Seed Sampling and Testing a Risk Reduction Stratagy for Sprouts

Seed Sampling and Testing:
A risk-reduction strategy for sprouts
Extracted from a PowerPoint Presentation by
Bob Sanderson
Jonathan's Sprouts

Parts of this presentation:

An example of seed sampling and testing

How well does seed sampling work?

Variables involved

Testing and process control

          A survey of published reports

How well does chlorine work?

          A survey of published reports

Advantages of seed sampling and testing


An example of seed sampling and testing

1.    The seed delivery arrives…

2.    A small amount of seed is sampled from each bag...

3.    We take a minimum 3kg composite sample from every seed lot, being sure to take at least one sample from every bag.

4.    3 kg will give a high probability of detecting pathogens in seed, based on observations of contamination levels in seed lots implicated in outbreaks.

  • The bigger the composite, the better.

  • The seed cost of even large samples is not a significant economic factor.

5.    The holes are covered with an identifying label...

6.    The seed is stored separately until the composite of seeds just sampled is sprouted, a runoff water sample from the sprouts is taken to an independent lab for testing, the lab returns the test results.

7.    Total sample from this arrival:
51.68 pounds...

  •  Possibly a world record seed sample.

  • Cost of seed:  about $25

  • Cost of labor for sampling:  2 people 1/2 day- $100

8.    The composite is grown out using normal sprouting procedures, but separated from ongoing production.

  •  It is important to grow the sample at temperatures that would support pathogen growth; if any pathogens were present...

  •  After 48 hours, the runoff is sampled...

  •  Paperwork detailing the chain of custody is prepared for the lab...

  •  The lab identification paper

  •  These seeds are now acceptable for sprouting.

This irrigation water sampling and testing will be repeated with every batch of sprouts before release to the stores.


Steps in Process:

  1. Every-bag seed sampling

  2. Sprouting of sample to encourage growth of pathogen if present

  3. Runoff water sampling

  4. FDA approved testing procedures at independent lab

  5. Confirming negative for Salmonella and Ecoli O157:H7


How well does seed sampling work?

Variables in detection confidence:

  1. Probability of picking up pathogens in the seed sample, if present in the lot

  2. Sprouting of seed sample

  3. Runoff water sampling

  4. Lab testing protocols

  5. Sensitivity of tests

(1) Probability of pathogens in sample:

With uniform distribution of pathogens in seed, probability of picking up at least one pathogen in seed sample described above is: 99.99999999999% at a contamination level of 1 cfu per kg. (29)


In other words, the chances are very good that this big a sample contains at least one pathogen, if present in the seed.


A much smaller sample, 3 kg, consisting of 200: 15g sample units, would have a 95% likelihood of including one pathogen, at a contamination level of 1cfu per kg, and a 99.99% chance at a contamination level of 4 cfu per kg. (1 cfu/kg is lower than levels found in seed implicated in outbreaks.)


(2) During 48 hours of sprouting, using warm (30C) irrigation water, any viable enteric organisms will proliferate.

Viable cfus in the sample will increase approximately 2-3 logs or more. (1,31)


(3) Runoff samples should be taken from all growing areas.

Runoff has been shown to have the same microbial profile as sprouts; approximately one log lower(10)


(4) Selective enrichment for pathogens in runoff water during lab testing increases viable pathogens by about 5 logs. (9)


(5) Following 5 log selective enrichment of sprout irrigation water, with initial pathogen concentration >10 cfu/l, there is high level of confidence using rapid testing methods. (9)

No seed, which has been sampled, enriched, and tested in this way, before use in production, has ever been implicated in an outbreak.


Following outbreaks, much less thorough methods have successfully isolated pathogens from implicated seed. (8,10,13,14,18,20,23,25,32,33)


If this had been done beforehand, it would have allowed the contaminated seed to be diverted to non-food uses. This would have averted:


                                                Bad press!


The answer seems to involve the relationship between testing and process control.


If seed sampling is potentially so effective, why hasn't the FDA issued a seed sampling guidance recommendation?

HACCP defines testing as merely a way to confirm the effectiveness of a process control, such as chlorination. Testing, in itself, is not considered a risk reduction step. This is regardless of the relative effectiveness of sampling and testing vs. treatment for any given product. Dozens of research efforts have demonstrated that sprouts are uniquely testable but nearly impossible to sanitize.


