Foodborne diseases are a major cause of illness in Canada. One of the main pathogens causing cases and outbreaks of foodborne illness in Canada is Escherichia coli O157:H7. From 2008 to 2018, 11 outbreaks of E. coli O157:H7 infection in Canada were linked to leafy greens, including 7 (63.6%) linked to romaine lettuce, 2 (18.2%) linked to iceberg lettuce, and 2 (18.2%) linked to other or unspecified types of leafy greens. The consumption of lettuce in Canada, the behavior of E. coli O157:H7 on lettuce leaves, and the production practices used for romaine and iceberg lettuce do not seem to explain why a higher number of outbreaks of E. coli O157:H7 infection were linked to romaine than to iceberg lettuce. However, the difference in the shape of iceberg and romaine lettuce heads could be an important factor. Among the seven outbreaks linked to romaine lettuce in Canada between 2008 and 2018, an eastern distribution of cases was observed. Cases from western provinces were reported only twice. The consumption of romaine and iceberg lettuce by the Canadian population does not seem to explain the eastern distribution of cases observed, but the commercial distribution, travel distances, and the storage practices used for lettuce may be important factors. In the past 10 years, the majority of the outbreaks of E. coli O157:H7 infection linked to romaine lettuce occurred during the spring (March to June) and fall (September to December). The timing of these outbreaks may be explained by the availability of lettuce in Canada, the growing region transition periods in the United States, and the seasonality in the prevalence of E. coli O157:H7. The consumption of romaine lettuce by the Canadian population does not explain the timing of the outbreaks observed.
More E. coli O157:H7 infection outbreaks were linked to romaine than to iceberg lettuce.
The difference in the shape of romaine and iceberg lettuce heads could be a factor.
E. coli O157:H7 infection cases linked to romaine lettuce were seen mostly in eastern Canada.
Most E. coli O157:H7 infection outbreaks linked to romaine lettuce occurred in fall and spring.
Foodborne diseases are a major cause of illness in Canada. An estimated 4 million cases of domestically acquired foodborne illness occur in this country each year (120). These estimates also indicate that Shiga toxin–producing Escherichia coli O157 (H-antigen not specified) causes the ninth greatest number of foodborne illnesses (120), the fourth greatest number of hospitalizations, and the third greatest number of deaths (119).
E. coli O157:H7 is a gram-negative bacterium characterized by the production of Shiga toxins. The intestinal tracts of ruminant animals, such as cattle, are reservoirs for E. coli O157:H7 (51), and bacteria are shed by these animals through their feces (11). Most E. coli O157:H7 infections in humans occur from consuming contaminated food, such as ground beef, dairy products, and produce (80). In Canada, beef is responsible for an estimated 47.4% of Shiga toxin–producing E. coli O157 (H-antigen not specified) foodborne illnesses, followed by raw produce (19.5%) (19).
In the past several years, E. coli O157:H7 infection outbreaks have been linked to leafy greens, most of them to romaine lettuce (103, 105). These outbreaks prompted Health Canada, the federal department responsible for helping Canadians maintain and improve their health, to look at this issue from a Canadian perspective.
The first part of this article is a literature review of information on leafy greens, focusing specifically on romaine and iceberg lettuce and describing the production of leafy greens in Canada and the United States, the consumption of leafy greens by the general Canadian population, E. coli O157:H7 in leafy greens, and outbreaks of E. coli (including E. coli O157:H7) infection linked to leafy greens in Canada and the United States since 2008. The second part of this article consists of a scientific opinion in which information presented in the literature review is analyzed from a Canadian perspective to better understand the issues at hand.
The following literature review is mainly focused on romaine and iceberg lettuce, covering production and consumption, details about E. coli O157:H7 in these products, and information on outbreaks of E. coli (including E. coli O157:H7) infections linked to leafy greens in Canada and the United States.
OVERVIEW OF THE PRODUCTION OF LEAFY GREENS IN CANADA AND THE UNITED STATES
Types of leafy greens
Leafy greens are leafy vegetables, including escarole, endive, spinach, cabbage, kale, arugula, chard, spring mix, various types of lettuce such as iceberg, romaine, green leaf, red leaf, butter (butterhead), and baby leaf (26, 104). Within the group of leafy greens referred to as lettuce, two general categories are recognized: head lettuce and leaf lettuce.
The most well-known type of head lettuce is iceberg. Head lettuce has leaves arranged in a dense rosette that ultimately develops into a compact head (134). At the rosette stage, the early leaves are elongated and gradually increase in width with each successive leaf until they are more broad than long. After about 10 to 12 leaves have formed, newer leaves are cup-shaped and begin to overlap, enclosing the newest leaves to form a head structure. As new leaves continue to appear and expand from inside, the head becomes large and firm (85).
Examples of leaf lettuce include romaine and green and red leaf. Leaf lettuce leaves branch from a single stalk in a loose bunch or rosette, which does not develop into a compact head before harvest (135). For romaine lettuce, the heads can either be closed and used for romaine hearts or relatively open at the top (85).
Leafy greens consumed in Canada are produced domestically and imported from other countries. From 2013 to 2017, the majority of lettuce available in Canada was imported (72.8%) and the rest (27.2%) was produced domestically (3). During this time, the United States was the principal supplier of lettuce to Canada (98.7%); the remaining product was imported from countries such as Mexico (1%) (3). Most of Canada's imports of lettuce are supplied in bulk (82.5%), and the remainder (17.5%) consists of packaged fresh-cut mixes and salads (3). The majority (91%) of all Canadian lettuce imports is conventionally produced in fields, 8.9% is certified organic, and only 0.2% is greenhouse grown (3).
Greenhouse production of leafy greens in Canada and the United States
In Canada and the United States, leafy greens can be produced in a greenhouse but are mainly grown in fields. Between 2013 and 2017, only 5.9% of lettuce grown in Canada was produced in greenhouses; 94.1% was produced in fields (3). Different growing practices are used to grow greenhouse lettuce; some is grown hydroponically in a recirculated flowing film of nutrient solution, and some is grow aeroponically with a nutrient solution sprayed on the bare roots (2, 4).
Québec, British Columbia, Ontario, and Alberta are the leading producers of greenhouse lettuce in Canada. Various types of lettuce are grown all year in greenhouses, including Boston, romaine, and red and green leaf (3).
Field production of leafy greens in Canada and the United States
The United States produces leafy greens on a large scale (Table 1) (114, 125). In 2017, the United States produced ca. 50 times more lettuce than did Canada in a planted area only ca. 28 times bigger.
U.S. PRODUCTION OF LEAFY GREENS IN THE FIELD
Growing regions and seasons
Lettuce, including romaine and iceberg, are produced year-round in the United States, mainly in California (72%) and Arizona (28%), in three growing regions: central coast region, San Joaquin Valley region, and desert region (Table 2) (125). Lettuce is harvested in the central coast region from April to November, in the desert region from November to March, and in the San Joaquin Valley region in spring and fall (109, 123). Therefore, spring and fall coincide with transition between growing regions (124).
In California and Arizona, 50.6% of the lettuce produced is head lettuce, 36.4% is romaine lettuce, and 13% is leaf lettuce (125).
Both leaf lettuce and head lettuce are cool-season crops, with an optimal growing temperature of 23°C during the day and 7°C during the night. During growing seasons, most growing regions in California have daytime temperatures of 17 to 28°C and night temperatures of 3 to 12°C (109, 123). Lettuce grows best in well-drained soil such as silt loams and sandy soil (134, 135). As lettuce grows, leaves sprout internally from the root system; therefore, inner lettuce leaves are younger than outer leaves (24).
