• Decrease font size
  • Return font size to normal
  • Increase font size
U.S. Department of Health and Human Services

Science & Research

  • Print
  • Share
  • E-mail

Peer Review Report

Risk Profile: Hepatitis A Virus Infection Associated
with Consumption of Fresh and Fresh-Cut Produce


Introduction

The Food and Drug Administration, Center for Food Safety and Applied Nutrition (FDA/CFSAN) contracted with Versar, Inc. (Springfield, VA), to conduct an external letter review of the report "Draft Risk Profile: Hepatitis A Virus Infection Associated with Consumption of Fresh and Fresh-Cut Produce" (Contract Number HHSF223200730231G, Task Order 2). On July 22, 2008, the report was sent to five experts with a request for comments by August 30, 2008. The peer reviewers were as follows:

Tomás J. Aragón, MD, Dr.PH
Center for Infectious Diseases & Emergency Readiness
University of California, Berkeley School of Public Health
Berkeley, CA 94704

Dean O. Cliver, Ph.D.
School of Veterinary Medicine
University of California, Davis
Davis, CA 95616

Phaedra S. Corso, Ph.D.
University of Georgia
College of Public Health
Athens, GA 30602

Robert B. Gravani, Ph.D.
Department of Food Science
Cornell University
Ithaca, NY 14853

Harold S. Margolis, MD
International Vaccine Institute
SNU Research Park
Seoul, Korea 151-919


Charge To Peer Reviewers

This document is a risk profile; it describes the current state of scientific knowledge of hepatitis A virus illness associated with the consumption of fresh/fresh-cut produce. It describes the public health problem and current control strategies. It also includes an evaluation of the strengths and limitation of future potential mitigation strategies as well as significant gaps in knowledge. The term risk profile has been defined by the Codex Commission for Food Hygiene (CCFH) as a decision tool which it uses to consider proposed new work in microbiological risk assessment/risk management. CFSAN has adapted this tool for the purpose of describing a food safety problem by identifying those elements of the hazard or risk that are relevant to a food safety issue and potential risk management options. A risk profile will also identify data gaps. The format of this risk profile is very different from that of a risk assessment in that it is not divided into: Hazard Identification; Hazard Characterization; Exposure Assessment; and Risk Characterization.

 

Charge Questions:

  1. Are the scope and purpose of the risk profile clearly identified? If not, what additional information should be provided?
  2. Considering the public health problem addressed, is the methodology used appropriate and is the report written in a transparent and clear manner? Does the report adequately address the questions and stated objectives? If not, how might the report be revised?
  3. Are there any additional scientific or technical studies available that were not considered that are relevant to the public health impact of hepatitis A virus in fresh and fresh-cut produce? For example, has FDA identified all sources of contamination throughout the farm-to-fork continuum? Are there any additional control measures or interventions that were not described?
  4. Regarding the economics of illness due to hepatitis A virus-infected produce:
    • Are the estimated costs for HAV illnesses associated with morbidity and death reasonable? If not, are there available published reports with more recent cost estimates?
    • Is the use of QALDs (Quality Adjusted Life Days) for describing loss of well-being due to HAV infection appropriate?
    • Do you have any additional information or estimates of any potential long term shifts in consumer demand away from fresh produce that result from produce-associated outbreaks?
    • Are there additional estimates of post-outbreak adjustment costs to industry in addition to that documented on page 37 of the document (e.g., law suits and business closures)?
  5. Does the report identify all potential future controls and does it adequately describe the strengths and limitations of each? If not, please provide additional suggestions of information to include in the report.
  6. Are the data gaps presented in Appendix 1 complete and adequately described? Are there additional research needs that should be considered in relation to this public health issue?
  7. Do you have any additional comments that might improve the document?

 


Summary of Comments

Introduction

FDA thanks the experts for taking the time and effort to provide us with their comments. We feel that these comments helped, to a great extent, to improve our risk profile. We have addressed the comments to the best of our ability and according to the available scientific data. Immediately below is a summary of the comments prepared by Versar, Inc., under contract to FDA (without any FDA response). Below the summarized comments is a table that sets forth the specific reviewer comments (the identities of the reviewers are masked), with FDA's response to each.
NOTE: the page numbers in the reviewers' comments do not necessarily reflect the page numbers in the revised document.

Summarized Comments

 General Impressions: In general, reviewers responded positively to the overall organization and scientific merit of the HAV risk profile. Specific comments credit the document's summary of the current state of knowledge and practices of Hepatitis A food protection, its broad body of information, its presentation of information in a logical sequence, and its sound conclusions. As a general critique of the HAV risk profile, one reviewer commented that the multi-authored document lacked a unified editorial direction and was inconsistent in both style and message. The reviewer wrote that the document should include all available information and not be limited by the year 2000. The reviewer also was concerned about the fact that the document seemed to hold all fresh-cut produce at an equal risk of becoming contaminated with HAV and presented an unbalanced view of the magnitude of the problem of foodborne HAV transmission, compared with other sources of HAV in the United States. Another reviewer was concerned that the cost/benefit analysis was applied only to the immunization of food workers and not to other proposed control measures. 

Charge Question 1: Are the scope and purpose of the risk profile clearly identified? If not, what additional information should be provided?

The majority of reviewers responded that the scope and purpose of the risk assessment were clearly identified. One reviewer commented that it might be helpful to modify the description of the risk profile found in Appendix 1 and to include it in the introduction of the report. Another reviewer believed that the authors should have developed a framework for assessing the true risk of hepatitis A in the United States. That reviewer also stated that the authors should address whether hepatitis A represents a unique enough issue to warrant its own prevention strategy, different from that of other foodborne diseases transmitted by the fecal-oral route. 

Charge Question 2: Considering the public health problem addressed, is the methodology used appropriate and is the report written in a transparent and clear manner? Does the report adequately address the questions and stated objectives? If not, how might the report be revised?

Responses to this charge question, regarding the appropriateness of the methodology used and the transparency/clarity of the report, were mixed. One reviewer listed a number of points that were not made clear in the document, including how articles and reports were selected and from which databases, selection criteria, number of articles and/or reports selected, evaluation of the quality of reports and articles, and whether unpublished and/or foreign language studies were included. That reviewer also noted that elucidating the above list would make the risk profile transparent, with reproducible methodologies. Another reviewer commented that while Sections III, V, and VI of the report seemed relevant and transparent, Sections IV and VII could be improved. A reviewer added that the risk profile seemed biased towards foodborne hepatitis A being a major public health problem without a convincing evidence-based argument. That reviewer went on to comment that viewing food handlers/workers/preparers at each stage of the process as conferring the same degree of risk in introducing HAV is illogical and not supported by the data. Another reviewer commented that the report accurately and transparently addresses the public health problem associated with HAV. 

Charge Question 3: Are there any additional scientific or technical studies available that were not considered that are relevant to the public health impact of hepatitis A virus in fresh and fresh-cut produce? For example, has FDA identified all sources of contamination throughout the farm-to-fork continuum? Are there any additional control measures or interventions that were not described?

In answer to this charge question concerning the availability of additional scientific or technical studies, many of the reviewers offered additional references not cited and listed them in the specific comment sections of their individual responses. One reviewer commented, more specifically, on the topics of molecular epidemiology and hepatitis A vaccination, pointing out certain shortcomings in the discussion with respect to its utility. 

Charge Question 4a: Regarding the economics of illness due to hepatitis A virus-infected produce: Are the estimated costs for HAV illnesses associated with morbidity and death reasonable? If not, are there available published reports with more recent cost estimates?

Overall, reviewers found the estimated costs for HAV illnesses associated with morbidity and mortality reasonable given the paucity of contemporary data in this area. One reviewer noted a 2002 paper not found in the risk profile entitled Valuation of Symptomatic Hepatitis A in Adults: Estimates Based on Time Trade-off and Willingness-to-pay Measurement. Another reviewer had much more to say in response to this charge question, first noting that the authors did not attempt to estimate aggregate economic burden within a given time period (such as annually). Adding such an estimate, the reviewer stated, would strengthen the report. That reviewer also commented on the authors’ use of data on medical and hospital costs from the studies Lasher et al., 2007, and Scharff et al., 2008. A reviewer pointed out that both papers were lacking details on the methodology for estimating costs. After reviewing some of the cited papers, the reviewer noticed more information from the Dalton et al. and Berge et al. papers on cost estimates for productivity loss and other cost per case estimates that should have been included in the report. 

Charge Question 4b: Regarding the economics of illness due to hepatitis A virus-infected produce: Is the use of QALDs (Quality Adjusted Life Days) for describing loss of well-being due to HAV infection appropriate?

Generally, reviewers agreed that the use of QALDs (Quality Adjusted Life Days) was appropriate in this setting. One reviewer noted that, while the use of QALDs is an acceptable way to measure burden of disease and injury, it is still controversial to place a dollar value on QALDs. That reviewer also added that justification for choosing one approach over another on how to value quality of life losses should have been included in the report and discussed. 

Charge Question 4c: Regarding the economics of illness due to hepatitis A virus-infected produce: Do you have any additional information or estimates of any potential long term shifts in consumer demand away from fresh produce that result from produce-associated outbreaks?

In answer to this question regarding the economics of illness due to hepatitis A virus-infected produce, one reviewer recommended keeping abreast of recently published consumer surveys conducted by the Food Marketing Institute (FMI) and the International Food Information Council (IFIC) as well as other trade associations and consumer organizations. None of the reviewers noted any additional information or estimates for potential long term shifts in consumer demand away from fresh produce resulting from produce-associated outbreaks. 

Charge Question 4d: Regarding the economics of illness due to hepatitis A virus-infected produce: Are there additional estimates of post-outbreak adjustment costs to industry in addition to that documented on page 37 of the document (e.g., law suits and business closures)?

As with the previous question (4c), none of the reviewers cited additional estimates of post-outbreak costs to industry. One reviewer suggested that perhaps further information of post-outbreak costs could be identified from attorneys who are actively engaged in produce-associated outbreak litigation. 

Charge Question 5: Does the report identify all potential future controls and does it adequately describe the strengths and limitations of each? If not, please provide additional suggestions of information to include in the report.

In regard to the identification of all potential future controls, reviewers commented that the risk profile seemed comprehensive, but made several suggestions. For example, one reviewer commented that the automation of produce handling could serve as a control to reduce contamination risk. Another reviewer added that the cost/benefit analysis should have been performed for each of the proposed control methods, rather than just for the immunization of food workers. That reviewer was skeptical of the cost/benefit rewards of surveillance testing. Another reviewer emphasized efficacy of control measures such as UV disinfection of HAV in wash and flume water, as well as keeping children out of fields. The same reviewer also commented on the merit of proposed control measures in his specific-comment section. Another reviewer added that the section of the report on control options is difficult to understand, because so much of the important information is given in Table 11a rather than explained in the text.

That reviewer also suggested that the authors rank the control options based on some weighting scheme. Another reviewer commented more specifically about hepatitis A vaccination in response to charge question 3. That reviewer noted that the document does not specify which food handlers should be vaccinated. A reviewer pointed out that the vaccination of food handlers in most steps of the farm-to-fork continuum is not likely to reduce risk or be cost-effective, because introduction of hepatitis A probably occurs during the harvest step. 

Charge Question 6: Are the data gaps presented in Appendix 1 complete and adequately described? Are there additional research needs that should be considered in relation to this public health issue?

Most of the reviewers pointed out that the data gaps were actually presented in Appendix 2 of the report rather than Appendix 1, as the charge suggested. Addressing the research gaps, one reviewer commented that the lack of randomized intervention studies to test the efficacy of produce protection interventions is a glaring research methodology gap. A reviewer suggested that, while it was helpful to identify the data gaps, it would also be helpful to the reader to rank them in terms of importance. That reviewer went on to comment on each knowledge gap presented in Appendix 2. 

Charge Question 7: Do you have any additional comments that might improve the document?

Reviewers thoroughly noted edits and particular concerns in the specific comment sections of their individual responses. One reviewer noted that it might be helpful to use Google Docs to collaboratively create reports. Another reviewer noted that the report could be improved with more cost per case estimates, more thorough descriptions of the methodology used to define cost per case in the cited literature, and the development of an overall estimate of the annual economic burden of illness. A reviewer stated that the document would benefit greatly from a glossary of terms and the addition of an executive summary. 