How well does seed sampling work? - A survey of published reports

In 1999, a researcher at CDHS (California Department of Health Services) published a report describing the recovery of Salmonella from seed, which had caused an outbreak.(14)  Numerous samples were taken, from different locations in the seed lot.  These samples were enriched by growing the seed into sprouts before taking the test sample.


Prior to this research, this would be considered a huge sample, and a very complicated enrichment method.


Until this research was published, pathogen detection in seeds was hit or miss.

Small samples were taken from few locations (2,4,13,18)  and/or, samples were not pre-grown.(24)  This resulted in much less effective pathogen detection. The poor results of these early detection methods became frequently cited in later research.


Typical evidence against the effectiveness of seed sampling:


  1.  "..the sprouting method is time-consuming and labor intensive" (14)

    This is the viewpoint of a lab researcher doing the whole process in a lab.  For a grower, this is a simple procedure.


  1. "The pathogen often cannot be isolated from the implicated seed" (33).

    Reference: "Our experience shows that Salmonella cannot be cultured from the seeds but only after sprouting." 

    In other words, the sprouting step must be an integral part of seed testing.


  1. "We obtained a 50 pound bag of seeds…. 500 grams were sampled directly; the same amount sprouted; neither yielded E. coli O157:H7" (4)

  • Only one bag of seed was available.

  • A small sample was taken from one location in the lot.

  • Only half of that sample used sprouting as pre-enrichment (500g is only 5kg).

4.    "Less than 5 kg of >20,000 kg of alfalfa seed implicated in the US trace-back investigation was available for culture.... A sample of these seeds and sprouts grown from them did not yield S. Stanley" (18)

  • A small sample was taken from one location; the specific protocol was not described.

  • In several such instances where pathogens were not found, no specifics were given for the methods used: (5,19,21,23,27)

These are typical examples of assertions, which question the effectiveness of seed sampling.  In every case, the sampling and testing protocol was either inadequate or not specified.


What is difficult in the lab is easy for a sprouter. The sprouter need only take care:

  • To avoid possible cross contamination of the sample.

  • To avoid possible cross contamination of the production crop by the sample.

Although exact quantification of seed sampling effectiveness has not been done, the empirical evidence indicates that it is an effective risk-reduction procedure.



How well does chlorine work?  - A survey of published reports


Research into the effectiveness of seed sanitation with chlorine has generally shown questionable results.   Pathogens have often been recovered from seed that has been treated with 20,000-ppm chlorine solutions. (5,25,32,34,36)


Following 20,000-ppm chlorine treatment, researchers have recovered pathogens on naturally contaminated seed, as well as on inoculated seed. (5,25,32)


On at least two occasions, sprouts grown from seed treated with 20,000-ppm chlorine have been implicated in illness. (5,25)


One published report suggests that chlorine was successful at eliminating Salmonella from seed.(33)   It is not mentioned in this report that the sampling used to measure the success of the chlorination, would have identified the presence of Salmonella in the first place.


A treatment which can be quantified (i.e. chlorine, which may give a 2- to 3-log-reduction, but rarely eliminates pathogens) is more acceptable than sampling and testing, which works very well, but is difficult to quantify.



Advantages of seed sampling and testing:

  • Can remove contaminated seed from production.

  • Can help identify farm practices that introduce pathogens into foods, and therefore encourage and reward GAPs.

  • Is safe for sprout workers.

  • Does not introduce possible risks from reduction of background flora. (30)

  • Does not introduce the possibility of selection for resistance.

  • Does not produce hazardous wastes.

  • Can be implemented by seed supplier, or sprout grower, or both.

  • Does not disenfranchise organic growers.

  • Would improve morale of sprouting industry.


Seed sampling and testing, used in conjunction with GMPs, and sampling and testing runoff water from every batch of sprouts produced, can provide the highest assurance of safe sprouts.

References cited in this presentation:

(1,2,4,5,8,9,10,13,14,18,19,20,21,23,24, 25,27,29,30,31,32,33,34,36)


Related articles: (3,6,7,11,12,15,16,17,22,26,28,35)

This list of references is annotated by the author to highlight evaluations of the efficacy of seed sampling in identifying contaminated lots of seed, and the efficacy of 20,000ppm chlorine in eliminating pathogens from contaminated seed.



