new england journal of medicine

The
established in 1812

november 10, 2011

vol. 365

no. 19

German Outbreak of Escherichia coli O104:H4 Associated with Sprouts

Udo Buchholz, M.D., M.P.H., Helen Bernard, M.D., Dirk Werber, D.V.M., Merle M. Bhmer, Cornelius Remschmidt, M.D., Hendrik Wilking, D.V.M., Yvonne Deler, M.D., Matthias an der Heiden, Ph.D., Cornelia Adlhoch, D.V.M., Johannes Dreesman, Ph.D., Joachim Ehlers, D.V.M., Steen Ethelberg, Ph.D., Mirko Faber, M.D., Christina Frank, Ph.D., Gerd Fricke, Ph.D., Matthias Greiner, D.V.M., Ph.D., Michael Hhle, Ph.D., Sofie Ivarsson, M.Sc., Uwe Jark, D.V.M., Markus Kirchner, M.D., M.P.H., Judith Koch, M.D., Grard Krause, M.D., Ph.D., Petra Luber, Ph.D., Bettina Rosner, Ph.D., M.P.H., Klaus Stark, M.D., Ph.D., and Michael Khne, D.V.M., Ph.D.

A bs t r ac t
Background
From the Department of Infectious Disease Epidemiology (U.B., H.B., D.W., M.M.B., C.R., H.W., Y.D., M. Heiden, C.A., M.F., C.F., M. Hhle, J.K., G.K., B.R., K.S.) and the Post graduate Training for Applied Epidemiology Program (H.W., C.A.), Robert Koch Insti tute; the Federal Office of Consumer Protec tion and Food Safety (G.F., P.L.); and the Federal Institute for Risk Assessment (M.G.) all in Berlin; the Governmental Institute of Public Health of Lower Saxony, Hannover (J.D., M. Kirchner), the Lower Saxony State Office for Consumer Protection and Food Safety, Oldenburg (J.E., U.J., M. Khne), and the Veterinary University, Hannover (M.G.) all in Germany; the Swedish Institute for Communicable Disease Control, Solna, Swe den (S.I.); and the Department of Epidemi ology, Statens Serum Institut, Copenhagen (S.E.). Address reprint requests to Dr. Buch holz at the Department of Infectious Dis ease Epidemiology, Robert Koch Institute, DGZRing 1, 13086 Berlin, Germany, or at buchholzu@rki.de. Members of the Task Force EHEC (Entero hemorrhagic Escherichia coli) at the Fed eral Office of Consumer Protection and Food Safety in Germany, the Danish HUS (HemolyticUremic Syndrome) Investiga tion Team, the Governmental Institute of Public Health of Lower Saxony HUS Investi gation Team, the Lower Saxony State Of fice for Consumer Protection and Food Safety HUS Investigation Team, the Robert Koch Institute HUS Investigation Team, and the Swedish HUS Investigation Team are listed in the Supplementary Appendix, available at NEJM.org. This article (10.1056/NEJMoa1106482) was published on October 26, 2011, at NEJM.org. N Engl J Med 2011;365:176370.
Copyright 2011 Massachusetts Medical Society.

A large outbreak of the hemolyticuremic syndrome caused by Shiga-toxinproducing Escherichia coli O104:H4 occurred in Germany in May 2011. The source of infection was undetermined.
Methods

We conducted a matched casecontrol study and a recipe-based restaurant cohort study, along with environmental, trace-back, and trace-forward investigations, to determine the source of infection.
Results

The casecontrol study included 26 case subjects with the hemolyticuremic syndrome and 81 control subjects. The outbreak of illness was associated with sprout consumption in univariable analysis (matched odds ratio, 5.8; 95% confidence interval [CI], 1.2 to 29) and with sprout and cucumber consumption in multivariable analysis. Among case subjects, 25% reported having eaten sprouts, and 88% reported having eaten cucumbers. The recipe-based study among 10 groups of visitors to restaurant K included 152 persons, among whom bloody diarrhea or diarrhea confirmed to be associated with Shiga-toxinproducing E. coli developed in 31 (20%). Visitors who were served sprouts were significantly more likely to become ill (relative risk, 14.2; 95% CI, 2.6 to ). Sprout consumption explained 100% of cases. Trace-back investigation of sprouts from the distributor that supplied restaurant K led to producer A. All 41 case clusters with known trading connections could be explained by producer A. The outbreak strain could not be identified on seeds from the implicated lot.
Conclusions

Our investigations identified sprouts as the most likely outbreak vehicle, underlining the need to take into account food items that may be overlooked during subjects recall of consumption.

n engl j med 365;19

nejm.org

november 10, 2011

1763

The New England Journal of Medicine Downloaded from nejm.org on February 14, 2013. For personal use only. No other uses without permission. Copyright 2011 Massachusetts Medical Society. All rights reserved.

