Because our study population is relatively small (approximately 4,100 residents ) and the total number of GI cancer deaths is low (67 deaths over 22 years), the annual SMRs exhibit high variance and extreme year-to-year fluctuations (as visualized in Figure 1). By aggregating the data into roughly equal, multi-year blocks, we increased the number of observed and expected cases per interval. This effectively smoothed out the random statistical noise, narrowed the confidence intervals, and allowed us to visualize and evaluate medium-term temporal trends more clearly. We added this clarification in our methods section (“Additionally, annual and 3-year SMRs for all GI cancers were plotted over the study period (2000-2021) to reduce the statistical noise inherent in annual mortality rates within a small population. This aggregation improved the statistical stability of the estimates while maintaining sufficient temporal granularity to observe medium-term trends and disease latency effects.” L222-226).

 

 

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

I thank the authors for the opportunity to review this manuscript and for their work on an important public health issue.

Please revise the terminology throughout the manuscript and avoid using “heavy metals”, as this term is scientifically imprecise and inconsistently defined. Use a more appropriate descriptor (e.g., trace elements, potentially toxic elements, metals and metalloids) and specify the analytes assessed.

Please clarify whether you mean gender or biological sex, and use consistent wording across the manuscript. If the variable comes from registries or administrative records, it is likely biological sex; in any case, describe how it was recorded.

The Introduction would benefit from a clearer Greece-specific context and a stronger justification for focusing on Oinofyta. Please expand the background on the casuistic in Greece and specifically in Oinofyta municipal unit (industrial setting, contamination history, relevance to the study question).

The Materials and Methods should include a proper description of Oinofyta municipal unit and how the study area was defined for mortality attribution (administrative boundaries, population context, water supply characteristics, and any relevant changes over time). The Presidential Decree mentioned should be properly referenced (full citation). In addition, the manuscript should explain, with references, how and why this situation was allowed to occur or persist (e.g., regulatory/enforcement context). If this cannot be supported with evidence, the text should be toned down accordingly.

Reference 15 appears to contain a non-working link; please correct it and ensure the referenced source is accessible.

Around line 103, the manuscript discusses US regulations for Cr(VI) in water; please add the relevant regulatory context for Greece and the EU, and avoid over-reliance on US standards without presenting what applies locally (including whether limits refer to total chromium or specifically Cr(VI), and how this changed over time if applicable).

There is an important temporal inconsistency: you cite Cr(VI) measurements in water from 2007–2008 while discussing mortality trends over 2000–2021. Please justify why those measurements are adequate to interpret a two-decade mortality period, clarify whether they are representative of earlier/later exposure, and temper the interpretation if exposure data are limited.

The causal narrative is currently overstated. The claim that increased GI cancer mortality in the second decade supports the possibility that decades of oral Cr(VI) exposure increased cancer mortality is not sufficiently supported by the analysis as presented. Please rephrase to clearly separate observation from attribution, strengthen the evidence if available, and present alternative explanations and uncertainty.

You mention confounding factors (e.g., smoking) but do not appear to have collected or incorporated relevant covariates. Please either include available covariate information in the analysis (even at area level, if individual data are not possible) or explicitly acknowledge this as a major limitation and avoid attributing the findings to Cr(VI) exposure.

Finally, the discussion of arsenic at the end also reads speculative relative to the analyses shown. Please either support this with relevant data and analysis for As in the study area/timeframe or remove/limit these statements to avoid unsupported linkage.

Author Response

Reviewer: 2

Issues:

1. Please revise the terminology throughout the manuscript and avoid using “heavy metals”, as this term is scientifically imprecise and inconsistently defined. Use a more appropriate descriptor (e.g., trace elements, potentially toxic elements, metals, and metalloids) and specify the analytes assessed.

Response

Corrected. (L19-20, L42-43, L45, L49-50, L441)

2. Please clarify whether you mean gender or biological sex, and use consistent wording across the manuscript. If the variable comes from registries or administrative records, it is likely biological sex; in any case, describe how it was recorded.

Response

We thank the reviewer for the insightful comment. We acknowledge the need for this clarification, as a certain variable was obtained from death registries. “Gender” was replaced with “biological sex” (L28, L184, L198-202,  L206-208, L217, L376).

