Figure IV-1. Schematic depiction of the Harvest Module of the V. parahaemolyticus (Vp) risk assessment model.
Exposure assessment is the determination of the likelihood of ingesting pathogenic V. parahaemolyticus by eating raw molluscan shellfish harboring the organism and the amount of pathogenic V. parahaemolyticus present when consumed. Exposure assessment is subdivided into Harvest, Post Harvest, and the epidemiology and consumption segments of the Public Health Module.
The Harvest Module identifies the parameters contributing to the likelihood that shellfish in a growing area will contain disease-causing (pathogenic) strains of V. parahaemolyticus and the levels in which they're found. These parameters are listed below.
Vibrio spp. are found in the estuarine environment in the tropical to temperate zones. Several studies have been published on the concentration of V. parahaemolyticus in shellfish growing areas (35, 38, 70, 73, 74, 76, 77, 89). There are several pathways by which V. parahaemolyticus strains may be introduced into shellfish growing areas. V. parahaemolyticus may originate or new strains may be introduced naturally by terrestrial and aquatic animals, or through human activities such as "relaying" shellfish or releasing ballast water. Terrestrial and aquatic animals (including plankton, birds, fish, reptiles) may harbor virulent strains of V. parahaemolyticus and may play a role as intermediate hosts and vehicles for spread (118). V. parahaemolyticus has been isolated from a number of fish species where it is associated primarily with the intestinal contents (101). V. parahaemolyticus can be introduced into non-contaminated areas by relaying shellfish prior to commercial harvesting.
Ship ballast release may be a potential mechanism of introducing V. parahaemolyticus into a particular environment. Most cargo ships must carry substantial quantities (millions of gallons) of ballast water to operate safely when they are not carrying cargo. Cargo ships take on ballast water from the body of water in which the ship originates. Having taken water on board, it is normally retained until the ship is about to load cargo, at which point ballast water is discharged. During de-ballasting, organisms picked up from one port could be introduced into the loading port. Ship ballast may have spread the epidemic strain of V. cholerae to the U.S. Gulf of Mexico (93). Strains of V. cholerae indistinguishable from the Latin American epidemic strain were found in non-potable water taken from a cargo ship docked in the Gulf of Mexico. The same could occur for V. parahaemolyticus.
Sewage discharge may indirectly influence the densities of V. parahaemolyticus present in shellfish growing areas (141). For example, densities of V. parahaemolyticus in the water column in Narragansett Bay, Rhode Island were correlated with fecal coliforms from sewage; however, the effect of sewage was an indirect one mediated by stimulation of zooplankton with which the V. parahaemolyticus were associated. Laboratory studies showed that nutrients in the sewage did not directly increase V. parahaemolyticus levels (141). Other reports have shown that organic matter does have an effect on growth and survival of the organism (121). In another study, the distribution of V. parahaemolyticus in sediments in Boston Harbor was independent of densities of fecal coliforms (120).
Once introduced, a number of factors are relevant to whether V. parahaemolyticus will become established. These include interactions of environmental conditions, species and physiology of the shellfish, and the genetics of the microorganism. Certain areas may have more favorable environmental conditions that support establishment, survival, and growth of the organism. Predictive factors to be considered in determining the prevalence of V. parahaemolyticus include temperature (including El Niño and La Niña weather patterns), salinity, zooplankton, tidal flushing (including low tide exposure of shellfish) and dissolved oxygen (4, 49, 70, 137).
Warmer temperatures and moderate salinities, especially those prevailing during the summer months, favor the growth and survival of V. parahaemolyticus (31, 65, 101, 147). Most of the shellfish-borne illnesses caused by this organism also occur in the warmer months. The Centers for Disease Control and Prevention (CDC) randomly selected seven of the 76 (nine percent) existing Texas Department of Health monitoring sites for environmental conditions in Galveston Bay, and compared water temperature and salinity levels before and during the 1998 outbreak, with environmental data recorded over the previous five years (34). They demonstrated a significant difference in mean values. During May 1998, water temperatures were 81° Fahrenheit (F) compared with 76°F for the previous five years. In June, water temperatures were 85°F, compared with 83°F for the previous five years. Significantly less rainfall than usual, during April (0.59 inches) and May (0.02 inches) preceding the outbreak, causing extreme drought conditions in Texas, resulted in markedly increased salinity levels in Galveston Bay. During May, salinity levels were 18.3 parts per thousand (ppt) compared with 8.4 ppt for the previous five years. During June, salinity levels were 21 ppt compared with 9.1 ppt for the previous five years. It is therefore possible that environmental factors such as increased temperature and salinity levels, known to promote growth of V. parahaemolyticus, may have contributed to this outbreak. Elevated temperatures were also suspected to have played a role in the 1997 outbreak on the West Coast (22).
Another variable that must be considered is that V. parahaemolyticus often "over-winters" (survives the winter) in the sediment and is absent from the water column and oysters during the winter months (69, 75, 136). During the summer, shellfish often have levels of V. parahaemolyticus from 10- to 100-fold greater than those in the water (38, 73); therefore, sediment should be the preferred samples for monitoring during the winter and shellfish should be the preferred samples for monitoring during the summer. Under extreme environmental conditions, Vibrio species, including V. parahaemolyticus, may enter a "viable but non-culturable (VBNC) phase" in marine waters and could be missed by traditional cultural methods (15, 24, 110, 146). This issue remains a controversial one. Methods such as gene probes developed by the FDA are capable of detecting most virulent strains and are useful in monitoring programs (51).
V. parahaemolyticus favors the presence of particulates, zooplankton and other chitin sources (70, 102, 137). Microorganisms are incorporated into shellfish by filter feeding. Factors that favor active filter feeding by shellfish increase the probability that shellfish in a given area will take up the pathogen (100). Shellfish species and physiology (e.g., sexual maturity, immune function, metabolic state) can affect survival and growth of disease-causing Vibrio spp. within shellfish. There is evidence that the immune status of the shellfish may play an important role in the prevalence and persistence of the microorganism (45, 85, 86, 111, 138). There also appear to be seasonal differences in the oyster's cellular defense system. A recent study showed that the bactericidal activity of hemocytes (oyster blood cells) was greater in summer than in winter (50). Certain factors, such as the oyster parasite Perkinsus marinus, play a role in the affinity of bacteria for shellfish tissue and the ability of oyster hemocytes to kill the internalized organisms (85, 86, 126). Factors, such as spawning or adverse environmental conditions (e.g., the presence of chemicals in the environment: tributyltin oxide, polycyclic aromatic hydrocarbons, wood preservative leachates), that reduce or stop filter feeding in shellfish, or cause selective feeding (e.g., new nutrient sources) may prevent or delay incorporation of V. parahaemolyticus into shellfish by affecting oyster physiology and possibly affect oyster-bacterial interactions (124, 142, 143).
Persistence of virulent strains of V. parahaemolyticus in shellfish in the environment may be dependent on several parameters. Whether virulent and non-virulent strains are affected in a similar fashion by environmental and other factors is unknown. The presence of the urease gene may provide a competitive environmental advantage over other strains allowing access to a wider range of nutrients (1). Urease-positive strains have been identified as a predominant cause of Vibrio-associated gastroenteritis on the U.S. West Coast and Mexico (1). The presence of a pathogenicity island (a physical grouping of virulence-related genes) in V. parahaemolyticus may foster rapid microevolution, promote growth and survival, and result in transmission of factors, such as those responsible for virulence, to other strains (horizontal gene transfer) (47, 62, 63). In addition, bacteriophages may genetically alter vibrios (14, 62).
The distribution and variation in levels of virulent V. parahaemolyticus in shellfish and among shellfish growing areas may need to be determined before harvest because many of the described factors may have contributed to higher concentrations in certain areas. During the 1998 outbreaks, the Pacific Northwest shellfish harvested from the Hood Canal area of Washington were responsible for 32 of 48 (67 %) of the illnesses in the State of Washington (132). In the Gulf Coast, 20 of 30 harvest sites were implicated. In the Atlantic Northeast region, Oyster Bay Harbor (Area 47) was the only area implicated in the outbreak of that region (21).
|
Figure IV-1. Schematic depiction of the Harvest Module of the V. parahaemolyticus (Vp) risk assessment model. |
Although a number of factors have been identified as potentially affecting the levels of pathogenic V. parahaemolyticus in oysters at time of harvest, there are not sufficient quantitative data available to incorporate all of these factors into a predictive model. To incorporate an environmental factor into the simulation, as a predictor of V. parahaemolyticus densities at harvest, it is necessary to identify both the relationship of V. parahaemolyticus densities to the parameter of interest and the regional and temporal variation of the parameter within the environment. Moreover, due to the relatively low prevalence of pathogenic V. parahaemolyticus and limitations of current methods of detection, the distribution of pathogenic V. parahaemolyticus is not well understood. A critical issue in the development of the Harvest Module simulation is the use of the estimated distribution of total V. parahaemolyticus densities to bridge this data gap and derive an estimate of the distribution of pathogenic V. parahaemolyticus densities in oysters at harvest. Figure IV-1 is a schematic depiction of the parameters considered in modeling the Harvest Module. Preliminary modeling demonstrated that the parameter, water salinity is not as strong a determinant of V. parahaemolyticus levels as water temperature, and therefore is represented as a dotted bubble.
Effect of water temperature and salinity on total V. parahaemolyticus densities
The best available data on the relationship of total V. parahaemolyticus densities in oysters (and water) versus water temperature and salinity is found in the study by DePaola et al, 1990 (38). This study was conducted throughout an entire year with collection of samples from all four regions of the country (i.e., Northeast, Gulf Coast, Mid-Atlantic, and Pacific Northwest). A total of 65 paired samples of oyster and water were analyzed for total V. parahaemolyticus by a membrane filtration method. While there have been several other surveys of V. parahaemolyticus between 1982 to 1995, these studies are typically limited to specific regions and/or seasons, and few have reported quantitative data. These studies are summarized below.
Kelly and Stroh (78) examined V. parahaemolyticus frequency in natural and cultivated oysters from British Columbia and isolated V. parahaemolyticus from 44% of natural and 21% of cultivated oysters under warm conditions (July and August) but did not find V. parahaemolyticus in March and April.