The

n e w e ng l a n d j o u r na l

of

m e dic i n e

uman infection with Shiga-toxin producing Escherichia coli is a major cause of postdiarrheal hemolyticuremic syndrome. This life-threatening disorder, which is characterized by acute renal failure, hemolytic anemia, and thrombocytopenia, typically affects children under the age of 5 years. Shiga-toxinproducing E. coli O157 is the serogroup that is most frequently isolated from patients with the hemolyticuremic syndrome worldwide.1 In May 2011, a large outbreak of the hemolytic uremic syndrome associated with the rare E. coli serotype O104:H4 occurred in Germany.2-5 The main epidemiologic features were that the peak of the epidemic was reached on May 21 and May 22 4,5 and that the vast majority of case subjects either resided or had traveled in northern Germany. Almost all patients from other European countries or from North America had recently returned from northern Germany.2,6,7 Of the affected case subjects, 90% were adults, and more than two thirds of case subjects with the hemolyticuremic syndrome were female.4 Early studies in Hamburg suggested that infections were probably community-acquired and were not related to food consumption in a particular restaurant. A first casecontrol study that was conducted on May 23 and 24 suggested that raw food items, such as tomatoes, cucumbers, or leaf salad,3 were the source of infection. The consumption of sprouts, which was previously implicated in outbreaks of Shiga-toxinproducing E. coli in the United States8 and Japan,9 was mentioned by only 25% of case subjects in exploratory interviews, so consumption of sprouts was not tested analytically. This report describes the investigations that were conducted by the federal agencies under the auspices of the German Ministry of Health and the Ministry of Food, Agriculture, and Consumer Protection, as well as by the respective state agencies, to identify the vehicle of infection of this international outbreak.

all three studies became available between June 2 and June 9, 2011. Results, even if preliminary, were communicated among the investigation groups and the Task Force EHEC (Enterohemorrhagic Escherichia coli) at the Federal Office of Consumer Protection and Food Safety in Berlin. If the results were judged to have appropriate validity, they were communicated as soon as possible to the public. The casecontrol and cohort studies were conducted within the framework of the Communicable Diseases Law Reform Act of Germany. Mandatory regulations were observed, and review by an ethics committee was not required.
CaseControl Study

From May 29 to June 4, we conducted a casecontrol study to further specify the type of raw vegetables associated with illness in this outbreak. A case was defined as clinically diagnosed hemolyticuremic syndrome in an adult who was hospitalized in one of three hospitals in northern Germany, located in the cities of Bremen, Bremerhaven, and Lbeck. Control subjects were individually matched with case subjects on the basis of age group and neighborhood. Case and control subjects were predominantly asked about consumption of fruit and vegetable items, including sprouts, during the 14 days before the onset of illness (for case subjects) or before the interview date (for control subjects) (for details, see the Supplementary Appendix, available with the full text of this article at NEJM.org).
Recipe-Based Restaurant Cohort Study

Me thods
Study Design

Three types of parallel studies were conducted: one casecontrol study, one recipe-based restaurant cohort study, and combined trace-back and trace-forward investigations. The main results of

Since the earlier studies had not identified a single source of infection, we conducted a cohort study at restaurant K in Lbeck, SchleswigHolstein. Preliminary information revealed that several visitor groups with subsequent cases of gastroenteric disease had eaten in restaurant K between May 12 and May 16, 2011, which was defined as the outbreak period in this study. Using the booking notes, we identified cohorts that had eaten in the restaurant during this period and asked all members about the menu items they had consumed. We interviewed the chef of the restaurant about the ingredients and their quantities used to prepare the menu items offered in the restaurant. A case was defined as an illness in a member of any of the cohorts that was associated with bloody diarrhea, self-reported laboratory-confirmed Shiga-toxinproducing E. coli O104 infection, or

1764

n engl j med 365;19

nejm.org

november 10, 2011

The New England Journal of Medicine Downloaded from nejm.org on February 14, 2013. For personal use only. No other uses without permission. Copyright 2011 Massachusetts Medical Society. All rights reserved.

