This S. equi subsp.zooepidemicus outbreak was caused by fresh goat cheese manufactured from unpasteurized milk in a small dairy on a farm. Six of seven cases had eaten fresh goat cheese manufactured in this dairy, and also the seventh case had probably eaten the same cheese. Human, cheese, bulk milk and goat samples yielded S. equi subsp.zooepidemicus. Molecular typing with two different methods showed indistinguishable patterns for all outbreak associated strains, providing strong evidence that goat cheese was the source of infections.
All cases with invasive S. equi subsp.zooepidemicus disease recovered in this outbreak. In several other outbreaks, the case fatality proportion (CFP) has been substantial. Among 11 patients hospitalized in Hong Kong because of septicemia from 1982 to 1986, two died (CFP 22%) . Two persons of sixteen (CFP 12%) died in an outbreak associated with homemade cheese in New Mexico in 1983 . In another outbreak affecting 11 persons caused by unpasteurized milk in England in 1984, seven patients died (CFP 64%) . In a series of three cases, two of which were septicemias, in England in 1985, one patient died . Single cases of invasive infection with variable outcomes have been reported from several countries [19–26]. Although the median age of cases in this outbreak was high, none of them had severe underlying illnesses, which may explain that all of them recovered. In addition, after previously reported outbreaks from the 1980s, substantial development has taken place in the treatment of severely ill patients. The clinical features of illness in this outbreak were severe, however, which is demonstrated by the long hospitalization required for most cases.
The cases with invasive disease likely represent only a small proportion of all persons infected in this outbreak, because after the outbreak became public in the media, several persons reported that they had suffered from pharyngitis after having eaten the implicated goat cheese. Of the contaminated batch from October 23, 13 kg were sold, which means that probably tens of people were exposed to cheese from this batch. In addition, our cases had eaten cheese from batches manufactured earlier in October, indicating that also other batches have been contaminated.
Invasive group C streptococcal infections are not common in Finland. In most clinical laboratories, the strains are routinely typed only to the group level. Identification of the strains to species level would better enable public health professionals to detect outbreaks. In addition, even in individual cases, typing to species level might help in determining the probable source of infection. Molecular typing in this outbreak provided strong evidence that cheese from the implicated dairy was the source of the outbreak. Because the outbreak had substantial economic consequences to the manufacturer, it was important to get as strong evidence as possible about the source of these infections, also by using two types of molecular fingerprinting methods.
The two persons manufacturing goat cheese may have acquired their infection from contaminated milk or cheese. However, the owner might also have acquired the infection through direct contact with goats as she milked the animals and took care of them. Transmission of S. equi subsp. zooepidemicus through direct animal contact has been described in several reports [7, 19, 22]. The other food handler participated only in the manufacture of cheese and might have become infected through the manipulation of contaminated cheese. Also person-to-person transmission might have occurred, because symptomless food handlers can carry the organism in the throat .
S. equi subsp.zooepidemicus is a normal commensal of mucous membranes of a variety of animal species, especially horses. It is not known to be a common cause of genital infections in goats. Some cases of mastitis and meningitis have been reported [27, 28]. The prevalence of carrier state is not known. In the present case, the S. equi subsp.zooepidemicus -positive animal was in heat during the first sampling, and abundant growth of S. equi subsp.zooepidemicus was detected on the primary plates. The vaginal discharge might have contaminated the udder and the milk during milking. In the second sample the growth was sparse, and could easily have been overlooked. This indicates that the number of organisms excreted may vary, and may be difficult to detect.
The isolation of the bacterium from cheese samples was performed without using a selective medium which might have weakened the sensitivity of the analysis as the background flora was abundant. The dilution of the samples, however, enhanced the detection of the organism. The vaginal and nasal samples therefore were not heavily contaminated and the only goat with an isolation of S. equi subsp. zooepidemicus had an abundant growth of the organism.
Pasteurization is effective in preventing transmission of bacterial infections through milk products, and nearly all milk products in Finland are manufactured of pasteurized milk. Several outbreaks reported in the literature have been associated with unpasteurized milk, including salmonellosis, campylobacteriosis, listeriosis, and S. equi subsp. zooepidemicus infections . After pasteurization of milk became common, some of these infections, including S. equi subsp. zooepidemicus have become rare. If unpasteurized milk is used for manufacture of foods, the facility should be monitored carefully, including frequent microbiological sampling of the lactating animals and the products. The spectrum of the microbes examined routinely in Finland may not be sufficient for these facilities, but it should be extended to include S. equi subsp. zooepidemicus or beta-hemolytic streptococci as an indicator group. Because delicacy products manufactured of unpasteurized milk, like the goat cheese in this outbreak, may become more common, the importance of rigorous routine control measures should be emphasized.