Salmonella and Vibrio cholerae in brackishwater cultured tropical prawns

International Journal of Fcusl Microbiolvgy. 16.(|qq212~3-.'~11 ~'3 1992 Elsevier Science Publishers B.V. All rights reserved 0168-1tdlS/q2/$05.110

293

FOOD tH153(1

Salmonella and Vibrio cholerae in brackishwater cultured tropical prawns P.J.A. Reilly and D.R. Twiddy Natural Res~mrces Institute. Chathmn. Kent. Estgland. UK

(Received 6 April 19112:revision received 10 June It,~2:accepted 29 June Iq92)

The occurrence of Salmcmella and Vibrio ¢-hcderae in brackishwater ponds was monitored over a 2-year perit~d in one of the major prawn exporting countries in Southeast Asia. The principal production areas were identified and regular samples taken for Salmonella and V. cholerae analysis, Results deraonstrafed that brackishwaler ponds and cultured prawns were inherently contaminated with both bacterial pathogens, Salmonella spp. were present in 16.0% of prawns and 22.l% of mud/water samples frum ponds: and V. rholvrat, present in 1.5% of prawns and 3.1% of mud/water samples. Culturing by intensive methods tended to favour contamination by these pathogens, which is most likely due to the accumulation of waste and increase in the volume of sediments in Ixmds. Typical environmental factors such as water temperature, pH. and ,salinity were all favonrable for gr~.vth of microorganisms.The incidence of the pathogens increased during the wet season and was marginally higher when ponds were located close to urban areas. S. wehecreden was identified as the principal serotype found in ponds, and to a lesser extent S. anamm (11%) S. wandsworth (8%) and S. potsdam (8%). The V. cholerae belonged to the nun-OI serogroup~ Key words: Cultured prawn: Shrimp: Pathogenic bacteria: Salmonella; ~Tbrio

Introduction Frozen prawns and shrimps arc one of the m a j o r export fisheries products from Southeast Asia. Failure of p r o d u c e to m e e t the bacteriological standards of importing countries has labelled prawn and shrimps in this region as inferior quality products and has r e d u c e d their competitiveness in world markets. In 1989, over 8300 tonnes of frozen raw prawns w e r e detained by the U S F D A due to contamination with S a l m o n e l l a species ( U S F D A , 19891. T h e detention at port of entry can lead to additional costs of shipping, cold storage, handling a n d processing which are all c h a r g e d to the exporter. It can also lead to total rejection or re-exportation of shipments, which results in severe economic losses. Exporters can insure against rejection, but costs are high at 7 % of shipment value. Similar

Corresl~zdence address: P.J.A. Reilly. Natural Resources Institute, Central Avenue, Chatham Maritime, Kent ME4 4TB. England, UK

204 problems, although on a smaller ,scale, are faced by exporters to the Japanese market. The Japanc~ will accept Salmonella in raw frozen produce but will reject all products which contain Vibrio cholerae. Data on recent rejections arc not available, but between 1983 and 1986 prawns valued at US$ 3 million from Southeast Asia were rejected and destroyed by health authorities in Japan due to the presence of pathogenic V. cholerae (Reilly, 19871. The USFDA standards for absence of Salmonella in raw frozen prawns arc based on the fact that prawns of marine origin do not naturally harbour this group of bacteria and that any contamination is likely to he of human origin as a result of poor sanitation and standards of hygiene during processing. Some exporting countries argue that Salmonella may be naturally present in pond cultured prawns and there may be a case for revising standards. The Natural Resources Institute (NRI) in collaboration with one of the main exporting countries of cultured shrimp in Southeast Asia initiated a two year survey on the incidence of Salmonella and Vibrio cholerae in brackishwatcr ponds and in the cultured products. The objective of the study was to survey the incidence of Salmonella and V. cholerae in brackishwater ponds, and to identify the influence of environmental filctors such as pond location, weather conditions, pond growth practices and water quality on the occurrence of these microorganisms.