(1) Andrews, W.H. et al. Microbial Hazards Associated with Bean Sprouting J Assoc. Off. Chem. (Vol. 65, No 2, 1982).

This report discusses how conditions for growing sprouts are supportive of microbial growth, including growth of Salmonella inoculated onto mung beans and alfalfa seed. There were no outbreaks discussed.



(2)Beneden Van et al. Multinational outbreak of Salmonella enterica, serotype Newport infections due to contaminated alfalfa sprouts. JAMA 1999: 281:158-162.

This report states, "The SN was grown from 1 of 52 samples of sprouts (an opened package) and none of 31 samples of lot-A seed taken from grocery stores, homes, restaurants, and sprout growers in Oregon and British Columbia. Salmonella Newport was also grown from unsprouted lot-A seed confiscated from a grower in California.


It is not specified how large the samples were, or how it was determined that they were all from seed lot-A. Later, pathogens were recovered from seed lot-A; but the protocol is not given. This mention of seed sampling is confusing, and tends to support the idea that it is hit-or-miss. But this is misleading without mentioning the specific protocols that were used.



(3) Beuchat, L.R. 1997. Comparison of chemical treatments to kill Salmonella on alfalfa seeds destined for sprout production. Inter J of Food Microbiol 34, Pages 328-333.

This research investigated treatment with various chemicals; "viable Salmonella cells were detected in seeds treated for 10 min in these solutions."



(4)Breuer et al.. A Multistate Outbreak of Escherichia coli 0157:H7 Infections Linked to Alfalfa Sprouts Grown from Contaminated Seeds. Emerging Infectious Diseases, Vol. 7, No. 6, November-December 2001 Pages 977-982.

This investigation used one 50-pound bag of seed, and, of that, sprouted 500g and plated 500g directly "…neither yielded E. coli O157:H7." "The removal of 6,000 pounds of remaining seeds from the marketplace likely prevented more illnesses, although cultures of implicated seeds and sprouts grown from them did not yieldE. coli 0157:H7. Contamination was probably not uniform in the lot of seed from which the implicated sprouts were grown, and since only a single bag, representing <0.003% of that seed lot was cultured, recovery of E. coli 0157:H7 may have been unlikely."


The failure to recover pathogens from the seed could have resulted from inadequate sampling, rather than inherent problems with sampling, and so it should not be taken as an example of the general ineffectiveness of sampling.



(5) Brooks et al. Salmonella Typhimurium Infections  Transmitted by Chlorine-pretreated Clover Sprout Seeds. Am J Epidemiol Vol. 154, No 11, 2001.

This report states, "No seed samples yielded the outbreak strain of S. Typhimuriuim. In most previously reported, sprout-associated outbreak investigations, the outbreak bacterial strain was not isolated from implicated seeds despite abundant epidemiologic evidence that seeds were the source." This statement is not referenced.


The report goes on to say, "Recovery of bacteria may be complicated by uneven and low levels of seed contamination and the limited number of samples tested. Salmonella has been recovered from sprouts grown under aseptic conditions using seeds from which bacteria could not be isolated before sprouting" This reference is incorrect. ((31)Splittstoesser) states that Salmonella was not foundin any samples of sprouts or seeds.


"The sprout, seed, and environmental samples were cultured at CDPHE and CDC using previously described methods" These methods are referenced to Cook K,. Dobbs TE, Hlady WG et al.Outbreak of Salmonella serotype Hartford infections associated with unpasteurized orange juice, JAMA 1998; 280: 1504-9 These are enrichment and testing protocols, but do not relate to appropriate seed sampling protocols.



(6) Brown, K.L., and C.A, Oscroft. Seed sampling recommendation from: Guidelines for the hygienic manufacture, distribution and retail sale of sprouted seeds with particular reference to mung beans. Campden and Chorleywood Technical Manual No 25, Sept. 1989.

These are the U.K. Guidelines for sprouting. The first recommendation: test the seed, does not give a very thorough protocol.



(7)Duynhoven, T.H.P. van et al. Salmonella enterica Serotype Enteritidis Phage Type 4b Outbreak Associated with Bean Sprouts. Emerg Infect Dis 8 (4) 2002 (CDC).

This report states, "The sprout grower applied a concentration of hypochlorite solution that was too low for seed disinfection." Although the grower used a very low concentration, it does not follow that a high concentration would have eliminated the Salmonella.