Escherichia coli O104:H4 Associated with Sprouts

the hemolyticuremic syndrome with an onset of diarrhea within 2 weeks after visiting restaurant K. Noncase subjects were those who remained healthy after visiting restaurant K. We excluded from the analysis all subjects who had diarrhea that was not bloody, who did not have laboratory-confirmed Shiga-toxinproducing E. coli O104 infection, or whose disease onset was later than 14 days after visiting restaurant K. For the analysis, we evaluated only data on ingredients that were used in the dishes that were served to the restaurant guests (i.e., data that were solely based on the information obtained from the chef) (for details, see the Supplementary Appendix). To estimate the total number of cases accrued at restaurant K during the study period, we collected information on the total number of main dishes purchased in this period from billing data. Using the attack rate among persons who were served sprouts from the cohort study, we calculated the number of ill persons as the number of main dishes served containing sprouts times the attack rate among sprout eaters.
Environmental, Trace-Back, and TraceForward Investigations

Table 1. Vegetables or Fruits Evaluated in a CaseControl Study in the German Outbreak.* Case Subjects Control Subjects Exposed Exposed no./total no. (%) Sprouts Cucumbers Apples Peppers Strawberries 6/24 (25) 22/25 (88) 22/24 (92) 16/24 (67) 19/26 (73) 7/80 (9) 52/79 (66) 57/81 (70) 35/80 (44) 43/81 (53) 4.35 (1.0518.0) 3.53 (0.9612.9) 3.91 (0.8617.7) 2.66 (0.907.9) 2.33 (0.906.0) 0.04 0.06 0.08 0.08 0.08 Matched Odds Ratio (95% CI)

Food Item

P Value

* P>0.10 for raw onions, tomatoes, leaf salad, asparagus, carrots, and basil.

sprouts from suspect producer A to outbreak clusters, a study that was conducted by the food-safety authorities of the respective counties, compiled by the pertinent federal food-safety authorities, and analyzed by the task force. The task force also initiated a trace-back investigation from producer A.

R e sult s
CaseControl Study

From physicians, patients, county and regional health departments, and foreign national public health institutes, we received information on clusters or apparently sporadic cases of illness that occurred in the context of the overall epidemic among persons who had probable exposure at only one location or venue (see the Supplementary Appendix). Information was continuously reported to national and local food-safety authorities and to the task force. On the basis of findings in the early studies, food-safety authorities initially concentrated their investigations on tomatoes, cucumbers, and leaf salads, as well as on other vegetables eaten raw and salad ingredients, including toppings. Local and state food-safety authorities assessed distribution channels of raw food products connected with clusters or single case subjects with single exposures. In addition, both epidemiologically suspected and other raw food items were sampled and, after specific enrichment procedures, were tested by means of immunoassay for Shiga toxin and polymerase-chain-reaction assay for the Shiga-toxin stx2 prophage gene cluster and for genetic markers of the O104:HA strains. The task force initiated a trace-forward investigation for
n engl j med 365;19

We included 26 case subjects (9 male and 17 female) and 81 control subjects in the study. On univariable analysis, the only significant variable was sprouts (Table 1). Other food items, such as raw minced beef and milk and other dairy products, were not significantly associated with illness. The sequential addition and removal of other variables resulted in a multivariable model containing only sprouts (matched odds ratio, 5.8; 95% confidence interval [CI], 1.2 to 29.0) and cucumbers (matched odds ratio, 6.0; 95% CI, 1.1 to 31.0). Before the date on which the public was advised not to consume sprouts (June 10, 2011), only 6 of 24 case subjects (25%) remembered having consumed them (Table 1, and the Supplementary Appendix). After that date, we wished to ascertain the possible degree of false recall among the case subjects. We tried to recontact all case and control subjects who had not reported sprout consumption previously. Of 8 case subjects who could be reached, 3 (38%) remembered having eaten sprouts in the 14 days before the onset of illness. By contrast, all 37 control subjects who had not reported sprout consumption in previous interviews continued to report that they had not eaten sprouts.
november 10, 2011

nejm.org

1765

The New England Journal of Medicine Downloaded from nejm.org on February 14, 2013. For personal use only. No other uses without permission. Copyright 2011 Massachusetts Medical Society. All rights reserved.

