);
freezing (
); heat treatment
(
); rapid cooling
(
).
In response to the 1997 and 1998 outbreaks of V. parahaemolyticus infections in the United States, the Food and Drug Administration (FDA) conducted a risk assessment to characterize the public health impact associated with consumption of raw oysters harboring pathogenic V. parahaemolyticus. This risk assessment focused specifically on oysters, because this was the food predominantly linked to the outbreaks. The risk assessment structures our knowledge of V. parahaemolyticus in a systematic manner, and includes sophisticated, mathematical models developed to estimate exposure to this microorganism and the dose-response relationships between the consumer and V. parahaemolyticus.
This interpretive summary is a non-technical version of the formal risk assessment document. The formal risk assessment document is designed to communicate the details of how the analysis was carried out. It frequently requires the use of technical terms and advanced statistical concepts that are sometimes not familiar to non-scientists. This interpretive summary includes the essential elements of the risk assessment in a manner that can be understood by non-scientists. It states simply why the risk assessment was conducted, what was required of the risk assessment team and what was done in response, what the results were, and what these results signify. It is recommended that those who wish to do an in depth evaluation of the work read the complete risk assessment document. The entire assessment document may be found on the FDA/CFSAN website: http://www.foodsafety.gov/~dms/fs-toc.html.
We have attempted to seek the best scientific information available. However, to ensure that we have both identified all key data sources and submitted the assessment to a rigorous peer review, we are releasing the assessment in draft form. A comment period has been established during which we will be actively seeking comments, suggestions, and additional data sources. Written comments should be submitted to the Dockets Management Branch (HFA-305), Food and Drug Administration, 5630 Fishers Lane, Room 1061, Rockville, MD 20852 within 60 days after publication of the document. Two copies of comments are to be submitted, except that individuals may submit one copy. Comments must be identified with the Federal Register Docket No. 99N-1075. The information acquired during the comment period will be reviewed and used, as appropriate, to further enhance the risk assessment and decrease uncertainty to the greatest degree possible. As stated in the Federal Register, Docket No. 99N-1075, the preliminary results of the draft V. parahaemolyticus risk assessment will be presented at a public meeting during the comment period.
In general, the risk assessment process includes four steps: 1) Hazard Identification, which identifies a potential hazard, 2) Exposure Assessment, which characterizes the level of exposure of people to the hazard, 3) Hazard Characterization/Dose-Response, which provides information needed to relate exposure to the hazard to the occurrence and severity of disease, and 4) Risk Characterization, which characterizes the impact of the hazard on public health, given the hazard characterization data and the size of the consuming population. This risk assessment is structured according to these four steps. The steps are further divided between three modules that reflect the chain of events from oyster harvest to consumption: Harvest, Post Harvest, and Public Health. In addition, because V. parahaemolyticus levels may be affected by climate and by region-specific oyster harvesting practices, the quantitative modeling was done separately for each season and for five separate geographic regions (Northeast Atlantic, Mid-Atlantic, Pacific Northwest, Louisiana Gulf Coast, and the remaining Gulf Coast).
Sporadic (isolated) cases of V. parahaemolyticus infections are primarily reported by the Gulf Coast states. Sporadic cases occur throughout the year, peaking in spring and summer. The Centers for Disease Control (CDC) estimates that the total number of foodborne V. parahaemolyticus cases in the United States for 1996, 1997, and 1998 were 2730, 8596, and 5525, respectively. These estimates include a projection on the part of CDC that due to under diagnosing and underreporting, the total number of cases is 20 times greater than the number of cases that are actually reported in each year.
Although V. parahaemolyticus infections have been linked to crayfish, shrimp, and crab consumption, the recent outbreaks in the United States and Canada were attributed predominantly to consumption of raw oysters. Food intake surveys indicate that raw shellfish is not a commonly consumed food in the U.S., with only about 10 to 20 percent of the population consuming raw shellfish at least once a year. Among raw shellfish consumers, a serving of raw oysters is eaten approximately once every six weeks, normally with a range of 6 to 24 oysters per serving.
