Journal Search Engine
Search Advanced Search Adode Reader(link)
Download PDF Export Citaion korean bibliography PMC previewer
ISSN : 1229-1153(Print)
ISSN : 2465-9223(Online)
Journal of Food Hygiene and Safety Vol.33 No.4 pp.248-254
DOI : https://doi.org/10.13103/JFHS.2018.33.4.248

A Risk Assessment of Vibrio parahaemolyticus for Consumption of Shucked Raw Oyster in Korea

Jong-Kyung Lee *, Ki-Sun Yoon1, Hyang Lee2, Hyun Jung Kim3
Department of Food & Nutrition, Hanyang Women’s University, 200 Salgoji-gil, Seongdong-gu,Seoul, Korea
1Department of Food and Nutrition, Kyunghee University, 26 Kyungheedae-ro, Dongdaemoon-ku, Seoul, Korea
2The Research Institute of Nursing Science, Seoul National University, 103 Daehak-ro, Jongno-gu, Seoul, Korea
3Korea Food Research Institute, 245 Nongsaengmyeong-ro, Iseo-myon, Wanju-gun, Jeollabuk-do, Korea
Correspondence to: Jong-Kyung Lee, Department of Food & Nutrition, Hanyang Women’s University, 200 Salgoji-gil, Seongdong-gu, Seoul 04763, Republic of Korea Tel: 82-2-2290-2183, Fax: 82-2-2290-2199. E-mail: jklee@hywoman.ac.kr
February 19, 2018 March 20, 2018 July 19, 2018

Abstract


To assess the risk of V. parahaemolyticus infection caused by consumption of raw oysters in Korea, contamination levels during the retail-to-table route of oysters was modeled to predict V. parahaemolyticus growth based on temperature and time. The consumed amount data of the KNHANES and the standard recipe of RDA were applied. A consumption scenario for exposure assessment was developed and combined with a Beta-Poisson dose-response model. The estimated probability of illness from consumption of pathogenic V. parahaemolyticus in raw oysters during three separate months (April, October, and November) was 5.71 × 10−5 (within the 5th and 95th percentile ranges of 2.71 × 10−8 to 1.03 × 10−4).The results of the quantitative microbial-risk assessment indicated that the major factors affecting the probability of illness were the initial contamination level at the retailer, the consumed amount, the prevalence of pathogenic strains [tdh or trh genes], and exposure temperature and time.



초록


    Korea Food Research Institute

    Vibrio parahameolyticus is closely related to seafood safety. V. parahaemolyticus is a Gram-negative, mesophilic and halophilic foodborne pathogen that causes gastroenteritis, with symptoms including severe stomach cramps and diarrhea. Raw fish and shellfish are the most common sources of illness caused by V. parahaemolyticus, in addition to contaminated utensil materials or improper storage. The incidence and severity of V. parahaemolyticusassociated gastroenteritis are influenced by the dose of bacteria and virulence characteristics of isolated strains1,2). The major virulence factors of V. parahaemolyticus are thermo-stable direct hemolysin (tdh) and tdh-related hemolysin(trh) genes3). V. parahaemolyticus is the main foodborne pathogen in Asia including Korea4).

    The Codex Alimentarius Commission recommended that a quantitative microbial risk assessment be carried out based on scientific information, including hazard identification, exposure assessment, hazard characterization (dose-response relationship), and risk characterization5). A risk assessment of V. parahaemolyticus in raw oysters from the Pacific and Atlantic oceans was previously carried out in the U.S. in 20056). However, we previously showed that contamination levels, processing methods, and consumption patterns differ between the U.S. and Korea7,8). To estimate the extent of exposure to pathogen due to food intake, risk factors specific to a country or region such as pathogen contamination rate and levels, eating habits, consumer behaviors during preparation, cooking method, and consumed amounts must be considered9).

    In Korea, raw oyster meat is typically distributed without the shell. Oyster shucking is carried out in coastal harvest areas, with some steps of the shucking process for domestic distribution taking place at ambient temperature. Accordingly, shucked oysters are the most useful shellfish model for predicting the worst possible outcome in terms of the safety of oyster consumption in Korea. Koreans normally consume raw oysters from October to April of the following year. During this period, V. parahaemolyticus outbreaks mostly occur in October, November, and April4). Raw oyster is consumed in a variety of ways in Korea, including as seasoned oyster or in mixed dishes with vegetables and meats; calculating the amount of raw oyster consumed is therefore complicated and requires a standard recipe.