The

n e w e ng l a n d j o u r na l

of

m e dic i n e

Table 2. Characteristics of 10 Cohorts of Subjects Who Visited Restaurant K between May 12 and May 16, 2011.* Subjects with LaboratoryConfirmed STEC Infection no. 10 1 11 2 1 1 0 3 9 3 41 9 1 4 0 1 1 0 3 9 1 29 10 2 5 1 2 1 0 4 6 2 33 4 0 1 0 0 0 0 1 1 1 8 Attack Rate for Subjects Subjects Subjects Subjects Evaluated Fulfilling Fulfilling with for Case Case Case HUS Definition Definition Definition no. (%) 34 (92) 2 (100) 25 (81) 9 (90) 9 (100) 17 (89) 9 (90) 14 (100) 25 (100) 8 (73) 152 (90) no. 9 1 5 1 1 1 0 3 9 1 31 % 26 50 20 11 11 6 0 21 36 13 20

Cohort Subjects No. Size with Data no. 1 2 3 4 5 6 7 8 9 10 Total 37 2 31 11 12 19 10 17 25 13 177 37 2 31 10 9 19 10 14 25 11 168

Male Sex % 5 50 55 0 0 37 50 57 40 36 32

Median Age (IQR) yr 49 (4557) 49 (4652) 57 (4568) 54 (5355) 46 (4348) 32 (1549) 40 (4042) 68 (6573) 74 (6975) 24 (2148) 53 (4267)

Subjects with Diarrhea

Subjects with Bloody Diarrhea

* HUS denotes hemolyticuremic syndrome, IQR interquartile range, and STEC Shigatoxinproducing Escherichia coli. The case definition was the presence of bloody diarrhea, laboratoryconfirmed infection with Shigatoxinproducing E. coli, or the hemolytic uremic syndrome with an onset of disease within 2 weeks after visiting restaurant K. According to the case definition, 16 subjects were ex cluded either because they had diarrhea that was not bloody or because the date of the onset of diarrhea was more than 14 days after visit ing restaurant K or was unknown.

Recipe-Based Restaurant Cohort Study

We identified 10 cohorts with a total of 177 persons who had eaten at restaurant K (Table 2). Of these persons, we interviewed 168 (95%), including 161 who were interviewed directly and 7 for whom information was obtained from a proxy. Among the 152 persons who could be evaluated for the case definition, 31 (20%) had an illness that fulfilled the case definition. Among these subjects, the hemolyticuremic syndrome developed in 8 (26%) (see the Supplementary Appendix). In univariable analysis of all raw food items, only visitors who had been served sprouts were significantly more likely to become ill (Table 3). The P value for the risk ratio for all other items was greater than 0.15. Of 115 persons who had been served sprouts, 31 (27%; 95% CI, 19 to 36) became case subjects, whereas none of 37 persons who had not been served sprouts reported having gastrointestinal symptoms that fulfilled the case definition. Thus, all 31 case subjects had been served menu items containing sprouts. Nearly half the menu items served in the restaurant contained raw sprouts as a garnish or were served with

a side salad containing raw sprouts. Side salads contained radicchio, Chinese cabbage, lettuce, cucumber, tomato, and sprouts. No menu item contained cooked sprouts. During this period, the restaurant used only one type of sprout assortment, which was received from a distributor in Schleswig-Holstein and contained four types of sprouts: lentil sprouts, alfalfa sprouts, fenugreek sprouts, and adzuki bean sprouts. During the outbreak period, 884 main dishes containing sprouts were served to the guests of the restaurant. Applying the attack rate of 27% among sprout eaters, we extrapolated that a total of 239 cases of bloody diarrhea (95% CI, 168 to 318) occurred among customers of this restaurant.
Environmental, Trace-Back, and TraceForward Investigations

The task force identified 41 clusters or cases with single exposures. One of the clusters was a hotel in lower Saxony in which Swedish citizens were affected (cluster 1) (Fig. 1). A spicy sprout mixture that was sampled from the distributor (distributor 1) of the hotels restaurant on June 2, 2011,

1766

n engl j med 365;19

nejm.org

november 10, 2011

The New England Journal of Medicine Downloaded from nejm.org on February 14, 2013. For personal use only. No other uses without permission. Copyright 2011 Massachusetts Medical Society. All rights reserved.