Kelly and Stroh (79) also reported an association with V. parahaemolyticus illness and V. parahaemolyticus density in the estuarine waters of British Columbia. V. parahaemolyticus was isolated in 11-33% of water samples collected during the summer with peak densities of 70 cfu/ml. Oysters were not examined.
Kaysner et al. (74) sampled water, sediment and oysters of Willapa Bay, WA during August when salinity ranged from 23.6-30.5 ppt and temperature 15.5-22.6°C. Highest densities (log10 MPN/g) were found in sediments (1.6-5.4), followed by oysters (1.5-4.0) and water (0.5-3.0); a similar trend was observed with frequency of isolation.
Tepedino (130) surveyed Long Island oysters from October to June and found 33% to contain V. parahaemolyticus with an MPN range of 3.6-23/g.
Hariharan et al. (55) conducted a year long survey of Prince Edward Island, Canada mussels and oysters and V. parahaemolyticus was isolated from 4.7% and 6.7%, respectively.
Chan et al. (23) examined V. parahaemolyticus levels in seafood from Hong Kong from June through October. Mean V. parahaemolyticus densities in oysters (harvest), mussels (market) and clams (market) were 3.4 x 104, 4.6 x 104, and 6.5 x 103 per gram, respectively.
Kiiyukia et al. (81) enumerated V. parahaemolyticus in water and sediments of Japan. They isolated V. parahaemolyticus in 2/8 market oyster samples but did not enumerate V. parahaemolyticus in oysters.
Ogawa et al. (108) investigated the ecology of V. parahaemolyticus in Hiroshima Bay from July 1987 through June 1988. The highest incidence of detectable V. parahaemolyticus (68.8%) was found from May to October when water temperature ranged from 19.3 to 22.0°C. V. parahaemolyticus levels in oysters were seasonal and ranged from 103 - 101/100g (108). This study also compared favorably with the DePaola et al. 1990 study (38).
DePaola et al. (40) had previously evaluated 4 methods for enumeration of V. parahaemolyticus in natural seawater and oysters and found considerable variability between methods for V. parahaemolyticus recoveries; highest recoveries were obtained with a method using filtration through a hydrophobic grid membrane.
DePaola et al. in 1990 (38) enumerated V. parahaemolyticus (hydrophobic grid method) in seawater and oysters samples collected seasonally from May 1984 through April 1985 from shellfish growing areas from the Pacific, Gulf and Atlantic Coasts. Seasonal and geographical distributions of V. parahaemolyticus were related to water temperature, with highest densities in samples collected in the spring and summer from the Gulf Coast.
We considered the study by DePaola et al. (38) to be most appropriate for the purpose of quantitative risk assessment of V. parahaemolyticus illness from consumption of U.S. oysters. This study, which is the most comprehensive regional/seasonal study available, examined seasonal changes in V. parahaemolyticus density in oysters from major oyster producing areas representative of the Pacific, Gulf and Atlantic Coasts (38). Studies reporting only presence or absence of detectable V. parahaemolyticus are of limited value for quantitative risk assessment (55, 78, 79, 81). Of the additional studies available reporting quantified V. parahaemolyticus densities in oysters, samples were either obtained from a single estuary (23, 55, 74, 78, 79, 81, 130), were not seasonal (23, 74, 130), or did not report salinity and temperature (23, 130). Differences in methodology used by the various investigators may also have affected V. parahaemolyticus recoveries and complicate comparisons between studies. V. parahaemolyticus levels observed in oysters from Long Island, NY (130) were similar to those reported by DePaola et al. (38) from the Northern Atlantic Coast during the fall, winter and spring. Kaysner et al. (74) observed higher V. parahaemolyticus densities in Willapa Bay, WA in August than reported on the Pacific Coast during the summer by DePaola et al. (38). This difference may have been due to the small number of samples (N=4) collected from the Pacific Coast during the summer by DePaola et al. (38), or to favorable environmental conditions for V. parahaemolyticus abundance in Willapa Bay during the study by Kaysner et al. (74). Including the data from other studies of the Atlantic (130) and Pacific Coasts (74) would increase the sample size. However, these studies employed different methodology than that used by DePaola et al. (38) and inclusion of these data could bias comparisons between other seasons or regions that did not include data from these individual estuaries using different methods. Specifically, due to differences in method error associated with various analytical methods, statistical analysis of pooled data must account for the differences in variation of observed measurements according to the analytical methods used. Although this may be readily accomplished, the precision of estimating trends is not necessarily increased due to the necessity of estimating multiple sources of variation. Furthermore, lack of precise estimates of method error makes it difficult to estimate the population variation of V. parahaemolyticus densities (i.e., true variation in the absence of method error). Consequently, to maintain consistency only data from DePaola et al. (38) was used in the harvest module of this risk assessment.
The distributions of total V. parahaemolyticus densities in water and oyster samples were positively skewed. This is consistent with the almost universal observation that microbial populations in foods are lognormally distributed. Therefore the logarithm of the density being more normally distributed was regressed against temperature and salinity. V. parahaemolyticus was not detected in a relatively large proportion of samples (e.g. 19 of 61 oyster samples (31%)). Some of these samples are likely to have been false-negatives due to limitations of the method. In order to avoid upward bias of predicted levels at low temperatures the estimate of the regression line of log10 total V. parahaemolyticus/g oyster meat was obtained by the censored or Tobit regression method. The Tobit regression is a maximum likelihood procedure with likelihood reflecting both the probability of obtaining a nondetectable outcome at a given temperature as well as the probability distribution of observable densities given that a sample has detectable V. parahaemolyticus. The effect of this likelihood structure is to weight the influence of nondetectable outcomes on estimated trends differently in comparison to samples with quantifiable densities. The influence of nondetectable outcomes is based on the probability of the density of a sample falling below a fixed limit of detection rather than the assumption that a nondetectable measurement corresponds to an observed and quantifiable density at the limit of detection or one-half the limit of detection as is commonly assumed.
In the reanalysis of the DePaola et al. study (38), the effect of temperature on mean log10 total V. parahaemolyticus densities was found to be approximately linear over the range of environmental water temperatures. The presence of a quadratic effect in temperature was not evident (i.e., not significant). With regard to salinity, a quadratic effect was found to be significant, suggesting that V. parahaemolyticus increase with increasing salinity up to an optimal level and then decrease with increasing salinity thereafter. There was no significant interaction between temperature and salinity evident based on the data. Consequently, the best fitting model obtained was of the form
where TEMP denotes temperature in °C; SAL denotes salinity in parts per thousand (ppt); 
, 
,
1, and 2 are regression parameters for temperature and salinity effects on mean log10 densities,
and 
is a random normal deviate with zero mean and variance 
2 corresponding to the combined effects of population and method error variation.
The resulting parameter estimates were
= -2.6
= 0.12
1 = 0.18
2 = -0.004
2 = 1.0
The estimated relationships between total V. parahaemolyticus densities in oysters versus water temperature and salinity are shown in Figures IV-2 and IV-3, respectively.
Figure IV-2. Observed log10 V. parahaemolyticus (Vp) densities in oysters versus water
temperature at different salinities.
Figure IV-3. Observed log10 V. parahaemolyticus (Vp) densities in oysters versus salinity at different temperatures.
|
Both salinity and temperature effects were significant based on the regression. The variation of observed values about the predicted mean regression line shown in Figure IV-2 is attributable to the effects of salinity as well as the variation about the mean due to population variation and method error. This regression line gives the predicted mean levels versus temperature at a predicted optimal salinity of 22 parts per thousand (ppt). Similarly the variation of the observed data about the regression curve (parabola) for salinity effect shown in Figure IV-3 is partially attributable to differences in water temperature in addition to population and method error variation about the mean.
Extremes of salinity below 5 ppt are known to be detrimental to survival of V. parahaemolyticus. However, the influence of salinity within a range of moderate environmental salinities (i.e., 5-35 ppt) is not as clear. Based on the regression analysis, a quadratic relationship for V. parahaemolyticus densities versus salinity within the 5-35 ppt range is consistent with the DePaola et al. data (38). However, this projected effect of salinity is not as strong as that of temperature. Within a broad range around the optimal salinity of 22 ppt, the results of the regression suggest that the differences in salinity actually encountered in oyster harvesting have relatively little effect on the V. parahaemolyticus population (Figure IV-4).
Figure IV-4. Effect of salinity on predicted mean log10 V. parahaemolyticus (Vp) density in oysters and water relative to predicted density at optimal salinity (22 ppt). |
Clearly, in order to predict the distribution of V. parahaemolyticus densities at harvest based on the regression and the projected influence of water temperature and salinity in the environment, representative distributions of water temperature and salinity need to be estimated. Based on near-shore buoy data available from the National Buoy Data Center, regional and seasonal distributions of water temperature were available. However, representative data concerning the variation of salinity in shellfish growing areas were not identified. Consequently the effect of salinity was not incorporated into the present simulation.
Two considerations suggest that neglecting the effect of salinity does not adversely affect the predictive value of a model based on temperature alone. First, as shown in Figure IV-4, predicted mean V. parahaemolyticus densities vary by less than 10% from the optimal (maximum) density as salinity varies from 15 to 30 parts per thousand (ppt). Secondly, measurements of oyster liquor salinity at the retail level (44), which are strongly correlated with salinity of harvest waters (44), suggest that oysters may be harvested from the more saline areas of the estuaries year round. The mean oyster liquor salinity in the ISSC/FDA survey was found to be 24 ppt with a standard deviation of 6.5 ppt based on 249 samples. The study was conducted year round with samples obtained from all regions of the country. These two considerations suggest that the effect of variation of salinity on predicted distributions of V. parahaemolyticus densities would be minor. Variations in salinity between 15 and 30 ppt would increase the variance of the predicted distribution by only a small amount.