FDA Response to Reviewer Comments

 

General Impressions
Reviewer Reviewer Comments FDA Response
Reviewer 1 I am a medical epidemiologist with training in internal medicine, infectious diseases, and epidemiology --- but not food safety. My comments reflect this perspective. Overall, this profile is outstanding. I learned a tremendous amount about how hepatitis A virus contaminates fresh and fresh-cut produce. This concise document summarizes the primary contamination mechanisms and control measures. I found especially useful the summary of hepatitis A virus survival in the environment and its resistance to heat and disinfection, including hand sanitizer. This risk profile also, indirectly, addresses an important risk of fecal contamination of harvested produce from human or animal sources (e.g., Salmonella, E. coli, etc.). Improvement of these control measures will reduce the risk from multiple pathogens. We appreciate the comment.
From my perspective, the greatest value of this risk profile is as an entry document that summarizes the state of knowledge and practices of hepatitis A food protection from "farm-to-fork." When necessary, readers can then pursue other resources in more depth. We agree.
Reviewer 2 Overall, this is a scholarly and well-organized document. It brings together a great body of information and arrays it in orderly fashion. We appreciate the comment.
There are points at which knowledge of the hepatitis A virus (HAV) might have been applied more inductively; on the other hand, some of the cited information about fresh and fresh-cut produce probably does not apply to HAV. Some of the cited information on fresh and fresh-cut produce does not directly apply to HAV, but helps us to better understand how potential control options impact fresh and fresh-cut produce, as well as microbes in general.  This information can help us determine the potential applicability of control options where direct research with HAV does not exist.
I have some concerns with where the cost-benefit principle is applied and where it is not. We agree with the reviewer that there is a lack of information in the document regarding cost/benefit analyses. We could not find any additional data in the literature on cost/benefit analysis of the other control measures provided in the document. Comparison of cost and benefit was included for immunization for consideration of the feasibility of immunizing the different population groups.
Reviewer 2 There are some rhetorical lapses in the writing, as well as grammatical failures, such as use of a singular verb with plural subjects (e.g., Charge Question 1: "Is the scope and purpose ...?"). The report has undergone editing.
Reviewer 3 The FDA report titled Risk Profile: Hepatitis A Virus Infection Associated with Consumption of Fresh and Fresh-Cut Produce is intended to address the potential for contamination of produce with the Hepatitis A Virus (HAV) and to identify potential measures, and the impact of those measures, to prevent contamination. Overall, the report is well written and addresses the objectives of the report outlined on page 83. We appreciate the comment.
Reviewer 4 The Risk Profile for Hepatitis A Virus Associated with Consumption of Fresh and Fresh Cut-Produce written by FDA scientists is complete, accurate and well written. All of the major topics are adequately addressed and the report represents an excellent review of the current literature and knowledge in the field. The document is well organized and the information is presented in a very logical sequence, providing the reader with an opportunity to fully understand the risk profile of this organism in these products. The data contained in the document is fully explained, well illustrated with key tables and figures and the use of photos gives the reader a better understanding of the issues presented. The scope and rationale, as well as the definition and explanation of a risk profile, sets the tone for the document. The report is written in a clear and concise manner, and provides sound conclusions. We appreciate the comment.
Reviewer 5 This multi-authored document seems to suffer from what appears to be lack of a unifying editorial direction provided by a senior editor or editorial team, and often comes across as inconsistent in both style and message. In many instances (or sections), an appropriate context is not provided for the problem or issue being addressed, the reader is easily confused by inconsistent terminology or definitions of terms (the document would benefit greatly from a glossary of terms), and often the available data or information is not synthesized into a unifying concept or conclusion. A person not familiar with the field would have a difficult time coming away with a reasonable understanding of the issues addressed by the document. The document has been edited to make it more consistent in style and message. A glossary of terms has been added.
An example of poor writing that occurs throughout the document, which raises concerns about the knowledge of the authors, is use of the terms "hepatitis A virus (HAV)", "hepatitis A", and "cases". It is well established that HAV refers to hepatitis A virus and NOT to the disease caused by infection with HAV (i.e., hepatitis A -- see further comments under the section - Specific Observations). The document has been edited to clarify these terms and make sure that they are used consistently as suggested by the reviewer.
Another observation is that the ‘horizon' for the risk profile has been limited to the year 2000. As stated by the authors, this was due to the hepatitis A outbreaks from green onions which made the headlines and resulted in a great deal of pressure on the FDA. However, similar events had occurred previously (e.g., the frozen strawberry hepatitis A outbreaks among school children). If the purpose of the document is to address the risk of HAV transmission from fresh produce, all available information should be used. While some of the ‘older' information has been incorporated into the document, it is not done in a consistent manner. In addition, a number of lessons were learned from outbreaks of hepatitis A and efforts to prevent transmission of HAV from filter feeding shellfish such as clams and oysters.

Although the citations for most of the outbreaks cited in Table 3 are from 2000, most of the outbreaks occurred before 2000. Much of the information was obtained from published, peer reviewed CDC outbreak review articles.

Regarding lessons learned from the shellfish industry, their control approach is akin to control measures that are also applicable, and are already being implemented by the produce industry.

Shellfish harvest area are classified generally on a two tier approach- 

Potential Pollution Source Identification/ Mitigation

Shellfish Program: Pollution sources are identified and evaluated prior to establishment of shellfish harvest areas- buffer areas established around pollution sources.  Attempts are made to mitigate pollution sources if  possible

Produce Industry:

They have implemented mitigation procedures to minimize the likelihood of viral contamination i.e. defecation in the fields, disallowing children from entering the fields. 

Sanitary quality of Water

Shellfish Program: Water quality is routinely monitored to establish sanitary quality.  If reestablished standard is exceeded (fecal coliform indicator) harvest area is closed until water quality improves.

Produce Industry:   In many instances the sanitary quality of the water used for irrigation is monitored for potability. If water quality is in violation irrigation from that water body is ceased.

Outbreak Investigations: In both instances  either when a particular field or shellfish growing area can be identified in causing illnesses all remaining materials from those harvest areas are recalled and destroyed.  In addition to the recall, follow-up investigations are conducted to identify potential sources of the contamination and to mitigate those sources accordingly.

However the big difference between the two Industries is traceback

In the shellfish industry, tags are attached to each bag of shellfish which identifies the harvest area location, date of harvest, and the licensed harvester.

This tag follows the shellfish to the retail level and must be retained by the retailer for a predetermined amount of time.  Also, shellfish lots can not be co-mingled.

A similar approach by the produce industry however has not been implemented.  Co- mingling is the norm- produce from multiple fields and even multiple growers is often mixed together. This comingling issue in addition to no tagging mechanism makes trace back during an outbreak investigation very difficult. The shellfish industry approach is now included in the document where the need for produce traceback is discussed.

Reviewer 5 Outbreaks of infectious disease transmitted by food are always a public health problem and concern. However, the document does not provide a balanced view of the magnitude of the problem (hepatitis A) as it relates to foodborne transmission compared to other sources in the United States. In many places throughout the document, especially in the economic analysis, the authors seem ascribe the overall burden of hepatitis A to foodborne transmission, when this is not the case. We have clarified, in the text, the differences between the magnitude of foodborne hepatitis A vs. hepatitis A from other sources.
Reviewer 5 A final observation is that the document seems to hold all fresh and fresh-cut produce at equal risk of becoming contaminated with HAV and transmitting disease. However, the available data would suggest that this is not the case. While all fresh foods have the potential to become contaminated with HAV (or other infectious agents for that matter), a clear statement of the relative risks of HAV transmission from foods based on available data would go a long way in determining where to focus risk mitigation efforts. I think the reader of the document gets lost in all the possible things that can be done to lower the risk of HAV transmission.

The source of infection is not determined in approximately 50% of HAV outbreaks (Mead et al., 1999).

We do not have sufficient information to compare the relative risk of contamination between the different produce items. As shown in Table 3, almost every type of produce has been associated with an outbreak.


 

Charge Question 1: Are the scope and purpose of the risk profile clearly identified?
If not, what additional information should be provided?
Reviewer Reviewer Comments FDA Response
Reviewer 1 Yes, the scope and purpose of the risk profile is clearly identified. We appreciate the comment.
Reviewer 2 The definition and explanation of risk profile (Appendix 1) are extremely helpful. The authors have done a reasonably good job of following their own guidelines. We appreciate the comment.
Reviewer 3 The scope and purpose of the report are clearly defined in Section II of the report. The authors follow a logical framework by first assessing the epidemiologic burden (Section III), followed by the economic burden (Section IV). The next three sections highlight the sources of contamination (Section V), followed by current (Section VI) and proposed (Section VII) approaches to preventing contamination. We appreciate the comment.
Reviewer 4 The scope and purpose of the risk profile were clearly identified, however; I believe that the description of a risk profile contained in Appendix 1 should be modified slightly and included in the introduction of the report. Then Appendix 1 can be eliminated. Since the concept of a risk profile is relatively new, we felt that we needed to provide a detailed description in the Appendix. We did provide a more abbreviated description of a risk profile in the introduction.
Reviewer 5 According to the Introduction and Appendix 1, "FDA's Center for Food Safety and Applied Nutrition (CFSAN) continues to seek new risk-based ways to describe food safety issues and assist in regulatory decision-making. One new approach is in the development and application of risk profiles. A risk profile, originally developed by CODEX (CAC, 2004), is a document that describes the current state of scientific knowledge of the food safety problem …" It describes the public health problem and current control strategies …" We agree.
Reviewer 5 The Appendix goes on to describe the elements of the "public health problem" as the: 1) description of the pathogen, 2) characteristics of the disease, and 3) economic impact. However, what is missing in the description of the public health problem is the burden of disease (hepatitis A) attributable to transmission through consumption of food, and in particular, fresh produce. Although the document addresses the cost-per-case of hepatitis A, it leaves the reader with the impression that all these costs are attributable to disease acquired through consumption of food. However, we know from epidemiologic studies performed prior to introduction of hepatitis A immunization in the US, that attributable fraction of cases of hepatitis A due to foodborne transmission was quite low. We have revised the text to include the percentage of illnesses attributed to foodborne HAV, in the public health section of the document.
Reviewer 5 Since this is a risk profile, one would expect that there would be an extensive discussion and assessment of the true risk of hepatitis A in the United States from both food consumption and fresh produce in particular. While it is recognized that the data are limited, the authors should have developed a framework for estimating this risk. For instance, risk of disease from an exposure is usually expressed as a rate per number of exposure events (disease per person-years of exposure; e.g., lung cancer per pack/years of smoking). This document is a risk profile, neither a quantitative nor a qualitative risk assessment, and expression of risk is beyond its scope. In Appendix 1 we describe the difference between risk profiles and risk assessments, as defined by FDA/CFSAN.
Reviewer 5 In the case of risk of acquiring hepatitis A from fresh produce, this could be expressed as risk of disease per average amount of high risk produce consumed by an adult per year -- e.g., risk of hepatitis A per amount [grams, kilos] of green onions consumed annually by the average adult. While the numbers for each of these parameters would have a lot of uncertainty (i.e., wide confidence intervals), such a model would provide the reader and decision makers with an estimate of the relative risk of getting hepatitis A from fresh produce consumed in the US. Also, having this type of model would allow the user to perform sensitivity analysis around each of the input parameters to determine, for instance, how many kilos of green onions a person would have to consume annually to likely contract hepatitis A from that source. At the time this risk profile was commissioned, insufficient information was available in the literature to develop the type of risk assessment model described. Since we started developing the risk profile, several new studies have been published with much information. However, as noted in the data gaps section, we still need additional information before we can conduct a fully quantitative risk assessment.
Reviewer 5 The absolute risk of a person in the US acquiring hepatitis A from fresh produce is quite (extremely) low, in-spite of the occasional newsworthy outbreak. I think a major issue that the authors must address is whether foodborne hepatitis A represents such a unique problem that it warrants a prevention strategy that is different from that other foodborne diseases transmitted by the fecal-oral route. While HAV has some characteristics make it somewhat different from other microbial agents transmitted by food, a convincing case has not been made that HAV/hepatitis A warrants a prevention strategy uniquely different from existing strategies used to prevent transmission of foodborne disease. The one area where hepatitis A differs from all other foodborne diseases is that it is vaccine preventable. However, the present risk profile does not address this potential prevention strategy in a way that is likely change the disease outcome.

In general, anything that prevents contamination by fecal-oral route is a control measure. An additional control measure, specifically for HAV discussed in the risk profile, is to modify certain agricultural production guidelines to include measures specifically intended for prevention or reduction of HAV-associated foodborne illnesses. One of these options specific to HAV is to exclude children from the food production area. The only known reservoir for HAV is humans. There is anecdotal evidence associating HAV-associated cases with the presence of children in the produce-growing fields during time of harvest (see figures 4a and 4b). Children can have asymptomatic or unsuspected HAV infection and, therefore, play an important role in HAV transmission and serve as a source of infection for others (Smith et al., 1997; Staes et al., 2000). Children and infants can shed HAV in their stools for longer periods than adults, up to several months after the onset of clinical illness (Rosenblum et al., 1991).

To address the issue that HAV is vaccine-preventable, we have incorporated additional references supporting this viewpoint. A previous study had demonstrated that vaccinating the entire population against HAV was not cost-effective (O'Connor et al., 1999). Another study suggested that vaccinating children would prevent hepatitis A in adults (Rein et al., 2007).


 

Charge Question 2: Considering the public health problem addressed, is the methodology used appropriate and is the report written in a transparent and clear manner? Does the report adequately address the questions and stated objectives?
If not, how might the report be revised?
Reviewer Reviewer Comments FDA Response
Reviewer 1 Essentially, this document is a systematic review and synthesis of the existing scientific and technical literature on this important topic. The synthesis and summary is primarily determined by the articles and reports that were selected for review. From this risk profile, the following are not clear:
  • How were the articles and reports selected? What databases were used?
  • What were the selection criteria (inclusion and exclusion criteria)?
  • How many articles and reports were selected for inclusion?
  • How many articles and reports were excluded from consideration?
  • How was quality of the article and reports evaluated?
  • Were unpublished studies elicited and how?
  • Were foreign language studies included? If yes, which languages?

This methodology is important in order for this risk profile to be transparent, and so that the methods can be replicated. This can be included as Technical Notes or in a separate document.

Before developing the risk profile, the risk profile team conducted a comprehensive search of all of the literature on foodborne viruses that was available at the time. Additional references were included as new data became available. We have attached an appendix in the revised document (Appendix 3) describing the criteria for selection of the articles used in the risk profile.
Reviewer 3 The public health methodology, relevancy of the scientific literature, and transparency of writing seem appropriate for Sections III, V and VI. Specific suggestions for improving sections IV and VII are provided below. The document has been edited.
Reviewer 4 I believe that the report adequately addresses the public health problem associated with HAV in fresh and fresh-cut produce. The document was written in a clear and transparent manner and effectively addresses the stated objectives. We appreciate the comment.
Reviewer 5 My concerns with the document are primarily those stated above. I really think this document presents a strong slant/bias towards foodborne hepatitis A being a major public health problem. However, the document does not make a convincing evidence-based argument that that this is the case and warrants a prevention strategy that differs from that used to prevent transmission of disease from other fresh produce. The document does not intend to portray foodborne hepatitis A as a major public health concern. As stated in the introduction, we acknowledge that with the advent of immunization in certain states, the incidence of HAV has decreased. Therefore, the 2000-2003 produce-associated outbreaks raised concern among food regulators, which led to the request for a risk profile on consumption of HAV and fresh and fresh-cut produce to help identify potential control options. We trust that the revised version clarifies this issue.
Another area of concern is that in the farm-to-fork continuum, food handlers/workers/preparers at each stage of the process are viewed as conferring the same degree of risk in introducing HAV onto the food. However, logic would indicate that is not the case, and this is supported by data from a number of the hepatitis A foodborne outbreaks. Also, food handlers/workers/preparers at each of the different stages in the continuum have different behavioral/sociological and oversight characteristics which come into play in devising and implementing prevention strategies and practices - a field worker harvesting strawberries is different from a salad chef in a restaurant. We are aware that behavior patterns of field workers are likely different from those of food preparers at retail. However, we feel that general hygienic principles and practices apply to all who handle food/produce at any time. Guidelines already in place for field or retail food workers can be modified as needed to relate to the appropriate type of food worker.
Reviewer 5 Following the early strawberry associated hepatitis A outbreak, I came to the realization that a field worker harvesting fresh fruit had to be considered a "food handler" with respect to their potential to introduce infections transmitted by the fecal-oral route. However, having worked both in growing fields and in restaurants during my younger days, I realize the challenges in trying to implement public health practices and minimize disease transmission. My concern is that the risk profile appears to consider each of the areas in the farm-to-fork continuum on almost equal footing with respect its level of risk in introducing and subsequently transmitting HAV. Similarly, proposed prevention practices seem to be considered in a similar manner for each of the areas on the farm-to-fork continuum, although there appears to be the recognition that introduction of these practices might be more difficult in some areas. I think that by assigning a relative importance/ranking/priority to the respective areas, based on available evidence, would go a long way toward implementing the proposed prevention activities. We agree that a ranking of the different control options would be helpful. However, we do not have sufficient information on all the potential options to add such a ranking.
One of the challenges in adoption and implementation of public health practices is to convince the participants of their importance. This needs to be done at a number of levels, including economic, behavioral, engineering and regulatory. The risk profile needs to better identify the key prevention practices that should be adopted to prevent HAV transmission from fresh produce, and make a convincing, evidence-based case for the adoption of these practices. The document also needs to make a more evident and convincing case that these practices need to be different from those already in place to prevent transmission of there infectious agents from fresh produce. We trust the revised version of the risk profile has addressed this issue and makes a more convincing case for identification and implementation of modified practices to prevent transmission of HAV to and from fresh produce.