Planting to harvest of both mature leaf and head lettuce takes 65 to 80 days for midsummer plantings and as long as 130 days for late fall and winter plantings (109, 123). Prior to planting, soils are typically amended to loosen clods and improve overall quality. Lettuce beds are formed and then preirrigated with overhead sprinklers. Typically, plants are grown from pelleted seed placed in the soil with a precision planter pulled behind a tractor (134, 135).
Irrigation sources can vary depending on the growing region; in California, most growers use ground water, whereas in Arizona most growers use surface water from the Colorado River that travels to fields through irrigation canals (134, 135). Until seedlings emerge, lettuce fields are sprinkler irrigated every 2 to 3 days. After the plants are established, crops are irrigated less frequently with various types of irrigation systems, including furrows, surface drips, and sprinklers. Overall, the majority of the water is applied during the last 30 days before harvest. After the final irrigation, irrigation equipment is removed (134, 135).
To provide nutrients to lettuce, growers can use chemical fertilizers or biological soil amendments of animal origin that have been properly treated either physically (e.g., thermal treatments) or biologically (e.g., composting) (6, 134, 135). Untreated manures are not used in lettuce production because of food safety concerns (109, 123).
Many modes of lettuce harvest are in use, and the major types are mechanical and hand. For mechanical harvesting, the type of equipment used depends on the harvester and the final product. For example, a top-and-tail mechanical harvester may be used for romaine lettuce. When harvesting by hand, a special long-blade knife with an angled cutting edge is commonly used. Lettuce is not usually washed in the field (134, 135), but some processes in the field can involve water.
For most types of leaf lettuce, the outer leaves are removed in the field before the lettuce is packed (135). Leaf lettuce can be packed naked, wrapped in film, or as hearts in plastic bags (i.e., romaine lettuce) (109). For most types of head lettuce, outer leaves also are removed in the field. For head lettuce destined for direct distribution to the consumer, heads can be sealed in plastic bags and then field packed in cartons. For the head lettuce destined for further processing, heads may be cored in the field and/or stacked in bins for transport to the processing facility (134).
Lettuce is highly perishable and must be cooled as soon as possible after harvest. Cooling types include forced-air cooling, hydrocooling, and vacuum cooling. Hydrocooling and vacuum cooling are often used for leaf lettuce (27, 135), and vacuum cooling is frequently used for head lettuce (23). Forced-air cooling is generally less commonly used (134, 135).
For hydrocooling, the lettuce is immersed directly in cool water or cool water is sprayed on the lettuce as it travels through a tunnel. The cool water rapidly reduces the temperature of the lettuce. For vacuum cooling, the lettuce is placed in a closed vacuum chamber in which the reduced pressure causes water to evaporate from the lettuce and reduces its temperature For forced-air cooling, lettuce is placed in a refrigerated room with large fans to draw cold air through the lettuce and quickly drop the temperature (134, 135).
Some lettuce also can undergo further processing, including coring, trimming, washing, and precutting into various sizes (123, 134). Most commercial processing lines for production of fresh-cut leafy greens include steps for chopping or shredding, conveying, fluming, shaker table dewatering, and centrifugal drying (17, 18).
Distribution and storage conditions
Lettuce is shipped in temperature-controlled trucks throughout the United States and Canada (134, 135) and is generally kept under a well-controlled commercial cold chain system during distribution (78, 79, 142). Lettuce imported into Canada from the United States can travel a long distance (Table 3). For example, a shipment of lettuce from Salinas, CA, would travel 4,411 km to Toronto, Ontario, which represents a ca. 40-h nonstop drive. The western part of Canada is closer than the eastern part to the lettuce-growing regions in the United States (49) (Table 3).
The ideal conditions for storing or holding lettuce are temperature of 0 to 1.1°C (32 to 34°F) and relative humidity of 98 to 100% (134, 135). The U.S. Food and Drug Administration (FDA) recommends that consumers store perishable fresh fruits and vegetables such as lettuce in a clean refrigerator at ≤4.4°C (40°F) (131).
In California and Arizona, shippers of leafy greens can be members of their respective state Leafy Green Handler Marketing Agreement (LGMA), but LGMA membership is not mandatory. The LGMA program is designed to prevent contamination of leafy green crops from planting to harvest. Members are required to comply with a set of science-based food safety practices (5, 22).
In April 2019, new water safety requirements under the LGMA were put in place, including the need to categorize the source of the water, consider how and when water is applied to the crop, conduct testing to assure the water is safe for the intended use, sanitize the water when necessary, and verify that all of the above precautions have been taken (21).
CANADIAN PRODUCTION OF LEAFY GREENS IN THE FIELD
In Canada, field production of all types of lettuce is done in a manner similar to that used in the United States (82).
Growing regions and seasons
According to data from 2013 to 2017, field lettuce is mostly produced in Québec (86.6%), with some production in British Columbia (7.2%) and Ontario (4.5%) (Table 4) (3). In Québec, most of the lettuce is produced in Montérégie (ca. 85%) and in the Capitale-Nationale and Chaudière-Appalaches regions (82). Approximately 54.5% of the Canadian production is leaf lettuce (i.e., romaine, green leaf, and red leaf), and the remainder is head lettuce (i.e., iceberg, Boston, and butterhead) (83).
Depending on the year and province, lettuce is generally planted in fields after the last frosts of the season (April to May), with harvest commencing in June. Peak harvest for lettuce in Canada occurs between June and September, and imported supplies during that time drop by ca. 30% compared with the rest of the year. Harvest can continue well into October in some provinces when the weather and ground conditions are favorable (3).
In Canada, lettuce plants in fields are commonly grown from seedlings. The earliest seedlings are started in greenhouses in early February and are planted out as soon as the fields can be prepared. Lettuce can also be planted directly in the field with a precision seed planter (83, 99). Lettuce in Canada can be planted on beds in sandy peat mucks, deep black sandy loams, or loam soil (83).
In Québec, it takes ca. 53 days in the spring and 70 days in the summer for lettuce planted from seeds to reach maturity. When lettuce is planted in the fields as seedlings, 40 to 55 days are needed to reach maturity (99).
Distribution and storage conditions
According to data from 2013 to 2017, ca. 20% of the lettuce produced in Canada was exported. The largest market was the United States, which accounted for >99% of all exported product. Other destinations have included Hong Kong, Taiwan, France, Greenland, the Saint Pierre and Miquelon archipelago, and Japan (3).
Lettuce is a highly perishable product and cannot be stored for an extended period of time. Upon harvest, it is sold directly to consumers via retailers, farmers markets, or food service establishments or to wholesalers or processors for custom packaging and redistribution to retail outlets (3).
Different regulations, guidance, or practices regarding the display temperature of foods at retail apply across Canada. In Québec, the Regulation Respecting Food from the Food Products Act states in article 1.4.1 regarding holding temperature: “The products must be kept at a temperature which will ensure their preservation. Products subject to deterioration by heat except for fresh whole fruits and vegetables must be cooled without delay and kept at a constant internal and surrounding temperature not exceeding 4°C until delivered to the consumer, except during the time needed for processing or treatment requiring a higher temperature and recognized by the food industry” (71). Interpretation of this article by the Ministère de l'Agriculture, des Pêcheries et de l'Alimentation du Québec specifies that room temperature is a proper holding temperature for whole fresh fruits and vegetables such as whole romaine lettuce, whereas fresh-cut lettuce must be kept at 4°C (82). Anecdotal evidence indicates that in Canada whole lettuce for retail sale can be found either at room temperature or in refrigerators.
Health Canada's (53) recommendation to consumers is that leaf and iceberg lettuce should be kept in a refrigerator at 4°C or lower for 3 to 7 days or 1 to 2 weeks, respectively.
CanadaGAP is designed to help implement effective food safety procedures within fresh produce operations and is one of the food safety certification programs recognized by the Global Food Safety Initiative (48).