Neglecting the effect of variations in salinity in the simulation can be accomplished in either of two ways. Either salinity can be fixed to a mean value (i.e., 22 ppt) in the regression relationship derived above or the prediction of V. parahaemolyticus densities can be based on a regression analysis of the DePaola et al. data (38) with water temperature as the only effect in the model. With water temperature as the only effect the regression equation is:
where TEMP denotes temperature in °C, and are regression parameters for temperature effect on mean log10 densities, and is a random normal deviate with zero mean and variance 2 corresponding to the
combined effects of population and method error variation.
Parameter estimates obtained based on the Tobit estimation method are
= -1.03
= 0.12
2 = 1.1
Based on the data, the estimate of the variance about the mean (2) is an inflated estimate of
population variation due to method error. An estimate of population variation about the mean is
obtained by subtracting out an estimate of the method error. The membrane filtration method
used in the DePaola et al. study (38) was the HGMF procedure developed by Watkins et al.
(140) and latter revised by Entis (41). When all suspect colonies are tested for confirmation, the
precision of the HGMF method has been shown to be somewhat greater than the 3 tube MPN
(most probable number) procedure (41, 140). In the DePaola et al. study (38), enumeration of
V. parahaemolyticus colonies was based on testing of five suspect colonies. Consequently,
enumeration was not as precise as possible and overall method error associated with estimating
V. parahaemolyticus densities may have been more comparable to that of a 3 tube MPN
procedure. An estimate of the method error variance of the 3 tube MPN procedure is 0.35 (39)
and this value was considered a reasonable estimate of the method error for the DePaola et al.
study (38).
Figure IV-5. Observed log10 V. parahaemolyticus (Vp) densities in oysters versus water temperature at different salinities. (<10 ppt (  ), 10 to 20 ppt ( ) and >20 ppt ( ) in comparison to predicted log10 densities (solid line) and 95% confidence limits (dashed lines) based on temperature only regression model).
|
The predicted mean log V. parahaemolyticus level versus temperature for the temperature only regression is shown in Figure IV-5. Clearly, this relationship is comparable to that which would be obtained by fixing the salinity to a near optimal value (22 ppt) in the prediction equation based on both water temperature and salinity. The temperature only regression was used to model the relationship between temperature and density of total V. parahaemolyticus at time of harvest.
Water Temperature Distributions
Regional and seasonal distributions of water temperatures were developed based on accumulated records from coastal water buoys (National Buoy Data Center (NBDC) data). Seasons were defined by calendar month: winter (January through March), spring (April through June), summer (July through September), and fall (October through December). For each region and season a shallow water buoy was selected as being representative of the water temperature distribution for oyster harvest areas within that region/season combination. The available database for most buoys has hourly water temperatures from 1984 up to the present, with occasional data gaps due to instrumentation malfunction. The correlation between water temperature and the ambient air temperature that oysters are subject to after they are harvested was accounted for by selecting buoys for which air temperature records were also available.
Considering that oyster harvesting outside of the Pacific Coast region commences early in the morning and ends mid or late afternoon, the daily water temperature recorded at noon was considered to represent an average daily temperature.
The distribution of these "average" temperatures within a given region and season varies from year-to-year with wider variations occurring during the transitional seasons of spring and fall.
Within a given year, the distribution of the noontime water temperature was found to be unimodal within a given range. This empirical distribution is adequately approximated as a normal distribution provided that no weight is given to implausible values outside the historical range of values that may be expected. Differences in these distributions from one year to the next are evident in the buoy data. We have characterized this year-to-year variation in the water temperature distributions by calculating the central tendency and variation in both the mean and standard deviation of these distributions. The buoys selected and the summary statistics calculated are shown in Table IV-1.
| Region | Seasonal Water Temperature Distributions (°C) | |||
|---|---|---|---|---|
| Winter (Jan - March) |
Spring (April - June) |
Summer (July - September) |
Fall (October - December) | |
| Northeast Atlantic (Ambrose buoy, NY harbor) |
mean( ) a = 4.51
mean( ) = 1.23
variance( ) = 1.04
variance( ) = 0.23
corr( ,) = -0.14 |
mean() = 12.0
mean( ) = 4.2
variance( ) = 0.74
variance( ) = 0.34
corr( ,) = 0.57 |
mean() = 20.7
mean( ) = 1.34
variance( ) = 0.86
variance( ) = 0.22
corr( ,) = -0.25 |
mean() = 12.0
mean( ) = 3.37
variance( ) = 0.73
variance( ) = 0.36
corr( ,) = -0.08 |
| Mid-Atlantic (Thomas Point Lighthouse buoy, Chesapeake Bay) |
mean() = 3.92
mean( ) = 1.92
variance( ) = 1.0
variance( ) = 0.21
corr( ,) = -0.31 |
mean() = 16.8
mean( ) = 5.1
variance( ) = 0.56
variance( ) = 0.34
corr( ,) = -0.16 |
mean() = 25.0
mean( ) = 1.8
variance( ) = 0.25
variance( ) = 0.12
corr( ,) = 0.47 |
mean() = 11.6
mean( ) = 5.1
variance( ) = 1.0
variance( ) = 0.85
corr( ,) = -0.28 |
| Gulf Coast (Dauphin Island, AL buoy) |
mean() = 14.2
mean( ) = 2.7
variance( ) = 1.54
variance( ) = 0.27
corr( ,) = -0.08 |
mean() = 24.5
mean( ) = 3.5
variance( ) = 0.98
variance( ) = 0.27
corr( ,) = -0.55 |
mean() = 28.9
mean( ) = 1.5
variance( ) = 0.11
variance( ) = 0.11
corr( ,) = -0.41 |
mean() = 17.9
mean( ) = 4.5
variance( ) = 3.2
variance( ) = 0.55
corr( ,) = -0.53 |
| Pacific Northwest (Washington State Shellfish Specialists) |
mean() = 8.1
mean( ) = 1.62
variance( ) = 0.76
variance( ) = 0.13
corr( ,) = 0.01 |
mean() = 13.7
mean( ) = 2.4
variance( ) = 1.0
variance( ) = 0.24
corr( ,) = 0.7 |
mean() = 17.4
mean( ) = 2.4
variance( ) = 0.60
variance( ) = 0.16
corr( ,) = -0.13 |
mean() = 10.7
mean( ) = 2.8
variance( ) = 0.16
variance( ) = 0.13
corr( ,) = 0.36 |
| Source of data: National Buoy Data Center (NBDC)
http://www.seaboard.ndbc.noaa.gov/Maps/Wrldmap.shtml and Washington State shellfish specialist N. Therien, personal communication (131) NBDC measures surface water temperature (sensors are generally 1.0 to 1.5 meter deep) a and denote mean and standard deviation of within region/season temperature distribution, respectively; mean(), variance(), and corr() denote the mean, variance and correlation between the parameters and across different years | ||||
In Table IV-1, and denote the population mean and standard deviation of the distribution of
water temperatures within any particular year for different region and season combinations.
The extent of year to year variation of these distributions is summarized by the mean and the
variance of the parameters and . The mean and variance of these parameters are denoted in
the table as mean(), variance(), mean() and variance(), respectively. The correlation
between and is denoted by corr(,). A positive correlation between parameters and
summarizes the observation that when the mean water temperature is higher than normal the
variation in temperatures from one day to the next is generally greater than that observed when
the mean temperature is lower than normal. Similarly, a negative correlation summarizes the
observation that temperatures are less variable when the mean water temperature is higher than
normal.
For example, the NBDC buoy located at Dauphin Island, Alabama was chosen as representative
of water temperatures for the Gulf Coast. Among other meteorological parameters, this buoy
has recorded water and air temperatures from 1987 to the present time. In reference to Table
IV-1, for the spring season (defined as April through June), the distribution of noontime water
temperature was found to vary from year to year with a typical (or average) mean of 24.5o C
[mean()]. The variance of the mean from one year to the next was 0.98o C [variance()] which
corresponds to a standard deviation of 0.99o C. Similarly, for the standard deviation of the
within year temperature distributions, the central tendency across different years was an average
of 3o C [mean()] with a variance of 0.27o C [variance()]. The correlation between and
[corr(, )] was -0.55 indicating that the day-to-day temperatures were less variable when the overall mean temperature was higher than that of a typical year.
For the Pacific Coast there were no near-shore NBDC buoys recording water temperatures that could be considered representative of oyster growing areas. Consequently, for this region, seasonal and year-to-year variations in water temperature distributions were developed based on compiled data from WA State shellfish specialists (Washington State Department of Health) from 1988 through 1999. These water temperature data were recorded in association with collection of samples for monitoring of vibrios and fecal coliforms and are therefore directly representative of temperatures for oyster growing areas. Averages of water temperature were substituted when multiple measurements were recorded for any given day. Year-to-year variations in the water temperature distributions for the Pacific Coast were developed in the same manner as that for the other regions.
Additional sources of information concerning water temperatures (and salinity) in oyster growing areas include the EPA STORET (Storage and Retrieval of U.S. Waterways Parametric Data) database (http://www.epa.gov/OWOW/STORET/) and the National Estuarine Reserve Sites (NERR) program (http://inlet.geol.sc.edu/cdmoweb/home.html). In comparison to the NBDC sites, STORET and NERR are more specific to estuaries as opposed to open coastal waterways. Some NBDC sites such as Thomas Point Lighthouse (Chesapeake) are located within estuaries but similar sites could not be identified for the Gulf Coast and Northeast Atlantic within the NBDC database. Comparison of NERR data for Weeks Bay, AL versus that of the Dauphin Island NBDC buoy suggests that shallow water estuaries may be slightly warmer than open coastal waters but that the difference is not substantial (i.e., ~1°C difference on average). An additional consideration is the availability of enough long-term historical data to determine extent of year-to-year variation. As already indicated, data is available from most NDBC buoys from 1988 to the present. The NERR program only started data collection in 1995. Although STORET has considerable long term historical data associated with monitoring of water quality dating back to 1964, access to STORET records is not readily available at present and the data could not be accessed during the time frame of the risk assessment. Also, STORET records do not necessarily correspond to fixed locations, as is the case for NBDC and NERR.
Additional data on water temperature (and salinity) measurements specific to oyster harvesting areas were made available to the risk assessment team by State agencies in Texas, Alabama, New York, and Connecticut. Water temperatures provided were not substantially different from the NBDC data selected for each region.