Escherichia coli O104:H4 Associated with Sprouts

Table 3. Relative Risk of Infection Associated with Sprouts and Other Raw Food Items in Univariable Analysis. Subjects Exposed (Percent of Cohort) 115 (76) 50 (33) 50 (33) 45 (30) 45 (30) 45 (30) Cases among Subjects Exposed (Attack Rate) 31 (27) 14 (28) 14 (28) 13 (29) 13 (29) 13 (29) Subjects Not Exposed (Percent of Cohort) 37 (24) 102 (67) 102 (67) 107 (70) 107 (70) 107 (70) Cases among Subjects Not Exposed (Attack Rate) 0 17 (17) 17 (17) 18 (17) 18 (17) 18 (17)

Food Item Sprouts Tomatoes Cucumbers Chinese cabbage Radicchio Lettuce

Total Subjects Evaluated no. 152 152 152 152 152 152

Relative Risk (95% CI) 14.23 (2.55) 1.68 (0.773.62) 1.68 (0.773.62) 1.72 (0.773.71) 1.72 (0.773.71) 1.72 (0.773.71)

P Value 0.001 0.18 0.18 0.17 0.17 0.17

no. (%)

tested positive for Shiga toxin on a commercial immunoassay. Although confirmatory testing later proved to be negative, immediate trace back led to producer A in Lower Saxony, at which a total of 452 water, seed, sprout, and surface samples tested negative for Shiga-toxinproducing E. coli O104:HA. Producer A was licensed as a horticultural farm and produced 18 different sorts of sprouts at the time. Protective measures consisted of regularly instructing employees on the application of proper hygiene necessary for the production of sprouts and the frequent testing of sprouts for salmonella, according to European Union regulations, as well as for coliforms. All employees of producer A were interviewed, and 5 of 15 had become ill in May 2011 or tested positive for O104:H4. Employees frequently ate sprouts produced at their company. Preferred types were fenugreek, broccoli, and garlic sprouts. Tracing forward from producer A led to four distributors (Fig. 1). Distributor 1 was connected not only to the restaurant in which cluster 1 had occurred but also to restaurant K. Distributor 4 delivered food to a caterer in Frankfurt that was linked to a cafeteria outbreak that occurred early in the epidemic.3 Subsequently, 22 more distributors (for a total of 26) that obtained sprouts from producer A were identified. Distributors were located in 7 of the 16 federal states. Each of the 41 case clusters could be linked with at least one of the identified sprout distributors (Fig. 2).10 Fenugreek or lentil sprouts were suspected as the outbreak vehicle because these types of sprouts were the common ingredients in two different sprout mixtures that were packaged for distribution by producer A and had been supplied to most of the 41 clusters.
n engl j med 365;19

In mid-June, investigations on the origin of sprouts that were consumed by additional case subjects revealed that two case subjects from Lneburg, Lower Saxony, had eaten a homegrown sprout mix that included fenugreek sprouts. The seeds for these sprouts had been purchased at a retail store that had the same supplier of seeds (supplier X) as producer A (Fig. 1).

Discussion
We report evidence from epidemiologic, microbiologic, and food trace-back and trace-forward investigations that incriminates sprouts as the vehicle of infection in this large outbreak of the hemolyticuremic syndrome associated with Shigatoxinproducing E. coli. Although definitive molecular evidence is lacking, the argument that sprouts were responsible for this outbreak is strong on the basis of the following five factors: both epidemiologic studies implicated sprouts, the restaurant study showed that 100% of cases of illness could be explained by the consumption of sprouts, no other food ingredient consumed at restaurant K was associated with the risk of illness, all 41 clusters or cases of single exposure could be linked to sprout producer A and its distribution channels, and several employees of sprout producer A who frequently consumed sprouts at the company became symptomatically ill early in the outbreak period or tested positive for Shiga-toxinproducing E. coli O104:H4. Information obtained during the outbreak investigation in Germany already hinted at an outbreak source before producer A in the seed and sprout distribution chain. Seeds that were used by producer A and by the two case subjects from
november 10, 2011

nejm.org

1767

The New England Journal of Medicine Downloaded from nejm.org on February 14, 2013. For personal use only. No other uses without permission. Copyright 2011 Massachusetts Medical Society. All rights reserved.

