Many of the hot meal samples exceeded the limit of the AEA standard for E. coli counts (8.2%). The number of samples having a higher total count than 1.0 x 10 6 cfu/g was 9.2%, which is the limit set by the AEA for food items that have been handled after heat treatment. A previous study reported an even higher proportion of hot meals exceeding these values, for total counts (24%) and E. coli counts (25%) (Roberts et al. 1989). The following reasons may result in high bacterial counts in hot meals. Critical aspects controlling the bacterial level in hot meals are chilling, the time-temperature combination during portioning and packing, the temperature during storage in the flight kitchen, transport to the aircraft and the storage on board before serving. Considerable differencies in the means of total bacteria and E. coli counts may indicate differencies in the hygienic levels between countries. However, undercooked food items as deep-frozen blanched vegetables and steaks are commonly used in hot meals. They may be one important factor contributing to the high counts of total bacteria, E. coli , coliforms and Enterobacteriaceae , too.
The frequency of S. aureus, B. cereus and C. perfringens was lower than reported in previous studies (Roberts et al. 1989, Lambiri et al. 1995). B. cereus was the most common pathogen in this study, as was the case in a previous study, too (Roberts et al. 1989). However, quite few samples of hot dishes (0.7%) exceeded the AEA standard limit (10 3 cfu/g). Present knowledge suggests that B. cereus is highly underreported, and that any food containing more than 10 3 cfu/g can not be considered completely safe for consumption (Granum 1996). B. cereus is widely distributed in nature. The organism is present in most raw materials used in food manufacture and its ability to form spores ensures its survival through all stages of food processing. Cooking food items which are then used as ingredients for hot meals leaves a residual flora of spores of B. cereus as well as C. perfringens . Temperature abuse and inefficient heat treatment on board may lead to food poisoning.
The occurrence of S. aureus was rather low in this study (0.6%). However, the occurrence in processed foods shows contamination via the hands, indicating inadequate personal hygiene among food handlers during the preparation of food. Cooked food can be contaminated by a colonised person during handling and portioning in the flight kitchen. The storage of contaminated foods at an inappropriate temperature (7 to 46°C) could possibly have led to the multiplication of S. aureus and the formation of enterotoxins, which are very resistant to heat and will survive cooking, even some sterilisation processes (Mossel and van Netten 1990). Because S. aureus is a poor competitor, it seldom causes problems with raw products. Heat-treated proteinaceous food items are good media for their growth. S. aureus cells are salt-tolerant and may be selected for in salt-containing products or products with lowered water activities (Genigeorgis 1989). Ham included in a hot breakfast was traced to be the vehicle of one large aircraft-meal-associated outbreak (Eisenberg et al. 1975).
The occurrence of Salmonella in hot meals during the present surveys (0.3%) was similar to that of a previous study dealing with hot meals (Roberts et al. 1989), but lower than that reported by a Greek study (1%) (Lambiri et al. 1995). The source of contamination can be raw material, cross-contamination via raw materials, surfaces, utensils and infected food handlers. Although the meals undergo a final re-heating procedure on board before serving, the risk of foodborne disease is associated with contaminated hot meals. Proper heat treatment is sufficient to destroy surface contamination, but a malfunctioning oven used on an aircraft has been suspected as being the reason for an outbreak of S . Infantis (IV). Tauxe (1987) has also reported that one S . Enteritidis outbreak was associated with a hot meal served on an aircraft. Several hot meals contaminated by the same serotype during a period of one week in a Beijing flight kitchen may indicate that there was a carrier among the food handlers or that the surfaces and facilities of the kitchen were contaminated. However, no foodborne illnesses were reported after these findings.