The report goes on to say, "Although the seeds tested negative for Salmonella in the routine control operations of the producer, testing is probably ineffective, as contamination may be intermittent and low level." One reference for this is ((35) Taormina), which lists 15 outbreaks up to 1998. Of these 15, pathogens were found in seed in 4 instances; in 5 instances the contamination may have come from the facility, and in the remainder, either small amounts of seed were available, or no information about the sampling is provided. All these outbreaks occurred prior to the publication of appropriate sampling protocols as described in ((13,14) Inami)


Also referenced is ((2) Beneden, van) which states both that the pathogen was recovered from the seed, and that it was not.



(8) Feng, Peter, A Summary of Background Information and Foodborne Illness Associated With the Consumption of Sprouts. 1997, Division of Microbiological Studies, OSRS, CFSAN, FDA, Washington, DC.

This article summarizes the following research into outbreaks associated with sprouts:


(8a) A 1973 B. cereus outbreak associated with home sprouting kits. Pathogens were recovered from the seed: Portnoy, B.L., J.M. Goeffert and S.M. Harmon. 1976. An outbreak of Bacillus cereus food poisoning resulting from contaminated vegetable sprouts. Am. J. Epidemiol. 103:589-594.


(8b) A Salmonella Saint-Paul outbreak associated with bean sprouts recovered from seeds and samples of sprouts from stores: O'Mahony, M., J.Cowden, B. Smith, D. Lynch, M. Hall, B. Rowe, E.L. Tearle, R.E. Tettmar, A.M. Rampling, M. Coles, R.J. Gilbert, E. Kingcott, and C.L.R. Bartlett, 1990, An outbreak of Salmonella St. Paul infection associated with bean Sprouts. Epidemiol. Infect. 104:229-235


(8c) An outbreak, in Sweden and Finland, associated with alfalfa sprouts: 1994. Lancet article given as reference. "pathogen was not isolated from seeds, but only from germinated sprouts" The Lancet article actually states, "Our experience shows that Salmonellae cannot be cultured from the seeds but only after sprouting."


(8d) Alfalfa sprouts implicated in a Denmark outbreak in1995: "Incriminated seed lots contained from 0.1 to 0.6 CFU of S. Newport/25 g of seeds." Aabo, S. and D.L. Baggesen. 1997. Growth of Salmonella Newport on naturally contaminated alfalfa sprouts and estimation of infectious dose in a Danish Salmonella Newport outbreak due to alfalfa sprouts. Salmonella and Salmonellosis, 425-426.


(8e) Alfalfa sprouts implicated in an outbreak caused by Salmonella Enterica serotype Newport, in Oregon and B.C. The serotype was the same as in the Danish outbreak (above) where pathogen was recovered from seed. This is probably the same seed lot in ((2) Beneden), where field samples didn't detect the pathogen in seed, but it was recovered from confiscated seed.


(8f) A large international outbreak, caused by Salmonella Stanley, in both Finland and the USA in 1995; no pathogen was found on any seed. "Only single bag of the implicated seed was available for analysis." This outbreak is described in ((18) Mahon).


(8g) The world's largest outbreak attributed to E.Coli O157:H7 was associated with radish sprouts. Pathogens were never found on seeds. No definitive research report is available. There was a lot of controversy over this. Recently "The Osaka District Court, on Friday, was cited as ordering the state to pay 6 million yen in compensation to a radish farmer; as government announcements over a food-poisoning case damaged shipments of the product." Associated Press, March 15, 2002



(9) Fu, T.J.  FDA, NCFST, IIT (2001) personal communications.



(10) Fu, T.J., D. Stewart, K Reineke, J Ulaszek, J. Schluesser, and M. Tortorello, 2001. Use of Spent Irrigation Water for Microbiological Analysis of Alfalfa Sprouts. J Food Prot. Vol 64, No. 6, Pages 802-806.

In this research, spent irrigation water is described as a useful way to detect pathogens in growing sprouts. Pathogens were detected in three different lots of seed that had been implicated in outbreaks.



(11) Glynn, M Kathleen et al. When Health Food Isn't So Healthy- An Outbreak of Salmonella Serotypes Anatum and Infantis Associated with Eating Contaminated Sprouts, Kansas and Missouri, 1997. Programs and abstracts of the 47th Annual Epidemic Intelligence Service Conference. Atlanta, GA 1998:16.