The

n e w e ng l a n d j o u r na l

of

m e dic i n e

Cluster 1 Hotel and restaurant, District Lneburg, Lower Saxony Distributor 1, SchleswigHolstein Cluster 2 Restaurant K, Lbeck, Schleswig Holstein

At least 18 cases

At least 50 cases

Sprout producer A

Distributor 2, Lower Saxony

Cluster 3 Restaurant, District Rotenburg/ Wmme, Lower Saxony

At least 4 cases

Distributor 3, North Rhine Westphalia

Cluster 4 Cafeteria, Bochum, North RhineWestphalia

At least 6 cases

Distributor 4, Hesse

Cluster 5 Cafeteria, Frankfurt, Hesse

At least 20 cases

Sprout seeds

Supplier X

Distributor D Homegrown sprouts, District Lneburg, Lower Saxony

Several distributors, Germany

At least 2 cases

Distributors, United Kingdom, France

Homegrown sprouts, Bordeaux, France

15 cases

Figure 1. Anatomy of the German Outbreak. Shown are the trading connections from supplier X to sprout producer A and through four distributors to five outbreak clusters, as well as to two other distributor groups leading to a cluster of illnesses caused by Shigatoxinproducing E. coli in Lneburg and an outbreak in France unrelated to producer A. Known case subjects at restaurant K included those who were part of the cohort study and others who were not.

Lower Saxony who grew their own sprouts originated from supplier X. In June 2011, an outbreak of the hemolyticuremic syndrome associated with Shiga-toxinproducing E. coli O104:HA occurred in Bordeaux, France.11 The E. coli responsible for the outbreak was genetically related to that in the German outbreak, and there was an epidemio1768
n engl j med 365;19

logic association with consumption of homegrown fenugreek sprouts. These findings spurred trace-back investigations by a task force (set up by the European Food Safety Authority (EFSA), which consisted of experts from the European Commission, relevant European Union member states, the European Center for Disease Prevennovember 10, 2011

nejm.org

The New England Journal of Medicine Downloaded from nejm.org on February 14, 2013. For personal use only. No other uses without permission. Copyright 2011 Massachusetts Medical Society. All rights reserved.

Escherichia coli O104:H4 Associated with Sprouts

tion and Control, the World Health Organization, the Food and Agriculture Organization of the United Nations, and EFSA staff members. It concluded that a certain lot of fenugreek seeds that was imported in late 2009 from Egypt was the most likely common link for the outbreaks in Germany and France.12 With respect to the point of contamination, it is possible that it occurred at the site where seeds were produced, during transportation, or at the importer. By August 2011, this question had not been resolved. The account of the outbreak was dramatic: 4321 outbreak cases, including 3469 cases of Shigatoxinproducing E. coli and 852 cases of the hemolyticuremic syndrome, had been reported by July 26, 2011, when the outbreak was declared to be over.13 By that time, 50 patients had died. After epidemiologic and food-safety investigations had concurred in identifying sprouts as the outbreak vehicle, the public had been advised on June 10 to abstain from the consumption of raw sprouts, to eliminate raw sprouts in possession at that time, and to remove any sprouts stemming from producer A.14 In addition, producer A was temporarily closed. Cases of illness still occurred until the end of July 2011, partially as a consequence of secondary transmission,15 but the number of cases dropped substantially. Early in the outbreak investigation, raw food products other than sprouts had been suspected as the vehicle. The three studies that we present here built on these findings and complement one another. The early epidemiologic findings helped food-safety authorities to streamline their investigations, which led them to turn their attention to producer A. Since the casecontrol study was conducted before sprouts came into focus, the significant association between sprouts and illness is therefore revealing and important. Nevertheless, the findings received strong support only through the restaurant study and the food traceback and trace-forward investigations. The restaurant study provided an idea of why tomatoes, cucumbers, and leaf salad had been suspected early on. The one dish that frequently exposed guests to sprouts was the side salad, which contained tomatoes, cucumbers, three sorts of leaf salads, and sprouts. Sprouts may have been the ingredient that visitors recalled least in such a mixed salad. This hypothesis is also suggested by the results of the repeated interviews in the case control study. Because the earlier studies had attempted to find a vehicle that explained the
n engl j med 365;19