    To this end, the present study investigated the consumption patterns of Koreans in order to establish a model for raw oyster consumption that can be used assess the risk of V. parahaemolyticus-related illness for Korean consumers. We also propose a risk management plan to reduce the risk associated with V. parahaemolyticus in Korea.

    Materials and Methods

    Assumption for risk assessment

    For risk assessment, the simplest probable retail-to-table pathway of raw oyster meat from its source to Korean consumers was determined (Fig. 1). The model incorporated initial contamination at the retailer and projected growth based on time and temperature during transportation prior to consumption. The risk during the months of December through March was excluded because V. parahaemolyticus has not been detected during this period in microbial analysis studies7,11-13). In addition, the risk from May to September was excluded because Koreans do not eat raw oysters during this period.

    The following assumptions were made in this study due to the lack of data: 1) tdh- and trh-positive pathogenic V. parahaemolyticus have similar growth rates; 2) similar amounts of raw oyster are consumed in April, October, and November; 3) consumption patterns of the surveyed group are representative of the Korean population; 4) oyster temperature is the ambient temperature during transportation; 5) there is no reduction in V. parahaemolyticus due to consumer behaviors such as oyster washing or mixing with other ingredients during foot preparation; and 6) raw oysters were consumed immediately upon purchase by the consumer.

    Exposure assessment: Collection of microbiological data on the prevalence and pathogenicity of V. parahaemolyticus in raw oyster

    There are not many released quantitative data of V. parahaemolyticus in seafood requiring MPN methods with identification of pathogenic strains by PCR methods14). Especially quantification data of V. parahaemolyticus in retail oysters are limited in Korea7,11). Collection of quantitative monitoring data were restricted to April, October, and November in this work, since these are the most common periods of V. parahaemolyticus outbreak, V. parahaemolyticus detection in raw oysters, and raw oyster consumption in Korea (Table 1). The two studies7,11) also supported this period risk model, as V. parahaemolyticus was detected from April to November but not in December through March. The V. parahaemolyticus contamination rate (14/37 samples, 37.8%) was higher in October (91.7%) than in November (16.7%) or April (7.7%) (Table 1). A beta distribution (r + 1, n − r + 1) was used to model the prevalence of V. parahaemolyticus, where n (trial number) and r (success number) were 37 and 14, respectively, and modelled as Beta (15, 24). Previous microbial analyses12,13) of raw oysters were not included in the retail-to-table model since in those studies, oyster samples were collected at the site without retail shucking process.

    Best fitting distribution of the [log most probable number (MPN)/g] of V. parahaemolyticus contamination level of in raw oyster (Table 1) was applied by @Risk v.6.0 software with Chi-square, A-D and K-S statistics and was shown to be Lognormal (3.066, 0.922).

    Data on the virulence of V. parahaemolyticus isolated from oysters and seafood in the Pacific region of Korea were obtained from three studies7,13,15) (Table 1); two of the datasets were combined to determine the presence (%) of the virulence factor among 155 samples and the number of samples positive for the virulence factor (n = 3)7,15). In one study, 11 pathogenic trh strains were isolated from a total of 11513). Data from the three studies were used for modeling and a beta model (15, 257) was used to establish virulence.

    Exposure assessment: Model of V. parahaemolyticus growth according to temperature and time variables

    A predictive model for temperature variables were used to predict V. parahaemolyticus growth at different steps along the pathway (Fig. 1). A previous study comparing the growth patterns of trh-positive pathogenic V. parahaemolyticus and non-pathogenic strains isolated from oysters showed that the former grew more slowly than the latter16). Since gastrointestinal illness is caused by consumption of pathogenic strains, we modeled the growth of trh-positive V. parahaemolyticus per serving of oysters during transportation using the following primary growth model of cells grown in nutrient broth with 3% NaCl:

    Yt = No + Cmax × exp [−exp ([2.718 × SGRt/Cmax] × [LT − t] + 1)]
    (1)

    where Yt is the log count (cfu ml−1) at time t, No is the initial level of bacteria, Cmax is the growth from inoculum to stationary phase, SGRt is relative growth rate, and t is time. A Cmax value of 9.0 from the 15°C growth curve was used in this study. The values of lag time (Lt) and specific growth rate (SGRt) of pathogenic strains as a function of temtemperature 16) were obtained by equations 2 and 3.