This report states, "Cultures of this producer's sprouts and alfalfa seeds from the same lot used to produce contaminated sprouts yielded S. Infantis and S. Anatumstrains.


This outbreak could probably have been averted with seed sampling



(12) Gordenker, A. Better Detection Methods Needed for Pathogens on Fresh Sprouts. Inside Lab Management, AOAC International, Jan. 1999.

This article contains a section titled, "Isolation of Pathogens From Seeds Very Difficult."

The basis for this statement is a number of interviews and/or unreferenced reports. Dr. Peter Feng of the FDA is identified as the source of the statement "no one has yet been able to isolate E. coli 0157:H7 from seeds."


There have been two reported outbreaks involving E. coli 0157:H7. One, in 1997, was traced to a seed lot used by growers in Michigan and Virginia; ((4)Breuer). Little seed was available for sampling.


In the other outbreak, there was no epidemiological evidence that the problem originated with the seed, since all the people who got sick ate sprouts from a single producer. It is pointed out in ((18) Mahon) that the growing conditions were very unsanitary.


There may not be any inherent difficulty in isolating E. coli O157:H7 from seeds, since there has only been a single instance where the evidence suggests the problem was seed-related, and inadequate sampling was done in this instance.

((12) Gordenker) continued...


The article also refers to a number of failures to isolate Salmonella from implicated seed.
The author states, "In some… outbreaks… investigators were able to isolate Salmonella from both the seeds and the sprouts, but in other outbreaks detection proved impossible." quoting Dr. Inami of the CDHS.


Mention is made that an investigation of seed implicated in a 1995 outbreak in Finland, Arizona, and Michigan, involving Salmonella Stanley, did not find the pathogen, however "…only a single bag, representing <0.003% of that seed lot was cultured." ((18) Mahon).


Two other outbreaks, in California in 1996, involving Salmonella Montevideo and Salmonella Meleagridis, were mentioned. Both involved a single grower, and were therefore not epidemiologically linked to seed. It is mentioned in ((21) Mohle-Boetani) that the grower "used soil" and that the facility had evidence of rodents. Although the seed used was traced to a field that was open to contamination, it seems possible that the contamination could also have been introduced by the grower, and so would not have shown up in the seed.


Although this AOAC article contributes to the general attitude of skepticism toward seed sampling, it does not provide any strong evidence to support its position.



(13)Inami et al. Two Processing Methods for the Isolation of Salmonella from Naturally Contaminated Alfalfa Seeds. Journal of Food Protection, Vol 64, No. 8, 2001, Pages 1240-1243.

In this research, thirty 100g sample units revealed Salmonella in each of three different contaminated seed lots; lot A- 3 out of 30; Lot B- 10 out of 30; Lot C- 27 out of 30.


This report provides a description of the success of detecting pathogens in naturally contaminated seed when proper protocols are followed. It states, "Isolation of a microbial pathogen from seed is laborious, but it is a necessary function for laboratories supporting epidemiological investigations or a regulatory agency." This statement overlooks the fact that the sampling would have been much more useful if it had been used on the seed prior to the outbreak, and thus averted it.



(14) Inami et al. Detection and Isolation of Salmonella from Naturally Contaminated Alfalfa Seeds Following an Outbreak Investigation. Journal of Food Protection, Vol 62, No. 6, 1999, Pages 662-664.

In this research, ten 100g samples of implicated seed were grown out and tested; 4 out of 10 were positive using EIA; 5 out of 10 were culture positive. This report contains the statement: "Sprouting method is time-consuming and labor intensive"


This report was published prior to ((13) Inami), and may be the first description of an effective method for sampling seed to detect pathogens. It contains a reference to ((19) Ponka): "Ponka et al. isolated Salmonella serotype Bovismorbificans from naturally contaminated alfalfa seeds after sprouting and growth but not from the ungerminated seeds."



(15) Itoh, Y et al. Enterohemorrhagic Escherichia coli O157: H7 present in radish sprouts. Appl. Environ Microbiol 1998; 64:1532-5.

This report deals with inoculation and absorbance into tissue. It does not mention seed sampling, but it followed the big E. Coli O157:H7 outbreak in Japan. There was apparently an unresolved question of the source of this contamination, and this research introduced the possibility that it could have been the irrigation water.