Materials and Methods

Sampling A total of 304 samples of prawns and mud/water samplcs were taken from 131 ponds in major aquaculture regions between February 1988 and January 1990. The ponds sampled in this survey varied in size from 0.5 to 1.5 hectares. Samples were taken throughout the prawn harvesting period in order to obtain it representative sample of the total pond population. Thc size of the sample depended upon the pond area and the quantity of prawns harvcstcd. The actual size of samples takcn from ponds is shown in Table I. Composite subsamples of ;00 g were taken for bacteriological analysis by aseptically sampling 2-3 g from individual prawns.

TABLE I Sample sizc taken rclntivc to thc cstimnted volumcof prawn harvcst Estimated volume of prawns harvcstcd (kg)

Sample sizc (kg)

< I {XMI [ I00-13(~)

l.O 1.5

1400-1800

2.0 2.5 3.0 3.5

I t~X)-2 I(X)

22[X)-26{Xl 2700-2~)

205 Samples wcrc taken randomly and included part of the shell, gut, flesh and head of prawns. Gut samples were prepared by aseptically rcmoviilg the gut from individual prawns. Representative samples of water and mud from the bottom of ponds were taken for analysis, which consisted of a I-litre composite sample from at least flve different locations in the ponds. Samples contained approximately I: I mud : water. In order to obtain a representative cross-section of the brackishwater-aquacalturc industry, samples wcrc coUcctcd from ixmds which used extensive, semi-intensive and intensive culturing systems. Various environmental factors were also monitored during the course of this survey and these factors were correlated with the presence/absence of h~th pathogenic bacteria. The factors included the type of feeds used in ponds, data on wet and dry sea,~ms, pond location, pond reconditioning, water salinity, temperature and oxygen concentration.

Exmninatkm ~Jr Sabnonclla Prawns, mud/water and gut samples were analysed for Salmonella spp. using the standard methods recommended by the USFDA (1978). To 25 g of prawns was addod 225 ml of lactose broth. For mud/water samples, ql} ml was added to 450 ml of lactose broth. Each gut sample was weighed and added to an appropriate amount of lactose broth to maintain a 1:9 ratio of sample to pre-enrichment broth. After pre-enriehment for 18 h at 37°C, I ml of the culture was transferred to l0 ml of selenite-cystine broth and a further 1 ml to 10 ml of tetrathionate broth. The enrichment media were incubated for 18 h at 35°C. Cultures were plated onto Bismuth-sulphite agar and xylose-lysine-desoxycholate agar plates, which were incubated for 18-20 h at 35°C. Suspect Salmonella colonies were picked and inoculated into Triple-sugar iron agar and Lysine-iron agar for identification. The identification of suspect colonies was confirmed at the NRi, and isolates were serotypcd by the Public Health Laboratory Service, UK.

Examination for Vibrio cholerae Samples were analysed for Vihrio cholerae using the standard methods recommended by the USFDA (1978). For each prawn sample 25 g of prawns was added to 225 ml of alkaline peptone water and 25 g to 225 ml of Gelatin Phosphate Salt (GPS) medium. For mud/water samples 90 ml was added to 450 ml of alkaline peptone water and 90 ml to 450 ml of GPS medium. Each gut sample was weighed and added to an appropriate amount of alkaline peptone water and GPS medium to maintain a 1 : 9 ratio of sample to enrichment broth. Each sample was blended with the enrichment broth and incubated for 6-8 h at 35°C. After incubation Ioopfuls of inoculum from the pellicle of the broths were plated onto three media, Thiosulphatc citrate bile salts (TCBS), Gelatin and GPS agars, and incubated for 18-24 h at 35°C. Suspect colonies were purified and identified at NRI using the methods of the USFDA (1978). Confirmed isolates were scrotypcd by the Public Health Laboratory Service, UK. With the exception of gelatin and GPS agars which were prepared from basic ingredients, all the media used wcre supplied by Oxoid, UK.