    SGRt = [0.00219 (T − 6.128)]2
    (2)

    Lt = 90.35 + (−4049/T) + (45493/T2)
    (3)

    The average daily temperatures for April, October, and November in Korea17) were 7.9°C ~ 29.3°C (the most likely temp was 20.1°C), and modelled as Pert (7.9, 20.1, 29.3). Changes in log MPN/g of V. parahaemolyticus during retail, transportation, and consumption of oysters were modeled based on temperature and time variables (Fig. 1).

    Exposure assessment: Consumer survey for determination of transport time of raw oysters

    Data on transport time of raw oysters after purchase were obtained through a consumer survey18). The variability in transport time (tt) was modeled as Uniform (0.144, 4.68) using average transport time (0.006-0.195 days or 0.144- 4.68 h).

    Exposure assessment: Analysis of data on amount of raw oyster consumed

    The amount of raw oysters consumed was calculated using Korea National Health and Nutrition Examination Survey (KNHANES) data conducted in April19) and Rural Development Administration (RDA) standard recipe information 19). This survey of 9,047 subjects over 12 years of age was carried out and it contains information regarding food items and amount of food consumed by respondents during 24 retrospective hours per serving per person. To account for variations in consumption caused by the cooking method, the amount of oyster (g) containing other food materials such as vegetables and meat that was consumed was calculated by multiplying the consumed amount of raw oyster-containing food (g) from the KNHANES data by the raw oyster ratio of the standard RDA recipe (Table 2). The ratio of oysters in the food was calculated from the standard RDA recipe for each person?i.e., raw oyster = 100%, seasoned oyster = 71.4%, oyster jotgal = 44.5%, seasoned oyster with radish slice = 18.2%, bossam kimchi = 2.3%, and bossam = 1.5% (data not shown). The consumed amount (Ca) of each food portion containing raw oysters (KNHANES data) was multiplied by the raw oyster ratio (RDA data of standard recipe) to calculate the amount of raw oyster (g) per serving.

    Best fitting distribution of the consumed amount of raw oyster (Table 2) was applied and was shown to be Exponential (8.248). V. parahaemolyticus exposure levels in humans (expressed as log MPN/serving) was used to calculate the amount per serving by multiplying by the final contamination level in log MPN/g of V. parahaemolyticus (Table 2).

    Dose-response relationship

    A beta-Poisson dose-response curve21,22) was used to estimate the risk of illness caused by pathogenic V. parahaemolyticus in raw oysters:

    P(ill|d) = 1 − (1 + d/β)−α
    (4)

    where d is dose and P(ill|d) is the probability of illness due to the pathogenic strain. When the conditions β >> α and β >> 1 were satisfied, The beta-Poisson distribution was determined to be suitable for modeling the V. parahaemolyticus dose-response relationship23).

    Simulations

    The input parameters for V. parahaemolyticus risk assessment are shown in Table 3. @Risk v.6.0 software was used for Monte Carlo simulation with 10,000 iterations.

    Risk characterization

    The dose-response model was combined with the output of exposure assessment to estimate the probability of exposure to pathogenic V. parahaemolyticus per serving of raw oyster meat. A sensitivity analysis was carried out to compare input factors contributing to risk.

    Results and Discussion

    To predict the risk by consumption of raw oyster contaminated by pathogenic V. parahaemolyticus in Korea, input and output variables were described and summarized in Table 3. The input values of Lv.p (Level of V.p in raw oyster at retail), Ppatho (Prevalence of pathogenic genes (tdh or trh)), Tt (Transport temperature), tt (Transport time) and Ca (Consumption amount) were used estimate the output values of Gt (Cell population density at consumption), d (Dose of pathogenic V.p in consumed meal), and Pill (Probability of illness) (Table 3).