Pathogenic strains of V. parahaemolyticus may have one or more distinctive traits compared with nonpathogenic V. parahaemolyticus strains, including the ability to produce thermostable direct hemolysin (TDH, a toxic substance that breaks down red blood cells). The vast majority of V. parahaemolyticus strains isolated from the stools of people with V. parahaemolyticus infections are TDH-positive, and recently improved methodology has enabled scientists to detect these strains in oysters when they are present in sufficient amounts. Other traits have also been linked to pathogenic strains, including the ability to invade the intestine, the ability to produce an enterotoxin (a toxin that causes diarrhea in animals), and the ability to produce urease, an enzyme that can help microorganisms survive the acidic conditions of the stomach. However, the role of the traits other than TDH in pathogenicity is still unclear.
Once present in the environment, V. parahaemolyticus levels can be affected by such factors as the amount of zooplankton in the shellfish growing area, the rate of tidal flushing, levels of dissolved oxygen in the water, the presence of pollutants, water temperatures and salinity levels. Oyster-specific factors, such as the physiology and health of the oyster, may also contribute to the ability of V. parahaemolyticus to infect and grow in the oysters. Toxins or other proteins produced by bacterial strains that infect oysters at the same time as V. parahaemolyticus may also affect the survival of the V. parahaemolyticus. Estimates from several studies suggest that the average percentage of pathogenic V. parahaemolyticus is higher on the West Coast (approximately 3%) than in other areas of the country (0.2 to 0.3%).
In summary, a review of published studies on V. parahaemolyticus suggests that a number of factors can affect the presence and growth of V. parahaemolyticus in oysters at the time of harvest. However, only water temperature and percent pathogenic V. parahaemolyticus were incorporated into the model because there was little published quantitative data available to introduce the other harvest-related factors. The model demonstrated that although both water temperature and salinity affect V. parahaemolyticus levels, salinity has a minor effect. Water temperature, on the other hand, has a major effect, with total V. parahaemolyticus levels increasing as water temperature increases. The model is constructed so that it reflects this relationship between V. parahaemolyticus levels and water temperature. Likewise, as noted above, the percentage of V. parahaemolyticus that is pathogenic varies from region to region in the U.S; the model also takes this variation into account. The final output of this component of the model is a set of regional and seasonal estimates of V. parahaemolyticus levels at time of harvest.
The model developed for the Post Harvest Exposure Assessment was also used to test the effects of several interventions that have been proposed for decreasing growth or presence of V. parahaemolyticus. Proposed intervention measures include rapid cooling of oysters immediately after harvest, and mild heating (5 minutes at 50°C) or freezing (35 days at -20°C) of oysters. The resulting data were subsequently used to determine the effect of these measures on the probability of illness.
To develop a quantitative dose-response model, this risk assessment relied on epidemiologically based estimates of illness rates and human clinical trial data from studies performed prior to 1974, in which human volunteers received specific doses of V. parahaemolyticus with concurrent antacid administration (which eliminates the protective effect of stomach acid). Analysis of the data indicated that the presence of antacid may have had a substantial effect on dose-response. Consequently, the dose-response model was modified using epidemiological data to adjust for the fact that the clinical trial participants were fed V. parahaemolyticus doses with antacids.
The risk assessment also reviewed a variety of human and animal studies of infection with V. parahaemolyticus (or with other Vibrio species that can serve as surrogates for V. parahaemolyticus infection). The review suggests that a number of factors can affect the infectious Vibrio dose. These factors include route of exposure (e.g., by mouth or by injection), food matrix factors (e.g., fat content), bacterial virulence factors (e.g., TDH production, enterotoxin production, invasion, presence of iron-limiting conditions), and host factors (immune status of the host).
Another aspect of dose-response relationships is the severity of infection, i.e., will the patient just develop gastroenteritis or will the patient go on to develop septicemia. People with underlying medical conditions (e.g., the immunocompromised population) are more likely to develop septicemia than people without underlying conditions. Using CDC data, the risk assessment model predicts the likelihood that a V. parahaemolyticus patient will develop septicemia, and estimates the expected number of cases of septicemia.