Prediction of the distribution of pathogenic V. parahaemolyticus densities
Table IV-2 shows estimates of the percentage of total V. parahaemolyticus isolates that have been found to be pathogenic in several studies. The estimate based on studies by Kaysner and colleagues applies to the Pacific Northwest (76) with the other estimates in Table IV-2 being appropriate for all other areas of the country. The estimates suggest that the average percentage of V. parahaemolyticus that are pathogenic relative to total V. parahaemolyticus, on the West Coast is ~3% and that the average percentage pathogenic in the Gulf Coast and other areas of the country is 0.2 to 0.3%.
| Oyster samples (12 oyster composites) containing detectable pathogenic V. parahaemolyticus (TDH+ or KP+)a | V. parahaemolyticus isolates that are TDH+ or KP+ |
Source | ||||
| Number oyster samples with pathogenic Vp | Number tested | Percent oyster samples with pathogenic Vp | Number isolates pathogenic | Number isolates tested | Percent isolates pathogenic | |
| 8 TDH+ | 193 | 4.1 | 9 TDH+ | 3233 | 0.3% | ISSC/FDA retail study (unpublished) (44) |
| NDb | 153 oyster, water, & sediment samples tested for KP+ | ND | 4 KP+ | 2218 | 0.18% | Galveston Bay , TX (133) |
| 4 TDH | 25 | 16 | 10 TDH | 308 | 3.2% | Grays Harbor, WA (73)
Puget Sound, WA (76) |
| 3 TDH | 96 | 3.1 | 10, 140, and 10 cfu/g in three samples | ND | 0.3% | FDA study of Texas outbreak Galveston Bay, TX (37) |
|
a KP+ - Kanagawa-positive; TDH+ - thermostable direct hemolysin-positive, a toxin produced by Vibrio
parahaemolyticus that lyses red blood cells in Wagatsuma agar. These terms are interchangeable in defining
pathogenicity of V. parahaemolyticus
b ND = not determined | ||||||
There is considerable uncertainty with regard to the average percentage of total V. parahaemolyticus isolates that are pathogenic due to the relatively small sample sizes for estimating such a small percentage. Furthermore, this percentage is likely to vary somewhat from one year to the next. Even if an average percentage were known with certainty, this information together with the estimated distributions of total V. parahaemolyticus densities is not sufficient to identify the distribution of pathogenic V. parahaemolyticus densities. It is likely that the density of pathogenic strains is spatially and temporally clustered in the environment to some degree. The average number of isolates that are pathogenic does not identify the extent of this clustering.
To account for the probable spatial and temporal clustering of pathogenic strains relative to total V. parahaemolyticus densities, we have assumed a beta-binomial distribution for the number of pathogenic V. parahaemolyticus at the time of harvest. Under a beta-binomial distribution the percentage of total V. parahaemolyticus which are pathogenic varies from one sample of oysters (e.g. 12 oyster composite) to the next. Given the occurrence of outbreaks this appears to be a reasonable assumption but cannot be validated directly since extensive quantitative surveys of pathogenic V. parahaemolyticus densities are not available. Specifically, based on the number of total V. parahaemolyticus (Vptotal), within a given composite, the number of pathogenic (Vppath) present is assumed to be distributed as a binomial random variable with Vptotal trials (size parameter) and a probability of success (p) distributed as a beta random variable. The distribution of the probability parameter p is called a mixing distribution and the variation of this parameter across composites of oysters induces a clustering of pathogenic strains relative to total V. parahaemolyticus.
Formally this beta-binomial model is expressed as:
The notation here indicates that the distribution of the number of pathogenic V. parahaemolyticus present is conditional on the number of total V. parahaemolyticus present (n). The mean and variance of this conditional distribution are:
![E[Vp path | Vp total = n] = alpha / alpha + beta * n and Var[Vp path | Vp total = n] = n * [alpha * beta / (alpha + beta) squared (1 + 1 / alpha + beta + 1 (n - 1))] = n * [alpha * beta / (alpha + beta) squared (1 + phi * (n-1))]](vrskeq4.gif)
where E[•] and Var[•] denote the mean and variance, respectively. The parameter 
is called the overdispersion parameter. The parameters and of mixing distribution in the beta-binomial can expressed in terms of the average percentage of isolates which are pathogenic (P),
which is the mean of the mixing distribution, and the dispersion parameter :
From Table IV-2, best estimates of the parameter P are 0.03 for the West Coast and 0.002 for
other regions of the country. The information is more limited with respect to the value of the
shape parameter . This parameter pertains to the variation of frequency of pathogenic V.
parahaemolyticus across different oyster samples or composites. Based on the data on
frequency of pathogenic isolates, Bayes estimates of the parameters and are:
where r is the number of pathogenic isolates and n is the total number of isolates. These estimates differ for the West Coast versus other regions of the country in the same manner as does average percentage pathogenic (P). An estimate of the dispersion parameter is:
Based on the data shown in Table IV-2, estimates of the dispersion parameter are 0.0032 for the West Coast and 0.00045 for the other regions of the country.
To the extent that the average percentage of isolates that are pathogenic is uncertain, and may
vary from year to year, P was evaluated as an uncertainty parameter in the Monte Carlo
simulations. The uncertainty was modeled as a triangle distribution with a different mean and
range for the Pacific Northwest than for other regions of the country. For the Pacific Northwest
the average percentage pathogenic was estimated to be 3% and the minimum and maximum of
the distribution was taken to be 2% and 4%, respectively. For all other regions of the country
the average percentage pathogenic was estimated to be 0.2% and the corresponding minimum
and maximum of the distribution was 0.1% and 0.3%, respectively. Uncertainty with regard to
the shape parameter was not evaluated.
Overall, the Monte Carlo simulation of the distribution of pathogenic V. parahaemolyticus present in oysters was preformed as follows. For each region and season the mean and standard deviation of water temperature distributions were sampled based on the bivariate normal distributions given in table IV-1. Each random sample from these distributions represents a distribution of water temperature (i.e., for different years). Given a water temperature distribution, the distribution of total V. parahaemolyticus densities in composites of 12 oysters at harvest was simulated by (a) sampling from the distribution of water temperature; (b) using the regression relationship to calculate a mean density corresponding to each sampled water temperature; and (c) perturbing the calculated means by a random normal deviate corresponding to the estimate of population variation of the densities. The distribution of pathogenic V. parahaemolyticus densities was derived from that of total V. parahaemolyticus assuming a beta-binomial model for the extent of clustering of pathogenic relative to total counts. Multiple simulations were run with different values of average percentage of isolates pathogenic in order to evaluate the uncertainty with regard to this parameter.
The Post Harvest Module describes the effects of typical industry practices, including transportation, handling and processing, distribution, storage, and retail, from harvest to consumption, on V. parahaemolyticus densities in oysters harvested from various locations and seasons. Factors considered as possible influences on the levels of pathogenic V. parahaemolyticus at consumption include: ambient air temperatures at time of harvest; time from harvest until the oysters are placed under refrigeration; time it takes the oysters to cool once under refrigeration, and length of refrigeration time until consumption. This module also describes possible intervention strategies, such as mild heat treatment, freezing, hydrostatic pressure, depuration, and relaying, which could reduce V. parahaemolyticus densities.
Although the ecology of V. parahaemolyticus has been studied extensively (69, 70), little is known about the growth and survival of V. parahaemolyticus in shellstock oysters (30) or the effectiveness of mitigations aimed at reducing V. parahaemolyticus levels. The effects of post harvest storage on V. vulnificus growth in oyster shellstock (26, 27), and the effectiveness of various mitigation strategies for reducing V. vulnificus have been studied more extensively (29, 42, 99, 113, 122). Similar approaches are currently under investigation for V. parahaemolyticus and some preliminary data are included in this section.
The National Shellfish Sanitation Program (NSSP) time/temperature matrix for control of V. vulnificus requires oyster harvesters from any state, which previously had two or more confirmed cases of V. vulnificus to refrigerate oysters within 10 hours (h) after harvest during summer months, depending on water temperature. This provides approximately 10-fold reduction in V. parahaemolyticus growth relative to 20 h required in other months and on other coasts.
Proposed mitigation processes such as mild heat and freezing, which have been shown to be effective in reducing V. vulnificus levels, would probably have a similar effect on V. parahaemolyticus but only limited data is currently available (29). Other possible strategies include irradiation, high pressure, depuration and relaying. However, we have no relevant data on their effectiveness on V. parahaemolyticus in shellstock oysters, and had to rely on data from studies on V. vulnificus.
Reducing time to refrigeration
It has been shown from the literature that a reduction in the extent of growth of 0 to 10,000-fold in V. parahaemolyticus densities could be achieved depending on the initial V. parahaemolyticus levels, ambient air temperature and time to refrigeration (30, 51, 66, 67).
Mild heat treatment
A 6-log reduction of natural V. vulnificus population was achieved by heating shucked oysters for 5 min at 50°C (29). Similar heat sensitivity was observed between V. parahaemolyticus and V. vulnificus (51). Assuming that V. parahaemolyticus responds similarly to heat as. V. vulnificus, a 4.5 to 6-log (1,000,000-fold) reduction of V. parahaemolyticus densities could be expected by treating oysters for 5 min at 50°C.
Freezing treatment
A 1973 study reported a two-stage mortality for V. parahaemolyticus with an initial stage of cold shock followed by a second stage related to the frozen storage conditions (66). Estimates of effect of cold shock and frozen storage conditions were obtained by regression analysis of the observed data. Based on the analysis, freezing combined with frozen storage for 30 days at -30°C and -15°C is projected to result in a 1.2 and 1.6 log10 reduction of V. parahaemolyticus numbers in oysters, respectively. A similar decline (2 to 3 logs) of V. parahaemolyticus (natural population and dosed with pathogenic O3:K6 serotype) was observed in oysters frozen 35 days at -20°C (25). Freezing combined with frozen storage for 30 days would be expected to produce approximately a 2 log reduction of pathogenic V. parahaemolyticus. Both pathogenic strains (TDH+) and non pathogenic (TDH-) V. parahaemolyticus respond similarly to freezing (25).