296 Results and Discussion Results from this survey showed that Salmonella spp. were present in 16.0% of prawns and 22.1% of m u d / w a t e r samples; and that V. cholerae was present in 1.5% of prawns and 3.1% of m u d / w a t e r samples from ponds (Table II). These results clearly demonstrate that Salmonella is present in brackishwater ponds and that it can also be found as part of the natural mieroflora of cultured prawns. Import standards for raw frozen prawns for the USA are based on the fact that Salmonella species are not naturally present in the marine environment, and therefore are unlikely to be detected in freshly harvested produce. Additionally Salmonella in raw frozen prawns are thought to be due to unhygienic handling and poor manufacturing practices. Results from this survey indicate that bacteria present on cultured prawns are very different to those reported in the marine species, and that the bacterial flora of prawns is a reflection of the environment. The occurrence of both Salmonella and V. cholerae in the environment is a serious cause for concern for both processors and exporters. There is a high risk that contaminated prawns will be accepted into processing plants and that pathogens will be detected in exported raw frozen products. Data from Table II show this to be the case. There is also an additional risk that cross-contamination between raw and cooked products will occur in processing plants, particularly when processors are unaware that produce are likely to be naturally contaminated with both pathogens. Public Health Authorities in the UK have found the same scrotypcs of Salmonella as have been isolated from ponds during this survey, in cooked prawns from Thailand and Malaysia (C.D.R., 1989, 1990). During the course of this survey 18 gut samples were tested for the presence of Salmonella and V. cholerae in order to identify' if either were established in the digestive systems of prawns. Although all samples were negative, data from a

TABLE I1 Incidence of pathogenic bacleria in prawns and pond (mud/water) samples from major production areas Region

No. of ponds Surveyed 1988/90

Percentage of samples Sabnondla present

V. cholerae present

Prawns

Mud/Water

Prawns

Mud/Water 0 12 0 0 3 0 0

A

Ii

18

36

0

B

25

24

28

4

C D E F G

29 7 39 8 12

7 0 IS 13 25

17 29 21 13 17

0 0 3 0 0

Total all regions

131

1O.{|

22.1

1.5

3.1

297 TABLE III Serotypes of Salmonella isolated from prawns ~*ndmud/water samples Serolype

% I.~lation

S. ~'lterreden S. anatum S. wand~worth S. ~ l s d a m S. nt'wport S. abony S. lansing S. Ihom.~mz S. htmten S. bmm,i S. hpitting~o,~

47 1I 8 8 5 5 3 3 3 3 3

previous study (Reilly, 1987) on a larger number of samples have shown that Salmonella can be present in thc gut of prawns. Table Ill shows the Salmonella isolates identified to species levels, in total, 11 different scrotypcs wcrc isolated, some of which arc well known human pathogens. S. weltevreden was the most common s ~ c i c s isolated and this organism has been previous!y reported as a common contaminant of freshly harvested cultured fish (iyer and Shrivastava, 1989). The normal cooking process of prawns will destroy all the Salmonella species isolated. D-values for S. weltevreden in a brown shrimp homogenate have been reported to be 0.68, 0.61 and 0.38 min at 60, 61.4 and 62.8°C respectively (McPherson and Zywno, 1985). ~ i w c c n 1980 and 1982 approximately 200 Salmonella isolates per year were recovered by the USFDA, and the most frequently isolated serovar has been S. weltel~reden (McPherson and Zywno, 1985). Although data on the serotypes of Salmonella isolated by the USFDA are not available, our results suggest that the contamination in cultured species is coming from ponds and not from poor standards of hygiene and processing. The occurrence of pathogens in aquaculture products has been extensively reported. A survey in Japan has shown that Salmonella species wcrc present in 21% of eel culture ponds (Saheki et aL, 1989). In the USA freshwater cultured catfish have been found to harbour Salmonella species with an overall incidence rate of about 5% (Wyatt et al., 1979). Aquatic birds are known to harbour pathogenic strains of !/. cholerae and Salmonella, and have been suggested as one possible means of spreading this organism from area to area (Ogg et al., 1989; Fenlon, 1983). In an attempt to identify the sources of pathogens in ponds, various factors were monitored and correlated with positive results for both Salmonella and V. cholerae. Table IV shows the relationship between the type of feeds and the occurrence of both pathogens. Natural feeding relies on the growth of algae and plankton on which prawns feed and is generally used in extensive culture systems. Fertilisers arc employed during pond preparation. Chicken manure is the most commonly

298 TABLE IV Relationship belween prawn feeds and the incidence ~f bacterial pathogens Feed type

No. of Ponds

Commercial Natural Natural/commercial Commercial/supplemental Natural/supplemental

q8 " 12 4 5 12

% Incidence of Sahmmella

"