CanadaGAP (23) is a program for companies that produce, pack, repack, store, wholesale, and broker fruits and vegetables. Membership is not mandatory, but lettuce producers, handlers, and others in the leafy greens industry can voluntarily join the program.
CONSUMPTION OF LEAFY GREENS BY THE GENERAL CANADIAN POPULATION
Consumption of leafy greens according to the CCHS
In 2004 and 2015, Statistics Canada and Health Canada (115, 116) conducted a national health survey of 35,107 and 20,487 Canadians, respectively, called the CCHS with a nutrition focus. The survey was designed to collect information on food consumption through a 24-h dietary recall process.
In 2004, 11,681 Canadians (33.3% of the participants) reported consuming leafy greens in the past 24 h, whereas in 2015, 7,299 Canadians (35.6% of the participants) reported this consumption. In both years, iceberg lettuce was the most popular type of leafy green consumed, followed by romaine lettuce, with approximately four times more iceberg lettuce eaters than romaine lettuce eaters. However, in both years, romaine lettuce eaters reported eating bigger portions (measured in grams) than did iceberg lettuce eaters (Table 5).
Consumption of leafy greens according to Foodbook
Foodbook, a Canadian population-based telephone survey, also was conducted to evaluate the consumption of leafy greens by Canadians; however, in this survey consumers were asked to recall their consumption over a 7-day period. A total of 11,016 persons participated in the survey, which was conducted for 12 months in 2014 to 2015 (100).
According to this survey, 82.4% of the participants reported the consumption of leafy greens in the past 7 days; 41.1% reported eating iceberg lettuce and 48.8% reported eating romaine lettuce. Thus, more Canadians consumed romaine than iceberg lettuce in the past 7 days, in contrast to the results of the CCHS. However, the Foodbook survey included the following categories: lettuce or leafy greens on a sandwich, burger, or taco at a restaurant or fast food establishment and prepackaged lettuce or leafy greens, which could include different types of leafy greens such as iceberg and/or romaine lettuce. These survey participants reported consuming romaine more frequently than iceberg lettuce (Table 6). The Foodbook survey also collected data on the consumption of leafy greens per province and per month.
E. COLI O157:H7
E. coli O157:H7 is a gram-negative, rod-shaped bacterium characterized by the production of Shiga toxins. The intestinal tracts of ruminant animals are well-established reservoirs for E. coli O157:H7, and cattle are considered the primary maintenance reservoir host (51). The organism is shed by these animals through their feces, and shedding varies with the seasons and between individual animals. Shedding E. coli O157:H7 levels in cattle have been reported as 1,000 (3 log) CFU/g (11), but some individuals animals can shed higher levels (≥4 log CFU/g) (7). Various fresh water sources have been identified as sources of E. coli O157:H7 infections in humans, including municipal wells and recreational waters. However, most E. coli O157:H7 infections in humans are acquired from consuming contaminated food.
The infectious dose of E. coli O157:H7 is estimated to be very low, at 10 to 100 cells (132). Illness begins within 3 to 9 days after exposure and is marked by 1 or 2 days of stomach cramps and diarrhea followed by 5 to 7 days of bloody diarrhea (hemorrhagic colitis). Additional symptoms may include nausea, vomiting, fever, chills, and headache. The majority of cases self-resolve, but 5 to 15% of cases can progress to life-threatening hemolytic uremic syndrome (HUS) or thrombotic thrombocytopenic purpura (64, 97). In adults, 12% of HUS cases may result in death, and 25% of survivors are left with long-term health problems (47). Although HUS can occur in all age groups, the incidence is highest in infants, young children, and elderly individuals (97).
Health Canada considers E. coli O157:H7 a severe pathogen because infection with this bacterium can result in at least one of the following scenarios: life-threatening illness, illness requiring urgent care, symptoms that are long lasting (>1 week), or chronic sequelae. As a result, the severity of E. coli O157:H7 infection is ranked as high.
E. COLI O157:H7 IN LEAFY GREENS
A large study was performed in California to estimate the prevalence of pathogenic E. coli in lettuce. Samples of leafy greens were taken directly at farms and tested for the presence of enterohemorrhagic E. coli (EHEC) strains such as E. coli O157:H7. The EHEC prevalence significantly increased from 2007 to 2013, especially in leafy greens from California's central coast region (<0.01 to 2.5%). EHEC was also more prevalent in leafy greens from farms surrounded by grazing land (65).
A survey performed by the FDA from 2009 to 2014 revealed a 0.01% overall prevalence of E. coli O157:H7 in leafy greens at the retail level in the United States. E. coli O157:H7 was detected in one sample of romaine lettuce and one sample of spinach but was not detected in the 3,310 samples of iceberg lettuce tested (143).
In Canada, from 2009 to 2013, a targeted Canadian survey undertaken by the Canadian Food Inspection Agency (CFIA) indicated that the prevalence of E. coli O157:H7 in leafy greens at retail in Canada was 0%; E. coli O157:H7 was not detected in any of the 11,392 samples of leafy greens tested (38). Through various surveillance programs and food safety investigations, the CFIA (25) tested 11,947 samples of lettuce, including 3,478 samples of romaine lettuce, over the past 9 years. Only one sample of romaine lettuce collected under a CFIA surveillance program tested positive for E. coli O157:H7. The Ontario Ministry of Agriculture, Food and Rural Affairs (93) tested fresh vegetables from Ontario growers for the presence of Shiga toxin–producing E. coli. Of the 556 lettuce samples tested from 2013 to 2017, only 1 sample (0.17%) of leaf lettuce tested positive.
E. COLI O157:H7 ATTACHMENT TO LEAFY GREENS
E. coli O157:H7 can attach to both the leaf and the root of leafy greens (16, 37, 106, 112). Under laboratory conditions, internalization of E. coli O157:H7 through the roots of young lettuce plants has been reported (58, 84, 112). However, in other studies, no internalization of E. coli O157:H7 into leafy greens plants was found (62, 144).
Patel et al. (96) found that attachment of E. coli O157:H7 to iceberg and romaine leaf surfaces was similar. E. coli O157:H7 can attach to the surface, trichomes, and stomata of lettuce leaves (106). When lettuce is processed (e.g., cut or shredded), E. coli O157:H7 attaches preferentially to cut edges compared with intact tissues (96, 117).
In two field studies, romaine lettuce leaves were inoculated with E. coli O157:H7 via contaminated water, and the pathogen was detected primarily on the outer leaves rather than the inner leaves (86, 91). In contrast, Brandl and Amundson (16) found that E. coli O157:H7 can grow to higher levels on inner (i.e., young) romaine lettuce leaves than on the middle leaves and that young leaf exudates were richer in total nitrogen and carbon.
CONTAMINATION OF LEAFY GREENS BY
E. COLI O157:H7
Contamination of leafy greens can occur at any point in the chain of food production (e.g., preharvest, harvest, and postharvest) from various sources (52). In a study performed in Belgium, pathogenic bacteria including EHEC such as E. coli O157 (H-antigen not specified) were found more frequently in samples taken from open field farms than from those taken from greenhouse farms (56). Samples tested included lettuce, soil, and water. Potential routes of contamination for field production are described below.
ROUTE OF CONTAMINATION AT PREHARVEST
Multiple routes of E. coli contamination have been identified at the preharvest stage, including animals and insects, manure and manure-amended soil, water, seeds, and dust.