(16) Jaquette, C.B. et al. (1996) Efficacy of Chlorine and Heat Treatment in Killing Salmonella Stanley Inoculated onto Alfalfa Seeds and Growth and Survival of the Pathogen during Sprouting and Storage. Appl Environ Microbiol July 1996 Pages 2212-2215.

This report states, "These studies indicate that while populations of S. Stanley can be greatly reduced, elimination of this organism from alfalfa seeds may not be reliably achieved with traditional disinfection procedures"



(17)Mahon, B., A. Ponka, et al. An International Outbreak of Salmonella Infections Caused by Alfalfa Sprouts Grown from Contaminated Seed. Abstracts of the 36th ICAAC, ASM, Sept. 15-18, New Orleans, LA.

This report states, "Cultures of implicated seed did not yield S. Stanley, but little seed was available." (Same outbreak as described in ((18) Mahon).



(18) Mahon BE, Ponka et al.  An international outbreak of Salmonella infections caused by alfalfa sprouts grown from contaminated seeds, J. Infect. Dis 1997, 175: 876-82. 3.

This report states, "Less than 5 kg of > 20,000 kg of alfalfa seed implicated in the US trace-back investigation was available for culture. A sample of these seeds and sprouts grown from them did not yield S. Stanley.


No details are given on the protocol for sampling the available 5kg. As with ((4) Breuer), the failure to recover pathogens from the seed could as well be attributed to inadequate sampling protocols, as to inherent problems with sampling.


The Mahon report also contains the statement: "These outbreaks have usually been linked to fecal contamination of fruits and vegetables during post harvest handling, shipping, or processing in circumstances that permitted bacterial multiplication.... Alfalfa seeds….may come in contact with salmonella from the feces of birds, rodents, or other animals during growth, harvest, processing, storage, or shipping; S. Stanley has been isolated from a wide variety of animal species."


The implication that pathogens are universally present avoids the question of how soils and seeds actually become contaminated and tends to justify treatment as a necessary component of food safety.



(19) MMWR August 15, 1997, Vol 46, No 32. "Outbreaks of Escherichia coli O157:H7 Infection Associated with Eating Alfalfa Sprouts- Michigan and Virginia. June-July 1997.

This report states that "microbiological examination of seed samples is underway" -no particulars are given. No response to email inquiry 3/25/02 asking for results



(20) MMWR January 11, 2002. 51(01); 7-9 Outbreak of Salmonella serotype Kottbus infections associated with eating alfalfa sprouts- Arizona, California, Colorado, and New Mexico. February-April, 2001.

This report proposes that the grower's use of an alternative seed sanitizing treatment, instead of the FDA recommended 20,000ppm chlorine soak, was the reason for the outbreak. This proposition is unfounded, since chlorine has usually been shown to not eliminate pathogens from seed, and no comparative study was done.


The report states "A culture of a sample of this seed lot yielded S. Kottbus." Although the levels of S. Kottbus were very low, this suggests that appropriate seed sampling could have identified the problem and kept the seed out of production. The report makes no mention of the grower's production sampling. Rumor has it that there was a flaw in the grower's testing program.



(21) Mohle-Boetani, J.C., J.A Farrar, S.B. Werner, D.Minassian, R. Bryant, S. Abbott, L. Slutsker, and D.J. Vugia. 2001. Eshereschia coli O157 and Salmonella infections associated with sprouts in California, 1996-1998. Ann. Intern. Med. 135:239-247.

This report contains a number of references to the inability of researchers to isolate E. coli and Salmonella from seed that had been implicated in 6 separate outbreaks in California.


Two of the outbreaks occurred in 1996. Although the report states that pathogens were not recovered from seed, no details of the seed sampling and testing are given. It is mentioned that this grower "used soil" in his sprout production, but no details are given about the nature of this soil. Contamination by the same Salmonella strain was found in two different varieties of sprouts at the grower's facility.


Two other outbreaks, which occurred in 1997-1998, were traced to another grower. One involved Salmonella and one involved E. coli O157:H7. The implicated strain of Salmonella was recovered from a sprouting drum. Although the report states that the pathogens were not recovered from the implicated seed lots, no details are provided about the seed sampling and testing protocols which were used. Salmonella was successfully recovered from seed implicated in two other outbreaks that occurred in 1998.