 

Charge Question 3: Are there any additional scientific or technical studies available that were not considered that are relevant to the public health impact of hepatitis A virus in fresh and fresh-cut produce?
For example, has FDA identified all sources of contamination throughout the farm-to-fork continuum?
Are there any additional control measures or interventions that were not described?
Reviewer Reviewer Comments FDA Response
Reviewer 1 I did a PubMed search and identified additional articles not in this risk profile. I put these at the end of this document under the specific comment section. We thank the reviewer for the additional publications. We reviewed these articles and incorporated them in the document where appropriate.
Reviewer 2

I have cited some additional references and made some additional recommendations and suggestions among the Specific Observations.

From Specific Comment Section:

We thank the reviewer for the additional publications. We reviewed these articles and incorporated them in the document where appropriate.
Reviewer 2 15, 4-6: This is an important point. The HAV that has been inactivated by UV, chlorine, or heat will still give a positive RT-PCR unless the sample is pretreated to eliminate the inactivated virus (Nuanualsuwan, S., and Cliver, D. O. 2002. Pretreatment to avoid positive RT-PCR results with inactivated viruses. J. Virol. Methods 104:217B225. - listed as a reference, but evidently not cited in the text). We agree with the reviewer that this is an important point, and the information has been added to the document.
Reviewer 2 19, 5-21: The comparative resistance of HAV during pasteurization in water and milk has also been studied (Mariam, T. W., and Cliver, D. O. 2000. Small round coliphages as surrogates for human viruses in process assessment. Dairy Food Environ. Sanit. 20:684B689.) The following statement and reference have been included in the section on heat resistance: Mariam and Cliver (2000) have shown that inactivation of hepatitis A virus in milk is incomplete after 30 sec at 72°C.
Reviewer 2 34, 24 - 35, 11: It is surprising not to see "Consumption of raw or undercooked shellfish" as one of the individual bullets in this list. Shellfish were one of the first known sources of HAV infection and may still predominate (Roos, B. 1956. Hepatitepidemi, spridd genom ostron. Svenska Läkartidningen 53:989-1003.). The focus of this risk profile was HAV in produce, and there had been one HAV outbreak associated with shellfish since the 1990's in the U.S. (Shieh et al., 2007; 2008). Therefore, we did not include shellfish as a contributing source of HAV in the bullets. However, we agree that to provide a more comprehensive document, we should, and have now, included shellfish as a potential contributing source of HAV, especially in light of the occurrence of HAV outbreaks resulting from certain European nations.
Reviewer 2 71, 9: The Mariam and Cliver (2000) paper, cited on 76, 7, also contains data on the action of chlorine dioxide versus HAV. We were aware of this study, but we did not initially include these data in the risk profile, as ClO2 did not appear to be effective in reducing HAV levels under "real world" conditions. However, based on the reviewer's comment, we have now incorporated this information in the document.
Reviewer 3 The epidemiologic evidence provided seems appropriate, but this is not my area of expertise. Additional studies relevant to the economic impact of hepatitis A have been identified in the section below. The control measures described in section VII seem exhaustive and appropriate. We appreciate the comment and have reviewed the additional references provided and included them in the document, as appropriate.
Reviewer 4 I am not aware of any additional studies that could have been considered in developing this risk profile. The sources of contamination throughout the farm-to-fork continuum that are detailed in the report appear to be complete. The control measures and interventions identified in the document are also adequately described. We appreciate the comment.
Reviewer 5 Molecular epidemiology. While the report mentions the identification of HAV by nucleic amplification and minimally discusses the characterization of HAV strains by genotype, it makes no mention of the utility of further characterization of HAV strains by genetic sequencing. HAV has only a relatively small degree of genetic variation and does not accumulate the high frequency of mutations observed in other RNA viruses. As the report indicates, most human isolates belong to three genotypes. However, since HAV genotypes tend to occur across large geographic regions and the genetic variation within genotypes is large, genotyping of HAV is not useful for differentiation of what might be considered ‘strains'. However, HAV has sufficient genetic variation to allow for the differentiation of HAV strains or clades circulating in communities, different geographic regions, and among groups at risk of infection.

Most HAV strain identifications to date are based on comparative sequence analysis of a single region of the viral genome that demonstrates a higher frequency of nucleotide variation than the genome as a whole. While the VP1-2A junction represents just such a region (of popular choice), the 3C/D, VP1-2B and 5' NCR sequences are also being investigated as regions providing genotype/strain identification. As a result, current strain identification requires the inclusion of qualifiers, such as the primers used and the region amplified.

However, increasing the precision of strain identification requires multiple region amplifications and subsequent sequence analyses of these regions (if not the entire genome). Such a detailed discussion and elaboration of these methods of analysis are beyond the scope of the risk profile.

The reviewer is correct to point out that "strain" identification requires viral genome sequencing. Indeed, this sequence information is required (the document has been appropriately modified to obtain information about the genotype/subtype of a particular virus strain.

Reviewer 5 The use of ‘molecular epidemiology' has been particularly powerful in identifying the source of HAV contamination or transmission, particularly in identification and characterization of foodborne outbreaks. It has shown that what appeared to be a common source outbreak was comprised of several outbreaks, each with a different virus source. Molecular epidemiology has shown that multiple outbreaks in different locales were caused by the same HAV strain and transmitted by the same food source. Because of the availability of a database at CDC with the sequences of a large number of HAV isolates from the US and other countries, the geographic source of an HAV isolate from a foodborne outbreak can be approximated quickly. In fact, most of the hepatitis A outbreaks described in the risk profile were characterized using molecular epidemiologic methods. The risk profile discusses a molecular epidemiology approach to detection and identification of HAV in foods. Without detection and identification of the virus in the culprit food, the source contamination and the virus genotype/strain responsible cannot be confirmed. This is different from instances where "molecular epidemiology" really means (most probable) assignment of likely (contaminant) source, in conjunction with analysis/identification of virus in serum and/or stool. We have, however, included a section acknowledging the power of molecular epidemiology for linking outbreaks and for its potential to determine the role of food and water in contributing to the burden of HAV illness.
Reviewer 5 With the dramatic decline in cases of hepatitis A in the US primarily due to hepatitis A immunization, molecular epidemiologic methods should become an even more tool in determining whether small, sporadic disease outbreaks are related, possibly through a common source such as food, including fresh produce. The comment has been incorporated in the document.
Reviewer 5 So, it is odd that these methods were not fully described (I don't think I am being ‘CDC-centric' in these comment). Some pertinent references include: 1) Nainan OV, et al. Diagnosis of hepatitis A virus infection: a molecular approach. Clin Micro Reviews 2006, 19: 63-79; 2) Nainan OV, et al. Hepatitis A molecular epidemiology in the United States 1996-1997: sources of infection and implications for vaccination policy. J Infect Dis 2005; 191:957-963. The methods described by the reviewer have been included in the final document, to make it more comprehensive.
Reviewer 5 Hepatitis A vaccination. Foodborne hepatitis A is unique in that it is a potentially vaccine preventable disease. Several options for the use of vaccine in preventing foodborne hepatitis A are discussed in the risk profile and include: 1) the effect of current recommendations for widespread vaccination of children to lower the overall susceptibility of the US population to hepatitis A from all sources, including food, and 2) the potential use of vaccine among food handlers. In the document, we discussed the application of HAV vaccination of food workers, children, and the entire population.
Reviewer 5 Although the risk profile recommends vaccination of food handlers, it does not specify which food handlers in the farm-to-fork continuum should be vaccinated (as mentioned above, one of the weakness of not looking at each of the steps in the product pathway). Hepatitis A vaccination of food handlers in the food service (e.g., restaurants) component of the product pathway has not been shown to be cost-effective or likely to result in lowering the risk of foodborne hepatitis A. We agree that published studies have shown the non benefit of vaccinating food workers at retail; however, as we stated in the document, none of these studies considered the potential loss of business that a specific food service company would face if a hepatitis A outbreak were attributed to that particular company. This loss would include not only the company's reputation, but also cost of discarded food and the potential loss of millions of dollars due to loss of business and to lawsuits that may transpire as a result of the outbreak. Moreover, by accounting for the loss in business, tort liability, reduced food service dining, and reduced fresh produce consumption that may occur during and following a hepatitis A outbreak could increase estimates of expected benefits from this intervention, as suggested from the analysis by Meltzer et al. (2001). Also, to our knowledge, there have been no cost/benefit analyses for vaccinating field workers.
Reviewer 5 Also, vaccination of food handlers in most of the component steps in the farm-to-fork continuum (see Figure 2 Section V) is not likely to be cost-effective or result in lowering the risk of foodborne hepatitis A since introduction of HAV probably does not occur, except during the harvest step. We disagree with the reviewer regarding the role of food workers at retail in transmission of HAV. At least one publication indicates that outbreaks due to HAV-infected food handlers at retail/catering, which can transmit HAV to dozens or even hundreds of persons, cause a substantial economic burden to public health (Dalton et al., 1996). Additional articles, such as the Hedberg et al. (2006), showed that food workers at retail contributed significantly to foodborne outbreaks. According to Heymann et al. (2004), the source of most reported foodborne HAV, however, has been HAV-infected food workers present at retail (such as in a restaurant) or who prepare food for social events. There have also been several HAV-associated outbreaks in the past couple of years implicating food preparers at retail. See Appendix 4, which shows that 18/26 (70%) exposures/ outbreaks/ cases reported by FSNET and PROMED from 2005 to 2007 were related to food preparers.
Reviewer 5 So, thinking about the workers in the harvest step. Most of the HAV strains associated with hepatitis A outbreaks associated with produce have come from Mexico and it is most likely that the contamination occurred during the harvest step. One area that was not discussed in the risk profile is hepatitis A vaccination in countries with high endemic rates of HAV infection (e.g., Mexico and most Latin American countries) - countries which are a major source of fresh produce in the US. If one ‘connects the epidemiologic dots' related to some of the fresh produce hepatitis A outbreaks, several things seem apparent. The growing source for the food (e.g., onions, strawberries) was a country with a high endemic rate of HAV infection. This means that >80% of teenagers or young adults have already been infected with HAV and that most infection occurs among children, especially young children who are at high likelihood of having an asymptomatic infection. As they say, a picture is worth a thousand words (and also a small amount of feces with HAV goes a long way). Just look at the pictures in the risk profile of young children sitting on top of piles of green onions and it doesn't take much to figure out how the HAV contamination was likely to have occurred. We thank the reviewer for the comment.
Reviewer 5 One possible control measure is routine early childhood hepatitis A vaccination in countries with high endemic rates of HAV infection and a large component of their economy that is dependent on export of fresh produce. Yes, the disease burden and economic analyses need to be performed. However, in a recent symposium on the global control of hepatitis A in Miami in 2007, several Latin American countries are considering this prevention strategy for a number of reasons. In addition, this prevention strategy was under discussion with health authorities in Mexico during the last several years. No response is needed.
Reviewer 5 While it would be difficult to require hepatitis A immunization in another country to protect the US food supply, these macroeconomic forces have impacted other national immunization strategies. For instance, in the Philippines, overseas workers contribute to a substantial proportion of its GNP. Recently, the country mandated universal infant hepatitis B immunization to prevent chronic hepatitis B virus (HBV) infection because a high proportion of their adult overseas workers were being rejected by the host country because of chronic HBV infection, which resulted in a substantial loss to the GNP of the Philippines. The information has been incorporated in the document.