Animals and insects
Domestic and wild animals can shed E. coli O157:H7 in their feces (98). This can lead to the contamination of leafy greens directly when animals enter the field and deposit feces on the leafy greens or indirectly through contamination of soil and water (8, 70). Ruminant animals such as cattle are well-established reservoirs for E. coli O157:H7 (51). In California, E. coli O157:H7 was detected in 2.6 and 2.7% of cattle fecal samples collected on 8 and 20 cattle farms, respectively (10, 141). A higher prevalence of E. coli O157:H7 was detected in the spring (6.8%) and fall (5.8%) than in the summer (0.5%) and winter (3.4%). However, the prevalence of E. coli O157:H7 on farms was highly variable, ranging from 0 to 90% (141). In the California central coast region, E. coli O157:H7 was also detected in wildlife feces (1.16%) but less often than in feces of domestic ruminants (6.63%) (33). The pathogen was found in various species of wild animals such as deer, elk, and feral pigs (70).
Animals and insects may also play a role in the dispersal of E. coli O157:H7 (36, 45). Flies can carry E. coli on their body surfaces and transfer it to a clean surface (36). In a study on fields of leafy greens near a cattle feedlot, results indicated that pest flies can carry E. coli O157:H7 up to 180 m (13). Thus, flies can be a vehicle for contamination of leafy greens (13, 36).
Manure and manure-amended soil
E. coli O157:H7 can survive for extended periods in manure and manure-amended soil (57, 69). The edible portions of leaf lettuce plants grown in soil containing contaminated manure can become contaminated (57, 84, 112). E. coli O157:H7 was detected on leaf lettuce for up to 77 days after seedlings were planted in soil amended with inoculated compost (57). Low levels of E. coli O157:H7 (101 to 104 CFU/g) in the soil or in manure-amended soil were enough to contaminate leaf lettuce (84).
Oliveira et al. (91) found that E. coli O157:H7 transfers from soil amended with contaminated compost mostly to the outer leaves of romaine lettuce rather than the inner leaves because the outer leaves are more exposed to the environment and are more likely to become contaminated through direct contact with the soil.
Contaminated irrigation water has been identified as a source of E. coli O157:H7 in outbreaks linked to lettuce. In Sweden, the use of contaminated irrigation water drawn from a small stream was identified as the cause of a large outbreak of Shiga toxin–producing E. coli O157 (H-antigen not specified) infections linked to iceberg lettuce (110). Irrigation water was also identified as the likely source of contamination for the E. coli O157:H7 infection outbreaks linked to romaine lettuce in Canada and the United States during spring 2018 (130) and fall 2018 (133).
Contaminated water is considered one of the major sources of leafy greens contamination (143). The microbiological quality of the irrigation water used for crops is important, and groundwater is generally safer to use than surface water. However, groundwater may become contaminated, especially in areas with extensive livestock production and/or manure application to soil (59, 72). In general, the likelihood of contamination with pathogens, such as E. coli O157:H7, increases according to the following ranking (safest to least safe) (45, 72): (i) potable water or rainwater, (ii) groundwater collected in deep wells, (iii) groundwater collected in shallow wells (due to inadequate installation or improper maintenance), (iv) surface waters (particularly in proximity to animals, human habitation, and their wastes), and (v) raw or inadequately treated wastewater.
In the United States, the microbiological quality of surface water irrigation supplies (i.e., water intended for field production) was assessed in six districts: two in California and four in Washington. E. coli O157 (H-antigen not specified) was detected in ca. 4% of the samples, with significantly higher prevalence in California than in Washington (95).
In some studies, the number of lettuce samples contaminated with E. coli O157:H7 depended on the type of irrigation system used (44, 111). In a field study performed in the desert region (Yuma, AZ), sprinkler irrigation resulted in a higher number of contaminated samples of both romaine and iceberg lettuce compared with furrow and drip irrigation (44). Similar results were obtained in another study performed in a greenhouse (111). E. coli O157:H7 was transferred to green ice lettuce, a type of leaf lettuce, during both spray irrigation (mimicking sprinkler irrigation) and surface irrigation (mimicking furrow and drip irrigation). The number of lettuce samples that tested positive for E. coli O157:H7 following spray irrigation was significantly higher than the number that tested positive following surface irrigation. In contrast, in another field study no consistent difference in the number of romaine lettuce samples contaminated by E. coli O157:H7 was noted after sprinkler or drip irrigation (86).
The timing of the final irrigation before harvest can influence the extent of contamination. In field studies performed in Canada (British Columbia and Nova Scotia) and the United States (California) on romaine lettuce plants inoculated with E. coli O157:H7, a rapid decline was observed in the E. coli O157:H7 level on the lettuce leaf surface shortly after inoculation (14, 86). However, low levels of E. coli O157:H7 were still detected 4 weeks postinoculation (86). Environmental conditions such as solar radiation, UV light, and desiccation were hypothesized as reasons for the rapid rate of population decline when pathogens are grown in a low-nutrient medium, such as the lettuce leaf surface (54).
Water used for the dilution of agricultural chemicals
Flooding events have been identified as possible contributing factors in some E. coli O157:H7 infection outbreaks that were traced back to California (9, 20). This pathogen can survive for extended periods in water (31, 140); therefore, waterways such as rivers, lakes, ponds, and streams can be central reservoirs for E. coli O157:H7 (32). The water can become contaminated from a variety of sources, such as wildlife, sewage, and agricultural runoff from animal operations (46, 139). When these natural waterways flood, such as after large rain events, pathogens can be washed into produce fields (32).
In an environmental survey of watershed sites (lakes, rivers, streams, and ponds) in the central coast region of California, E. coli O157:H7 was detected in 7.91% of the 860 watershed samples tested (122). In some areas, the prevalence of E. coli O157:H7 reached 28.4%. E. coli O157:H7 was also detected more frequently in the spring (9.50%) and fall (11.75%) than in the summer (4.02%). In this survey, E. coli O157:H7 was also detected significantly more often in areas impacted by animals (10.10%), such as cattle operations, than in areas impacted by humans (5.95%) (34, 122).
In British Columbia, Canada, the overall prevalence of Shiga toxin–producing E. coli in surface water samples collected from four watershed sites was 19.1% (87). Serogroups identified in that study included O157, O26, O103, and O111. In another study performed in Canada from 2006 to 2012, the prevalence of Shiga toxin–producing E. coli O157 (H-antigen not specified) in watershed sites was 1.4% (63). As in the United States, a higher prevalence of E. coli O157:H7 was detected in agricultural sites (3%) than in reference sites (i.e., sites located away from agricultural and human fecal waste sources) (1%) (42).
When evaluating studies of the prevalence of E. coli O157:H7 in watershed sites in Canada and the United States (34, 42, 63, 122), interpretation and direct comparison of results should be done with care because differences in study design (e.g., sampling methods, sampling sites, and analytical methods) can affect these results.
Contamination of seeds and seedlings can result in the presence of pathogens on plants such as leafy greens, but the probability of seed contamination is considered minimal when seeds are produced under controlled conditions (45). E. coli O157:H7 can attach to lettuce seeds (35) and survive for a long time (136). E. coli O157:H7 inoculated onto butterhead lettuce seeds survived on 8.7% of the seeds for 2 years. During that time, the E. coli O157:H7 level decreased from 8.63 to <1.3 log CFU/g. The pathogen also was found in 12.5% of the seedlings 2 years after seed inoculation; however, mature lettuce was not tested.
Dust can carry bacteria. In one study, E. coli O157 (H-antigen not specified) was recovered from 6.66% of air samples collected in the loading area of a beef cattle feed yard after dust generation (81). All samples collected before dust generation were negative for the presence of the bacteria. E. coli O157 (H-antigen not specified) also was found in 27 samples of water, sediment, cattle and pig feces, and dust tested during an outbreak investigation in 2006 (32).
In a field study, Berry et al. (12) examined the impact of proximity to a beef cattle feedlot on E. coli O157:H7 contamination of leafy greens. E. coli O157:H7 was detected more frequently in leafy greens samples taken at 60 m (3.5%) than samples taken at 180 m (1.8%) from the feedlot. Although E. coli O157:H7 was not detected in any air samples tested, the results suggested that airborne dissemination of the bacteria from the feedlot can occur.