    By Monte Carlo Simulations, the estimated mean and 5th and 95th percentile distributions of output values are shown in Table 4. According to the retail-to table model, retail oysters were contaminated with V. parahaemolyticus at a level of 0.49 log MPN/g (the mean value). Approximately 93.5% retail oyster in this season was estimated to be less than 2 log MPN/g by simulation (data not shown). When the oysters were transported to the location of consumption at ambient temperature, the concentration of V. parahaemolyticus was estimated to be 1.00 log MPN/g (the mean value) (Table 4). Given the exposure to ambient temperatures during transportation, V. parahaemolyticus growth was only 0.5 log MPN/g, although growth in nutrient-rich broth may be overestimated. The risk in the range of 5th to 95th percentiles (2.71 × 10−8~ 1.03 × 10−4) emphasizes the worst case of contamination level, consumption amount, the presence of virulence factor, etc. The predicted mean risk of illness for oyster consumers caused by pathogenic V. parahaemolyticus per serving of raw oyster (Pill) was 5.71 × 10−5, which was comparable to the mean risk determined in Brazil (4.7 × 10−4 in spring and autumn) and the Gulf Coast of the U.S. (1.7 × 10−4 and 4.3 × 10−5 in spring and autumn, respectively)24). Since the consumer survey by KNHANES was conducted in April, only 1.9% of the respondents were categorized as for the oyster consumer.

    Data gaps were found as the prevalence of oyster consumers, limited sample size, consumer behaviors during preparation in this study. In addition, some of the box package containing ice prohibits the growth of V. parahaemolyticus during transportation, the risk can be reduced in this case.

    The major factors influencing risk estimates for subjects consuming raw oyster in the sensitivity analysis were initial concentration of V. parahaemolyticus in retail oysters in log MPN per g (Lv.p); amount of raw oyster consumed (Ca); prevalence of pathogenic strains (i.e., those harboring tdh ortrh); prevalence of V. parahaemolyticus in oysters (Pv.p); and exposure temperature and time (t) (Fig. 2). These are the major points to reduce the risk. The variation in risk due to different pathogenic strains requires further investigation. One study reported a case of Vibrio alginolyticus harboring the trh gene of V. parahaemolyticus25) that was associated with a food poisoning outbreak. Thus, the risk of illness may be underestimated in the present study.

    For risk management options, the quantification of V. parahaemolyticus in seafood with seawater/ambient temperature monitoring needs to be conducted to decide oyster harvest season. Controlling transport/storage temperature and time is a realistic management strategy for both the raw oyster industry and consumers to prevent growth of Vibrio spp. It was previously reported that V. parahaemolyticus can proliferate in harvested raw oysters by 2.9 log CFU/g in 24 h at 26°C in the U.S.26). On the other hand, V. parahaemolyticus does not grow at temperatures under 14°C27). Therefore, the risk of illness caused by V. parahaemolyticus is low between November and March since the atmospheric and seawater temperatures are < 15°C in Korea during this period. There are several days in April and October when the average ambient temperature in Korea is > 20°C; therefore, in warmer months, and the shucking process needs to be conducted under 15°C. The oyster samples at the harvest site without shucking process12,13), showed lower microbial levels than shucked oysters at the same month.

    In addition, the uncertainty of predictions based on age and immune state must be addressed. It was previously shown that the rate of raw oyster consumption was higher among older individuals in Korea8), implying that elderly people who have weaker immune systems have a higher risk of illness, which must be taken into consideration not to underestimate the risk. The public should be educated on sources of and risk associated with V. parahaemolyticus to ensure that shellfish are properly handled, including limiting their exposure to ambient temperatures prior to consumption and thorough heating during cooking28), especially for the weak immune people.