Figure 2. Trading Network Leading to the German Outbreak. The trading network for the contaminated sprouts led from producer A in Lower Saxony (yellow circle) to 26 sprout distributors (black dots) and 41 identified out break clusters (red squares), established by combined back and forward tracing.

majority of cases, sprouts were missed. Although international guidelines16 generally recommend otherwise, this experience suggests that food items or ingredients that are deemed to be hard to remember should be included in analytical studies, even if such items are mentioned by less than 50% of those surveyed. Producer A was licensed as a horticultural company. Although hygienic measures were satisfactory and local food-safety authorities had inspected the company routinely under the same conditions as a food-processing company, it became apparent that European legislation has important deficits regarding Shiga-toxinproducing E. coli. Production of food that is vulnerable to contamination with this pathogen, such as sprouts or sprout seeds, should be monitored for this organism so that hygienic measures prevent amplification. In addition, both incoming seeds and outgoing food products should be tested for Shiga-toxinproducing E. coli. In general, focused restaurant studies provide a favorable situation to identify the vehicle, even in large, geographically dispersed outbreaks, because place and time of exposure are known and
november 10, 2011

nejm.org

1769

The New England Journal of Medicine Downloaded from nejm.org on February 14, 2013. For personal use only. No other uses without permission. Copyright 2011 Massachusetts Medical Society. All rights reserved.

Escherichia coli O104:H4 Associated with Sprouts

menu cards can be used to spur visitors memories. If one or few specific menu items are identified, ingredients can be further analyzed.17 If not, the reason may be that a common ingredient is contained in many menu items, resulting in the necessity to collect detailed ingredient information of the whole or a large part of the menu,18,19 information that needs to be obtained from those preparing the food. In conclusion, we have presented investigative results regarding an outbreak of the hemolytic uremic syndrome associated with Shiga-toxin producing E. coli. Under favorable circumstances, the recipe-based restaurant cohort study proved to be a quick method for detecting suspected food
References
1. Tarr PI, Gordon CA, Chandler WL.

ingredients with high reliability in a complicated setting of exposures. Recommendations regarding sprout use and consumption may need to be strengthened or adjusted as a consequence of this outbreak.
Disclosure forms provided by the authors are available with the full text of this article at NEJM.org. We thank the senior and junior owner and particularly the chef of restaurant K in Lbeck for their support and patience; physicians, patients, and colleagues in the hospitals in Bremen (Klinikum Bremen Mitte), Bremerhaven (Klinikum Bremerhaven), and Lbeck (Universittsklinikum Schleswig-Holstein, Campus Lbeck), as well as county health departments of Bremen, Bremerhaven, Cuxhaven, and Lbeck for their support in conducting the casecontrol study; and Johan Giesecke, Stephen Palmer, Kre Mlbak, Henriette de Valk, Roland Salmon, Lyle Petersen, Robert Tauxe, and Hubert Deluyker for their input into the interpretation of study and surveillance data and the management of the outbreak.

Shiga-toxin-producing Escherichia coli and haemolytic uraemic syndrome. Lancet 2005; 365:1073-86. 2. Askar M, Faber MS, Frank C, et al. Update on the ongoing outbreak of haemolytic uraemic syndrome due to Shiga toxin-producing Escherichia coli (STEC) serotype O104, Germany, May 2011. Euro Surveill 2011;16:19883. 3. Frank C, Faber MS, Askar M, et al. Large and ongoing outbreak of haemolytic uraemic syndrome, Germany, May 2011. Euro Surveill 2011;16:19878. 4. Frank C, Werber D, Cramer JP, et al. Epidemic profile of Shiga-toxinproducing Escherichia coli O104:H4 outbreak in Germany. N Engl J Med 2011. DOI: 10.1056/ NEJMoa1106483. 5. Wadl M, Rieck T, Nachtnebel M, et al. Enhanced surveillance during a large outbreak of bloody diarrhoea and haemolytic uraemic syndrome caused by Shiga toxin/ verotoxin-producing Escherichia coli in Germany, May to June 2011. Euro Surveill 2011;16:19893. 6. European Centre for Disease Prevention and Control. ECDC rapid risk assessment: outbreak of Shiga toxin-producing E. coli (STEC) in Germany. June 14, 2011 (http:// www.ecdc.europa.eu/en/publications/ Publications/1106_TER_Risk_Assessment _EColi_revised.pdf). 7. World Health Organization. International Health Regulations: EHEC outbreak in Germany (http://www.euro.who.int/en/ what-we-do/health-topics/emergencies/