ROUTE OF CONTAMINATION AT HARVEST AND POSTHARVEST
Leafy greens can also become contaminated during harvest and postharvest through various routes.
Infected persons may transmit E. coli O157:H7 via the fecal-oral route (102). At the time of harvest, contamination of lettuce may occur when workers handling the produce have poor hygiene practices. Direct contact with human hands can occur at several different steps in lettuce production (e.g., harvest, removal of outer leaves, coring, and packing) (45).
Contamination of lettuce also can occur in restaurants, at retail establishments, or in the home (137). When food is not being handled properly, E. coli O157:H7 can be transferred to lettuce from sources such as raw ground beef via contaminated hands and/or cutting boards (138). Food handlers have been suspected as the route of contamination in some foodborne illness outbreaks associated with leafy greens (39, 52, 137).
Machinery and equipment may also be a source of E. coli O157:H7 contamination of lettuce. When equipment used for field trimming and coring was contaminated with E. coli O157:H7, the processed lettuce also became contaminated (118). In one study, a single contaminated coring knife transferred E. coli O157:H7 to at least 19 lettuce heads (77).
Studies performed with a pilot-scale processing line and sanitizer-free water revealed that E. coli O157:H7 can readily transfer to and persist on various surfaces used in lettuce production (e.g., shredder, conveyor, flume tank, shaker table, and dewatering centrifuge) (17) and then contaminate large quantities of previously uncontaminated leafy greens during subsequent processing (18).
Postharvest water use
After harvesting, water can be used for cooling (i.e., hydrocooling) and washing of leafy greens and can be a source of contamination or cross-contamination (126). In a fresh-cut processing operations, improperly sanitized processing wash water can become a source of microbiological contamination for every piece of produce that passes through a facility (126). Lopez-Galvez et al. (74) found that cross-contamination occurred when uninoculated lettuce was washed in the same water as inoculated fresh-cut lettuce. This cross-contamination occurred both when no sanitizers were added to the water and when some commercial sanitizers were used. However, properly implemented washing and sanitizing procedures can reduce the overall microflora, including pathogens, present on leafy greens (45).
IMPORTANT POSTCONTAMINATION FACTORS
Once leafy greens are contaminated, several factors will influence the survival, growth, and levels of E. coli O157:H7, including temperature, further processing, washing, and consumer handling practices.
Lettuce that has been processed (e.g., cut or shredded) can promote the growth of E. coli O157:H7. In one study, population sizes increased by 4.54- and 11.05-fold on cut and shredded leaves, respectively, but by only 1.95-fold on intact leaves (15). Cut surfaces of lettuce exude nutrients, which bacteria such as E. coli O157:H7 can utilize for growth (54).
Washing, even with sanitizers, can at best reduce but not eliminate pathogens on produce (45, 94). Chlorine-based sanitizers have been tested on leafy greens, and the reduction of E. coli O157:H7 was typically <1.5 log CFU/g (45, 67, 68, 73). To better understand the effects of washing, some studies have looked at the locations of E. coli O157:H7 on lettuce after being washed with chlorine solutions (66, 75). E. coli O157:H7 was mostly located in stomata, in bacterial biofilms, or in small clusters on romaine and iceberg lettuce leaves (66, 75). Other treatments can also reduce E. coli O157:H7 levels on leafy greens such as the use of ozonated water, acids (acetic acid, citric acid, and lactic acid), irradiation (gamma ray), ultrasound, and UV radiation (45).
Some fresh-cut lettuce products are packed under passive or active modified-atmosphere packaging (MAP) mainly to extend their shelf life (43). Passive MAP is the alteration of the gaseous environment resulting from produce respiration, and active MAP is the addition or removal of gases in a food packages to modify oxygen and carbon dioxide concentrations (90).
Results from several studies suggested that gas composition in a package has no direct effect on the growth of E. coli O157:H7 on fresh-cut lettuce (1, 37, 40). However, higher E. coli O157:H7 levels were reached under modified atmospheres because of the longer shelf life (37). Storage at 15°C under near-ambient air atmospheric conditions can promote greater expression of E. coli O157:H7 virulence factors on fresh-cut iceberg lettuce (107).
Temperature is an important factor contributing to survival and growth of E. coli O157:H7 on leafy greens, although refrigeration will not eliminate microbiological contamination that has already occurred (45). Storage of minimally processed leafy greens under refrigerated conditions will lead to either a decrease or no change in E. coli O157:H7 populations (1, 37, 41). E. coli O157:H7 on fresh-cut lettuce can survive for several days under refrigerated conditions (1, 76).
In general, the commercial cold chain distribution system in Canada and the United States are well controlled. However, temperature abuse can occur (Table 7) and lead to the growth of E. coli O157:H7 on leafy greens (78, 79, 89, 142). The temperature of lettuce during commercial transportation can be influenced by the location of products within the refrigerated truck (78, 79) and the season. In Canadian studies, leafy greens spent more time at temperatures >5°C during transport in the summer than during the winter (79); temperatures recorded in the winter were typically <5°C (78). The location within a retail display can also influence the temperature to which lettuce is exposed. Bagged lettuce in the bottom front area of a display were at 7 to 11°C, and bags at the top or middle of the display were at 1 to 4°C (89).
Using the temperatures recorded in their studies and predictive modeling techniques, McKellar et al. (78, 79) found that in some instances E. coli O157:H7 populations could significantly increase in fresh-cut lettuce during transportation and retail storage, specifically in the summer. The maximum increase in E. coli O157:H7 predicted was 5.2 log CFU in summer and 0.76 log CFU in winter. Zeng et al. (142) evaluated pathogen growth temperatures during commercial transport, retail storage, and display of fresh-cut romaine lettuce samples inoculated with E. coli O157:H7. No E. coli O157:H7 growth was observed during transport and retail display; however, an increase from 0.1 to 3.1 log CFU/g was observed during retail storage.
Tian et al. (121) found that E. coli O157:H7 grew to similar levels on romaine and iceberg lettuce kept at 15°C. For romaine and iceberg lettuce, E. coli O157:H7 levels were 2.33 and 2.02 log CFU/g, respectively, on day 0 and 5.20 and 5.21 CFU/g, respectively, on day 7. No significant difference in survival of E. coli O157:H7 was reported for romaine versus iceberg lettuce stored at 4°C (145).
Consumer handling practices
At home, consumers likely use a variety of different handling practices, which may include washing the lettuce, removing outer leaves, and keeping the lettuce under refrigeration. These practices can influence the survival, growth, and levels of E. coli O157:H7 on lettuce.
E. COLI INFECTION LINKED TO LEAFY GREENS IN CANADA AND THE UNITED STATES
E. coli infection linked to leafy greens in Canada and the United States since 2008
From 2008 to 2018, 57 outbreaks of E. coli infection linked to the consumption of leafy greens were identified in Canada and the United States through multiple sources (29, 30, 101, 103, 105). Types of leafy greens identified were romaine lettuce, iceberg lettuce, leaf lettuce, spinach, mesclun mix, and spring mix (Supplemental Tables S1 through S3), which indicates that outbreaks of E. coli infection can be associated with nearly all types of leafy greens.
Of the 57 E. coli infection outbreaks identified, 48 were attributed to E. coli O157 and most of the causative agents (45 of the 48 outbreaks) were identified as E. coli O157:H7. For simplicity, this review will refer to all 48 outbreaks as being caused by E. coli O157:H7. The other nine outbreaks were attributed to non-O157 E. coli; three of these outbreaks were linked to romaine lettuce and six were linked to other or unspecified types of leafy greens. No outbreaks of non-O157 E. coli infection linked to iceberg lettuce were reported. Serogroups of non-O157 E. coli identified in these outbreaks were O26, O126, and O145 (data not shown).