 

Charge Question 4(a): Regarding the economics of illness due to hepatitis A virus-infected produce:
Are the estimated costs for HAV illnesses associated with morbidity and death reasonable?
If not, are there available published reports with more recent cost estimates?
Reviewer Reviewer Comments FDA Response
Reviewer 1

I did not see this article in your bibliography:

Jacobs RJ, Moleski RJ, Meyerhoff AS. Valuation of symptomatic hepatitis a in adults: estimates based on time trade-off and willingness-to-pay measurement. Pharmacoeconomics. 2002;20(11):739-47. PMID: 12201793

We thank the reviewer for the reference. However, the focus of the risk profile is to identify potential control options to prevent or reduce hepatitis A associated with fresh and fresh-cut produce. Nonetheless, we have incorporated that information in the document.
Reviewer 2 This is not my field. Overall, I find the figures credible; however, I am aware that the government has recently devalued human life, which suggests that some such "magic numbers" are entirely arbitrary. No response is needed.
Reviewer 3 The economic impact section highlights the cost per case of HAV. Unlike the epidemiologic impact section (Section III), the authors of this report do not attempt to estimate aggregate economic burden within a given time period (annual, for example). Such an estimate would strengthen this section and the report overall. Text describing the economic burden has been incorporated in the document.
Reviewer 3 The majority of the data provided in the original section on "Medical and Hospital Costs" were derived from an unpublished study (Lasher et al., 2007). Use of an unpublished study was a major cause of concern particularly because the details on the methodology used in that study were lacking. The revised version of the section on "Medical and Hospital Costs" excludes data from the Lasher 2007 paper and adds data from a different paper that has been accepted for publication (Scharff et al.). Unfortunately, the details on the methodology used in the Scharff paper have also been omitted from this report, which makes interpretation of the data challenging. However, a review of the Scharff paper reveals that like the Lasher 2007 paper, the authors use questionable methods for estimating costs, at least in terms of placing a dollar value on quality-adjusted life days (QALDs) lost [see 4b below for more comments on this topic]. We acknowledge that the economic impact section is not detailed. However, the focus of the risk profile is to identify those elements of the hazard or risk that are relevant to the food safety problem and identify potential control options to prevent or reduce hepatitis associated with fresh and fresh-cut produce. Therefore the risk profile team does not think it necessary to describe the methodology used by Scharff et al., which now has been published. The FDA accepts the use of QALDs to help calculate pain and suffering caused by foodborne illnesses.
Reviewer 3 Reporting on cost data from the literature could be improved. For example, if the authors report cost estimates from other sources, at a minimum the year in which those estimates are reported should be provided, and ideally, all costs should be adjusted to the same base year. This report does not specify a base year at all, and if adjustments were made, this is not specified nor is the methodology for adjustment defined. Finally, a review of some of the literature cited suggests that there were other data that would have been relevant to cite in this report. For example, the Dalton et al. paper also included a cost estimate for losses in productivity losses. The Berge et al. paper also included cost per case estimates. And a quick search of the literature found a number of other papers that might be applicable to this report, including Rein et al., 2007, Pediatrics, and Jacobs et al., 2000 Arch Pediatr Adolesc Med. Additional information provided by the references mentioned by the reviewer has been incorporated in the document.
Reviewer 4 The estimated costs of HAV associated illnesses appear to be reasonable given the scarcity of data in this area. We agree.
Reviewer 5 Having been involved in several of these studies, the costs seemed reasonable. It would be hard to obtain more contemporary data since the disease rate has dropped so dramatically and recent estimates indicate that fulminant hepatitis A has also dropped. We agree.


 

Charge Question 4(b): Regarding the economics of illness due to hepatitis A virus-infected produce: Is the use of QALDs (Quality Adjusted Life Days) for describing loss of well-being due to HAV infection appropriate?
Reviewer Reviewer Comments FDA Response
Reviewer 1 Yes. Because the hepatitis A case fatality risk is so low, then quality-adjusted life metrics are appropriate. We agree.
Reviewer 2 Again, this is not my field. However, I have been exposed to the QALD system before and find its application here credible. We appreciate the comment.
Reviewer 3 While the use of quality-adjusted life days (QALDs) is an acceptable way to measure epidemiologic burden of diseases and injury, it is still controversial to place a dollar value on QALDs. Fortunately Sharff et al (but not this report) recognize this concern and offer several variations on how to value quality of life losses in a sensitivity analysis. Scharff et al use the Viscusi and Aldy VSL estimates (based on labor market rates for riskier jobs) to derive QALDs lost in their main analysis. Then they conduct sensitivity analysis on these estimates also using the human capital approach coupled with a QALD derived from a $100,000/QALY threshold (which is really a hypothetical, not scientifically derived, estimate itself). Although either of these approaches might be acceptable, justification should be provided for choosing one approach over another, and should be explained in this report. For example, the IOM has recently suggested that VSL estimates as derived from revealed or stated preference methods (as used in Viscusi and Aldy) should not be used to value quality-adjusted life years, and consequently QALDs (Miller et al., Oxford University Press, 2006). At a minimum, this report should have discussed the methodology used to derive the QALD estimates, and at a maximum this report should have provided some variability in the estimates based on the sensitivity analyses of the original papers. Details on methodology are not similarly lacking in the epidemiologic sections, and this section should be no exception. We used the information that was available at the time. The focus of the risk profile is to identify potential control options to prevent or reduce hepatitis A associated with fresh and fresh-cut produce. Therefore, we did not think it necessary to describe the methodology used to derive QALDs or to go into depth on these estimates and sensitivity analyses. The objectives of the risk profile are to provide a literature review of the food safety issues and options for reducing illnesses due to consumption of fresh and fresh-cut produce contaminated with HAV; i.e., identify elements of the hazard or risk that are relevant to the food safety problem. Description of the methodology used to derive QALDs is beyond the scope of the risk profile.
Reviewer 4 Although it is very difficult to assess the loss of well-being due to HAV infections, the QALDs concept provides a reasonable and appropriate approach to this important issue. We appreciate the comment.
Reviewer 5 I cannot comment on this measure. I am only familiar with QUALYs and DALYs and their substantiation. No response is needed.


 

Charge Question 4(c): Regarding the economics of illness due to hepatitis A virus-infected produce:
Do you have any additional information or estimates of any potential long term shifts in consumer demand away from
fresh produce that result from produce-associated outbreaks?
Reviewer Reviewer Comments FDA Response
Reviewer 1 I am not aware of any studies that associate long term fresh produce avoidance and produce-associated outbreaks. If anything, the increasing year-round demand for fresh produce has required retailers to import fresh produce from farms that may not practice good agricultural and management practices. We appreciate the comment.
Reviewer 2 No. No response is needed.
Reviewer 3 None. No response is needed.
Reviewer 4 Keeping abreast of recently published consumer surveys conducted by the Food Marketing Institute (FMI) and the International Food Information Council (IFIC) as well as other industry trade associations and consumer organizations, may provide some useful insights into whether there is a long term shift in consumer demand away from fresh produce that is associated with outbreaks. Also, seeking information from the produce industry and food retailers about the sales of fresh and fresh-cut produce associated with outbreaks over specific time periods, can provide some useful consumption trend patterns. We appreciate the information.
Reviewer 5 I have no experience in this area. No response is needed.


 

Charge Question 4(d): Regarding the economics of illness due to hepatitis A virus-infected produce:
Are there additional estimates of post-outbreak adjustment costs to industry in addition to that documented on page 37 of the document (e.g., law suits and business closures)?
Reviewer Reviewer Comments FDA Response
Reviewer 1 Not my area of expertise. No response is needed.
Reviewer 2 I am not knowledgeable on this subject. No response is needed.
Reviewer 3 None. No response is needed.
Reviewer 4 There may be some additional estimates of post-outbreak costs to the produce industry (costs related to outbreak investigations, product recall and testing, product losses, lost business, law suits and business closures, etc.) that may be obtained from attorneys who are actively engaged in produce associated outbreak litigation. We appreciate the information.
Reviewer 5 I have no information on this area. No response is needed.


 

Charge Question 5: Does the report identify all potential future controls and does it adequately describe the strengths and limitations of each? If not, please provide additional suggestions of information to include in the report.
Reviewer Reviewer Comments FDA Response
Reviewer 1 To the best of my knowledge, this risk profile is very comprehensive. However, I did notice that automation of produce handling can reduce human handling of produce and thereby decrease contamination risk from human. (See Richards GP. Enteric virus contamination of foods through industrial practices: a primer on intervention strategies. J Ind Microbiol Biotechnol. 2001 Aug;27(2):117-25. Review. PMID: 11641770) This article may suggest interventions not mentioned in this report. From my review I did not see more control measures. We agree that automation of produce handling could significantly reduce the risk of contamination posed by human handling, and have included this as a possible prevention strategy. However, we must emphasize that the automated equipment also needs to be frequently cleaned and disinfected to prevent cross-contamination between the various lots of produce going through the machine.
Reviewer 2 I have made several suggestions regarding controls in the Specific Observations section. I am concerned that cost-benefit analysis is applied only to the option of immunizing food workers. It seems to me that cost-benefit analysis would be equally applicable to any other proposed control measure. I am especially skeptical of the cost-benefit rewards of surveillance testing of produce for HAV contamination. As a general comment, we do not have sufficient information to provide any evidence of exactly where contamination occurs.
Reviewer 2

From Specific Comment Section:

24, 12-15: This sentence makes a very important point. From my own experience working with HAV in our laboratory, every person from a developing country in our group was already immune, whereas the rest of us were required to be vaccinated. Where produce imported from developing countries is concerned, there is very little likelihood of hands-on contamination by infected workers. If contamination does occur, it is more likely to come from fecally contaminated water or from the hands of workers who have been caring for infants that accompanied them in the field. Although the latter possibility is mentioned further on in the draft, it deserves more emphasis than it receives.

We have tried to emphasize contamination of produce by children in the growing fields in the revised version of the document.
Reviewer 2 50, 2 & 3: It should be noted here that some classes of produce (e.g., strawberries and raspberries) are not amenable to washing at harvest, or even before sale, as they spoil rapidly after washing. We agree with the reviewer and have included wording in the document to describe this.
Reviewer 2 52, 16-19: This sentence asserts that HAV attaches to food contact surfaces, but does not say that the virus can readily come off again. Since people don't eat food contact surfaces, this seems to suggest a way of decontaminating food. This section is part of the chapter on routes of contamination. It continues further to say that the virus also survives well on these surfaces (Mbithi et al., 1990; 1991; 1992), making these surfaces potential sources for HAV transmission. Although there are no published data demonstrating contaminated utensils as a source of HAV-related foodborne illness outbreaks, several previous studies have implicated contaminated equipment and utensils as an important contributing factor in bacterial foodborne illness outbreaks (Bryan, 1978; Bryan, 1988; Weingold et al., 1994).
Reviewer 2 66, 15: It is unfortunate that use of potable water for irrigation is mentioned "in the same breath" with use of potable water for washing and misting. The idea that water might need to be treated to EPA drinking-water standards to permit its use for irrigation seems quite unrealistic. We are aware that it might not be economically possible to treat irrigation water to meet EPA drinking water standards; however, we believe there is a need for standardization of irrigation water.
Reviewer 2 69, 13 & 14: The prospect that gloves may be used as an alternative to hand washing is not considered here. It happens. The use of gloves has been mentioned as an alternative to hand washing. The team also notes that putting on gloves with unwashed hands could transfer the virus to the gloves.
Reviewer 2 70, 2-10: This is an extremely important section. For reasons stated earlier in this review, children are a much more likely source of HAV in developing countries than are the field workers themselves. The importance of children in the field in HAV transmission has been emphasized in the document.
Reviewer 2 73, 15-23: UV can also be used to disinfect HAV in water (such as wash water and flume water), if turbidity is low enough (Nuanualsuwan, S., Mariam, T., Himathongkham, S., and Cliver, D. O. 2002. Ultraviolet inactivation of feline calicivirus, human enteric viruses, and coliphages. Photochem. Photobiol. 76:406B410.). We thank the reviewer for this additional information; it has been incorporated in the document.
Reviewer 2 77, 10-15: There is no doubt of the safety and efficiency of the licensed HAV vaccines. When I and members of my laboratory group were immunized, the cost per person was substantially greater than the $20 mentioned (78, 20). The most important aspect -- not mentioned in the draft -- is that two doses are required, at least 6 months apart. In the high-transiency portions of the food system, the average worker's stay on one job is substantially less than 6 months. Thus, an employer who is willing to pay for immunization may well find that the employee is not available for the second dose. In the case of a worker who is reluctant to leave a paper trail, it may be difficult to document completion of the immunization series. We have added to the document that the vaccine is administered in two doses and that the high transience of food workers needs to be considered in this regard.
Reviewer 2 78, 14 - 79, 18: This is a scholarly analysis of the cost-benefit relationship in immunizing food workers. It seems a proper lead-up to the argument (beginning at 79, 19) that the US would do well to immunize all children against HAV, etc. The cost-benefit analysis might have included the option of hiring food workers from developing countries, who are almost certainly immune on arrival. Even if immunization were required, the cost of a test for pre-existing immunity is substantially less than the cost of vaccination. It might also have been useful to compare the cost of immunization to the costs of other, HAV-specific control measures. The option to test for HAV in migrant workers has been included in the document as a control option. As mentioned earlier, cost/benefit analysis of the different control options is beyond the scope of this risk profile. This was the only option for which data were available, and it is also specific to eating facilities, for comparing whether workers should be immunized. It is not meant to be compared with other control options.
Reviewer 3 91, 4 & 5: Keeping children out of fields may indeed be difficult in developing countries; but this may well be the most effective measure available, aside from avoiding fecally contaminated water. We realize the difficulty associated with implementing this option.
Reviewer 4 92, 8-10: Guidance on use of effective disinfectants will be most useful once the disinfectants have been identified and optimal means of applying them have been established. We agree with the reviewer and had already included this in the document sent for review as a data gap requiring additional research.
Reviewer 5 93, 11-15: As suggested earlier, the cost of hiring food workers who are already immune, by virtue of growing up in a developing country, should be compared. As mentioned earlier, cost/benefit analysis of the different control options is beyond the scope of this risk profile. This was the only option for which data were available, and it is also specific to eating facilities, for comparing whether workers should be immunized.
Reviewer 5 101, 4: Doses of permitted electron-beam radiation are measured in the same units (kGy) as doses of gamma rays. At this point, there seems to be no reason to believe that electron beams will be more antiviral than gamma rays. While we agree with the reviewer, this application remains a data gap until its efficacy is tested.
Reviewer 5 The section on control options is quite difficult to comprehend because so much of the important information is provided in Table 11a rather than in the text. For example, Table 11a provides critical information on three criteria by which to evaluate the control options: breadth and depth of coverage, factors influencing costs, and factors influencing benefits. The next obvious step in the policy analysis would then be to rank order the control options (at least within their respective sub-categories) based on some weighting scheme of these criteria. It seems like this important next step in the analysis was missed. The intention of Table 11a was intended to simplify the description of the control options.
Reviewer 5 I believe that the report identifies all potential future controls and adequately describes the strengths and limitations of each of those listed. We appreciate the comment.


 

Charge Question 6: Are the data gaps presented in Appendix 1 complete and adequately described?
Are there additional research needs that should be considered in relation to this public health issue?
Reviewer Reviewer Comments FDA Response
Reviewer 1 Yes. In general, there are almost no randomized intervention studies to test the efficacy of produce protection interventions. This is a glaring research methodology gap. We agree with the reviewer, and this has been listed as a data gap.
Reviewer 2 "Knowledge" gaps are presented in Appendix 2, rather than 1. I have critiqued these in the Specific Observations section. The report was edited to correct the typographical error in the Appendix designation.
Reviewer 2

From Specific Comment Section: 

41, "11": This section has surely identified more substantive Knowledge Gaps than "potential long-term shifts in consumer demand."