Depuration
In the United States, depuration is conducted exclusively with UV light disinfection (113). There is a broad spectrum of conditions under which shellfish are depurated. Optimal times, temperatures and salinities for effective depuration vary among shellfish species. Published literature has shown that depuration appears to have no significant effect on decreasing the level of Vibrio spp. in naturally infected oysters or clams, and these microbes may even multiply in depurating shellfish, tank water, and plumbing systems (42, 53). A 1 log reduction of V. parahaemolyticus was observed in the hardshell clam, Mercinaria mercinaria, after 72 h of depuration at room temperature (53), and >2 log reduction at 15°C (52). Son and Fleet (122) observed a 5 log reduction in lab-infected oysters (from 9x107 to 8x102 in 72 h). Eyles and Davey (42) showed no difference (p<0.1) before and after depuration in naturally infected oysters.
Relaying
Relaying is the process by which shellfish are cleansed by transferring them to clean shellfish growing areas. There is little data available on this approach, which is also problematic as V. parahaemolyticus is ubiquitous in estuarine environments. Son and Fleet (122) demonstrated a decrease from 18 V. parahaemolyticus/g to < 5 V. parahaemolyticus/g after 6 days.
Figure IV-6. Schematic depiction of the Post Harvest Module of the V. parahaemolyticus (Vp) risk assessment model. |
The purpose of modeling the Post Harvest Module is to simulate the effects of typical industry practices on the levels of V. parahaemolyticus in oysters from harvest to consumption for various locations and seasons. The module also simulates the effect of intervention strategies. The input to the module is the regional and seasonal distributions of total and pathogenic V. parahaemolyticus at harvest. The output of the module is a series of predicted distributions of the total and pathogenic densities at time of consumption. Figure IV-6 represents a diagrammatic representation of the parameters modeled in this section. The baseline prediction is the distribution of density of V. parahaemolyticus (in 12 oyster composites), assuming current industry practices and no intervention.
The principle assumption used to develop the relationships between densities at harvest and densities at time of consumption is that the growth and survival of pathogenic V. parahaemolyticus is the same as total V. parahaemolyticus. Although no definitive studies of the growth characteristics of pathogenic V. parahaemolyticus are available, preliminary data suggest that there is little difference between growth characteristics of pathogenic versus nonpathogenic strains (37). Furthermore, observation of the growth of total V. parahaemolyticus in oysters is limited to only one temperature (26°C). To bridge this data gap we have used a model of V. parahaemolyticus growth in broth developed by Miles et al. (95). The predictions of this model have been adjusted to predict the growth rate in oysters, which is less than that of broth model systems possibly due to the influence of competing microflora.
Growth of V. parahaemolyticus from harvest to first refrigeration
The extent of growth that occurs during the period of time from harvest until the time that oysters are first placed under refrigeration is determined by three factors: (a) the growth rate of V. parahaemolyticus as a function of temperature; (b) the temperature of oyster meat following harvest and (c) the length of time held unrefrigerated.
Growth Rate Model
Miles et al. (95) modeled the growth rate of V. parahaemolyticus based on studies of four
strains at different temperatures and water activity, which is a measure of the availability of free
water in the broth model system. Worst case estimates of growth were obtained based on the
fastest growing of the four strains studied. For each combination of temperature and water
activity, the extent of bacterial growth observed was modeled using the Gompertz function.
This is a sigmoid growth curve with a growth rate (slope) monotonically increasing up to a
maximum and then falling to zero as the bacterial population reaches a steady-state. The
maximal rate of growth (m) is the most relevant summary of the fit because the growth rate
approaches the maximal growth rate rapidly and does not decline significantly until steady-state
is reached.
A secondary model was used to estimate the effect of environmental parameters on the maximal growth rate. This model was assumed to be of the square root type:
![square root of mu_m = b * (T - T_min) * [{1 - exp(c * (T - T_max))} * square root of (a_w - a_w,mix) * [1 - exp(d * (a_w - a_w,max))]] / square root of ln(10)](vrskeq8.gif)
where
m = maximal growth rate (log10 per minute)
aw = water activity
T = temperature (in degree Kelvin)
Based on the data from the fastest growing strain the estimates of the parameters were:
b = 0.0356
c = 0.34
Tmin = 278.5
Tmax = 319.6
aw,min = 0.921
aw,max = 0.998
d = 263.64
The parameters Tmin, Tmax, aw,min, and aw,max denote the range of temperatures and water activity over which growth can occur. The authors validated their model by comparison of model predictions with observed rates in eight other studies of growth in broth model systems obtained from the literature.
A plot of the resulting model prediction for m as a function of either temperature or water
activity is a unimodal function with a maximum value and zero growth rate outside of the
predicted range of temperatures and water activity favorable for growth. To use this
equation as a prediction of growth rate in oysters we assumed that water activity of oysters
does not vary substantially and have fixed this parameter to the optimal value of 0.985
predicted for the broth model system. At this water activity, the predicted growth rate in
broth at 26°C is 0.84 log10 per hour which is approximately a 7-fold increase in density per
hour. This is four times greater than the rate of growth observed for V. parahaemolyticus in
oysters held at 26°C (51).
Based on this observation, our best prediction of the growth rate in oysters at temperatures other than 26°C was obtained by dividing the predicted rate for broth model by a factor of four. This assumes that the growth rate in oysters is a constant fraction of the growth rate in broth at all temperatures. We have evaluated the influence of this assumption in the risk assessment by considering this factor as an uncertainty parameter varying according to a triangle distribution in the range of 2 to 8 with a mean of 4. This evaluates the sensitivity of our conclusions to the magnitude of the relative growth rate in oysters versus broth model but does not fully address the uncertainty in so far as it is conceivable that the relative growth rate could be temperature dependant.
The use of the Gompertz function by Miles et al. (95) to model bacterial growth in broth is appropriate. After transfer of an inoculum to different medium or environmental conditions there is a demonstrable lag phase during which the bacterial population adapts to different environmental conditions and growth is suboptimal (24). However, the Gompertz is not an appropriate model for growth of V. parahaemolyticus in oysters after harvesting, as changes in environment are typically gradual and do not arrest the growth rate and induce a lag phase. Consequently, for oysters, the extent of growth occurring over time at a given average temperature and predicted maximal growth rate is assumed to follow a simple three-phase loglinear model with no lag phase (19). This model is of the form:
log10(N(t)) = min{log10(N(0)) + m * t,A}
Figure IV-7. Predicted loglinear growth of V. parahaemolyticus (Vp) from initial density of 1,000 (3 log10) Vp/g as a function of ambient air temperature. |
where N(t) refers to the bacterial density at a given time (t) post harvest, A is the logarithm
of the maximum attainable density of V. parahaemolyticus in oysters, and the parameter m
is a function of ambient temperature as described above. At 26°C, the density of V.
parahaemolyticus in oysters was observed to approach a plateau of approximately 6.0 log10
per gram after 24 hours (51). We have assumed this value for the maximal density (A) at all
temperatures. Figure IV-7 shows predictions of the log10 increase in V. parahaemolyticus
density from an initial level of 1,000/g as a function of time for three ambient temperatures
(20, 26 and 32°C).
Ideally, the average temperature used to determine the parameter m in the above equation is
the temperature of oyster meat of shellstock. Clearly the temperature of oyster meat
depends on the temperature of both the air and water at the time of harvest. Temperature of
the oyster meat after harvest is expected to gradually equilibrate with the temperature of the
air and may be modified somewhat by evaporative cooling and the extent to which oysters
are properly shaded from direct sunlight aboard ship. In the absence of information to the
contrary, we have assumed that the temperature of oyster meat equilibrates rapidly with that
of the ambient air and have therefore used air temperature as a surrogate for oyster meat
temperature. Ambient air temperature data recorded at noon from the near-shore NBDC
buoys representative of various coastal regions were used for this purpose.
Distribution of ambient air temperature
Examination of water and air temperatures obtained from the NOAA/NBDC database showed a strong correlation between water and air temperature. This correlation has been incorporated into the risk simulation by modeling the distribution of the difference in water versus air temperatures based on the normal distribution within any given region and season. These distributions are then used to predict the air temperature that oysters would be subjected to depending on the water temperature at the time of harvest.
In the process of simulating the distribution of total and pathogenic V. parahaemolyticus at harvest by the Monte Carlo method, the water temperature associated with any given outcome is retained. A corresponding air temperature is predicted by sampling from the appropriate distribution for the difference in air versus water temperature. This difference is then added to the water temperature to derive a corresponding air temperature. The distributions of difference in air versus water temperature were obtained by pooling the data available for each near-shore buoy across all available years. The mean and variance of these distributions are shown in Table IV-3.
| Region | Mean (standard deviation) Distribution Differences between Air and Water Temperature | |||
| Winter (Jan-March) |
Spring (April-June) |
Summer (July-Sept) |
Fall (Oct-Dec) | |
| Northeast Atlantic (Ambrose buoy, NY harbor) |
-2.6 (5.0) | 2.2 (3.2) | 0.52 (2.7) | -3.2 (4.2) |
| Mid-Atlantic (Thomas Point Lighthouse buoy, Chesapeake Bay, MD) |
-0.25 (4.0) | 0.54 (2.9) | -1.4 (2.1) | -2.1 (3.1) |
| Gulf Coast (Dauphin Island, AL buoy) |
-1.07 (3.3) | -1.24 (1.63) | -1.66 (1.33) | -1.62 (3.3) |
| Pacific Northwest (based on 3 years of data from NOAA buoy on north end of Puget Sound, WA) |
-1.6 (1.8) | 1.3 (1.3) | 1.3 (1.5) | -0.8 (2.0) |
| Source of data: http://www.seaboard.nbdc.noaa.gov/Maps/Wrldmap.shtml | ||||
Distribution of time oysters are left unrefrigerated
The distribution of the length of time that oysters are held unrefrigerated was developed by using the distribution of duration of daily oyster harvesting operations (i.e., length of working day). The distribution of length of time oysters are left unrefrigerated is derived by assuming that oysters are harvested uniformly from the start of the harvest up to one hour prior to conclusion of the harvesting operation when oysters are landed and placed in cold storage.