Our data search may not have captured all of the outbreaks of E. coli infection linked to leafy greens that occurred in Canada and the United States from 2008 to 2018; for example, intraprovincial or intrastate outbreaks not publicly available on the Internet were not captured in our data search. Sporadic cases of E. coli illnesses were also not included. A detailed list of outbreaks of E. coli O157:H7 linked to leafy greens in Canada and the United States is presented in Table S1 (romaine lettuce), Table S2 (iceberg lettuce), and Table S3 (other or unspecified types of leafy greens).
E. coli O157:H7 infection linked to leafy greens in Canada since 2008
In Canada, 11 outbreaks of E. coli O157:H7 infection linked to leafy greens were identified: 7 linked to romaine lettuce (63.6%), 2 linked to iceberg lettuce (18.2%), and 2 linked to other or unspecified types of leafy greens (18.2%). One outbreak was linked to both romaine lettuce (Table S1, outbreak 11) and iceberg lettuce (Table S2, outbreak 6). These 11 outbreaks resulted in 260 illness cases (laboratory confirmed and/or clinical cases): 210 linked to romaine lettuce (80.8%), 34 linked to iceberg lettuce (13.1%), and 16 linked to other or unspecified types of leafy greens (6.2%) (Table 8).
E. coli O157:H7 infection linked to leafy greens in the United States since 2008
In the United States, 37 outbreaks of E. coli O157:H7 infection linked to leafy greens were identified: 11 linked to romaine lettuce (29.7%), 5 linked to iceberg lettuce (13.5%), and 21 linked to other or unspecified types of leafy greens (56.8%). These 37 outbreaks resulted in 1,070 illness cases (laboratory confirmed and/or clinical cases): 491 linked to romaine lettuce (45.9%), 144 linked to iceberg lettuce (13.5%), and 435 linked to other or unspecified types of leafy greens (40.7%) (Table 9).
E. coli O157:H7 infection linked to romaine versus iceberg lettuce
Of the 48 outbreaks of E. coli O157:H7 infection identified in Canada and the United States between 2008 and 2018, 18 were linked to romaine lettuce (37.5%) and 7 were linked to iceberg lettuce (14.6%). However, 23 outbreaks, representing 47.9% of the outbreaks identified, were linked to other or unspecified types of leafy greens (e.g., spinach and green leaf lettuce) and to leafy green mixes, which could include romaine and/or iceberg lettuce.
Shredded or ready-to-eat lettuce was associated with 2 (28.6%) of the 7 outbreaks linked to iceberg lettuce and 2 (11.1%) of the 18 outbreaks linked to romaine lettuce (Tables S1 and S2). This finding suggests that processing of lettuce, especially of iceberg lettuce, may be a factor in outbreaks of E. coli O157:H7 infection. However, in most of these outbreaks, specific information on the form of the lettuce (e.g., whole lettuce or shredded lettuce) was not provided. In some of the outbreaks, iceberg and romaine lettuce were distributed to hotels, restaurants, and institutions. Because specific information on the form of the lettuce involved was not always provided, some uncertainty remains around the actual contribution of each type or form of lettuce to the total number of outbreaks and outbreak cases in each country.
In both Canada and the United States, the total and mean numbers of cases per outbreak were higher in outbreaks of E. coli O157:H7 infection linked to romaine lettuce than in outbreaks linked to iceberg lettuce. In the United States, the mean number of cases per outbreak is influenced by the outbreak in fall 2017, which included 210 cases. There also seem to be more hospitalizations, cases of HUS, and deaths associated with the consumption of romaine lettuce than with the consumption of iceberg lettuce (Tables 8 and 9). However, depending on the data source, this type of information was not always available. In Canada, this lack of data may be due to voluntary reporting of outbreaks of E. coli O157:H7 infection and the fact that central reporting of some variables, such as the number of hospitalizations and deaths, is not mandatory. Therefore, uncertainty remains around these indicators of disease severity and public health burden (101, 103, 105).
Geographical distribution of outbreaks of
E. coli O157:H7 infection linked to romaine lettuce in Canada
Among the seven outbreaks of E. coli O157:H7 infection linked to romaine lettuce in Canada between 2008 and 2018, most cased occurred in the eastern provinces. Of the seven outbreaks, cases from western provinces were only detected twice (Table S1), and in one of those two outbreaks, all of the affected individuals from western provinces reported having traveled to Québec, Ontario and the United States during their exposure period. In the United States, no distribution pattern of cases was obvious; some of the outbreaks linked to romaine lettuce involved as many as 36 states (Table S1).
Timing of outbreaks of
E. coli O157:H7 infection linked to romaine lettuce
Information on which month the outbreaks occurred was available for 17 of the 18 outbreaks linked to romaine lettuce in Canada and the United States from 2008 to 2018. The majority of these outbreaks happened during two seasons: eight occurred in the spring (March to June) and eight occurred in the fall (September to December) (Table S1). One outbreak happened in December and January. A seasonal pattern was previously reported by Heiman et al. (55); most of the outbreaks of E. coli O157:H7 infection linked to leafy greens in the United States happened in the fall (September to November) followed by the spring. In Canada, five of the seven outbreaks of E. coli O157:H7 infection linked to romaine lettuce were traced back to the United States (California or Arizona) (Table S1). Traceback information was not available for the other two outbreaks.
Farm identification and source of contamination in past outbreaks linked to romaine lettuce
In some outbreaks of E. coli O157:H7 infection linked to romaine lettuce, traceback activities were able to identify the farm where the lettuce originated (Table S1). However, detailed information on harvest practices, whether good agricultural practices were followed at the farm, results of laboratory testing, and likely sources of contamination were not always available. More information was collected for the outbreaks of E. coli O157:H7 infection linked to romaine lettuce that occurred during spring 2018 (outbreaks 15 and 16 in Table S1) and fall 2018 (outbreaks 17 and 18 in Table S1).
In the spring 2018 outbreaks, traceback investigation identified 36 growing fields on 23 farms in the Yuma, AZ, growing region as potential sources of contaminated lettuce. Growers interviewed reported the following common elements: romaine lettuce was grown under conventional agricultural practices; in most fields, biological soil amendments from animals were not used; Colorado River water via an open irrigation canal was used to irrigate the romaine lettuce and to dilute agricultural chemicals; and overhead sprinkler irrigation was used during the germination of romaine lettuce followed in most fields by furrow irrigation (128). Environmental samples were collected, and three water samples tested positive for the outbreak strain of E. coli O157:H7. These samples were collected from an irrigation canal in Yuma County located near a large concentrated animal feeding operation. The outbreak strain of E. coli O157:H7 was not found in other samples collected in the Yuma growing region (Imperial County, CA, and Yuma County, AZ), including a limited number of samples taken directly at the feedlot identified. The FDA concluded that the water from the irrigation canal where the outbreak strain was found most likely was the source of the contamination of the romaine lettuce linked to this outbreak. The timing and mechanism by which the irrigation canal became contaminated with the outbreak strain of E. coli O157:H7 were not determined (129, 130). Other pathogens, such as non-O157 E. coli, were also detected in some water samples tested (128).
In the fall 2018 outbreaks, the traceback investigation identified multiple farms from three Californian counties. Several types of samples were collected: soil, animal excreta, biological soil amendments of animal origin, and agricultural water including subsurface water, surface water, and reservoir sediment. The outbreak strain of E. coli O157:H7 was found in the sediment of an on-farm water reservoir in Santa Barbara County, CA. This E. coli O157:H7 strain was not detected in any of the 150 other samples tested. The FDA investigation indicated that the contaminated water from the on-farm water reservoir was most likely not effectively treated with a sanitizer, probably resulting in the contamination of some romaine lettuce linked to this outbreak (127, 133).