international-health-regulations/ehecoutbreak-in-germany). 8. Taormina PJ, Beuchat LR, Slutsker L. Infections associated with eating seed sprouts: an international concern. Emerg Infect Dis 1999;5:626-34. 9. Michino H, Araki K, Minami S, et al. Massive outbreak of Escherichia coli O157:H7 infection in schoolchildren in Sakai City, Japan, associated with consumption of white radish sprouts. Am J Epidemiol 1999;150:787-96. 10. Bundesinstitut fr Risikobewertung (Bf R). Bedeutung von Sprossen und Keimlingen sowie Samen zur Sprossenherstellung im EHEC O104:H4 Ausbruchsgeschehen im Mai und Juni 2011 (Stellungnahme Nr. 023/2011 des BfR. July 5, 2011) (http://www.bfr.bund.de/cm/343/bedeutung _von_sprossen_und_keimlingen_sowie _samen_zur_sprossenherstellung_im _ehec_o104_h4_ausbruchsgeschehen _im_mai_und_juni_2011.pdf ). 11. Gault G, Weill FX, Mariani-Kurkdjian P, et al. Outbreak of haemolytic uraemic syndrome and bloody diarrhoea due to Escherichia coli O104:H4, south-west France, June 2011. Euro Surveill 2011;16:19905. 12. European Food Safety Authority. Tracing seeds, in particular fenugreek (Trigonella foenum-graecum) seeds, in relation to the Shiga toxin-producing E. coli (STEC) O104:H4 2011 outbreaks in Germany and France. July 5, 2011 (http://www.efsa .europa.eu/en/supporting/doc/176e.pdf). 13. Robert Koch-Institute. Informationen zum EHEC-/HUS-Ausbruchsgeschehen von

Mai bis Juli 2011 in Deutschland Ende des Ausbruchs. Epidemiol Bull 2011;31:295-6 (http://www.rki.de/cln_160/nn_2030884/ DE/Content/Infekt/EpidBull/Archiv/2011/ 31__11,templateId=raw,property= publicationFile.pdf/31_11.pdf ). 14. Idem. Neue Erkenntnisse zum EHECAusbruch: Gemeinsame Pressemitteilung von BfR, BVL und RKI. June 10, 2011 (http://www.rki.de/cln_169/nn_467482/DE/ Content/Service/Presse/Pressemitteilungen/ 2011/08__2011.html). 15. Hauri A, Gotsch U, Strotmann I, et al. Secondary transmissions during the outbreak of Shiga toxin-producing Escherichia coli O104 in Hesse, Germany, 2011. Euro Surveill 2011;16:19937. 16. Foodborne disease outbreaks: guidelines for investigation and control. Geneva: World Health Organization, 2008. 17. Mody RK, Greene SA, Gaul L, et al. National outbreak of Salmonella serotype saintpaul infections: importance of Texas restaurant investigations in implicating jalapeo peppers. PLoS One 2011;6(2): e16579. 18. Buchholz U, Mermin J, Rios R, et al. An outbreak of food-borne illness associated with methomyl-contaminated salt. JAMA 2002;288:604-10. 19. Wheeler C, Vogt TM, Armstrong GL, et al. An outbreak of hepatitis A associated with green onions. N Engl J Med 2005; 353:890-7.
Copyright 2011 Massachusetts Medical Society.

1770

n engl j med 365;19

nejm.org

november 10, 2011

The New England Journal of Medicine Downloaded from nejm.org on February 14, 2013. For personal use only. No other uses without permission. Copyright 2011 Massachusetts Medical Society. All rights reserved.