Additional knowledge gaps for this section have been added.
Reviewer 3 Yes the gaps are adequately defined. The authors of this report did a good job of identifying the additional research needs that should be considered in relation to this public health issue. We appreciate the comment.
Reviewer 4 The knowledge gaps identified in the risk profile and listed in Appendix 2 are complete and adequately described. I cannot think of any that were not listed in the report. We appreciate the comment and have corrected the typographical error.
Reviewer 5 While it is useful to identify data gaps, there should be some attempt to rank them in terms of importance to the question at hand, namely prevention of hepatitis A associated with consumption of fresh produce. My comments are given for each of the ‘gaps' listed in Appendix 2: Ranking the data gaps is beyond the scope of this document and would require development of criteria for ranking.
Reviewer 5 1. No ideal method for rapid detection of HAV in various types of produce. Adequate high throughput molecular methods exist to detect HAV, which can be adapted to various types of produce and can be used to investigate outbreaks or used for research purposes. Given the very low risk of infection, it is highly unlikely that HAV detection methods would be employed for screening food prior release into the market. FDA and others spent a great deal of effort to develop HAV ‘screening methods' for shellfish which were not widely implemented as the perceived risk became even smaller. We did not intend to recommend screening of produce for HAV prior to sale.
Reviewer 5 2. Frequency and levels of HAV in produce at all stages in the farm-to-fork continuum are unknown. This is true. However, in the face of apparent low rates of disease it is unclear why such expensive and complicated studies should be undertaken. Although HAV infrequently causes fulminant disease (i.e., less than 1 to 1.5% of patients experience hepatic involvement), fulminant hepatitis involves massive hepatic necrosis and has a high case fatality rate (>70 to 80%) (CDC, 2007a).
Reviewer 5 3. Most known disinfectants are not highly active against HAV. We agree with the reviewer, and this has been listed as a data gap.
Reviewer 5 4. The actual infectious dose of HAV is unknown; available dose-response data are not sufficient to understand the relationship between exposure to HAV from consumption of produce and illness in different sub-populations. These data could be derived from the inocula used by the laboratories (NIH, CDC) that performed the original infectious dose studies in chimpanzees and marmosets. Any subsequent studies would have to be performed in these animal models. However, from these original animal studies there did not appear to be a relationship between disease (as measured by liver enzyme elevations or histologic changes) and HAV dose. Extrapolating from animal to human dose-response is beyond the scope of this risk profile. Clinical studies with human volunteers and vaccine testing have been conducted in the past, from which dose-response can be retrospectively determined, but we do not as yet have that information available to us.
Reviewer 5 5. HAV illnesses are highly underreported. Yes, hepatitis A is underreported. However, there have been excellent estimates of the degree of underreporting based on models derived from NHANES serologic data. Given the declining incidence of hepatitis A due to immunization and the inability to differentiate antibody from HAV infection and immunization, future studies of disease underreporting could probably not be performed. We agree that HAV is decreasing in endemic areas, due to immunization; however, given the number of outbreaks and cases not yet published in peer-reviewed articles (27 - see appendix 4), and considering underreporting, we still think this remains a data gap.
Reviewer 5 The public health priority (and probably a gap) is the rapid reporting and investigation of hepatitis A cases to determine the source, particularly if it is foodborne. However, this would require a molecular epidemiologic investigation which cannot be performed by most local health departments. We agree; hence our need for better methodology, because this, too, contributes to underreporting.
Reviewer 5 6. The transmission routes for HAV infection are not always identified (unknown for 48% infections). There are only two routes of HAV infection - percutaneous (bloodborne- very rare) and oral. What I think the authors mean is that the source of the infection among persons with hepatitis A is not always identified. However, there have been population based surveillance studies (i.e., CDC's Sentinel Countries Study of Acute Viral Hepatitis) that have shown that only 20% of persons with hepatitis A do not have a known source of their infection.

We agree that the routes of infection include fecal-oral and subcutaneous. There was an error in terminology in the document. This has been changed to read "source of HAV infections is not always identified"…..

We did indeed mean that the source of hepatitis A infection is not always identified. The methodology is inadequate, and many clinical laboratories might not have the facilities to perform detection assays. We do not have the information to determine actual source of infection, unless, as in the household example, several areas within the house, including food and water and food contact surfaces, are tested to determine the source of infection.

We thank the reviewer for alerting us to the CDC's Sentinel Countries Study of Acute Viral Hepatitis. After perusing both the 2007 (2005 surveillance study) and 2008 (2006 surveillance study) publications, the latter one already cited in the risk profile, we realized that we had cited an older paper (Fiore, 2004) that stated a 48% unknown transmission route. The two CDC documents by Wasley et al. showed a 57.6% and 65.2% respectively, for 2007 and 2008 publications. We have since revised the document to reflect this change.

Reviewer 5

In the description of the epidemiology of hepatitis A in the risk profile, the authors really seemed to miss the point that this disease is primarily spread by person-to-person contact through the fecal-oral route. For instance, studies have shown that in households of adults with hepatitis A and no identifiable source of infection, a high proportion of their children were IgM anti-HAV positive, which indicated that they most likely acquired infection from an asymptomatic child.

Given the low disease incidence in the US because of hepatitis A immunization, it would be extremely difficult to ‘fill this gap' and it is not clear what purpose it would serve. What seems more important from the perspective of the FDA is to identify foodborne outbreaks of hepatitis A, which would be hard enough.

We have revised the document to include person-to-person spread within a household.
Reviewer 5 7. Survival of HAV in produce under different storage conditions, e.g., temperature, humidity, pH are not well understood. I think there are enough data to know that HAV is a relatively resistant/hardy virus. It might be helpful in framing prevention recommendations to know that exact combinations of storage conditions that minimize infectivity. These studies are difficult to perform since ascertaining infectivity with wild-type virus in animals is difficult because of the limited host range of the HAV (humans, chimps, macaques, marmosets). If cell-culture plaque forming HAV is used, the question always arises whether this represents the wild-type virus, since all cell culture adapted HAVs are attenuated. So, even though the studies could be performed in controlled conditions, the results would only provide estimated of survival and infectivity. The scientific literature demonstrates that there are differences of HAV survival on various types of produce. For example, Croci et al. (2002) showed that HAV survives differently on different produce types held under refrigeration. In addition, Bidawid et al. (2001) showed that HAV on lettuce is relatively resistant in inactivation under low CO2 atmospheres in which lettuce is normally stored for distribution.
Reviewer 5 8. The effect of potential treatments and techniques that could be applied to irrigation water to inactivate HAV, such as activated carbon, reverse osmosis, and membrane filtration, and diethylenetriamine, is not known. While this gap seems important, it is still not know whether fresh produce become contaminated with HAV through their vascular systems. However, it is know that other disease agents gain access to fresh produce through the plant's vascular system, so this might be of greater importance in prevention of all diseases transmitted via fresh produce. If that is the case, then HAV should be used as one of the ‘indicator' organisms in the evaluation of these methods. Also, treatment of irrigation water might be costly, which would make an economic assessment part of the evaluation of the various methods. Irrigation methods such as drip or spray, with contaminated water could contaminate the surface of the crops (Richards, 2001).
Reviewer 5 9. Although suspected, the role of contaminated utensils as a source of a HAV foodborne illness outbreak; i.e., transfer rates of HAV from utensils to food has not been determined. Personally, I don't see this as an important "gap" based on the presumptive sources of HAV contamination of the various types of fresh produce involved in outbreaks. Although not demonstrated in HAV-associated outbreaks, several investigations of other foodborne outbreaks have associated the outbreaks with cross-contamination of contaminated knives (Jacquet et al., 1995; Dorozynski, 2000).
Reviewer 5 10. The role of consumer in HAV-infected produce-associated illnesses has not been determined. Actual consumer behavior such as hand washing and cleaning of utensils and food contact surfaces is unknown, as is transfer rates of HAV from hands to food using gloves. Honestly, I don't see this as a gap. Have consumers really been the source of foodborne outbreaks of hepatitis A? I am discounting contamination that may occur in a household where there is a case of hepatitis A, or where an infected person contaminates food that is consumed by non-household members. While these are a number of these events in the ‘older' literature, these have become quite rare to non-existent. Although there have been no reports on consumers being the source of HAV outbreaks, there have been reports on consumer foodborne illness due to other microbial hazards. (Bloomfield et al., 2007; 2008; WHO/Euro, 2003).
Reviewer 5 11. Effect of hand washing on contamination of produce and subsequent illness with HAV specifically, has not been evaluated. There has been one published study on the transfer rates of HAV from hands to produce. Probably don't need additional studies. The study by Bidawid et al. (2000c) demonstrated the rate of HAV transfer from finger pads to lettuce, and demonstrated the effectiveness of hand washing and the use of disinfectants to clean the hands. After some discussion, we agreed that this may not necessarily be a data gap, and deleted this bullet from the data gap list.
Reviewer 5 12.    The effect of mucosal immunity on HAV is not as yet known. It is therefore unknown whether an immunized person, either by natural infection or by passive immunization, can be an asymptomatic carrier; i.e., it is not known whether HAV can replicate in the intestinal cells of an immunized person. There is really no evidence for the "immunized asymptomatic carrier" both in human and animal studies. It is well established that hepatitis A immunization prevents both disease and infection in almost all instances, even with post-exposure immunization. Also, I am not sure how one would conduct such experiments other than by human challenge studies. Such studies have been conducted in the chimpanzee and marmoset models of HAV infection. We have reviewed some of the studies on marmoset monkeys, which focused mainly on liver damage caused by HAV, but could not find evidence of whether or not the immunized animals were asymptomatic carriers.


 

Charge Question 7: Do you have any additional comments that might improve the document?
Reviewer Reviewer Comments FDA Response
Reviewer 1 This document needs the following:
  • Careful review for grammar and typos. I found a few, but I am sure there are more (there always are!) I also found misspelled words that were not picked up by your spellcheck.
  • Do not use MS Excel graphics---they are of very poor quality. You can produce high quality, publication ready graphics using freely available, open source software: we use R (http://www.r-project.org)
  • XML-structured documents is the industry standard for scientific or technical documents. You can do this in MS Word, but OpenOffice.org is much better. There are commercial products that accomplish the same, but the open source programs are more than sufficient.

An aside: We use Google Docs to collaboratively create reports. Then we use LaTeX to professionally typeset our report. LaTeX is open source and is available for all platforms. Here is a recent report I produced for San Francisco Health Department: http://www.sfdph.org/dph/files/reports/StudiesData/CHE_Rpt07242007C.pdf

The report has been edited.

We thank the reviewer for the advice concerning the graphics. However, we have not found a problem with using EXCEL graphics.

Reviewer 3 Overall the report is well written and meets its stated objectives. The report could be slightly improved with more cost per case estimates provided, more description of the methodology that was used to define cost per case in the literature cited, and the development of an overall estimate of the annual economic burden of illness. The identification of potential control options could be improved if the text provided a discussion of pros and cons of each option based on the criteria identified. The focus of this risk profile is not cost/benefit analysis. We have included in our research needs the acquisition of an overall estimate of burden of illness of HAV from fresh and fresh-cut produce.
Reviewer 4 I think that the report is well organized, clearly and concisely written and meets the stated objectives. It provides a very detailed risk profile for the HAV infection associated with consumption of fresh and fresh-cut produce and will be a useful resource for those interested in this topic. We appreciate the comment.
Reviewer 5 As I mentioned earlier, the document would benefit greatly from a glossary of terms. Also, I believe that such a long document would benefit from an executive summary. We agree with the reviewer that an executive summary would be helpful. Therefore, one has been included in the document.


 

Specific Observations
Reviewer Reviewer Comments FDA Response
Reviewer 1

From my PubMed on 8/18/2008, 1:00am, search using "Hepatitis A virus [mh] AND (food contamination [mh] OR food handling [mh])" is displayed below. I excluded articles that are already cited in the risk profile and articles about seafood.

Casas N, Amarita F, de Marañón IM. Evaluation of an extracting method for the detection of Hepatitis A virus in shellfish by SYBR-Green real-time RT-PCR. Int J Food Microbiol. 2007 Nov 30;120(1-2):179-85. Epub 2007 Sep 8. PMID: 17900731

Sánchez G, Bosch A, Pintó RM. Hepatitis A virus detection in food: current and future prospects. Lett Appl Microbiol. 2007 Jul;45(1):1-5. Review. PMID: 17594452

Hasegawa Y, Matsumoto F, Tanaka C, Ouchi Y, Hayashi K. Outbreak of hepatitis A virus infection caused by food served in a restaurant. Jpn J Infect Dis. 2007 May;60(2-3):150-1. No abstract available. PMID: 17515658

Frank C, Walter J, Muehlen M, Jansen A, van Treeck U, Hauri AM, Zoellner I, Rakha M, Hoehne M, Hamouda O, Schreier E, Stark K. Major outbreak of hepatitis A associated with orange juice among tourists, Egypt, 2004. Emerg Infect Dis. 2007 Jan;13(1):156-8. PMID: 17370535

Bialek SR, George PA, Xia GL, Glatzer MB, Motes ML, Veazey JE, Hammond RM, Jones T, Shieh YC, Wamnes J, Vaughan G, Khudyakov Y, Fiore AE. Use of molecular epidemiology to confirm a multistate outbreak of hepatitis A caused by consumption of oysters. Clin Infect Dis. 2007 Mar 15;44(6):838-40. Epub 2007 Feb 13. PMID: 17304457

Croci L, Losio MN, Suffredini E, Pavoni E, Di Pasquale S, Fallacara F, Arcangeli G. Assessment of human enteric viruses in shellfish from the northern Adriatic sea. Int J Food Microbiol. 2007 Mar 10;114(2):252-7. Epub 2006 Dec 28. PMID: 17196284

Kingsley DH, Guan D, Hoover DG, Chen H. Inactivation of hepatitis A virus by high-pressure processing: the role of temperature and pressure oscillation. J Food Prot. 2006 Oct;69(10):2454-9. PMID: 17066927

Zoni R, Bigliardi L, Pavoni E, Sansebastiano G. [Integrate cell culture--PCR (ICC/PCR) in viruses researches in environmental and food samples. Note I] Ann Ig. 2006 Jul-Aug;18(4):305-12. Italian. PMID: 17063629

Tekeuchi Y, Kobayashi G, Matui Y, Miyajima Y, Tanahashi S, Honma M, Takahashi M, Eguchi H, Tanaka M. Outbreak of food-borne infection with hepatitis A virus. Jpn J Infect Dis. 2006 Oct;59(5):346. No abstract available. PMID: 17060706