The scientific opinion presented here is based on analysis of the information from the literature review. This opinion serves to highlight topics that may be important to an understanding of the broader food safety context of this issue, including E. coli O157:H7 in leafy greens, the number of outbreaks of E. coli O157:H7 infection linked to romaine lettuce versus iceberg lettuce, and the geographical distribution and timing of outbreaks of E. coli O157:H7 infection linked to romaine lettuce in Canada.
E. COLI O157:H7 IN LEAFY GREENS
Leafy greens, including various types of lettuce, are often eaten raw (104); thus, no cooking step or other intervention is used to reliably reduce the bacterial load on these types of produce. Washing, even with sanitizers, can reduce but not eliminate pathogens (45, 94).
The estimated infectious dose of E. coli O157:H7 is very low (10 to 100 cells) (132), which suggests that E. coli O157:H7 populations do not need to increase on the lettuce to be able to cause illnesses. Should conditions allow E. coli O157:H7 to grow on lettuce, an increase in dose and the resulting increase in the likelihood of illnesses can be expected.
In Canada and the United States, the prevalence of E. coli O157:H7 in leafy greens is very low (25, 38, 93, 143). However, from 2008 to 2018, 48 outbreaks of E. coli O157:H7 infection linked to the consumption of leafy greens were identified in both countries: 18 linked to romaine lettuce (37.5%), 7 linked to iceberg lettuce (14.6%), and 23 linked to other or unspecified types of leafy greens (47.9%) (29, 101, 103, 105).
The very low prevalence of E. coli O157:H7 in leafy greens suggests that although these products are rarely contaminated by E. coli O157:H7, the frequency of consumption by the Canadian population and the very low infectious dose of this pathogen make leafy greens a important vehicle of E. coli O157:H7 illness.
NUMBER OF OUTBREAKS OF
E. COLI O157:H7 INFECTION LINKED TO ROMAINE LETTUCE VERSUS ICEBERG LETTUCE
In Canada, more E. coli O157:H7 infection outbreaks were linked to romaine than to iceberg lettuce in the past 10 years. To understand the difference in the number of outbreaks linked to romaine versus iceberg lettuce, the consumption of lettuce by the Canadian population, the behavior of E. coli O157:H7 on lettuce leaves, the production practices, and the difference in the shapes of iceberg and romaine lettuce were analyzed.
Consumption of leafy greens
Romaine and iceberg lettuce are two popular types of leafy greens consumed in Canada (100, 115, 116). Based on the data from the Foodbook surveys and CCHS reviewed above, differences in volume or frequency of consumption of these two lettuce types by the Canadian population probably does not explain the difference in the number of outbreaks of E. coli O157:H7 infection linked to these two food commodities.
E. coli O157:H7
In studies of the behavior of E. coli O157:H7 on romaine and iceberg lettuce leaves, no differences were found in bacterial attachment (96), survival at 4°C (145), or growth at 15°C (121). Therefore, these factors do not seem to explain the higher number of E. coli O157:H7 infection outbreaks linked to romaine compared with iceberg lettuce.
Growing, harvest, and postharvest practices differ among growers, which can help the industry to stay strong and competitive. Our review of the practices used for romaine and iceberg lettuce did not reveal specific components that could explain the difference observed in the number of outbreaks associated with these two food commodities. For most outbreaks of E. coli O157:H7 infection linked to leafy greens, detailed information on harvest practices and whether good agricultural practices were followed at the farm were not available.
Lettuce shape and leaf contamination
Within the United States and within Canada, romaine and iceberg lettuce are grown under similar environmental conditions (e.g., region, season, temperature, and soil composition). As lettuce grows, leaves sprout internally from the root system (24). Despite these similarities, romaine and iceberg lettuce heads differ in shape; romaine lettuce heads are more open than iceberg lettuce heads.
Because romaine lettuce heads are relatively open (85), both outer and inner leaves are exposed to E. coli O157:H7 coming from sources such as contaminated irrigation water; thus, the romaine lettuce open head shape likely increases the leaf surface area available for contamination by E. coli O157:H7 compared with that of iceberg lettuce. Two field studies confirmed the contamination of both outer and inner leaves of romaine lettuce following irrigation with contaminated water (86, 91). Although E. coli O157:H7 was detected primarily on the outer leaves, it was also detected on inner leaves. Because of the compact head formed by iceberg lettuce, the outer leaves are more likely to be contaminated by E. coli O157:H7 than the inner leaves, which are not directly exposed to the environment. Inner leaf contamination is still possible because the inner leaves would have been in contact with the environment at the beginning of the growth of the lettuce, before the head structure was formed, but this growth pattern and shape may explain the lower exposure to bacteria.
For both romaine and iceberg lettuce, outer leaves are removed in the field during harvest (134, 135). Thus, E. coli O157:H7 attached on any outer leaves would be removed, reducing the amount of bacteria present on the lettuce after harvesting. However, Brandl and Amundson (16) found that on romaine lettuce, young leaves supported the growth of E. coli O157:H7 better than did older leaves. Therefore, if E. coli O157:H7 were present on inner (i.e., young) leaves, the pathogen could potentially grow to higher levels.
Based on the information available at the time of this scientific opinion, the difference in the shape of iceberg and romaine lettuce heads could be a factor explaining why a higher number of outbreaks of E. coli O157:H7 infection have been linked to romaine than to iceberg lettuce.
Other types of leaf lettuce
Other types of leaf lettuces, such as red leaf and green leaf lettuce, also have an open shape, with both outer and inner leaves exposed to the environment. According to data from the United States, the proportion of total lettuce production represented by leaf lettuces (13%) is smaller than that for romaine (36.4%) and iceberg (50.6%) lettuce (125). Leaf lettuces are also less commonly consumed by Canadians (115, 116). These two factors could explain why fewer outbreaks of E. coli O157:H7 infection linked to other type of leaf lettuce (i.e., excluding romaine lettuce) were reported in the past 10 years (Table S3).
GEOGRAPHICAL DISTRIBUTION OF OUTBREAKS OF
E. COLI O157:H7 INFECTION LINKED TO ROMAINE LETTUCE IN CANADA
In the past 10 years, seven outbreaks of E. coli O157:H7 infection linked to romaine lettuce occurred in Canada. An eastern distribution of cases was observed (from Ontario, Québec, Nova Scotia, New Brunswick, and Newfoundland). To understand the geographical distribution of cases, the consumption of romaine and iceberg lettuce by the Canadian population, the commercial distribution, and travel distance of lettuces, and the storage practices were examined.
According to Foodbook surveys, 45.2 to 57.0% of Canadians from western provinces (British Columbia, Alberta, Saskatchewan, and Manitoba) reported consuming romaine lettuce in the past 7 days compared with 35.5 to 50.9% of Canadians from eastern provinces (Ontario, Québec, New Brunswick, Nova Scotia, Prince Edward Island, and Newfoundland) (100), which suggests that the eastern distribution of cases is probably not linked to more consumers of romaine lettuce living in these eastern provinces.
Commercial distribution and travel distance
A possible explanation for the eastern distribution of E. coli O157:H7 illness cases observed in Canada could be related to the commercial distribution of lettuce. Lettuce imported from U.S. lettuce-growing regions can travel long distances to reach Canada, and distances are farther to the eastern part of the country (Table 3). The distribution of within Canada of domestically produced lettuce is unknown.