There are no limits for aerobic colony counts in the AEA standards for cold dishes (1996). Total bacteria counts examined showed the proportion of cold meals as having higher counts than 10 6 cfu/g to be 10-41%. This was similar to a previous study (Roberts et al. 1989). A high total bacteria count fails to reflect the microbiological quality of cold meals, because appetisers, salads and desserts often include raw items such as fresh vegetables, fruits or garnishes, and they normally contain a high count of total bacteria. The use of sausages and cheeses produced using starter cultures as items in appetisers increases bacterial count, too. Many of the cold dishes (6-18%) in this study had higher E. coli counts than the AEA standard permits (1996). However, the results showed a better level than in the previous study, where 19-35% of the cold dishes exceeded the AEA limit of 10 cfu/g for E. coli . The occurrence of E. coli , especially in such high values as 1.0 x 10 6 cfu/g detected, indicates contamination and poor microbiological quality. Raw items are commonly used for appetiser and salad dishes. The highest contamination rates were found in these dishes.
The frequency of S. aureus (7%) and B. cereus (5%) in this study was higher compared to the previous study (Roberts et al. 1989), where it was 0.3% and 3% respectively. A considerably higher frequency of S. aureus (24%) was reported by Lambiri et al. (1995). Cold meals need a lot of manual handling and contamination via the hands is therefore possible. Contamination of cooked items may occur during handling and portioning. The storage of contaminated food items that are inadequately refrigerated permits the multiplication of S. aureus and enterotoxin formation. In respect to cold dishes, desserts such as custards and chocolate cakes have been implicated with aircraft outbreaks. Flight delays and subsequent temperature abuse was proved to be the final reason for two S. aureus outbreaks via desserts served on board (Munce 1978, CDC 1973). The frequency of B. cereus detected in this study indicates that it is difficult to avoid this microbe in cold meals. The source of contamination is raw materials, where low counts of B. cereus are commonly found (Kramer and Gilbert 1989). Inadequate refrigeration and storage at too high a temperature leads to a growth of bacteria and enterotoxin formation.
Salmonella was found in only one (0.1%) of the 1576 cold meals examined in present studies. It was lower than in the previous survey by Roberts et al. (1989), where it was 0.5% in cold meals. The Salmonella positive finding in one cold meal was subsequently found to be connected with an outbreak among air passengers. An appetiser prepared in Bangkok was shown to be the source of S . Ohio infection of five Finnish passengers in 1990 (Jahkola 1992). The contamination of a cold meal containing ham, together with the meals being transported unchilled during a long-haul flight, was considered to have been the reason for this outbreak. Vehicles of previous Salmonella outbreaks traced to meals served on board were mostly cold meals (Tauxe et al. 1987). In this study Salmonella cases indicated a higher risk of foodborne illness associated with cold than with hot meals. In order to avoid faecal contamination via raw vegetables, they are disinfected by soaking in a solution containing chlorine 60-100 ppm in the flight kitchen (Asplund 2000).
The results of this study show differences in the microbiological quality of meals between the countries where the food was prepared. Especially the means of E. coli and B. cereus refers to differencies between hygienic level in production.
The infection probably spread to the catering establishment staff via contaminated cold cuts sliced in the cold kitchen. The number of infected food handlers (23/118, 19%) was considerable. The fact that most of the infected food handlers were symptom-free carriers (17/23, 74%) and that the rest had mild symptoms, but worked normally, must have led to widespread contamination of the kitchen’s production, causing the infection to spread to railway and airline passengers, too. If the medical service of the airline company had immediately started to investigate the cause of the gastrointestinal illness of the food handlers in the beginning of August and excluded them from work, the spread of contamination of the products could possibly have been prevented. An exceptional heat wave in Finland at that time and shortcomings in the cold chain were contributing factors. There was neither any in-house control system in the kitchen nor any food hygiene education for employees at that time. As a consequence of the outbreak the airline company recruited a food hygienist a few years later. The results of a recent study dealing with food handlers’ (411) knowledge about foodborne diseases still strongly emphasised the need for educational hygiene courses (Angelillo et al. 2000).
Raw materials used in the catering establishment were widely tested, but they were Salmonella negative. Because none of the food handlers had recently been abroad, it was considered possible that one of them became infected via Finnish food during July in 1986. Salmonella bacteria were seldom found in Finnish foodstuffs at that time (Nurmi and Schildt 1987), nor are they today (Ministry of Agriculture and Forestry 1999, 2000). However, the source of the Salmonella outbreak remained unclear.