Hewitt J, Greening GE. Effect of heat treatment on hepatitis A virus and norovirus in New Zealand greenshell mussels (Perna canaliculus) by quantitative real-time reverse transcription PCR and cell culture. J Food Prot. 2006 Sep;69(9):2217-23. PMID: 16995527

Amon JJ, Devasia R, Xia G, Nainan OV, Hall S, Lawson B, Wolthuis JS, Macdonald PD, Shepard CW, Williams IT, Armstrong GL, Gabel JA, Erwin P, Sheeler L, Kuhnert W, Patel P, Vaughan G, Weltman A, Craig AS, Bell BP, Fiore A. Molecular epidemiology of foodborne hepatitis a outbreaks in the United States, 2003. J Infect Dis. 2005 Oct 15;192(8):1323-30. Epub 2005 Sep 12. PMID: 16170748

Weinberg M, Hopkins J, Farrington L, Gresham L, Ginsberg M, Bell BP. Hepatitis A in Hispanic children who live along the United States-Mexico border: the role of international travel and food-borne exposures. Pediatrics. 2004 Jul;114(1):e68-73. PMID: 15231975

Kukavica-Ibrulj I, Darveau A, Fliss I. Immunofluorescent detection and quantitation of hepatitis A virus in sewage treatment effluent and on agri-food surfaces using scanning confocal microscopy. J Virol Methods. 2003 Mar;108(1):9-17. PMID: 12565149

Dubois E, Agier C, Traoré O, Hennechart C, Merle G, Crucière C, Laveran H. Modified concentration method for the detection of enteric viruses on fruits and vegetables by reverse transcriptase-polymerase chain reaction or cell culture. J Food Prot. 2002 Dec;65(12):1962-9. PMID: 12495017 [PubMed - indexed for MEDLINE]

Richards GP. Enteric virus contamination of foods through industrial practices: a primer on intervention strategies. J Ind Microbiol Biotechnol. 2001 Aug;27(2):117-25. Review. PMID: 11641770

Leggitt PR, Jaykus LA. Detection methods for human enteric viruses in representative foods. J Food Prot. 2000 Dec;63(12):1738-44. PMID: 11131900

ADDITIONAL:

Papafragkou E, Plante M, Mattison K, Bidawid S, Karthikeyan K, Farber JM, Jaykus LA. Rapid and sensitive detection of hepatitis A virus in representative food matrices. J Virol Methods. 2008 Jan;147(1):177-87. Epub 2007 Oct 10. PMID: 17931710

Richards GP. Enteric virus contamination of foods through industrial practices: a primer on intervention strategies. J Ind Microbiol Biotechnol. 2001 Aug;27(2):117-25. Review. PMID: 11641770

Some of the references have been incorporated into the revised document as appropriate.
Reviewer 2

This exercise would have been facilitated by numbering the lines in the manuscript. Observations that follow are coded by the page number and a line number that I have supplied, ignoring the four-line running head and numbering only lines that have text in them.

7, 19: This speaks of "reported cases of HAV due to all causes …" In fact, HAV is the cause, and hepatitis A (HA) is the disease of which cases are reported. Rather than "causes," this should say "sources" or "modes of transmission." This distinction would seem to be critical in the context of a risk profile.