3. The Introduction would benefit from a clearer Greece-specific context and a stronger justification for focusing on Oinofyta. Please expand the background on the casuistic in Greece and specifically in Oinofyta municipal unit (industrial setting, contamination history, relevance to the study question).

Response

We acknowledge the need for a clearer Greece-specific context and a stronger justification for focusing on Oinofyta in the Introduction. We have expanded the Introduction to address this issue (“While the global regulatory… industrial pollution in the region. L71-118). Additionally, the “Study Design and Setting” part in the Methods section was changed accordingly, because some information is discussed in detail in the introduction part.

4. The Materials and Methods should include a proper description of Oinofyta municipal unit and how the study area was defined for mortality attribution (administrative boundaries, population context, water supply characteristics, and any relevant changes over time).

Response

We thank the reviewer for this recommendation. We have significantly revised the Methods Section to provide a clearer and more structured Oinofyta municipal unit. In more detail, we describe population context from both Oinofyta and Voiotia, water supply characteristics, and any relevant changes over time (i.e., “The study was conducted… levels of Cr(VI) (<0.01-1.53 μg/l).”; L151-178). We clarified the water supply characteristics (“the historical reliance on local groundwater wells that became contaminated”; L151-178) and highlighted the relevant temporal change when the water supply was permanently switched to the safe Mornos Lake system in June 2009 (L151-178). We explicitly state that mortality attribution was strictly defined by the administrative boundaries of the Oinofyta municipal unit (part of the Tanagra municipality) using data from the Hellenic Statistical Authority (“Our study population consisted of… Hellenic Statistical Authority.” L183-188).

5. The Presidential Decree mentioned should be properly referenced (full citation).

Response

Corrected.

6. In addition, the manuscript should explain, with references, how and why this situation was allowed to occur or persist (e.g., regulatory/enforcement context). If this cannot be supported with evidence, the text should be toned down accordingly.

Response

We agree that clarifying the enforcement and regulatory context significantly strengthens the manuscript. While formal, documented evidence explicitly detailing the administrative reasons for why this situation was allowed to persist remains scarce, we have substantially expanded the Introduction to provide a more comprehensive analysis of the historical and regulatory landscape that facilitated these conditions (L83-118).

7. Reference 15 appears to contain a non-working link; please correct it and ensure the referenced source is accessible.

Response

Corrected. Below the functional link:

https://www.elinyae.gr/ethniki-nomothesia/pd-841984-fek-33a-2131984

and also the pdf link

https://www.elinyae.gr/sites/default/files/2019-07/33-84.1110112961738.pdf

8. Around line 103, the manuscript discusses US regulations for Cr(VI) in water; please add the relevant regulatory context for Greece and the EU, and avoid over-reliance on US standards without presenting what applies locally (including whether limits refer to total chromium or specifically Cr(VI), and how this changed over time if applicable).

Response

We thank the reviewer for allowing us to further clarify the existing limits for both Chromium and Cr(VI). A paragraph in the Introduction was added to address this gap (“While the global regulatory… in drinking water. L71-82)

9. There is an important temporal inconsistency: you cite Cr(VI) measurements in water from 2007-2008 while discussing mortality trends over 2000-2021. Please justify why those measurements are adequate to interpret a two-decade mortality period, clarify whether they are representative of earlier/later exposure, and temper the interpretation if exposure data are limited.

Response

We appreciate the reviewer raising this important point regarding the temporal constraints of the exposure data. We agree that the lack of official water measurements before 2007 presents a limitation. However, the 2007-2008 measurements are representative of the culmination of a chronic, decades-long contamination process. As detailed in our manuscript, a 1969 ministerial directive and a 1979 presidential decree authorized the unrestricted discharge of industrial waste into the Asopos River. Furthermore, residents reported visible drinking water discoloration as early as the 1990s, indicating that severe contamination was present long before official measurements were taken. Following the reviewer's recommendation, we address this issue in our limitations: “A notable limitation is the absence…individual participant exposure levels.” (L404-410)

10. The causal narrative is currently overstated. The claim that increased GI cancer mortality in the second decade supports the possibility that decades of oral Cr(VI) exposure increased cancer mortality is not sufficiently supported by the analysis as presented. Please rephrase to clearly separate observation from attribution, strengthen the evidence if available, and present alternative explanations and uncertainty.