Train passengers on several routes became infected, whereas the gastrointestinal illness of air passengers was traced to one flight, only. Egg sandwiches for trains as well as hot aircraft ready meals were prepared at the same time in the flight kitchen as the infection had spread among kitchen staff. The batch of hot aircraft meals, from where Salmonella was isolated, had been delivered to other charter flights, too. The malfunction of the regeneration oven used to re-heat the dishes served hot on board may be the reason for the spread of infection on this particular flight.
In many national epidemiological registries, non-typhoid Salmonella spp. continue to figure prominently as the leading cause of bacterial foodborne disease (D’Aust 1994, Notermans and Borgdorff 1997). The growing importance of the international food trade between countries that maintain widely different levels of hygiene in their agricultural and food processing industries presents a health concern also for flight kitchens supplying airline companies. The ubiquitous distribution of Salmonella in the natural environment and its prevalence in the global food chain predicate the need for stringent controls at all levels of the food production process. From the 2500 Salmonella serovars currently known, only 10 to 15 are of epidemic importance, in the first place S. Typhimurium and S . Enteritidis. S . Enteritidis became the predominant serotype in western Europe, North America and South America in the late 1980s and early 1990s (Rodrique et al. 1990, Bean and Griffin 1990, Sockett et al. 1993). Eggs and raw meat and meat products were of prime importance.
The prevalence of S. aureus among Finnish flight catering employees (29% and 9% according to nasal and hand sampling, respectively) corresponded to many previous studies of healthy humans (Williams 1963, McBride et al. 1975, Oteri and Ekanem 1989, Caruso et al. 1992, Namura et al. 1995, Al Bustan et al. 1996). The present study and previous studies (Bergdoll 1989, Röder et al. 1995) showed that half of the strains from human carriers are enterotoxigenic. A person harbouring S. aureus must be considered as a potential source of enterotoxigenic strains. In addition, coagulase negative staphylococci isolated on the hands of food handlers may produce staphylococcal enterotoxins and be a potential cause of food poisoning (Danielsson and Hellberg 1984, Udo et al. 1999).
The majority of the food handlers studied had actively used hand disinfectants. This might reflect in the lower detection rate of S. aureus in the hands (9%) compared to nasal samples (29%). These results indicate that examining S. aureus from hand samples only is not always a reliable way to detect the carriage of S. aureus . Preparing food for aircraft is a highly vulnerable operation, and therefore testing carriage among food handlers is of valuable assistance in planning preventive measures. A potential risk of foodborne disease was shown by the results of the present study, which found S. aureus in both hot and cold aircraft meals, 0.6% and 7% respectively (I, II), as well as the study by Ewald and Christensen (1987) dealing with the occurrence of enterotoxin producing S. aureus strains in aircraft meals. This should be taken into account in hygiene training. Extra training, such as the proper use of disposable gloves, should focus on food handlers who are S. aureus carriers.
PFGE typing revealed a wide diversity in genomic types, showing 32 different types among 35 food handlers. In the present study one clone mainly colonised one person, but three persons were also found to carry 2 different types in the samples taken on the same day. In this study, in 4 cases of the 7 showing S. aureus on both the hands and in the nose, the strain was also found on the hands, clearly indicating transmission of the strain from the nose to the hand. As regards the types which were found in hand samples only, a person’s hands may naturally become contaminated with strains from a source other than the person him/herself. When the carriage of S. aureus in the nares was monitored in Japan (Hu et al. 1995), only one clone colonised one person and it persisted for a long period. This gives rise to major questions: is a persistently colonised individual always inhabited by the same strain, or can strain exchange occur. A recent study following S. aureus nasal carriage over eight years identified 47% non-carriers, 17% intermittent carriers, and 36% persistent carriers (Van den Bergh et al. 1999). Further characterisation of S. aureus strains by PFGE is very useful in case of tracing the source of contamination.