(22)Mouzin, E et al. When a Health Food Becomes a Hazard: Large Outbreak of Salmonellosis Associated with Alfalfa Sprouts-California. Programs and abstracts of the 46th Annual Epidemic Intelligence Service Conference. Atlanta, GA 1997:15.

This report contains the statements, "…alfalfa seed cultures were negative" (no protocol given) "Contamination at the sprouter facility, rather than the seeds, was the probable source."



(23) FDA National Advisory Committee on Microbiological Criteria for Food,  "White Paper". Microbiological Safety Evaluations and Recommendations on Sprouted Seeds, May 28, 1999.

Although the White Paper states, "In all of the reported outbreaks the likely source of contamination was seed." there is no seed sampling among the recommendations.


Several outbreaks were mentioned: in four cases, the pathogen was isolated from implicated seed; in two cases, where pathogens were not isolated from seed, there was poor sanitation at the facilities and no reason to assume it was seed-related; in two cases, the protocols for the seed sampling are not provided.


The report also mentions the big E. coli radish sprout outbreak in Japan, the cause of which remains controversial.



(24) Ponka, A, Y. Andersson, A. Siitonen, B. deJong, M. Jahkila, O. Haikala, A. Kuhmonen, and P. al. Feb 18, 1995 Salmonella in alfalfa sprouts. Lancet Vol 345, 462-463.

This report states, "Our experience shows that salmonellae cannot be cultured from the seeds but only after sprouting." This statement has been used as evidence that seed sampling is often ineffective, whereas it actually provides information as to how it should be done.



(25)Proctor et al. Multistate Outbreak of Salmonella Serovar Muenchen Infections Associated with Alfalfa Sprouts Grown from Seeds Pretreated with Calcium Hypochlorite. J Clin Microbiol, October, 2001), pages 3461-3465.

This report states, "Low numbers of culturable Salmonella may not be detected in seed even with methods that provide the greatest potential for detection and recovery."


The reference for this statement ((14) Inami) describes the recovery of Salmonella from the remainder of a lot of alfalfa seed that had been implicated in an outbreak. Although, in this report, Salmonella was detected in 5 out of 10 samples, the statement, which is apparently the basis of the reference, is "Negative-EIA/positive culture results have been previously observed in our laboratory, suggesting that low numbers of culturable organisms may not be detected by EIA."


The report also states, "…sprouters A and B routinely disinfected sprouts according to a then-current… (FDA) procedure. Specifically, after a prewash to remove visible dirt, 5 lb of seed was treated for 15 min in 1 gal of 2 % calcium hypochlorite, drained, and rinsed again with potable water."


(26) Prokopowich D, Blank G. Microbiological evaluation of vegetable sprouts and seeds. J Food Protection 1990; 54:560.

This report contains one of the early mentions of microbial increase with sprouting: "Seeds, pregerminated in potable water (16-18 h, 20-22C) contained an APC (aerobic plate count) ca. 1 logarith higher than the corresponding dry seeds."



(27) Puohiniemi R, Heiskanen T, Siitonen A. Molecular epidemiology of two international sprout-borne Salmonella outbreaks. J Clin Microbiol 1997; 35:2487-91.

This report states, ":..the culturing of salmonella from sprouts or seeds has commonly failed."

The references for this statement are:


-A reference from Scand. J. infect. Dis (1996) having to do with Salmonella and septicemia, which is not related to seed sampling.


-A report from the 36th Interscience Conference on Antimicrobial Agents and Chemotherapy, ASM (1996) which includes the statement "Cultures of implicated seed did not yield S. Stanley, but little seed was available."


-A report in Br. J. Rheumatol (1994) dealing with reactive arthritis, not related to seed sampling.


-A report in Epidemiol Infect (1990). This report could not be located, but in 1990 nobody was sampling seed very well.


-A letter to Lancet (1995), which includes the statement: "Ponka et al. isolated Salmonella serotype Bovismorbificans from naturally contaminated alfalfa seeds after sprouting and growth, but not from the ungerminated seeds." ((24) Ponka)



(28) Sanderson, R., Jonathan's Sprouts, Inc. Letter to FDA Dockets Management Branch, 12/8/99.

Recommends issuance of seed sampling Guideline.



(29) Sanderson, R., Jonathan's Sprouts, Inc. Seed Sampling Formula.

With uniform distribution of pathogens in seed, the probability of picking up at least one pathogen in a seed sample is: P = 1- (c/t)^n where c/t = ratio of clean sample units/total, and n = number of sample units taken.