Table IV-4 shows the minimum, maximum and mean duration of oyster harvesting that we have projected for the different regions and seasons. In the risk simulation, we have used Beta-PERT distributions based on these parameters to simulate the variation in the duration of harvesting. A Beta-PERT distribution is a translated and scaled Beta distribution with specified moments. It is commonly used for the purpose of simulating parameter variation within a defined range in Monte Carlo simulations. Figure IV-8 shows the probability density of the Beta-PERT distribution with minimum of 2, maximum of 11 and mean of 8 hours.
| Location | Duration of Harvest (hours) | |||
| Winter (Jan-March) |
Spring (April-June) |
Summer (July-Sept) |
Fall (Oct-Dec) | |
| Northeast Atlantic (assumed same as pre-NSSP Control plan in Gulf- TX (64)) |
max = 11
min = 2 mean = 8 |
max = 11
min = 2 mean = 8 |
max = 11
min = 2 mean = 8 |
max = 11
min = 2 mean = 8 |
| Mid-Atlantic (assumed same as pre-NSSP Control plan in Gulf- TX (64)) |
max = 11
min = 2 mean = 8 |
max = 11
min = 2 mean = 8 |
max = 11
min = 2 mean = 8 |
max = 11
min = 2 mean = 8 |
| Gulf Coast - LA (50% of harvest) (pre-NSSP Control plan in LA in winter; ICP otherwise (64)) | max = 13
min = 7 mean = 12 |
max = 11
min = 5 mean = 9 |
max = 11
min = 5 mean = 9 |
max = 13
min = 7 mean = 12 |
| Gulf Coast - FL, AL, TX (50% of harvest) (assumed same as pre-NSSP Control plan in Gulf- TX in winter, NSSP Control otherwise (64)) | max = 11
min = 2 mean = 8 |
max = 10
min = 3 mean = 7 |
max = 10
min = 3 mean = 7 |
max = 10
min = 3 mean = 7 |
| Pacific Northwest (139) | max = 4
min = 1 mean = 3 |
max = 4
min = 1 mean = 3 |
max = 4
min = 1 mean = 3 |
max = 4
min = 1 mean = 3 |
| Source of data: ISSC & FDA (ed.) 1997 National Shellfish (64)
Washington State Shellfish experts and Washington State Department of Health (139) | ||||
Figure IV-8. Beta-PERT probability density distribution for the duration of harvesting operations during the winter season (Mid-Atlantic, Northeast Atlantic, Gulf Coast, excluding Louisiana) (44). |
The parameters for these distributions were developed and based on a 1997 GCSL survey that included dealer reported statistics on the length of harvest (28). The study was conducted in several Gulf Coast states during the fall of two successive years; one season prior to initiation of the NSSP time to refrigeration requirements (for states whose product has been confirmed as the source of two or more V. vulnificus illnesses), and then the following year after implementation. Duration of harvest was longer in Louisiana than in Florida and Texas, during both years. This probably reflects more remote oyster harvesting areas in Louisiana. The practices of Florida and Texas were considered to be representative of other regions, and in the absence of conflicting information, the longer times were assumed for the other regions throughout the year.
For the Gulf Coast States, we assumed that current harvesting duration is limited in the spring, summer and fall due to the NSSP time to refrigeration requirements and that duration of harvest is generally longer in the winter when cooler water conditions prevail. Louisiana, representing roughly half of the Gulf Coast harvest was treated separately due to the longer duration of harvest year round. The distribution of harvest duration for the West Coast was not based upon the GCSL dealer survey in so far as oysters are generally harvested during intertidal periods and the length of time held unrefrigerated is substantially less. The Pacific Coast Shellfish Growers Association (PCSGA) stated that Pacific oysters are placed under refrigeration within four hours and this time is being assumed as the maximum for the Pacific Coast in the absence of survey data.
As indicated, harvesting of oysters was assumed to occur uniformly from start of harvest, up to one hour prior to end of harvest operation. The distribution of the duration of time oysters were held unrefrigerated, was simulated by first sampling from the distribution for duration of harvest operation and then sampling from a uniform distribution with a minimum of one hour and maximum corresponding to the randomly selected duration of harvest. Oysters are harvested at different times during the length of harvesting operations. Consequently the mean time that oysters remain unrefrigerated is much less than the maximum length of duration of harvesting might suggest.
Overall, the extent of growth occurring prior to time of first refrigeration (i.e., the time at which oysters are first placed in refrigerated storage) was simulated by: (a) sampling air temperature corresponding to the water temperature at harvest; (b) sampling duration of harvest; (c) sampling the length of time unrefrigerated given a particular duration of harvest; and then (d) calculating the extent of growth expected for the given duration of time unrefrigerated.
Excess growth of V. parahaemolyticus during cooldown time
V. parahaemolyticus will continue to grow in oysters after they are placed under refrigeration until the temperature of the oyster tissues falls below a certain threshold (e.g. 10°C). The time it takes for oysters to cool once under refrigeration is assumed to be quite variable depending on efficiency of the cooler, quantity of oysters to be cooled and their arrangement in the cooler. Data on cooling rates of commercial oyster shellstock could not be located. Preliminary GCSL experiments with a single in-shell oyster at 30°C in which a temperature probe was inserted into its tissue indicated a cooling rate of approximately 0.5°C/min when placed into a 3°C cooler (37). However, 24 oysters in an uninsulated plastic container required approximately 7 hours to drop from 26°C to 3°C. These data suggest considerable uncertainty for cooling times after oysters are refrigerated and it was concluded that a rectangular distribution between 1 and 10 hours would be appropriate to describe the current state of knowledge.
As oysters cool down to storage temperatures it is reasonable to expect that the growth rate of V. parahaemolyticus slows with the declining temperature of the oyster tissue. At the start of the cooldown period, when oysters are first placed under refrigeration, the growth rate is still equal to the initial rate as determined by ambient air temperature. At the end of the cooldown period, when oysters have reached storage temperatures, we assume that there is no further growth and that densities will decline slowly thereafter. Implicitly, this assumes that there is no appreciable temperature abuse after oysters have been placed in cold storage. The rate at which oysters cooldown during cold storage is not known. Therefore, in the absence of conflicting information, we have assumed that during the period of cooldown, the growth rate of V. parahaemolyticus drops uniformly down to zero.
A discrete approximation of the extent of growth that may occur during cooldown was simulated by first sampling from a discrete random uniform distribution between 1 and 10 hours (duration of cooldown). The extent of growth during each hour of the cooldown period was then approximated by an average growth rate during that hour times a duration of one hour. The average growth rates were dependant upon the growth rate of V. parahaemolyticus in oysters left unrefrigerated (i.e., as determined by the ambient air temperature for a given oyster lot) and the duration of cooldown. Total excess growth was the sum of these values over the cooldown period subject to the restriction that the maximum density of 6.0 log10 per gram could not be exceeded. These calculations are illustrated in the Table IV-5, where, for example, it takes k hours for a particular oyster lot to reach cooler temperature.
Table IV-5. Discrete approximation of variation in the growth rate of
V. parahaemolyticus during a cooldown period of k hours
| Hour of the cooldown period |
Average growth rate (log10/hr) during the hour of cooldown |
| 1 |  |
| 2 |  |
| 3 |  |
| ... | ... |
| k |  |
| k+1 | 0 |
Total excess growth is the sum of the growth over the k hours:
![sum (i=1 to k) mu_m * ((k + 1) - i)/k = mu_m * [(k + 1) - 1/k sum (i=1 to k) i] = mu_m * [(k + 1) - (k + 1)/2] = mu_m * (k + 1)/2](vrskeq13.gif)
Since the cooldown time k is a random variable with a mean of 5.5 hours, the average extent of
growth is 3.25* m, where m is the maximal growth rate determined by ambient air temperature
at time of harvest. Thus, for an initial growth rate of 0.19 log10 per hour (i.e., at 26°C), the
average growth occurring during cooldown is approximately 0.6 log10.
Die-off of V. parahaemolyticus during cold storage
Gooch et al. (51) showed that in oysters, V. parahaemolyticus declined 0.003 log10 per hour when stored 14-17 days at 3°C. This die-off rate was assumed to be typical of all refrigerated oysters. Error may be introduced because commercial oysters are typically stored at higher temperatures (5-10°C). Die-off may have been overestimated because chill-stressed V. parahaemolyticus may not be recovered by the methods used in the study. One of the enumeration methods employed a repair step in a medium containing magnesium, which has been shown to increase recovery of chill-stressed cells. This method did not give higher V. parahaemolyticus counts after refrigeration than did the other methods that were used to calculate die-off. Therefore, the effect of chill-stress on die-off rate was assumed to be negligible.
Data from the ISSC/FDA retail study for the time between harvest and sample collection were assumed to be a reliable estimate for the length of refrigeration time (time between refrigeration and consumption) (28). Summary statistics on the storage time for samples obtained during the study are shown in Table IV-6. A small degree of error may be introduced by assuming that these data are representative of storage time in so far as samples were generally collected on Monday or Tuesday and most servings are consumed in restaurants on weekends. Since this was a year long nationwide survey, the mean of 7.7 days and range of 1-21 days was assumed to be representative of all seasons and regions. In the simulation, we used a Beta-PERT distribution based on the overall mean, minimum, maximum and mode in order to obtain a smooth representation of the variation in the duration of storage time.
Storage Time |
Consumed locally (within the same region of harvest) |
Non local (transported outside region of harvest) |
Overall |
| Minimum | 1 | 2 | 1 |
| Maximum | 20 | 21 | 21 |
| Mean | 6.3 | 9.9 | 7.7 |
Mode |
6 | 5 | 6 |
| Source of data: (44) | |||
The predicted densities of V. parahaemolyticus at time of consumption were therefore simulated by randomly sampling from the distribution of storage times and multiplying by a die-off rate of 0.003 log10 per hour. The resulting distribution was then subtracted from the predicted distribution of V. parahaemolyticus densities in oysters initially reaching cooler (no growth) temperatures.
Mitigation Strategies
The effects of three possible post harvest mitigations were evaluated in the Monte Carlo simulations: (a) reduction of time to refrigeration (rapid cooling); (b) heat treatment and (c) freezing/cold storage.