The risk characterization results are in the form of distributions, or ranges of possible cases. For simplicity’s sake, this summary presents the mean of these ranges; i.e., the mean number of predicted cases. (A more detailed description of the distributions can be found in the technical document.) For the Gulf Coast, the model predicts a mean of 25 cases of V. parahaemolyticus infections in winter, 1,200 in spring, 3,000 in summer, and 400 in fall. For the Pacific Northwest, the model predicts a mean of 15 cases in spring and 50 in summer; for the Mid-Atlantic, 10 cases in spring and 12 in summer; and for the North Atlantic, 12 cases in spring, 30 in summer and 7 in fall. The predicted number of cases in the North and Mid-Atlantic is low even for the summer, because the harvest sizes and water temperatures in these regions are relatively low compared with other regions. In fact, because the number of expected cases was so low, exact estimates could not be made for the Mid-Atlantic fall and winter and the North Atlantic winter harvests. The model also predicts that an average of six septicemia cases per year will occur in the U.S.
|
| Figure 1. Effect of intervention measures on reducing the predicted risk of V. parahaemolyticus
illnesses from Gulf Coast summer harvests: No mitigation ();
freezing (); heat treatment
(); rapid cooling
(). |
|
| Figure 2. The tornado plot shows the relative importance of parameters that influence risk of V. parahaemolyticus (Vp) illness in the non-Louisiana Gulf Coast summer harvest. |
The model can also be used to evaluate which environmental or processing factors have the strongest influence on probability of illness. An example of this type of evaluation (sensitivity analysis) is shown in Figure 2. In this graph, a tornado plot, factors are ranked by the magnitude of their effects (whether positive or negative). For example, Figure 2 shows that the most important factor in determining the risk of V. parahaemolyticus-associated illness per serving for Gulf Coast summer harvests is the level of total (pathogenic and nonpathogenic) V. parahaemolyticus in oysters at time of harvest. However, the model is based on a correlation between total and pathogenic V. parahaemolyticus levels at time of harvest and assumes that pathogenic strains of V. parahaemolyticus grow at the same rate as non-pathogenic strains. Consequently, as the level of total V. parahaemolyticus increases so does the number of pathogenic V. parahaemolyticus. The second most influential factor for the Gulf Coast is the time between harvest and refrigeration (time unrefrigerated), the third is the weight or amount of oysters consumed, and the fourth is the temperature of the water at harvest time. Note that length of refrigeration time has the sixth largest effect in terms of size, but that the bar for this factor points in the negative direction because V. parahaemolyticus levels decrease as refrigeration time increases. As in the Gulf Coast, the most important factor for the Pacific Northwest was also V. parahaemolyticus levels in oysters at harvest. However, water temperature at harvest time was the second most important factor, followed by the weight of oysters consumed and the length of refrigeration time. Time unrefrigerated-the third most important factor for the Gulf Coast-was substantially less important for the Pacific Northwest. For the Mid-Atlantic, V. parahaemolyticus levels in oysters, water temperature, and grams of oysters consumed were the top three factors. Time unrefrigerated was the fourth most important factor for this region.
Model results were validated by comparing predicted V. parahaemolyticus levels in oysters to real world data from a 1998-1999 collaborative Interstate Shellfish Sanitation Conference (ISSC)/FDA survey on V. parahaemolyticus levels in oysters at retail. These data were deliberately not used in the development of the model to provide an unbiased check of model outcome. In general, the mean V. parahaemolyticus levels predicted by the model compared well with the mean levels from the ISSC/FDA survey, particularly for the Gulf and Mid-Atlantic summers when the risk of illness is highest. For the Pacific Northwest, the model estimates are higher for the summer than the ISSC/FDA estimates and lower for the spring. The differences for the summer appear to be within the range of expected year-to-year variation, but the spring estimates are farther off, perhaps because the spring mean is based on a very small number of samples.
The risk assessment predicted the number of V. parahaemolyticus illnesses likely to occur in the future given a certain range of environmental and oyster handling conditions. These predictions are generally consistent with the CDC estimates for annual V. parahaemolyticus incidence.
The risk assessment also reviewed published studies on factors that may affect pathogenic V. parahaemolyticus levels in harvest waters and in shellfish, including water temperature, water salinity, contamination of shellfish-growing waters with ballast water, environmental contaminants, and immune status of oysters. Temperature was the only such factor incorporated into the quantitative model of V. parahaemolyticus illness. The model confirmed that water temperature at time of harvest was a major factor influencing initial V. parahaemolyticus levels in oysters and the number of V. parahaemolyticus illnesses. Air temperature was also found to greatly influence the growth of V. parahaemolyticus in oysters after harvest and thus, V. parahaemolyticus levels in oysters at the time of consumption.
The model also showed that the effect of different parameters varied by region and by season; e.g., water temperature was a more important determinant of illness levels in the Mid-Atlantic and Pacific Northwest than in the Gulf. However, for all regions, as one would expect, the more oysters one eats, the more likely it is that one will become ill.