(30) Schoeller NP et al. Assessment of the Potential for Listeria monocytogenes Survival and Growth during Alfalfa Sprout Pruduction and Use of Ionizing Radiation as a Potential Intervention Treatment. J Food Prot. Vol. 65, No 8, 2002, Pages 1259-1266.

This report states, "The level of L. monocytogenes on sprouts irradiated at 1.5 kGy increased to a greater extent during refrigerated storage (a 2.5-loog CFU/g increase) than was seen with non-irradiated sprouts inoculated before packaging and subsequently refrigerated (a 1.0-log CFU/g increase…"


This suggests that an incomplete kill on seed may contribute to increased levels of pathogens, and so actually comprise an additional risk. The implications for the use of chlorine, which is known to frequently give an incomplete kill, are not mentioned.



(31) Splittstoesser et al. The microbiology of vegetable sprouts during commercial production, J. Food Safety 1983; 5,79-86.

This report describes finding similar populations of microbes on sprouts grown in aseptic conditions as in normal production conditions.



(32) Stewart et al. Growth of Salmonella during Sprouting of Alfalfa Seeds Associated wtih Salmonellosis Outbreaks Journal of Food Protection, Vol 64, No. 5, 2001, Pages 618-622.

In this experiment, Salmonella was detected in two naturally contaminated seed lots following 20,000ppm chlorine treatment by growing sprouts and testing runoff, in 48% of 100 g samples of one lot, and 31% of 100 g samples of the second lot.





(33) Suslow, T.V., Jaingchun Wu, William Fett, and Linda Harris. Detection and Elimination ofSalmonella Mbandaka from Naturally Contaminated Alfalfa Seed by Treatment with Heat or Calcium Hypochlorite. Journal of Food Protection, Vol 65, NO.3, 2002, Pages 452-458.

This is a recent research report in which no pathogens were recovered from samples of seeds or sprouts following 20,000ppm chlorine treatment.


The report contains the statement: "All tests reported to date concerning the efficacy of chemical seed treatments involving calcium hypochlorite have been performed on laboratory-inoculated seed." However, The Journal of Food Protection in which this report was published also published a report in 2001 ((32) Stewart), which investigated the efficacy of 20,000ppm calcium hypochlorite seed treatment on 2 lots of naturally contaminated seed, and found recovery and regrowth on both seed lots.


Since there was no observed survival of pathogens following treatment in this paper, the statement that it is the first and only research of its kind gives the incorrect impression that chlorine is highly effective at eliminating pathogens in naturally contaminated seed. This report understates the value of sampling and testing, and overstates the value of chlorine seed treatment, as effective risk-reduction procedures.


The report contains the statement, "…the pathogen often cannot be isolated from the implicated seed." The reference for this statement ((24) Ponka) states "Our experience shows that salmonella cannot be cultured from the seeds but only after sprouting…"


The fact that the seed sample must be grown out rather than tested directly is not a valid basis for questioning the usefulness of seed sampling and testing but merely describes necessary protocols.



(34) Taormina, P.J. and Beuchat L. (1999) Comparison of Chemical Treatments to Eliminate Enterohemorrhagic Escherichia coli  on Alfalfa Seeds. J Food Prot, Vol 62, No 4, Pages 318-324.

This report states, "Treatment with 20,000 ppm of active chlorine failed to kill 2.68 log/g of seeds."



(35) Taormina, Peter, L.R. Beuchat, L. Slutsker  1999  Infections Associated with Eating Seed Sprouts: An International Concern. Emerging Infect. Dis. Vol. 5, No. 5, Sept-Oct 1999.

This report describes 15 outbreaks up to 1998. Of these 15, pathogens were found in seed in 4 instances; in 5 instances it is mentioned that the contamination may have come from the facility, and in the remainder, either small amounts of seed were available, or no information about the sampling protocol is provided.


Since all these outbreaks occurred prior to publication of ((13,14) Inami): effective sampling protocols, it can be assumed that the methods used were mostly ineffective.



(36) Weissinger, W.R. and L. R. Beuchat. 2000. Comparison of Aqueous Chemical Treatments to Eliminate Salmonella on Alfalfa Seeds. J Food Prot. Vol 63, No 11, Pages 1475-1482.

"No treatment eliminated the pathogen, as evidenced by detection in enriched samples."