We appreciate the reviewer's comments, and they will be taken under advisement for future endeavors.
8, 6 & 7: This sentence is awkward. It could better read, "The CDC has identified various factors that contribute to the difficulty in determining the number of produce-related HAV cases, by means of routine surveillance." The third sentence in this paragraph is unnecessary, since what it asserts is a consequence of the recall problem stated in the first line of the text box. The sentence has been edited for clarity.
9, 2: Rather than "can be potentially transmitted," this should be "can potentially be transmitted." The document has been edited.
11, 6 & 7: Where two or more references are cited in a single set of parentheses, the references should consistently be in alphabetical order by author or in chronological order, depending on the publisher's stylebook. Although this set happens to be in chronological order, the references on line 14 and many other places in the manuscript are in neither alphabetical nor chronological order. The document has been edited to reflect the reviewer's comments.
Reviewer 2 16, 8: Aliquot (Oxford English Dictionary): "Contained in another a certain number of times without leaving any remainder; forming an exact measure of." Is the word used correctly here? Yes, it is used correctly; this is common terminology in scientific methodology.
Reviewer 2 16, 16-18: This is certainly a significant advance, in that it should demonstrate replication of non-cytopathic HAV; however, it may be that some wild strains of HAV are incapable of infecting cultured cells at all. The text has been revised to avoid confusion regarding wild strains of HAV.
Reviewer 2 18, 13 & 17: Unless the agency's stylebook requires that sentences not begin with an acronym, it would be better to use "HAV" here than "Hepatitis type A virus," to avoid confusing the reader. We prefer to use the style of most scientific journals that avoid the use of acronyms at the beginning of a sentence.
Reviewer 2 18, 16 & 17: This sentence is disjointed. It would be better to say, "On aluminum surfaces at 20 to 23°C, 50 days were required to reduce HAV by 2 logs." This has been revised.
Reviewer 2 18, 25: Groundwater cannot be infected by HAV. This should say "contaminated" or "inoculated." This has been revised.
Reviewer 2 20, 11: The anti-HAV effect of chlorine dioxide might also be mentioned here (Mariam, T. W., and Cliver, D. O. 2000. Hepatitis A virus control in strawberry products. Dairy Food Environ. Sanit. 20:612B616. -- paper cited on page 76, line 7). We did not include these data in the risk profile, as ClO2 did not appear to be effective in reducing HAV levels under "real world" conditions. However, based on the reviewer's comment, we have now incorporated this information in the document.
Reviewer 2 20, 24: Rather than "at acidic conditions," this should be "under acidic conditions." The document has been edited.
Reviewer 2 22, 16: fulminate => fulminant The document has been edited.
Reviewer 2 22, 21: HA has been known to cause permanent loss of such normal liver functions as production of alcohol dehydrogenase. Rather than a major disability, this is a quality-of-life issue. The text has been revised.
Reviewer 2 23, 17: It is better to speak of "shedding" of virus particles, reserving the word "excretion" for a specific, normal physiological function This has been revised.
Reviewer 2 24, 2: Use of "defines" is unfortunate here. It would be better to say "indicates" or "implies." This has been revised.
Reviewer 2 25, 1: The word "transpired" is most often used in connection with a discrete event; "occurred" would be a better choice here. This has been revised.
Reviewer 2 27, 6: It is not possible for food to be an "etiological agent"; food is a "vehicle." This has been revised.
Reviewer 2 27, 12: The incubation period given here is incorrect. Incubation may be as short as 2 weeks; CDC usually says "15-50 days." This has been revised.
Reviewer 2 27, 13: Distribution may be wide, both geographically and temporally. This has been revised.
Reviewer 2 28 & 29: This table is incorrectly organized. Since the order of the rows was clearly determined first by the year of occurrence, the year column should be at the left, possibly followed by the month column. Also, the title of the "Source of Transmission" column should be "Source [or Mode] of Contamination." This has been revised.
Reviewer 2 31, 6: The phrase "to.3/100,000" needs editing. This has been revised.
Reviewer 2 31, 24: Something has been omitted from the end of this sentence. This has been revised.
Reviewer 2 32, 20 & 21: It is redundant to say "specific attribution rates have not been calculated specifically ..." This has been revised.
Reviewer 2 34, 13: diseases => disease This has been changed.
Reviewer 2 34, 22: not => note This has been changed.
Reviewer 2 37, 14: which => who This has been changed.
Reviewer 2 39, 2: The value of a statistical life was recently reduced by the United States government. Arbitrary values such as this are very difficult to regard as valid. As mentioned above, the FDA accepts the use of QALDs to help calculate pain and suffering caused by foodborne illnesses.
Reviewer 2 39, 11-15: This sentence is quite awkward and should be reworded. This has been revised.
Reviewer 2 40, 2-7: This is a run-on sentence that is very difficult to comprehend. This has been revised.
Reviewer 2 40, 19 - 41, 3: These two sentences are also run-on to an extent that obscures their meaning. This has been revised.
Reviewer 2 41, 2: This case total is much greater than is shown in Table 1. Is the Associated Press as credible as CDC? We attempted to use all sources available to us. We found this information in FSNET.
Reviewer 2 46, 2 & 3: It is unnecessary to cite the same reference at the ends of two successive sentences. This has been revised.
Reviewer 2 47, 4-7: This sentence conveys some misconceptions. First, illness is not the greatest concern -- as was stated earlier, the majority of HAV is shed during the incubation period, before onset of symptoms. Second, as was pointed out above, if these are farm workers in a developing country, they have most likely been immune since childhood and are probably not actively shedding HAV. This sentence has been revised.
Reviewer 2 48, 8: that => who This has been changed.
Reviewer 2 48, 24 - 49, 5: As was pointed out earlier, this is a very important consideration. It deserves even more emphasis than it receives here. As noted above, we have tried to emphasize contamination of produce by children in the growing fields in the revised version of the document.
Reviewer 2 50, 9 & 10: How can "quality" be a "source of contamination"? This has been changed.
Reviewer 2 52, 27: for misting and washing produce and ice production => for misting and washing produce and for ice production [i.e., ice production is not being washed] This has been revised.
Reviewer 2 52, 28 - 53, 1: This phrase seems meaningless. The comment is noted.
Reviewer 2 53, 5 & 6: others is not generally considered to be food workers => others not generally considered to be food workers This has been changed.
Reviewer 2 57, 18 & 19: This is certainly a real risk; however, consumers are likely to want to handle fruits to determine their ripeness, perhaps a more critical function than selection of green beans (Figure 9). This figure was intended to depict a consumer touching (picking and choosing) produce; we were not focusing on a specific produce type. The figure has since been changed to a consumer picking tomatoes.
Reviewer 2 57, 23: rates ... has => rates ... have This has been changed.
Reviewer 2 63, 4: National Advisory Committee for the Microbiological Criteria for Food => National Advisory Committee on Microbiological Criteria for Food This has been changed.
Reviewer 2 63, 24: Institute for Food Technologists => Institute of Food Technologists This has been changed.
Reviewer 2 64, 19: potential ... are => potentials ... are This has been changed.
Reviewer 2 67, bottom figure: Note that the water provided for hand washing here is apparently not fit to drink. Photo has been removed to avoid confusion.
Reviewer 2 68, 15 & 16: The purpose of hand washing is to remove the virus from the skin. Therefore, HAV's resistance to inactivation is not relevant. From the literature, it appears that water alone is inadequate to totally rid the virus from skin. In addition, people often use sanitizing gels instead of washing with water and soap, so we feel that discussion of HAV resistance to disinfectants is relevant.
Reviewer 2 70, 21-24: It is certainly true that irrigation water should not contain HAV, which would be present as a result of human fecal contamination. This condition should be attainable without having to treat the water to EPA drinking-water standards. We also believe that there is a need for standardization of irrigation water.
Reviewer 2 71, 3-6: The estimate here of ~80 HA illnesses saved per year by disinfection of contaminated water could serve as the basis for another cost-benefit comparison. We do not have sufficient information to conduct a cost/benefit comparison.
Reviewer 2 71, 7: treatments techniques => treatment techniques (or just "treatments") This phrase has been revised to "treatment techniques."
Reviewer 2 72, 7: The call for "inclusion of sufficient effective disinfectant in wash water and flume water" should also address the fact that water used in these ways is recycled and builds up suspended matter that may increase disinfectant demand. The text has been revised to include the consequence of using recycled water.
Reviewer 2 72, 10: As was pointed out above, it should be noted that berries are exempted from the suggestion of treating all produce. As mentioned above, we agree with the reviewer and have included wording in the document to describe this.
Reviewer 2 78, 4-7: The first two sentences of this paragraph are identical. The second sentence has been deleted.
Reviewer 2 81, 9: Might the Texas entry in Table 10 have been meant to say, "(certain counties)"? This has been revised to "counties."
Reviewer 2 84, 1-3: This recommendation should be made as strongly as possible. We have emphasized the role of children in the growing fields where relevant in the document.
Reviewer 2 84, 6: At every opportunity, it should be recognized that washing strawberries that are to be sold fresh is not an option. As mentioned above, we agree with the reviewer and have included wording in the document to describe this issue.
Reviewer 2 85, 4-5: Given what has been said earlier about the difficulty of reversing HAV contamination of produce, it is hard to imagine what one would tell consumers that would contribute to their protection from HA. Contamination of produce is not always internal. Most of the contamination occurs during handling of the produce, which would be on the surface. Advising consumers to wash their produce before consumption would probably take care of superficial contamination.
Reviewer 2 85, 9-11: The HACCP approach should be invoked with great caution. On the one hand, if the HACCP system truly extends from farm to fork, a single, valid critical control point anywhere along this continuum should serve to protect the consumer. On the other hand, there is no  guarantee that a valid critical control point, which would truly prevent or eliminate HAV contamination, exists anywhere along the continuum. The third CCP option (reduce a hazard to an acceptable level) only works until the next outbreak. We agree with the reviewer's comments in that application of HACCP for the entire food supply chain is not appropriate. We are aware that HACCP has been adopted at many fresh-cut processing facilities, but NACMCF (1999) stated that HACCP at the farm is premature.  The statement in the report has been edited to include the term "HACCP-like," which would include assessments or preventative controls to be included at the farm.
Reviewer 2 85, 12-14: How will "safe handling" instructions avoid cross-contamination? If an item of produce (to be eaten fresh) is contaminated with HAV, it is hard to imagine a concern for cross-contamination. If the produce is not already contaminated with HAV, keeping it pristine may call for measures beyond the scope of "safe handling." We have tried to encompass all possible measures and provide options to decision makers to reduce or prevent illnesses associated with consumption of raw produce contaminated with HAV.
Reviewer 2 87, 3-5: This is an important consideration, but the example begs the question. In the case of HA, the majority of virus to be shed will already have been shed before onset of symptoms. A person recovered from HA will probably never shed HAV again in life, although some alternatives (prolonged shedding, inapparent carriage, relapse) are mentioned earlier in the draft. However, the symptoms described in the example resemble those of norovirus infection, in which shedding may continue for up to a week after recovery. This has been changed to better reflect HAV symptoms.
Reviewer 2 87, 15: high-risks => high-risk This has been changed.
Reviewer 2 87, 17-23: This is a very dubious recommendation. First, valid means of sampling and testing the great variety of produce for HAV have yet to be developed, as suggested in the following Research section, and available RT-PCR methods are likely to give false-positive results with HAV that has been inactivated. Second, the cost of extensive surveillance and testing would be huge and of limited benefit, possibly compared to immunization. There are no "dip-stick" tests for HAV or any other foodborne virus, so the cost of testing includes reagents, apparatus, and pay for highly skilled personnel. The risk profile has tried to provide all options, and it is up to the decision makers to determine the feasibility of the option.
Reviewer 2 88, 7: approaches ... is => approaches ... are This has been changed.
Reviewer 2 88, 14: The singular of "fomites" is "fomes." This has been changed.
Reviewer 2 88, 16-18: This recommendation deserves a hearty "Amen!" since it is likely to be achieved only by means of very powerful prayer. No response is needed.
Reviewer 2 88, 26 & 27: This should focus especially on produce that cannot be washed before sale. Where washing in processing is an option, washing is probably better done by professionals than by consumers. No response is needed.
Reviewer 2 89, 2: It must be recognized that an ill person in the home is a threat to other occupants in other ways than through foodhandling. One could argue that in-home preparation of food by an ill person is inappropriate behavior of itself; however, many households are lucky to have even one person who knows how to cook. Under the Consumer Section we say "Contamination of produce in the home can occur by unsanitary conditions, infected food preparer, improper hand washing, bare hand contact, cross contamination, or improper washing of utensils and other food contact surfaces."
Reviewer 2 89, 20: produced => produce This has been revised accordingly.
Reviewer 2 95, 19-25: Legislation to mandate sick leave compensation and address returning to work post-illness might well contribute to overall food safety, but is not particularly relevant to HA. Rule-making in these domains is probably better left to responsible agencies in the executive branch (e.g., FDA) than to legislation by lawmakers. There have been several reports in the past 5 years or so demonstrating outbreaks/illnesses/exposures to HAV via a sick food worker (see Appendix 4). Therefore we feel mandating sick leave to be an option for control. The risk profile is merely providing decision makers with several options for reducing foodborne illness due to HAV, one of which is vaccinating food workers. It is up to the decision makers to determine whether the proposed options are relevant and, if so, how to implement these options.
Reviewer 2 100, 20-22: potential ... remain a potential concern [singular subject, plural verb; redundant] This has been changed accordingly.
Reviewer 2 101, 13: Obviously, ill workers are anathema anywhere along the food continuum; however, this concern is least applicable to HA, because the majority of shedding occurs before onset of illness. As mentioned above, there have been several reports in the past 5 years or so demonstrating outbreaks/illnesses/exposures to HAV via a sick food worker (see Appendix 4).
Reviewer 2 102, 6: immunization ... have => immunization ... has This has been changed.
Reviewer 2 102, 27: See earlier discussion of the validity of critical control points. We agree with the reviewer that there is no guarantee that a valid critical control point that would truly prevent or eliminate HAV contamination, exists anywhere along the continuum. The risk profile also suggests that there is no single preventive measure; we need a combination of control measures to effectively reduce or prevent illness associated with HAV. However, we feel that there is a need to know what the critical points are, in order to develop and establish control measures for each step along the farm-to-fork continuum.
Reviewer 2 102, 31: risk options => risk management options "Risk options" has been changed to "options."
Reviewer 2 103, 9: As mentioned earlier, the likelihood that this would be cost-effective is extremely small. Changed to "when analyzing produce after an outbreak for the causative hazard" or similar language.
Reviewer 2 References: There are some format inconsistencies. On page 128, the Marena reference should appear before the Mariam reference. On page 138, Tjon should come before Tsai. This has been changed.
Reviewer 2 145, 12: The assertion that HAV illnesses are highly underreported seems tenuous in comparison to other foodborne illnesses in the United States. As mentioned above, we believe that, although the number of hepatitis A illnesses may be substantially less than that of other foodborne illnesses, given the number of outbreaks and cases not yet published in peer-reviewed articles (27 - see appendix 4), underreporting is still an issue.
Reviewer 2 145, 18: Activated carbon, reverse osmosis, and membrane filtration are unlikely to inactivate HAV. The best they might do would be mechanical removal of the viral particles. This has been changed.
Reviewer 2 145, 21: rates ... has => rates ... have This has been changed.
Reviewer 2 146, 1: as is transfer rates => as are transfer rates. This has been changed.
Reviewer 2 146, 1 & 2: hands to food using gloves => hands to food when gloves are used. [Food does not use gloves.] This has been changed.
Reviewer 3 page 34: "…since HAV can also be a travel related diseases." Change to: "…because HAV can also be a travel related disease." This has been changed.
Reviewer 3 page 34: "It is worthwhile to not, however, that contributing factors are unknown for 45% HAV infections." Change to: "It is worthwhile to note, however, that contributing factors are unknown for 45% of HAV infections." This has been changed.
Reviewer 3 page 37: "Ill adults lose an average of 27 days…" Change to: "Ill adults from HAV lose an average of 27 days…" This has been changed.
Reviewer 3 page 37: "Average costs, both direct and indirect, of hepatitis A range from…" Change to: "Average medical costs, both direct and indirect, of hepatitis A range from…" This has been changed.
Reviewer 3 page 40: "In the case the green onion outbreak…" Change to: "In the case of the green onion outbreak…" This has been changed.
Reviewer 3 page 50: "Seeing HAV survives in water, soil,…" Change to: "Seeing that HAV survives in water, soil…" This has been changed.
Reviewer 3 page 59: "prevention initial contamination of produce…" Change to "prevention of initial contamination of produce." This has been changed.
Reviewer 3 page 66: why is the first paragraph labeled as 8)? There does not appear to be a 1-7 preceding this paragraph. This has been changed.
Reviewer 3 The title of Section VII could better reflect the content, particularly to distinguish it from Section VI. Suggest: New/Additional Control Options This has been changed.
Reviewer 3 The last column of Table 11a is not entirely readable. This might be due to a formatting difference between computers.
Reviewer 4 There are some formatting issues (including page breaks, spacing in Figures and Tables, etc.) that will most likely be resolved as this draft is revised and printed in final form. The document has been edited and formatted.
Reviewer 4 Page 9, paragraph 1, line 6. Add an "s" to the word onion This has been revised.
Reviewer 4 Page 12, Section 2, Methods of Analysis. I suggest switching Section 2 (Methods of Analysis) with Section 3 (Characteristics) on page 16, so that the characteristics section comes before the methods of analysis section. The document has been revised.
Reviewer 4 Page 24, Section 3. Incidence of HAV, first sentence. Add "HAV" before the word cases. HAV has been added.
Reviewer 4 Page 25, Figure 1. Suggest centering this figure on the page. The figure has been centered.
Reviewer 4 Page 30, Regional differences section, last sentence. The Jacobs et.al., reference lists the year as 2006 but in the list of references on page 124, the year is listed as 2000. This has been revised in the final document.
Reviewer 4 Page 34, transmission section, first two sentences. Suggest revising the second sentence and using it to introduce the section….It is important to note that contributing factors are unknown for 48% of HAV infections (Fiore et al., 2004) but the known or potential susceptibility risk factors involved in transmission of HAV are listed below. This section has been changed.
Reviewer 4 Page 35, last bullet. The sentence that begins with "In the case of hepatitis A…" should be moved to another section of the document that addresses worker issues. This has been changed.
Reviewer 4 Page 37, Medical and Hospital Costs section, third line. The year on the Lasher et al., reference does not match with the Lasher citations on page 127. The text and content of this section have been changed, and a different reference was used.
Reviewer 4 Page 38, first line. The year of the Lasher et al. reference, doesn't match with the citation list on page 127; same on page 39, lines 6 and 11 and for Table 5 reference at the bottom of the page. The text and content of this section have been changed, and a different reference was used.
Reviewer 4 Page 40, Retail section, first line. Add an "s" to the word "other". This sentence is very confusing and should be revised. This has been changed.
Reviewer 4 Page 42, Primary Production section, last line. Omit the word "own" before hands. This has been changed.
Reviewer 4 Page 45, second paragraph, first two sentences. Suggest combining and revising these two sentences. This has been changed.
Reviewer 4 Page 45, second paragraph, third line. The is no citation for the Katzenelson et al.(1976) reference on page 125. The citation for Katzenelson et al. (1976) has been added.
Reviewer 4 Page 46, second paragraph, first two sentences. Suggest combining and revising the first two sentences. This has been changed.
Reviewer 4 Page 46, fourth paragraph, third line. The year of the Niu et al., 2002 reference does not match with the citation listed on page 131. This has been revised to match the year.
Reviewer 4 Page 47, photo credit. I would suggest the following wording for the credit given for all photos in the document on pages 47, 49, 50, 51, 52, 53, 55, 57, and 67. Photo courtesy of XXXXX. Wording of photo credits has been changed, as suggested.
Reviewer 4 Page 49, fifth line. The year of the Rosenblum et al., 1992 reference does not match the citations on page 134. The citation has been revised to match the 1991 reference in the reference list.
Reviewer 4 Page 50, first line. Suggest changing the word "Seeing" to "Since." This has been changed.
Reviewer 4 Page 52 and 53, Figures 6 & 7 Titles. I would suggest placing the Figure titles at the bottom of the photos instead of at the sides. This may just be a formatting issue in the draft document. This has been changed.
Reviewer 4 Page 55, source of photo. The photo is not clip art, so the source should be corrected. The photo has been replaced and the correct source provided.
Reviewer 4 Page 55, line 9. The year of the Parashar and Monroe, 1988 reference does not match the year in the citation on page 132. The citation and reference have since been omitted from the document.
Reviewer 4 Page 57, Figure 9. Suggest placing photo after the section on consumers instead of before and change the source of the photo. The figure has been replaced and moved, so that it follows the section on consumers, and a source of the figure has been provided.
Reviewer 4 Page 59, second line. Suggest changing the word "to" to "with." This has been changed.
Reviewer 4 Page 59, first bullet at bottom of page. Insert the word "of" after prevention. This has been changed.
Reviewer 4 Page 65, fourth paragraph, first line. Suggest deleting the word "above" and inserting the words "mentioned above" after the word "documents." This has been changed.
Reviewer 4 Page 67, first three lines. Suggest combining these sentences by inserting a comma after the references and adding the word "so" before the word "frequent" and add the word "strongly" before the word "encouraged." This has been changed.
Reviewer 4 Page 67, second line. There are two CDC citations listed on pages 110 and 111 that were published in 2005. These have been differentiated by the designations 2005a; 2005b.
Reviewer 4 Page 67, lines three through five. Suggest combining sentences three and four by beginning the sentence to read "Although studies…then placing a comma after the word "published" and deleting the word "accordingly." This has been changed.
Reviewer 4 Page 67, lines eight and nine. Delete the last sentence that begins with the word "However." The sentence has been deleted.
Reviewer 4 Page 67, second paragraph, second line. There is a reference for Kaplan et al, 1986 and there is no citation listed on page 125. The reference has been deleted.
Reviewer 4 Page 70, second paragraph, third line. Delete DHHS/ which appears to be a typo. DHHS has been deleted.
Reviewer 4 Page 72, first paragraph, lines six and seven. Suggest revising the sentence to read "In addition, it is of particular importance to include a sufficient amount of effective disinfectant in the wash water….. The sentence has been revised.
Reviewer 4 Page 72, second paragraph, lines seven and eight. It should read Table 9 not 8. I would also suggest placing Table 9 after that section on page 72 instead of on page 74. This has been changed.
Reviewer 4 Page 72, second paragraph, line twelve. The miscellaneous products mentioned are contained on page 64, not page 55. This has been corrected to reflect the correct page in reference to the miscellaneous products.
Reviewer 4 Page 72, end of third paragraph. Consider inserting Table 9 here instead of on page 74. Table 9 has been moved.
Reviewer 4 Page 75, second paragraph, lines four through six. Insert the word "a" before the word "processed" and note that the references to 21 CFR are not listed in the citations at the end of the document. This has been changed; the CFR reference has been added to the reference list.
Reviewer 4 Page 78, second paragraph, lines two through four. Delete the sentence that begins " The question….as it is mentioned twice. The sentence has been deleted.
Reviewer 4 Page 79, paragraph two, line seven. The CDC 2005a reference does not match with the citations on page110 and 111. The citations and references have been revised.
Reviewer 4 Page 81, Table 10. I believe that the Texas listing, should read "certain counties", instead of "certain states." This has been revised accordingly.
Reviewer 4 Page 81, First paragraph, line four. Insert the word "also" after the word "are." The word "also" has been inserted.
Reviewer 4 Page 83, first paragraph, line seven. The CDC, 2006 reference should be either a or b This has been changed.
Reviewer 4 Page 84, last bulleted item, line three. Delete the word "prevailing" and insert the words "for the." This suggestion has been implemented.
Reviewer 4 Pages 91, 92, 93, 94, 95, 96. Factors Influencing Costs Column. The word "entities" is used multiple times in this column and it is a bit confusing. I would suggest using another word such as companies, establishments, operations, etc. This has been changed.
Reviewer 4 Page 92, Breadth and Depth of Coverage column, line one. Delete the word "to" after the word "accrue." "To" has been deleted.
Reviewer 4 Page 94, Factors Influencing Costs column, Develop SSOPs. Suggest deleting the words "regulatory instrument" and inserting the word "regulations." This has been changed.
Reviewer 4 Page 94, Breadth and Depth of Coverage column, last box, third line. Delete the word "by" and insert the word "from" after HAV. This has been changed.
Reviewer 4 Page 94, Factors Influencing Benefits column, last box. There is a typo in the word "form." This has been changed.
Reviewer 4 Page 95, Breadth and Depth of Coverage column, first box, second line. Delete the word "by" and insert the word "from" after HAV. This has been changed.
Reviewer 4 Page 95, Factors Influencing Costs column, second box, fourth line. Delete the word "for" after the word "training." "For" has been deleted.
Reviewer 4 Page 100, lines one through three. Check CDC, 2006 reference in citation list; The year of the Calvin et al,.2005 reference does not match the one in the citation list on page 108; The Everson et al,. 2005 reference is not listed in the citations on page 119. The CDC reference has been revised; the Calvin et al. reference has been revised to match the citation; the Everson et al. reference has been deleted.
Reviewer 4 Page 100, second paragraph, line five. Suggest switching the words "primary production" and the "retail level", so the sentence reads…occurring at both the primary production and retail levels. This has been changed.
Reviewer 4 Page 101, last sentence. Suggest revising the sentence to read…Most establishments that prepare food for service do have….. The sentence has been revised.
Reviewer 4 Page 102, line one. Insert the "Decker" before the word "Griffith" in the reference. This has been inserted.
Reviewer 4 Page 102, second paragraph, line three. Delete the words we do have and insert the words, "There is." This has been changed.
Reviewer 4 Page 102, second paragraph, line five. Check CDC 2005a reference in list of citations on pages 110 and 111. The reference has been checked and revised accordingly.
Reviewer 4 Page 102, fourth paragraph, lines five through seven. Revise the sentence to read "The survival of HAV on produce needs to be determined and assessments conducted to identify critical control points along the farm to fork continuum before determining which risk management options have the most impact." The sentence has been revised.
Reviewer 4 Pages 104 to 141. The references that are cited need to be carefully checked for accuracy as there are a number that are missing from the list, have incorrect years or need to more clearly identified. Also, it appears that there are some citations listed that are not contained in the text of the document. There are also spacing issues which need to be corrected in the reference section. The references have since been checked and revised accordingly.
Reviewer 4 Page 145, item ten. Add the word "the" before the word "consumer." The sentence has been revised.
Reviewer 5 Introduction (p.7 as well as throughout the document). To a knowledgeable reader, the inconsistencies and incorrect use of the terms "HAV", "hepatitis A" and "cases" gives a poor introduction to the document. To a reader new to the field, these are very confusing. The document has been checked and edited accordingly.
Reviewer 5 An individual infected with HAV can either have an asymptomatic infection or a symptomatic infection; the latter being mild, moderate, severe or resulting in death. The term "hepatitis A" refers to a symptomatic HAV infection. Similarly, when one discusses "cases" of hepatitis A, this refers to symptomatic HAV infection (i.e., the disease) and NOT to all HAV infections, since a substantial proportion of HAV infections are asymptomatic. In the Introduction alone, these terms are used interchangeably and it is not appropriate. For instance: As stated above, the document has been edited to reflect the correct terminology.
Reviewer 5 "…HAV illness associated with the consumption of fresh/fresh-cut produce." should read "…hepatitis A associated with …. As stated above, the document has been edited to reflect the correct terminology.
Reviewer 5 The use of hepatitis A and illness in the following sentence is redundant since hepatitis A is an illness. "…contributing factors to produce-associated hepatitis A illnesses." As stated above, the document has been edited to reflect the correct terminology.
Reviewer 5 The statement "…including the economic burden associated with HAV infections." is in correct. Economic burden is associated with disease burden (i.e., hepatitis A) which is a subset of HAV infections. As stated above, the document has been edited to reflect the correct terminology.
Reviewer 5 The statement "…(CDC) reports that the total number of HAV cases in the United States…." is incorrect. The CDC conducts surveillance for hepatitis A, not HAV, and "cases of HAV" has no epidemiologic or clinical meaning; it's hard to imagine a case of a virus. If one were to discuss cases of HAV, the correct sentence structure would be "cases of HAV infection". The same is true in the following sentence and in the text box and many other places. Frankly, a good science editor / writer should go over the introduction and the remainder of the document. As stated above, the document has been edited to reflect the correct terminology.
Reviewer 5 p.8. Discussion attributed to CDC about factors contributing difficulty in identifying "produce-related HAV cases." The factors described relate to all foodborne hepatitis A and not just produce-related cases. The accompanying table has a number of problems. The title is not correct since it is cases of disease that are reported, not infections. In the 2nd box, "e.g." should be used and not "i.e." since an example is being provided for a class of events. The 3rd box shows confusion in terms - it is cases of disease (hepatitis A) that are reported, not HAV infections. Serologic methods must be employed to identify all HAV infections. The section has been revised.
Reviewer 5 p.9. Would be helpful to define "raw molluscan shellfish" (a glossary of terms would help) since outbreaks have only occurred from oysters and clams. The phrase "specifically oysters and clams" has been inserted after "raw molluscan shellfish."
Reviewer 5 p.11. In general, this entire section describing the pathogen takes a minimalistic approach. More could be included since this information is needed to better understand the detection methods described later in the document. A risk profile is not a review article. We attempted to summarize all the information relevant to provide control options that reduce or prevent illnesses associated with consumption of produce contaminated with HAV.
Reviewer 5 The term "hepatitis type A virus" was used many decades ago and is not the usual way HAV is described in the contemporary literature. We have tried to include all the different terminology for the virus (see glossary of terms).
Reviewer 5 More information about the animal models for infection is warranted since this is an impediment to conducting some of the experiments that the authors suggest later in their ‘gaps analysis'. We have added the use of nonhuman primates to provide information on whether an immunized person, either by natural infection or by passive immunization, can be an asymptomatic carrier; i.e., to determine whether HAV can survive and replicate in the intestinal cells of an immunized person/animal without causing illness.
Reviewer 5 In describing the cell culture growth of HAV, the word "unremarkable" is used. This seems to be a poor choice of words. What is remarkable cell culture growth versus unremarkable growth? The structure of the next sentence suggests that HAV is a cytolytic virus, but that it does not demonstrate a cytopathic effect in cell culture. However, HAV is considered a non-cytopathic virus based on its cell culture growth characteristics. The section has been revised.
Reviewer 5 A very poor description of "strains" and the utility of using this information (see my earlier comments). This section has been revised.
Reviewer 5 p.12. "The route of entry for foodborne HAV is typically via the intestinal tract…." The way the sentence is structured suggested that there are other ways that HAV enters. If so, what are they? The sentence has been revised.
Reviewer 5 p.15. In the discussion about the Detection, there is mention that RT-PCR is limited by its inability to distinguish between infectious and non-infectious HAV. One means to improve that differentiation has been through the use of immuno-capture RT-PCR since it relies on the detection of intact virions containing RNA. However, there is no discussion about this well-established method and its strengths and weaknesses. This section has been revised.
Reviewer 5 p.15-16. The section on "Tissue Culture" is very confusing, doesn't give a good review of methods and their strengths or weaknesses. The introductory sentences should briefly summarize what was presented earlier about the limitations of this method for primary isolation of HAV from clinical, environmental or food samples. Either here (preferably) or in the earlier section there should be a short review of the methods for primary isolation of HAV. HAV does grow in cell culture but its identification has been difficult because of lack of a cytopathic effect and identification of virus replication has been by identification of HAV antigen; these methods and their elative sensitivity should be summarized (e.g., radio- or immuno-focus assays). This would then logically lead to the descriptions of the newer methods that detect HAV nucleic acid. This could then be followed by a description of the plaque-forming HAVs which have been developed and have been used for a number of in vitro studies. The way the section is presently written, it leads the reader to believe that Cromeans developed a plaque assay that can be applied to any cell culture-derived HAV, when this is not true. This is not the case. The HM 175 strain described in the paragraph on p. 16 is the plaque-forming virus described by Cromeans, which has been used in various in vitro experiments. This section has been revised.
Reviewer 5 p.16. Antibody analysis. This section seems to be out of place since the previous section refers to detection of HAV in various environmental or food samples. The section on antibody analysis refers to detection of HAV infection - there should probably be a section on this since molecular diagnostics are also applicable to the diction of HAV infection in humans and non-human primates. This section is flat wrong and the reference cited is not appropriate. Since the correct diagnosis of hepatitis A among persons with jaundice is central to the issue at hand, this reviewer finds this lack of understanding quite disturbing. There is also another section where this is discussed (p. 24) and the description of the antibody response is also incorrect. This section has been revised.
Reviewer 5 The serologic diagnosis of acute infection with HAV is based on the presence of IgM anti-HAV, a diagnostic test which is commercially available. IgM anti-HAV persists for long periods of time following the acute infection. However, the commercially available tests have had their sensitivity lowered ("detuned") so that they become negative within 4-6 months following acute infection. This test can also be transiently positive following hepatitis A vaccination. An "increase in virus specific serum antibody titers" is not used to determine recent HAV infection as stated in the risk profile. This section has been revised.
Reviewer 5 The other commercially available test measures total (IgM and IgG) anti-HAV and is used to determine immunity to HAV infection, either from past infection or immunization. Depending on the test, it may not be positive following immunization because it may not detect the lower antibody levels generated from immunization. There are a number of papers on hepatitis A diagnostics, as well as information on web sites, text books, etc. This section has been revised.
Reviewer 5 p.18. Under survival in water, the statement "…artificially infected ground water samples." should be "artificially inoculated" since HAV cannot replicate in water. This sentence has been revised.
Reviewer 5 p.21. The Disease. The sections on disease, incidence and epidemiology of hepatitis A are not written at the in-depth level of academic review as afforded other sections of the document. A risk profile is not a review article. We attempted to summarize relevant information to describe the food safety problem and to provide control options that reduce or prevent illnesses associated with consumption of produce contaminated with HAV.
Reviewer 5 The section on infectious dose should review earlier chimpanzee and marmoset studies that established titered inocula that still exist (see my comments on Appendix 2) and established an infectious dose. These studies provide infectious dose for those species, not necessarily for human primates; the exact HAV strain used also is not well characterized.
Reviewer 5 p.22. There is a real disconnect between the statement "The illness is generally self-limiting…" and the economic analyses that indicate the relatively high morbidity and costs associated with hepatitis A. The section has been revised.
Reviewer 5 The next paragraph reads very poorly and the statement about fulminant hepatic failure and the low rate of liver involvement just makes no sense. The primary site of HAV replication is the liver and many histopathologic studies have shown the pattern and degree of liver pathology. While many factors have been postulated to be associated with fulminant hepatic failure, none have been shown to be associated. I think it is the responsibility of the authors in this type of document to give a best judgment as true association rather than list a mixture of primary papers and reviews. Also, there is no mention of the association of hepatic failure from acute hepatitis A among persons with underlying chronic liver disease, an association that seems to continue to hold up. The section has been revised.
Reviewer 5 p.23. A published paper from the NHANES describes seroprevalence studies of HAV infection in the US prior to introduction of immunization and should be cited - Bell BP, et al. Hepatitis A virus infection in the United States: Serologic results from the Third National Health and Nutrition Examination Survey. Vaccine 2005; 23: 5798-5806. We thank the reviewer for this reference; it has now been cited in the document.
Reviewer 5 In the description of fecal shedding of HAV, for this kind of review it is not appropriate to use a general review such as that of David Heymann. Primary references should be used, including the "old" ones where these observations were made, and the authors should try and contrast and compare them with the newer findings, and not leave the reader "hanging". The risk profile is not intended as an academic review of HAV; rather, we tried to include information pertinent to providing control options to reduce illness caused by consuming produce contaminated with HAV.
Reviewer 5 p.24-36. The section called "Incidence of HAV." Not only is this the wrong terminology (it should be "…hepatitis A or HAV infection") but the entire section should be called something like "Epidemiology of Hepatitis A in the United States", a title that is much more appropriate to the purpose and objectives of this risk profile. In my opinion, this entire section is missing crucial information and is poorly researched and written. Also, it immediately conveys a bias of the entire document. By placing "outbreaks" right after "incidence" it indicates to the previously uniformed reader that outbreaks (and foodborne at that) were the major source of hepatitis A in the US prior to implementation of immunization. And that they continue to be the major source of disease. However, this is not the case. This section has been rewritten and the title changed to "Epidemiology of Hepatitis A in the United States."
Reviewer 5 My recommendation is that the section should be rewritten to provide: 1) in-depth information/analysis on trends in disease incidence over time, to include information on age, sex ethnicity (some of this is in the section; 2) information on the epidemiology of this disease, including populations/communities with increased rates of infection (e.g., American Indians/Alaskan Natives, person of Hispanic ethnicity), and groups with risk factors for infection (e.g., men who have sex with men, drug users, recipients of clotting factor concentrates, travelers to countries with a high endemicity of infection); and 3) a summary of outbreaks, both foodborne and from other sources (e.g., drug use, community-acquired, MSM). While some of this information is present in the section it is scattered around in a manner that does not allow the reader to form a cogent picture of hepatitis A in the US. As indicated, this section has been rewritten to include the sections mentioned by the reviewer.
Reviewer 5 I would also suggest that there be a major section on the epidemiology of hepatitis A in countries with a high or intermediate endemicity of infection. This is critical to the profile since most of the fresh food being discussed, and many of the outbreaks, come from food imported from these countries. A section describing the epidemiology of hepatitis A in countries with a high or intermediate endemicity of infection has been included.
Reviewer 5 I am not sure where to put the present section called "international trade". It could serve as the introduction to the proposed section on the epidemiology of hepatitis outside the US. This section now serves as the introduction to the proposed section on the epidemiology of hepatitis outside the US.
Reviewer 5 There should be an in depth review/discussion that estimates (even crudely) the burden of disease attributable to foodborne hepatitis A and places it in perspective to hepatitis A from all other sources. A section describing the burden of foodborne hepatitis A has been included in the document.
Reviewer 5 p.25. The sentence at the top of the page seems to be "hanging" and doesn't match the figure below, since there was a rise in hepatitis A incidence between 1995-1999. The sentence has been moved to just above Table 1.
Reviewer 5 p.35. This list is really not helpful and only serves to confuse the reader. The section has been edited; the two lists have since been combined.
Reviewer 5