The risk assessment model also demonstrated that the single most important factor related to the risk of illness caused by V. parahaemolyticus is the level of V. parahaemolyticus in oysters at the time of harvest. Accordingly, intervention measures aimed at controlling or reducing levels of V. parahaemolyticus in oysters should have a direct bearing on controlling or reducing the risk associated with this pathogen. Model simulations indicated that several proposed intervention measures would be effective in decreasing viable V. parahaemolyticus counts in oysters and the probability of V. parahaemolyticus illness. Cooling oysters immediately after harvest and quick freezing of oysters both reduced V. parahaemolyticus counts and illness estimates, while the effect of mild heat treatment on oyster levels practically eliminated the likelihood of illness. Modeling also showed that V. parahaemolyticus densities decreased slowly but steadily during refrigerated storage, and that frozen storage also decreased viable V. parahaemolyticus oyster densities. Both measures thus decreased the probability of illness due to V. parahaemolyticus. As noted above, the risk assessment was asked to evaluate the FDA guidance level of 10,000 cells/g. The model made it possible to develop a mathematical means of relating potential microbiological criteria with both the predicted percentage of illness prevented and the predicted percentage of the oyster landings that would no longer be available to consumers if the criterion could be implemented with 100% efficiency (Figure 3). This analysis suggests that in the absence of subsequent post harvest mitigations, "at harvest" guidance levels of 100,000, 1,000 and 100 total V. parahaemolyticus per g could (potentially) reduce the illness rate by 2%, 50% and 90% with corresponding losses of 0.3%, 25% and 70% of the harvest, respectively.
 |
| Figure 3. Potential effect of control of total V. parahaemolyticus per g at harvest (Louisiana Gulf Coast summer harvest). |
The risk assessment also reviewed the effect of preexisting conditions on the risk of developing V. parahaemolyticus gastroenteritis or V. parahaemolyticus-related septicemia. Epidemiological studies indicate that people with certain preexisting conditions, such as renal or liver disease, peptic ulcer, diabetes, or immunodeficiency, and people who use antacids or have had gastric surgery are more likely to advance from gastroenteritis to septicemia once infected with V. parahaemolyticus than are healthy people. However, other epidemiological studies suggest that patients with preexisting conditions are not more susceptible than healthy people to contracting V. parahaemolyticus gastroenteritis.
The recently adopted ISSC interim control plan for monitoring levels of pathogenic V. parahaemolyticus at harvest is intended to prevent the harvesting of oysters (for raw consumption) from growing areas where relatively high densities of pathogenic V. parahaemolyticus are present. The potential effectiveness of this control plan could not be evaluated quantitatively, because not enough data are available for a quantitative model. However, the sensitivity analysis indicated that V. parahaemolyticus levels in oysters at harvest is the most important factor in determining the risk of illness in the absence of subsequent post harvest mitigations such as mild heating or frozen storage.
The risk assessment identified important data gaps. These gaps, which need to be addressed by research or further data collection, include the following topics: the prevalence of pathogenic V. parahaemolyticus in shellfish and in shellfish harvest waters; factors other than temperature that influence the presence of pathogenic strains; the role of oyster physiology and immune status in levels of V. parahaemolyticus; growth rate and survival of V. parahaemolyticus in oysters; rates of water turnover in shellfish harvest areas (based on levels of freshwater flow, tidal changes, winds, and depth of harvesting area); the infectious dose for pathogenic strains, i.e. how many V. parahaemolyticus organisms are required to cause illness; the effects of stomach acid production, food matrix, or immune status on dose-response relationships; and consumer handling practices of oysters prior to consumption. In addition, improved global public health surveillance of V. parahaemolyticus would identify new epidemic strains as they emerge.
In summary, this draft risk assessment significantly advances our ability to describe the current state of knowledge about V. parahaemolyticus, an important foodborne pathogen. It simultaneously provides a framework for integrating new scientific knowledge and evaluating its impact on public health. The results of the assessment are influenced by the assumptions and data sets that were used for Exposure Assessment and Hazard Characterization. The FDA is actively seeking new information, scientific opinions, or data during the public comment period. FDA anticipates that periodic updates to the risk model will continue to reduce the degree of uncertainty associated with risk estimates, and that these updates will assist FDA in making the best possible decisions and policies for reducing the risk posed by V. parahaemolyticus in raw molluscan shellfish.
Hypertext updated by dav/cjm/dms 2001-JAN-30