Duration of travel can become important from a food safety perspective because temperature abuse has been reported during commercial transport in refrigerated trucks in both Canada and the United States (78, 79, 142). Temperatures of lettuce in one study conducted during the Canadian summer reached 11.4°C inside refrigerated trucks (78, 79). Therefore, longer travel distances will increase the time lettuce may spend in a transport vehicle in which temperature control may not be adequate. When temperatures are >4°C, the likelihood of the growth of resident E. coli O157:H7 on the lettuce will increase.
Temperature abuse can also occur at retail establishments. A study conducted in the United States revealed that leafy greens can be exposed to a range of temperatures (1.1 to 19.2°C) inside refrigerated retail displays (89). Although Health Canada does not have data regarding actual retail practices across the country, anecdotal observations indicate that retailers, even within a region, store their lettuce under different temperature conditions (i.e., either at room temperature or under refrigeration). However, any information indicating that lettuce retail storage practices are associated with region or other factors could help explain the eastern distribution of E. coli O157:H7 illness associated with lettuce.
TIMING OF OUTBREAKS OF
E. COLI O157:H7 INFECTION LINKED TO ROMAINE LETTUCE
In the past 10 years, the majority of the outbreaks of E. coli O157:H7 infection linked to romaine lettuce occurred during two seasons: spring (March to June) and fall (September to December) (Table S1). To understand the timing of these outbreaks, the consumption of romaine lettuce by the Canadian population, the lettuce availability in Canada, the transition from growing regions in the United States, and the seasonal prevalence of E. coli O157:H7 were analyzed.
According to the Foodbook, the percentage of Canadians reporting consumption of romaine lettuce differs across months, from 43% in December to 58.2% in July (100). In the spring and fall, the consumption of romaine lettuce does not increase compared with other seasons. It is unlikely that the difference in the consumption of romaine lettuce per month can explain the timing of outbreaks of E. coli O157:H7 infection linked to romaine lettuce in Canada.
Lettuce availability in Canada
In Canada, lettuce available to consumers is produced domestically (27.2%) or is imported (72.8%); the United States is the principal supplier, accounting for 98.7% of all lettuce imported to Canada (3). In Canada, the peak agricultural harvest season for field lettuce is June to September, with some quantities harvested in October when weather and ground conditions permit (3). This harvest schedule suggests that most of the lettuce eaten during E. coli O157:H7 infection outbreaks in the spring and fall was not domestic in origin. Five of the seven outbreaks linked to romaine lettuce in Canada were traced back to lettuce imported from the United States (either California or Arizona) (Table S1).
Future outbreaks of E. coli O157:H7 infection linked to romaine lettuce that occur in the spring and fall will not necessarily be associated with lettuce produced in U.S. fields. During the spring and fall, lettuce grown in greenhouses is also available to Canadian consumers. In a study from Belgium, pathogenic bacteria (including EHEC such as E. coli O157, H-antigen not specified) were found more frequently in samples taken from open field farms than in samples from greenhouse farms (56); therefore, although less likely, greenhouse lettuce may also be contaminated with E. coli O157:H7, resulting in illnesses among consumers.
Transition from growing regions in the United States
In the United States, spring and fall coincide with the transition between growing regions, and these two seasons are when outbreaks of E. coli O157:H7 infection linked to romaine lettuce have occurred (124). During the transition periods, movement of machinery, equipment, and employees from one growing region to another can increase the likelihood of E. coli O157:H7 contamination in leafy greens.
Seasonality in the prevalence of
E. coli O157:H7
Another possible explanation for the seasonality of outbreaks of E. coli O157:H7 infection linked to romaine lettuce may be the seasonal differences in the prevalence of E. coli O157:H7. Worley et al. (141) reported a higher prevalence of E. coli O157:H7 in cow feces collected from 20 Californian farms in the spring (6.8%) and fall (5.8%) compared with the summer (0.5%) and winter (3.4%). However, the prevalence of E. coli O157:H7 on farms was highly variable, ranging from 0 to 90%. A seasonal pattern was also reported by Tian et al. (122) in water samples collected from watershed sites in California. E. coli O157:H7 was detected more frequently in the spring (9.50%) and fall (11.75%) than in the summer (4.02%).
DATA GAPS IDENTIFIED
The following data gaps and data sharing and research needs have been identified.
Gap 1. More detailed consumption information could be collected from outbreak cases: type of lettuce consumed, form of lettuce consumed (shredded versus whole lettuce), etc.
Gap 2. More detailed information could be collected on practices used on farms that were identified during outbreak investigations.
Gap 3. Information could be collected on what factors during the transition between U.S. growing regions may impact bacterial contamination.
Gap 4. More detailed information could be collected on practices used by large versus producers, including the practice of sending products from point of origin to distant processors.
Gap 5. More detailed information could be collected on the distribution of lettuce consumed in Canada, including domestic and imported lettuce.
Gap 6. More detailed information could be collected on Canadian retailers' practices (e.g., storage times and temperatures).
Gap 7. Information could be collected on Canadian consumer practices (e.g., storage times and temperatures).
Gap 8. Field and/or laboratory studies could be conducted with nontoxic E. coli O157:H7 strains to determine the fate of E. coli O157:H7 on whole lettuce leaves (e.g., effect of temperature).
Gap 9. To further study the effects of different lettuce shapes, field studies could be conducted with nontoxic E. coli O157:H7 strains to compare the contamination and survival of E. coli O157:H7 on iceberg and romaine lettuce (e.g., levels of the bacteria on inner and outer leaves and at different times during maturation).
Gap 10. To further study the effects of similar lettuce shapes, field studies could be conducted with nontoxic E. coli O157:H7 strains to compare the contamination and survival of E. coli O157:H7 on leaf and romaine lettuce (e.g., levels of the bacteria on inner and outer leaves and at different times during maturation).
Through the review of outbreaks of E. coli O157:H7 infection linked to romaine lettuce and other leafy greens in Canada, several elements related to food safety have been highlighted. In Canada, a higher number of outbreaks of E. coli O157:H7 infection have been linked to romaine than to iceberg lettuce; the difference in the shape of the heads of these types of lettuce could be an important factor explaining the number of outbreaks associated with these two food commodities. An eastern distribution of Canadian cases of E. coli O157:H7 illness linked to romaine lettuce was observed, which may be explained by commercial lettuce distribution and travel distances and storage practices. The majority of the outbreaks of E. coli O157:H7 infection linked to romaine lettuce have happened in the spring and fall. The lettuce availability in Canada, the transition periods between U.S. growing regions, and the seasonality in the prevalence of E. coli O157:H7 may be factors explaining the timing of these outbreaks. Several data gaps were identified with the hope that epidemiological and food safety investigation practices and microbiological research can be directed to help reduce the burden of illness as a result of contamination of leafy greens with E. coli O157:H7.
Thanks are given to Pascal Delaquis and Robert Blenkinsop for their critical review of the manuscript and to the following persons and organizations for their indispensable help: Romina Code, Pascal Delaquis, Abdul Isse, and Jean Mukezangango (Agriculture and Agri-Food Canada); Grant Campbell, Irina Frenkel, Guillaume Gagnon, Alia Ghiba, and Laura Reid (CFIA); Sonya Agbessi, Marie-Josée Bolduc, Forest Dussault, Alex Gill, Darren Leyte, Nicholas Petronella, Rebecca Rutley, Adrian Verster, Kuan Chiao Wang, and Connie Zagrosh (Health Canada); Sébastien Cloutier and Julie Nolin (Ministère de l'Agriculture, des Pêcheries et de l'Alimentation du Québec); Robert Blenkinsop (Ontario Ministry of Agriculture, Food and Rural Affairs); and Judy Greg, Elizabeth Hillyer, Anne-Marie Lowe, and Mariola Mascarenhas (Public Health Agency of Canada).
Supplemental material associated with this article can be found online at: https://doi.org/10.4315/JFP-20-029.s1