    국문요약

    본 연구에서 소비-섭취 시나리오와 온도-시간의 장염비 브리오 생육모델을 활용하여 국내 생굴의 병원성 장염 비 브리오균의 위해평가를 실시하였다. 장염 비브리오균의 오 염 수준 및 병원성 인자 데이터를 활용하였으며, 국민건 강영양조사와 농촌진흥청의 표준레시피를 활용하여 섭취 량을 조사하였고 용량반응관계는 Beta-Poisson모델을 활용 하였다. 국내 소비자가 생굴을 섭취할 때 병원성 장염 비 브리오균으로 발생하는 위해는 식중독이 주로 발생하는 4 월, 10월, 11월에 5.71 × 10−5 (5퍼센타일 2.71 × 10−8, 95퍼 센타일 1.03 × 10−4)로 추정되었다. 본 연구에서 생굴의 장 염비브리오 위해의 영향인자는 소비시점 생굴의 장염비브 리오균의 오염수준, 생굴 섭취량, 병원성 인자(tdh or trh 유전자)의 존재 여부, 상온의 노출온도 및 시간으로 나타 났으며 위해관리방안을 제시하였다.

    Acknowledgements

    This work was supported by the Korea Food Research Institute (KFRI) project.

    Figure

    JFHS-33-248_F1.gif

    Risk assessment pathways and parameters of pathogenic V. parahaemolyticus for raw oyster consumption in Korea.

    JFHS-33-248_F2.gif

    Effect of input variables on risk associated with consuming raw oyster in Korea, shown as a Tornado plot in the sensitivity analysis. Ca, Consumed amount; Lv.p, level of V.p in raw oyster meat at retail; Ppatho, prevalence of pathogenic gene (tdh or trh); PV.p: prevalence of V.p in raw oyster; temperature, transport temperature; time, transport time.

    Table

    Number of contaminated samples and V. parahaemolyticus levels in raw oysters at the retail market in April, October, and November and presence of pathogenic V. parahaemolyticus strain ratios in seafood in Korea

    Raw oysters consumed per serving as determined by a KNHANES study and RDA standard recipe

    Summary of variables, distributions, models, equations, and references used in this study