V. ctuderae

Prawn

I~a.id/Water

Prawn

Mud/Water

iF, 0 25

23 17 25 20 17

2 () t)

4 II t)

a It

I) 0

(1

17

used organic fertiliser. It is usually applied at a dose of 2 tonnes per hectare of pond ( S E A F D E C , 1990). This may be one source of pathogens in the ponds. Supplemental feeding involves the addition of small fish (dried or raw) and crushed shellfish to ponds in addition to other feeds. Most ponds surveyed use comr,~*:rcial feeds either exclusively or in combination with other feeds. Previous work by Reiily (unpublished data) over a 5-year period showed that Sahnonella is not a problem in commercial feeds. Table IV shows that the highest incidences of both Salmonella and V. cholerae were found in ponds using commercial feeds. It is very unlikely that these feeds are the ,source of pathogens because of the high t e m p e r a t u r e involved in their production. A more likely explanation is that when feeding rates are high, residual or waste feeds will accumulate in ponds. This will contribute to a high organic loading in ponds which in turn will favour an increase in the n u m b e r of microorganisms, such as Salmonella and V. cholerae. O t h e r factors such as faeces from aquatic birds, ~rawn excreta, and possibly that from farm animals and land drainage will all contribute to the organic loading of ponds. It is unlikely that this problem can be overcome without redesigning brackishwatcr ponds and changing firmly established aquaculture practices. The relationship between culture methods and the incidence of both pathogens is shown in Table V. In the extensive system ponds are stocked with about 10 fry per square metre. Aeration is not used and prawns feed on naturally occurring

TABLE V Relationship between culture methods and incidence of bacterial pathogens Culture methods

Extensive Semi-intensive Intensive

No. of Ponds

Sabnonella

% Incidence of

26 56 51

Prawns 4 I1 31

V. Choterae

Mud/Water 19 13 33

Prawns 0 2 2

Mud/Water 0 2 6

TABLE VI Effects of environmental filctors on the incidence of bacterial pathogens Environmental factors WealIRr Dry ,~a,~)n Wet Sea,'a)n Location of ponds Urban Rural Semi-Rural

No. of Ponds

el, Incidence ~f Sahn
L~ Cholerae

Prawns

Mud/Water

Prawns

SI 5()

14 3}

20 26

I} 4

2 4

51

12 27 17

24

2

b

18

22

'~ 0

~ (I

16 17

16 28

2 II

23 6

12

0 8

(I 5(i

II i}

{I (I

36

19

22

3

3

II 6 L)

Pond reconditioning time: 1-4 weeks 63 5- 6 weeks 18 7-8 weeks 2 Over 8 weeks N o t monilorcd

Mud/Water

organisms. The net result of this culture method is that yields are low and grow-out periods extended. The average yields observed in this survey were in the order of 400-600 kg per hectare over grow-out times of 6 - 8 months. In the semi-intensive systems fry were stocked in the range of about 10-17 m -~. Average yields were 2 - 3 tonnes per hectare per harvest and about 2 harvests per year were achieved. Stocking densities were increased to 18-30 f r y / m z in intensive systems. Feeding rates and aeration were also increased and yields of about 6 tonnes per hectare per harvest were attained. The extensive system can be regarded as a form of polyculture as other species grow and are commercially harvested. Both semi-intensive and intensive methods are monoculture systems where pisicides are employed to remove competing species. The highest incidence of Sabnonella in both prawns (31%) and m u d / w a t e r (33%) from ponds occurred when intensive culture systems were used. The isolation of V, cholerae was also highest with intensive systems. These observations are consistent with those regarding high organic Ioadings encouraging the growth of microorganisms. Accumulation of feed and sediments in ponds will lead to loss in water quality. Excessively high stocking densities will subject the prawns to stress and result in inferior quality products. Table VI shows the effects of different environmental factors on the incidence of both pathogens. A higher incidence was observed during the wet season, which is probably due to an increase in drainage to coastal areas and rivers, which are the sources of water in ponds. Direct land drainage into ponds is also a possibility. During the wet season Salmonella was isolated from prawns in 20% of the samples, as compared to 14% for the dry season. V. cholerae was not isolated from prawns during the dry season, only during the wet season.