Response

We completely agree that the causal language in the previous draft might be overstated, given the inherent limitations of an ecological study design. We have thoroughly revised the Discussion to strictly separate our demographic observations from causal attribution, addressing this issue in our limitations: “A fundamental limitation…variables (i.e., age, occupational exposures, cigarette smoking, nutritional habits, etc.).” (L36, L350, L388-396). Also changed the conclusion accordingly (L425-441).

11. You mention confounding factors (e.g., smoking) but do not appear to have collected or incorporated relevant covariates. Please either include available covariate information in the analysis (even at area level, if individual data are not possible) or explicitly acknowledge this as a major limitation and avoid attributing the findings to Cr(VI) exposure.

Response

The term confounding factors was mentioned only in the limitations of the study, addressing the lack of confounding factors as a limitation from our study “To draw definitive conclusions…variables (i.e., age, occupational exposures, cigarette smoking, nutritional habits, etc.).” (L393-396)

12. Finally, the discussion of arsenic at the end also reads speculative relative to the analyses shown. Please either support this with relevant data and analysis for As in the study area/timeframe or remove/limit these statements to avoid unsupported linkage.

Response

Corrected. We removed the discussion of arsenic from our paper (L419-423).

13. Figures and tables must be improved

Response

We have thoroughly reviewed and upgraded all visual elements in the manuscript to ensure they meet the high publication standards of Environments, according to another publication from the same special issue (DOI: https://doi.org/10.3390/environments12090291).

Specifically, we have completely redesigned the graphical representations. For instance, the temporal SMR trends (Figure 2) have been reprogrammed using R into a professional, high-resolution faceted plot to allow for a direct, side-by-side comparison of the demographic groups. Furthermore, Tables 1 and 2 have been reformatted to strictly adhere to the journal's publication style guidelines, ensuring a clean, readable layout without unnecessary borders (deleted all vertical borders inside the tables, deleted all horizontal lines except three (a horizontal line at the very top of the table, one under the column headers, and one at the very bottom of the table), headers are clearly bolded, table captions are above the table, and footnotes/abbreviations are below the table).

 

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript addresses an important and underreported issue: the long-term effects of hexavalent Chromium exposure on gastrointestinal cancer mortality. The topic is very relevant, especially in regions where chromium exposure remains a public health concern.

The study employs an appropriate ecological approach, and it covers more than 30 years. The study is interesting and well-substantiates the relationship between exposure and mortality from gastrointestinal cancer; however, it has important methodological limitations.

1. The population-level association cannot be directly correlated with the effect at the individual patient level.
2. How is individual water consumption evaluated in patients, with toxic effect, also monitoring studies of chromium levels in drinking water in the analysed area should be described over longer periods of time, not just for limited periods of the years 2007-2008
3. It would also be useful to describe chromium levels determined in biological samples over extended periods in the analysed area
4. How do you analyse the variability of patients' exposure to different contaminated water sources, and how can this influence
5. Important correlations should be developed about patients' co-exposure to: smoking, alcohol, and H. pylori infections to exclude their importance in the cause of gastrointestinal cancer death

It is recommended that conclusions be formulated more cautiously and more closely aligned with the data, avoiding generalisations.

Author Response

Reviewer: 3

Issues:

1. The population-level association cannot be directly correlated with the effect at the individual patient level.

Response

We fully agree that this is a primary limitation of our ecological study design, as our findings are inherently susceptible to the ecological fallacy. We now explicitly state in our limitations that the observed population-level associations cannot be directly correlated with individual patient-level risks: “A fundamental limitation of this design … Cr(VI) exposure and cancer mortality.” (L388-393)

2. How is individual water consumption evaluated in patients, with toxic effect, also monitoring studies of chromium levels in drinking water in the analysed area should be described over longer periods of time, not just for limited periods of the years 2007-2008.

Response

We acknowledge that our ecological study design lacks the evaluation of individual water consumption. We addressed this in our limitations, “A notable limitation … by the residents.” (L404-415) in our limitation.