The mitigation of reduction in time to refrigeration was modeled by assuming that oysters would be cooled to no growth temperatures immediately following harvest. Immediate cooling would involve icing or otherwise refrigerating oyster shellstock aboard ship while oyster harvesting operations continued. Assuming that this mitigation practice was followed without exception, post harvest growth of V. parahaemolyticus in oysters would occur only during the period of cooldown required for the oyster meat to reach no growth temperatures. In the simulation this is accomplished by assuming that the time unrefrigerated is zero (i.e., a degenerate distribution or constant). However, some growth is still projected to occur during cooldown as described above.
The effects of heat treatment and that of freezing/cold storage were evaluated by adjusting the simulated output of the baseline simulation (no mitigation) downward by factors of 4.5 log10 (the lowest level which caused a substantial reduction in illness after mild heat treatment) and 2 log10, respectively. Thus, random sequences of values for total and pathogenic densities produced in the course of Monte Carlo simulation were divided by 31,623 and 100, respectively. The implicit assumption here is that the effect of treatment on log10 V. parahaemolyticus densities is uniform with no induced change in the variance of log10 densities. The effects of these mitigations on the probability of illness are shown in the Risk Characterization Section.
The Public Health Module estimates the distribution of the probable number of illness which may be expected to occur within any given region and season based on the predicted distribution of pathogenic V. parahaemolyticus densities at time of consumption, and the resulting effects on members of the public eating these oysters. Factors taken into account include the number of V. parahaemolyticus infections, the level of pathogenic V. parahaemolyticus at consumption, the probability of V. parahaemolyticus infection at different dose levels, and the number of diarrheal cases as opposed to more serious outcomes such as septicemia.
Food surveys and oyster landing statistics provide a basis for estimating extent of exposure in the population. Dose-response relationships can be developed from epidemiological investigations of outbreaks and sporadic case series, human feeding trials or animal models of V. parahaemolyticus and related (surrogate) pathogens. The relevant parameters relating to extent of exposure and dose-response relationship are summarized under three sections: epidemiology, consumption, and dose-response (hazard characterization).
Gastroenteritis due to V. parahaemolyticus infection is usually a self-limiting illness of moderate severity and short duration (11, 12, 89). However, severe cases requiring hospitalization have been reported. A summary of clinical features associated with V. parahaemolyticus gastroenteritis infection is presented in Table IV-7 (12, 89). Symptoms include explosive watery diarrhea, nausea, vomiting, abdominal cramps, and less frequently headache, fever and chills. On rare occasions, septicemia, an illness characterized by fever or hypotension and the isolation of the microorganism from the blood, can occur. In these cases, subsequent symptoms can include swollen, painful extremities with hemorrhagic bullae (57, 83). Duration of illness can range from 2 hours to 10 days (12, 13).
Table IV-7. Clinical symptoms associated with gastroenteritis
caused by V. parahaemolyticus
| Symptoms | Incidence of symptoms | |
| Median | Range | |
| Diarrhea | 98% | 80 to 100% |
| Abdominal cramps | 82% | 68 to 100% |
| Nausea | 71% | 40 to 100% |
| Vomiting | 52% | 17 to 79% |
| Headache | 42% | 13 to 56% |
| Fever | 27% | 21 to 33% |
| Chills | 24% | 4 to 56% |
| Source of data: (12, 89) | ||
In Japan, after a decrease in V. parahaemolyticus infections, the incidence started to rise again in 1994 (9). There were 292 incidents of V. parahaemolyticus involving 5,241 cases in 1996. In 1997, the incidence increased to 568, with 6,786 cases, and in 1998, there were 850 incidents, second only to Salmonella infections, but involving more cases than Salmonella (9). In the United States, outbreaks involving over 700 cases in the Gulf Coast, the Northeast, and the Pacific Northwest, in 1997 and 1998, were caused by consumption of raw molluscan shellfish, predominantly oysters, harboring pathogenic V. parahaemolyticus (7, 17, 18). During the outbreaks, certain serotypes, linked to the consumption of raw molluscan shellfish, particularly oysters, were identified as important emerging pathogens.
Outbreaks
An outbreak is defined as the occurrence of 2 or more cases of a similar illness resulting from the ingestion of a common food. The incubation period ranges from 12-96 hours with a median of approximately 15-24 hours. The number of raw oysters consumed ranges from 1-109 (median of 12); however, the duration of consumption is not known. The typical prevalence of symptoms for cases with gastroenteritis parallels those that were identified during the Pacific Northwest outbreak of 1997. These symptoms include diarrhea (99%), abdominal cramps (88%), nausea (52%), vomiting (39%), fever (33%), and bloody diarrhea (12%). The first confirmed case of foodborne illness-associated V. parahaemolyticus infection in the United States occurred in Maryland in 1971 with an outbreak caused by contaminated steamed crabs (32). Between 1973 and 1998, forty outbreaks were reported to the CDC from 15 states and the Guam Territories (34). These outbreaks were all associated with seafood or cross-contamination with raw or undercooked seafood. In 1997, V. parahaemolyticus infection was confirmed in 209 persons who consumed raw oysters harvested from California, Oregon and Washington in the United States and from British Columbia in Canada (22). Prior to this outbreak, the last large outbreak of V. parahaemolyticus infections in North America occurred in 1981 with 6 culture-confirmed cases (107). In 1998, the largest outbreak in the United States occurred in Texas in which a total of 416 V. parahaemolyticus infections were associated with consuming raw oysters harvested from Galveston Bay (34). The first reported outbreak associated with raw shellfish harvested in New York occurred in 1998 as well, involving 23 culture-confirmed cases (21). For these recent outbreaks, the dates of onset of illness ranged from May-December with a peak in July-August. Although V. parahaemolyticus outbreaks are less frequent in occurrence, sporadic cases are not infrequent, as further described below.
Case Reports
Several case reports have been published that outline clinical presentations and outcomes of patients with V. parahaemolyticus. One such case report describes a 35-year-old woman who sought medical attention for abdominal pain after she had consumed raw fish (127). She presented with gastrointestinal symptoms, redness on lower extremities, fever, polyarthritis and weakness. V. parahaemolyticus was isolated in the stool culture. She was diagnosed as having reactive arthritis induced by V. parahaemolyticus infection. Another clinical case report describes a 31-year-old female with a history of alcohol abuse, hepatitis C virus infection, and cirrhosis (54). She presented with diarrhea, weakness, leg pain, and urine retention. The patient had ingested raw oysters and steamed shrimp 72 hours prior to admission. V. parahaemolyticus was isolated from blood samples. The patient developed cardiac arrest and died six days after presentation.
A suspected case of a laboratory-associated infection was reported in 1972 (117). One day prior to the development of diarrheal disease the laboratory worker had been handling V. parahaemolyticus strains for the first time. The illness was associated with severe upper abdominal pain, bloody stools, nausea and fever. Weakness and abdominal discomfort continued for 2 days beyond the onset of illness. No other source of V. parahaemolyticus could be identified, and it was believed that the infection was caused by a relatively small inoculum (117).
Case Series
A case series is a study of sporadic cases over a period of time. Sporadic cases of V. parahaemolyticus infections are commonly reported by many states but are primarily reported by Gulf Coast states. Most V. parahaemolyticus infections present clinically as gastroenteritis, which has a low case fatality rate. Life threatening septicemia can occur, especially in patients with underlying medical conditions. The case series has a range of infection throughout the year, with a peak in September to October. A case series of Vibrio infections related to raw oyster consumption was reported in Florida from 1981-1994 (57). Culture-confirmed case reports of Vibrio infections, reported to the Florida Department of Health and Rehabilitation Services, were investigated to determine the epidemiology of raw oyster-associated Vibrio infections. Clinical and epidemiological information from patients was compiled using standardized Vibrio illness case report forms. Oyster-associated Vibrio infection was defined as a history of raw oyster consumption in the week prior to onset of gastroenteritis or septicemia. Incidence rates were calculated using population data from the Florida Office of Vital Statistics. Estimates of raw oyster consumption were obtained from the Florida Behavioral Risk Factor Survey, 1988.
The average annual incidence of raw oyster-associated illness from any Vibrio species among raw oyster-consuming adults over 17-years-of-age was estimated to be 10.1/1,000,000 (95% CI: 8.3-11.9). The annual incidence of fatal raw oyster-associated infections from any Vibrio species was estimated to be 1.6/1,000,000 oyster-consuming adults (95% CI: 1.3-1.9). In two epidemiological studies, V. parahaemolyticus accounted for 77 of 339 reported Vibrio infections (Table IV-8) (57, 83). Of those 77 persons, 68 reported gastroenteritis and 9 had septicemia. Twenty-nine persons were hospitalized for gastroenteritis with no deaths reported. Eight patients were hospitalized for septicemia and four of those patients died. Patients with septicemia had underlying illness including, but not limited to cancer, liver disease, alcoholism and diabetes mellitus (57, 83).
Table IV-8. Clinical syndromes of raw oyster-associated Vibrio infections in Florida, 1981-1994
| Vibrio Species | Total Cases | Gastroenteritis | Septicemia |
| V. vulnificus | 95 | 13 | 82 |
| V. parahaemolyticus | 77 | 68 | 9 |
| V. cholera Non-O1 | 74 | 8 | 66 |
| V. hollisae | 38 | 35 | 3 |
| V. mimicus | 29 | 29 | 0 |
| V. fluvialis | 19 | 19 | 0 |
| Source of data: (57, 83) | |||
In another study, Hlady and Klontz (58) reported that of patients with infections, 25% had pre-existing liver disease or alcoholism. These included 75% of the septicemia patients, and 4% of the gastroenteritis patients. Of the remaining septicemia patients, 9 reported having a history of at least one of the following: malignancy, renal disease, peptic ulcer disease, gastrointestinal surgery, diabetes, antacid medication and pernicious anemia. Among the gastroenteritis patients, 74% had none of the above preexisting medical conditions or had insufficient information to classify. Thus, while the prevalence of underlying illness was high in the septicemia patients the majority of patients with raw-oyster associated Vibrio gastroenteritis had no underlying conditions. Case series data is available through the Gulf Coast Vibrio Surveillance system, which is a unique regional surveillance system that began in 1989 (89). Four states participate in this program (AL, FL, TX, LA). Investigators in state and county health departments complete standardized Vibrio illness investigation forms on all patients from whom Vibrio isolates are reported. Vibrio reporting comes from individual physicians, hospitals, or laboratories. Illness investigation forms contain clinical data concerning signs and symptoms, underlying illnesses, use of medications, as well as epidemiological information concerning seafood consumption in the week prior to illness. Information is then forwarded to the CDC.