300 The location of ponds sampled during the survey was recorded in o r d e r to study if there was a relationship between the distance of ponds from farms, houses, villages, etc., and contamination by pathogens. T h e r e was a slightly higher incidence of Salmonella in ponds which were located in urban and semi-rural areas (Table VI). The highest occurrence of pathogens in ponds was recorded in the most densely populated area of the country where the survey was carried out. The length of time between harvesting and restocking ponds is referred to as pond reconditioning time. I n a d e q u a t e reconditioning of ponds, overstocking, poor quality fced, misuse of antibiotics and poor sanitation were some of the reasons for the failure of the prawn aquaculture industry in Taiwan during 1988/89. Poor environmental conditions in the ponds led to disease of the shrimp from the virus Monodon baeulovirus and the rapid growth of bacteria (Underwood, 1989). Table VI relates pond reconditioning time to the isolation of both pathogens, but the data is inconclusive. The results presented here indicate that both Salmonella and V. cholerae are present as part of the natural flora of brackishwater cultured prawns, and these findings arc a serious cause for concern for both processors and exporters. Prawn processors are generally unaware that cultured products mliy harbour both Salmonella and V. cholerae and of the high risks of cross-contamination between these and raw marine shrimp or o t h e r processed products. The authors suggest that raw brackishwatcr cultured prawns should be placed in a different public health risk category to marine species. Alternatively, raw brackishwater cultured shrimp could bc classified in the same classification system as whole raw chicken, which usually contains Salmonella.

References C.D.R. (1989) .gal, umelhis from fc~d, Isl quarter 1989.Communicable Disease Report week 22-W/E 2nd Jane, 1989,p. 4. C.D.R. (1990) Salmvm, lhls from f0~, Isl quarter 199(I.Communicable Di~ase Report week 20-W/E 18 May, 1990.p. 4.

Fenlon, D.R. (1983) A comparison of Subnonellu serolypes found in the faeces of gulls feeding at sewage works with serotypes in the sewage. J. Hyg. 91, 47-52. lyer, T.S.G. and Shrivaslava, K.P. (1989) On the pattern of Sabnonella ~rotypes in fishery products, froglegs and processing environments. Fish. Technol. 26. (2) 131-136. McPhcarson, R.M. and Zywno. S.R. (1985) Heat resistance of Salmonella weltevreden in brown shrimp homogenate. In: Proceedings of the Tenth Annual Tropical and Subtropical Fisheries Conference of the Americas, Texas A.&It. University U.S.A., pp. 213-220. egg, .I.E., Ryder, R.A. and Smith, ILL. Jr. (1989) Isolation of V. cho/erae from aquatic bird in Colorado and Ulah. Appl. Environ. Microbiol. 55 (I), ¢5-99. Reilly. P.J.A. (1987) Sources of contamination of tropical prawns and shrimp. Ft~,:l Lab. Newslcn. IO, 40-48. Saheki, K., Kobayashi, S. and Kawanishi, T. (1989) Sahnonella contamination of eel culture ponds. Nippon Suisan Gakknishi 55, (4) 075-079. SEAFDEC (199(I) Grow-out management fnr marine species, production and maintenance of natural food. Aqua Farm News 8, (6), 1-4. Aquaculture Dept., Tigbaran, IIoilo, Philippines.

Undcrw(~.~, L, (19~9) Is there life after the crash? Taiwun's shrimp liarmcrs ~earch for solutions (Edil~)rial). Seafi~vA I..eader Nov./Dec., 66-72. USFDA (1978) Baclcrk~logical Analytical Manual F.D.A.. Bureau of Foods, Div. Microbiolngy. AOAC Washington. U.S.A. USFDA (1989) Monthly Impor! Detention List, Dcpt. ttcalth and Human Services. F.D.A, Washington. U.S.A. Wyatt. L.E.. Nichglson, R. and Vandcrzant. C. (197q) Occurrence :rod contrl~l {~[ Sabmmella in [rcshwatcr catfish. J. Fl~'~l Sci. 44. 1|)h7-1069: L{)73.