p.37-40. Economic Impact: This entire section is written in a manner that implies these figures are attributable to foodborne hepatitis A, when these costs have been derived from all cases of hepatitis A in the US, of which only a very small proportion would seem to be attributable to foodborne transmission. There needs to be a very clear statement of what fraction of hepatitis A costs might be attributable to foodborne transmission.

The introduction to this section discusses the broad range of costs that can be attributed to hepatitis A (i.e., direct and indirect). The section then presents data on direct medical costs as well as indirect costs, as derived through QALDs). Economic analyses are usually performed to inform a decision making process - - that is, to make a choice. Such analyses were performed to inform the decision making process related to hepatitis A immunization of food handlers and the childhood population. In the present risk profile, the reader would be better informed if there were comparison data for other foodborne diseases (e.g., diarrhea). Also, since economic analyses are performed to inform a decision making process, what decision is being informed -- the cost of control measures, the costs of immunization, the cost of preventing foodborne hepatitis A versus all foodborne illness (i.e., the marginal cost)?

This section has been changed.
Reviewer 5 p.49. The study by Rosenblum showed prolonged shedding of HAV in premature infants, which is not relevant to the risk profile. However, there is a study by Ticehurst et al. in Greek children that showed a longer period of HAV shedding in children compared to adults. This study has been incorporated in the document.