During the first year of Vibrio surveillance in 1989, V. parahaemolyticus accounted for 27 of the 85 reported Vibrio illness characterized by gastroenteritis or septicemia (89). V. parahaemolyticus was the most prevalent of the Vibrio species reported. Twelve of the 27 persons with V. parahaemolyticus were known to have eaten raw oysters. One person had septicemia while the remaining 26 persons had gastroenteritis. Oyster-associated infections occurred throughout the year with the peak occurrence in October.
Based upon CDC surveillance data on V. parahaemolyticus from 1988-1997 in Alabama, Florida, Louisiana and Texas, the six most common underlying medical conditions associated with infection include diabetes, peptic ulcer, heart disease, gastric surgery, liver disease and immunodeficiency (6). For gastroenteritis, 24% of respondents reported one or more of these six conditions compared with 71% of respondents who had sepsis. In 263 gastroenteritis cases: 7% had diabetes, 6% had peptic ulcer disease, 6% had heart disease, 4% had undergone gastric surgery, 3% suffered from alcoholism, 3% suffered from some form of immunodeficiency, 3% had liver disease, 2% had hematological disease, 2% had some form of malignancy, and 1% had renal disease. Out of 20 septicemic cases, 63% had liver disease, 18% had some form of immunodeficiency, 18% had peptic ulcer disease, 17% had diabetes, 14% suffered from alcoholism, 13% had hematological disease, 12% had undergone gastric surgery, 12% had heart disease, 12% had renal disease, and 11% had some form of malignancy. Among 88 patients with sporadic V. parahaemolyticus infection and known food histories, 77 (88%) reported eating raw oysters in the week before illness (34). Of 11 patients with septicemia and known food history, 10 (91%) had eaten raw oysters. Data from the CDC Gulf Coast Surveillance System from 1997 to 1998, were limited only to those cases that are both culture confirmed and ingestion confirmed and resulted in a subset totaling 107 cases. Of these 107 cases, 5 (5%) involved septicemia in which all five were hospitalized with one death. This is believed by CDC to be a fairly accurate estimation of the overall incidence of septicemia among culture-confirmed V. parahaemolyticus infections (6). The presence or absence of underlying conditions was reported by 4 of the 5 septicemic patients; 3 (75%) of whom had underlying conditions. Among the 102 cases with gastroenteritis alone, 27 of 90 (30%) respondents reported being hospitalized for the illness. Patient outcome was reported for 83 patients; one of whom died. The presence or absence of underlying conditions was reported by 79 persons; 29% of these reported underlying conditions. The underlying conditions included liver disease, alcoholism, diabetes, malignancy, renal disease, immunodeficiency, hematological disease, gastric surgery and heart disease.
Based upon active FoodNet data surveillance, CDC estimates that the total number of foodborne V. parahaemolyticus cases in the United States for 1996, 1997, and 1998 were 2,683; 9,807, and 5568, (rounded to 2,700; 9,800; and 5,600, respectively) (129). These estimates were derived from the numbers of Vibrio cases reported to FoodNet. For the calculations, the reports of Vibrio cases with unknown species were included for the estimate of the total number of Vibrio cases but was not included in the estimate of the percentage of all Vibrio spp. that were V. parahaemolyticus. This assumes that the isolates of unknown species are distributed the same as the isolates of known species. The percentage of V. parahaemolyticus cases attributed to being foodborne was estimated at 65%. The 1997 estimates are higher as a result of the increased reporting of cases during the Pacific Northwest outbreak. The variation in estimated cases from year to year is expected since the numbers obtained from FoodNet are very small. During the 1972 shrimp-associated V. parahaemolyticus outbreak, a survey revealed that of 72 persons with diarrhea only one sought medical attention (13). Due to underdiagnosing and underreporting of cases of V. parahaemolyticus, the CDC estimates that the total number of cases is equal to 20 times the reported cases (94). In a CDC random survey of Gulf Coast clinical laboratories, only 20% of the laboratories routinely used selective agar for isolating Vibrio species (34).
Geographic distribution
As mentioned earlier, V. parahaemolyticus was first identified as a foodborne pathogen in Japan in the 1950s (48). By the late 1960s and early 1970s, V. parahaemolyticus was recognized as a cause of diarrheal disease worldwide. Prior to 1994, the incidence of V. parahaemolyticus infections in Japan had been declining, however, from 1994 to 1995 there were a total of 1,280 reports of infection due to V. parahaemolyticus (9). During this time period, the incidents of V. parahaemolyticus food poisoning outnumbered those of Salmonella food poisoning. For both years, the majority of the cases occurred in the summer, with the largest number appearing in August. Food poisoning due to V. parahaemolyticus in Japan is usually restricted to relatively small-scale outbreaks involving fewer than 10 cases. From 1996-1998, there were 496 outbreaks, 1,710 incidents and 24, 373 cases of V. parahaemolyticus reported. The number of cases of V. parahaemolyticus food poisoning cases doubled in 1998 as compared to 1997 and again exceeded the number of Salmonella cases (9). Similar to the 1994-1995 period, outbreaks were more prevalent in the summer with a peak in August with few outbreaks during winter months. Boiled crabs caused one large-scale outbreak, involving 691 cases. The majority of outbreaks were small in scale but occurred frequently. The increased incidence during 1997-1998 has been attributed to an increased incidence of serovar O3:K6.
A hospital-based active surveillance study of V. parahaemolyticus infections in Calcutta, India, was conducted from 1994-1996, and identified 146 patients (109). The incidence suddenly increased in February of 1996 and remained elevated until August of that year when surveillance ended. The increased incidence of V. parahaemolyticus infections was associated with an increased prevalence of O3:K6 strains. This serovar had not been isolated in Calcutta prior to February of 1996. The incidence of diarrhea due to V. parahaemolyticus strain O3:K6 accounted for 63% of the strains isolated from patients in Calcutta between September 1996 and April 1997. The virulence of the O3:K6 strains isolated from travelers arriving in Japan from Southeast Asian countries was indistinguishable from O3:K6 strains found in Calcutta, India (92).
Implicated Foods
Vibrio organisms concentrate in the gut of filter-feeding molluscan shellfish such as oysters, clams, and mussels where they multiply and cohere. Although thorough cooking destroys these organisms, oysters are often eaten raw and are the most common food associated with Vibrio infection in the United States (57). However, there have been reports of V. parahaemolyticus infections associated with other seafood, including crayfish, lobster, shrimp, and crab. One such report was a case-controlled study of sporadic Vibrio infections in two coastal areas of Louisiana and Texas conducted from 1992-1993, in which crayfish consumption was reported by 5 of 10 persons affected with V. parahaemolyticus infection (16). Outbreaks of V. parahaemolyticus gastroenteritis aboard two Caribbean cruise ships were reported in 1974 and 1975 (87). The outbreaks were most likely caused by contamination of cooked seafood by seawater from the ships' seawater fire systems. In 1972, an estimated 600 of 1,200 persons who attended a shrimp feast in Louisiana became ill with V. parahaemolyticus gastroenteritis (13). Samples of uncooked shrimp tested positive for the organism. Three outbreaks occurred in Maryland in 1971 (32). Steamed crabs were implicated in two of the outbreaks after cross-contamination with live crabs. The third outbreak was associated with crabmeat that had become contaminated before and during canning. Recently, sampling studies in the Adriatic Sea demonstrated the presence of V. parahaemolyticus in fish, mussels and clams (10).
The purpose of this segment is to delineate the factors concerning the consumption of raw molluscan shellfish containing V. parahaemolyticus.
Frequency of Consumption and Amount of Raw Molluscan Shellfish Consumed
Intake data for molluscan shellfish are readily available from a number of governmental and non-governmental sources. However, because raw shellfish is not a commonly consumed food (~10- 20% of the population will consume shellfish raw at least once during a year), the data are typically based on very few eaters reporting consumption. The USDA Continuing Survey of Food Intake by Individuals (CFSII) (135) and the food frequency survey conducted by the Market Research Corporation of America (MRCA) (36) suggest that raw oysters are consumed on average approximately once every 6 weeks. The mean amount of raw oysters consumed at a single serving is 110 grams, approximately one-half dozen raw large Eastern oysters (128). The distribution of shellfish intake will be derived from food intake surveys, food frequency surveys, and from reported landings of shellfish and industry estimates of the percentage of shellfish consumed raw.
Population at Risk
Anyone who consumes shellfish raw is "at risk" for infection by V. parahaemolyticus. An FDA telephone survey completed in 1993 and repeated in 1998 has shown that consumption of raw shellfish is not uniformly distributed (90). A higher percentage of men consume raw oysters than women (16% vs. 7%), and raw shellfish consumption is higher for those living along the coastline of the United States than for those living inland (22% vs. 13%). The trends in raw shellfish consumption, as evidenced in the 1998 FDA survey is toward lowered consumption of raw shellfish. This may be the result of education efforts by the Agency concerning the risks associated with the consumption of raw or undercooked protein foods, such as beef, chicken, eggs, and shellfish. Paradoxically, raw shellfish consumption is highest among those with the highest education levels, and the trend toward reduction in raw shellfish consumption over the last 5 years is smallest in this education group.
Oyster Landings Data
The time of year of consumption was considered in the risk assessment, as most infections occur during warm months, that is, a person consuming the raw oysters in July is at higher risk than the same person consuming the same amount in December. The location of harvest is also important, with most landings of oysters occurring in the Gulf, particularly off the coast of Louisiana.
Hypertext updated by bap/kwg/cjm/dms 2001-JAN-30