Regarding the second point, we fully agree that longitudinal monitoring data over a longer period would be ideal. Unfortunately, official measurements before 2007 are nonexistent. However, the 2007-2008 measurements represent the culmination of a chronic, decades-long contamination process. As detailed in our manuscript, a 1969 ministerial directive and a 1979 presidential decree authorized the unrestricted discharge of industrial waste into the Asopos River, and residents reported visible drinking water discoloration as early as the 1990s, indicating severe contamination long before official measurements were taken. To ensure maximum transparency, we have updated the Limitations section to reflect this context: “A notable limitation is the … exposure levels.” (L404-410). Also changed the conclusion accordingly (L425-441).

3. It would also be useful to describe chromium levels determined in biological samples over extended periods in the analysed area

Response

We fully agree that longitudinal biomonitoring data (such as blood or urine chromium levels) provide the most accurate assessment of internal exposure and greatly strengthen our ecological findings. Unfortunately, no biological sampling was ever conducted or recorded for the residents of Oinofyta during the decades of active contamination or over the subsequent follow-up period. Because this biological data is entirely unavailable, we are unable to describe it in our analysis. This was addressed in our limitations: “Another critical limitation is the absence of a dose-response analysis; … the residents” (L410-415)

4. How do you analyse the variability of patients' exposure to different contaminated water sources, and how can this influence

Response

We thank the reviewer for raising this important methodological question. Because this is an ecological study based on aggregate population data, we are unable to quantitatively analyze individual variability in water consumption. However, we have explicitly addressed the potential variability of exposure to different water sources and its influence on our findings in the Limitations section (L399-402).

5. Important correlations should be developed about patients' co-exposure to: smoking, alcohol, and H. pylori infections to exclude their importance in the cause of gastrointestinal cancer death

Response

We did not have access to individual confounding factors such as smoking, alcohol consumption, and H. pylori infections in our study. Due to the nature of our ecological study, we could only gather aggregate data on the total number of deaths, categorized by age and sex groups. We acknowledged the absence of these individual-level confounders and emphasized that population-level associations cannot be directly correlated with risks at the individual patient level in our limitations section (L391-396).

6. It is recommended that conclusions be formulated more cautiously and more closely aligned with the data, avoiding generalisations.

Response

We agree that the previous conclusion contained generalizations that extended beyond the scope of an ecological study design. We have thoroughly revised the Conclusions section to ensure our statements are strictly aligned with our data (L425-441).

7. Figures and tables must be improved

Response

We thank the reviewer for this feedback. We have thoroughly reviewed and upgraded all visual elements in the manuscript to ensure they meet the high publication standards of Environments, according to another publication from the same special issue (DOI: https://doi.org/10.3390/environments12090291).

Specifically, we have completely redesigned the graphical representations. For instance, the temporal SMR trends (Figure 2) have been reprogrammed using R into a professional, high-resolution faceted plot to allow for a direct, side-by-side comparison of the demographic groups. Furthermore, Tables 1 and 2 have been reformatted to strictly adhere to the journal's publication style guidelines, ensuring a clean, readable layout without unnecessary borders (deleted all vertical borders inside the tables, deleted all horizontal lines except three (a horizontal line at the very top of the table, one under the column headers, and one at the very bottom of the table), headers are clearly bolded, table captions are above the table, and footnotes/abbreviations are below the table).

 

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have revised the manuscript in response to the reviewers’ comments. However, the Materials and Methods section remains insufficiently robust, and key methodological limitations inherent to the ecological design persist. The absence of individual-level exposure assessment, biomonitoring data, and dose–response analysis substantially limits causal inference. In my opinion, further work is required before publication. Specifically, the authors should consider designing a more rigorous epidemiological framework that includes systematic environmental monitoring of Cr(VI) in water and soil matrices, as well as biomonitoring of exposed populations. Particular emphasis should be placed on assessing internal chromium burden in individuals diagnosed with gastrointestinal cancers potentially associated with Cr(VI) exposure, in order to strengthen the biological plausibility and exposure–outcome linkage.