    Risk estimates

    Reference

    1. DePaola, A. , Hopkins, L.H. , Peeler, J.T. , Wentz, B. , McPhearson, R. : Incidence of Vibrio parahaemolyticus in US coastal waters and oysters . Appl. Environmental Microbiol., 56, 2299-2302 (1990).
    2. Zhang, X.H. , Austin, B. Haemolysins in Vibrio species . J. Appl. Microbiol., 98, 1011-1019 (2005).
    3. SakazakiR. Vibrio. pp. 127-136. In: Food-borne disease, CliverDO , RiemannHP (eds). Academic Press, London. (2002).
    4. MFDS (Ministry of Food and Drug Safety) (2016) http://www.foodsafetykorea.go.kr/portal/healthyfoodlife/foodPoisoning-Stat.do?menu_no=519&menu_grp=MENU_GRP02. Accessed Oct. 25, 2016.
    5. Codex Alimentarius Commission (1999) Principles and guidelines for the conduct of microbiological risk assessment, CAC/GL-30, Geneva.
    6. FDA (2005) Quantitative risk assessment on the public health impact of pathogenic Vibrio parahaemolyticus in raw oysters. http://www.fda.gov/downloads/Food/ScienceResearch/ResearchAreas/RiskAssessmentSafetyAssessment/UCM196915.pdf
    7. Lee, J.K. , Jung, D.W. , Eom, S.Y. , Oh, S.W. , Kim, Y.J. , Kwak, H.S. , Kim, Y.H. : Occurrence of Vibrio parahaemolyticus in oysters from Korean retail outlets . Food Control, 19, 990-994 (2008).
    8. LeeM.A. , LeeJ.K. , ChaS.M. Analysis on the consumer’s attitude and purchase behavior of oysters . Korean J. Cookery Sci.24, 919-930 (2008).
    9. Lamerding, A.M. Fazil, A. Hazard identification and exposure assessment for microbial food safety risk assessment . Int. J. Food Microbiol.58,147-157 (2000).
    10. Toyofuku, H. Harmonization of international risk assessment protocol . Mar. Pollut. Bull.53, 579-590 (2006).
    11. LeeH. Microbiological population of Vibrio parahaemolyticus in oysters of wholesale seafood markets . J. Food Hyg. Safety21, 238-243 (2006).
    12. YuH , OhE.G. , ShinS.B. , ParkY.S. , LeeH.J. , KimJ.H. , SongK.C. Distribution and antimicrobial resistance of Vibrio parahaemolyticus isolated from Korean shellfish . Korean J. Fish. Aquat. Sci.47, 508-515 (2014).
    13. KimS. , AnS. , ParkB. , OhE.G. , SongK.C. , KimJ.W. , YuH. Virulence factors and antimicrobial susceptibility of Vibrio parahaemolyticus isolated from the oyster Crassostrea gigas . Korean J. Fish. Aquat. Sci.49, 116-123 (2016).
    14. Bacteriological Analytical Manual http://www.fda.gov/Food/FoodScienceResearch/LaboratoryMethods/ucm070830.htm. Accessed Dec. 1, 2016.
    15. ChungY.T. , SoB.T. , KimY.K. , O Y.H., Ham H.J., Cha Y.S. et al. Detection and pathogenicity of hemolysin gene in Vibrio parahaemolyticus . The Report of Seoul Metropolitan Government Research Institute of Public Health and Environment38, 56-61 (2002).
    16. YoonK.S. , MinK.J. , JungY.J. , KwonK.Y. , LeeJ.K. , OhS.W. A model of the effect of temperature on the growth of pathogenic and nonpathogenic Vibrio parahaemolyticus isolated from oysters in Korea . Food Microbiol.25, 635-641 (2008).
    17. KMA (Korea Meteorological Administration) (2005) http://www.kma.go.kr/weather/observation/past_cal.jsp?stn=108&yy =2005&mm=4&obs=1&x=24&y=6. Accessed Jul. 26, 2016.
    18. YoonK.S. Risk Assessment of pathogenic Escherichia coli and development of pathogen modelling program. Korea MFDS reports. (2012).
    19. KCDC (Korea Center for Disease Control and Prevention) The Third Korea national health and Nutrition Examination Survey (KNHANES III) (2005).
    20. RDA (Rural Development Administration) (2010) http://koreanfood.rda.go.kr/inctfd/total_srch_more1.aspx. Accessed Feb. 16, 2010.
    21. FAO/WHO. Risk assessment of Vibrio parahaemolyticus in seafood . Microbiological Risk Assessment Series16, 1-200 (2011).
    22. ParkM.S. , ChoJ.I. , LeeS.H. , BahkG.J. A study on doseresponse models for foodborne disease pathogens . J. Food Hyg. Safety29, 299-304 (2014).
    23. FAO/WHO. Hazard identification, exposure assessment and hazard characterization of Campylobacter spp. in broiler chickens and Vibrio spp. in seafood. Geneva, Switzerland, July 23-27, (2001).
    24. Sobrinho, Pde S , Destro, M.T., Franco, B.D., Landgraf, M. A quantitative risk assessment model for Vibrio parahaemolyticus in raw oysters in Sao Paulo State, Brazil . Int. J. Food Microbiol.180, 69-77 (2014).
    25. GonzálezEscalona, N., Blackstone, G.M., DePaola, A. Characterization of a Vibio alginolyticus strain, isolated from Alaskan oysters, carrying a hemolysin gene similar to the thermostable direct hemolysin-related hemolysin gene (trh) of Vibrio parahaemolyticus . Appl. Environmental Microbiol.72, 7925-7929 (2006).
    26. Gooch, J.A. , DePaola, A. , Bowers, J. , Marshall, D.I. Growth and survival of Vibrio parahaemolyticus in postharvest American oysters . J. Food Protect.65, 970-974 (2002).
    27. NZFSA. Risk profile: Vibrio parahaemolyticus in seafood. http://www.foodsafety.govt.nz/elibrary/industry/Risk_Profile_Vibrio-Science_Research.pdf (2003).
    28. Yamamoto, A. , Iwahori, J. , Vuddhakul, V. , Charernjiratragul, W. , Vose, D. , Osaka, K. , Shigematsu, M. , Toyofuku, H. , Yamamoto, S. , Nishibuchi, M. , Kasuga, F. Quantitative modeling for risk assessment of Vibrio parahaemolyticus in bloody clams in southern Thailand . Int. J. Food Microbiol.124, 70-78 (2008).