Author Response

Reviewer 2

The authors have revised the manuscript in response to the reviewers’ comments. However, the Materials and Methods section remains insufficiently robust, and key methodological limitations inherent to the ecological design persist. The absence of individual-level exposure assessment, biomonitoring data, and dose–response analysis substantially limits causal inference. In my opinion, further work is required before publication. Specifically, the authors should consider designing a more rigorous epidemiological framework that includes systematic environmental monitoring of Cr(VI) in water and soil matrices, as well as biomonitoring of exposed populations. Particular emphasis should be placed on assessing internal chromium burden in individuals diagnosed with gastrointestinal cancers potentially associated with Cr(VI) exposure, in order to strengthen the biological plausibility and exposure–outcome linkage.

We fully acknowledge that the ecological design inherently limits causal inference at the individual level. As known, ecological studies are defined by analysis at the population level and are particularly appropriate when individual-level exposure data are unavailable, when exposure affects an entire community, or when the objective is to evaluate population-level health monitoring and generate hypotheses in environmental contexts [1]. In the case of Oinofyta, historical individual exposure measurements and biomonitoring data were not available, and contamination of the municipal water supply constituted a geographically defined, community-wide exposure. Therefore, the ecological framework was intentionally selected to evaluate population-level mortality patterns rather than to establish individual-level causality.

We agree that the absence of individual-level exposure assessment, biomonitoring, and direct dose-response modeling precludes definitive causal attribution. However, such analyses would require a fundamentally different analytical epidemiological design, including prospective environmental monitoring and biological sampling, which extends beyond the scope of the present ecological investigation.

Importantly, ecological studies remain a well-recognized and valuable approach in environmental public health research for identifying population-level signals and contextual effects of widespread exposure [1]. Their purpose is not to replace cohort or case-control designs, but to inform and justify subsequent analytical investigations. In this context, our findings should be interpreted as hypothesis-generating and indicative of a potential population-level association rather than proof of individual-level causation.

In response to the reviewer’s comments, we have:

1. Strengthened the justification of the ecological design in the Methods section by adding a short paragraph at the end of 2.1 Study Design and Setting explaining why an ecological design is appropriate here with supporting literature (L154-161).
2. Explained the residence-based exposure, justifying that exposure classification is reasonable, by adding a short paragraph in the Methods section 2.2. Study Population (L167-173).
3. Acknowledged the lack of a direct dose-response association, and the attempt of an indirect method, using time as proxy, by adding a short paragraph in the Results section  (L206-211) and we have already addressed this in our limitations paragraph.
4. Expanded the Limitations section to more explicitly discuss dose-response association, explaining that individual-level epidemiological data collection is beyond the scope of the present investigation (L406-411).
5. Clarified throughout the manuscript that findings represent population-level/ecological associations (L36, 157, 301, 383).
6. Expanded the Conclusions to explicitly recommend future cohort studies incorporating systematic environmental monitoring and biomonitoring of exposed individuals, including assessment of internal chromium burden (L424-433).

Reference
[1] Levin, K.A. Study Design VI - Ecological Studies. Evid Based Dent 2006, 7, 108–108, doi:10.1038/sj.ebd.6400454.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

It seems that the authors approved and accordingly completed the asked items. In my opinion, it can be published in the present form.

Author Response

We sincerely thank the reviewer for the careful evaluation of our manuscript and for the positive feedback. We greatly appreciate your acknowledgment that the requested revisions have been satisfactorily addressed and that the manuscript is suitable for publication in its present form.

Thank you again for your valuable comments and support.

Round 3

Reviewer 2 Report

Comments and Suggestions for Authors

I appreciate the efforts made in revision and the additional rationale provided. However, the central methodological weaknesses persist and are likely to materially bias the findings. As a result, the current data and analyses do not adequately support the conclusions as written. To move forward constructively, please reconsider the scope of the manuscript: present the work as preliminary/observational evidence, limit interpretation to descriptive findings, and ensure that limitations and potential sources of bias are clearly and prominently stated.

Author Response

We have added a dedicated subsection entitled “Study design limitations” (L382-459) in the Discussion section of the revised manuscript, explicitly describing the key methodological constraints of our study. We also clarify that the findings should be interpreted cautiously as preliminary, population-level observational evidence, and that causal inference cannot be established based on the current data. We again mention that our findings “should be interpreted cautiously and may reflect multiple factors” in our conclusion (L466-468).