Introduction

Large-colony Lancefield group C streptococci are grouped into two species, each with two subspecies: Streptococcus dysgalactiae subsp. equisimilis, S. dysgalactiae subsp. dysgalactiae, Streptococcus equi subsp. equi, and S. equi subsp. zooepidemicus [1]. S. dysgalactiae subsp. equisimilis is the subspecies most frequently detected in humans, either as a commensal organism or as a pathogen. The other subspecies are more commonly isolated in animals.

Of all group C streptococci subspecies, S. equi subsp. zooepidemicus is probably the most aggressive human pathogen, and several outbreaks and sporadic cases of severe infection due to this microorganism have been reported, including nephritis, arthritis, sepsis, meningitis, and pneumonia [2–8]. Human infections are linked to close contact with domestic animals [2, 9], with consumption of unpasteurized milk [9–14], and even with consumption of pork [15]. This report describes a clinical and epidemiological study of an outbreak of S. equi subsp. zooepidemicus infections on the island of Gran Canaria (Spain) associated with the consumption of inadequately pasteurized cheese between February and April 2003.

Patients and methods

Clinical and microbiological methods

A confirmed case of infection by S. equi subsp. zooepidemicus was defined as the isolation of S. equi subsp. zooepidemicus from a clinical sample. Samples obtained from normally sterile sites and from the respiratory tract were processed according to conventional microbiological methods. Blood cultures were processed using the Bactec 9000 system (Becton Dickinson Diagnostic Instrument Systems, Sparks, MD, USA). Dairy product samples were processed according to conventional microbiological methods [16].

Characterization of isolates was carried out by means of the following conventional tests: Gram stain, catalase activity, beta-hemolysis on blood agar plates, colony size, ability to ferment sorbitol and trehalose, and detection of Lancefield group C antigen (Phadebact Streptococcus Test; Boule Diagnostics, Huddinge, Sweden). Identification to species level was performed using the API Rapid 32 Strep System (bioMérieux, Marcy l’Toile, France) and the BBL-Crystal GP ID (Becton Dickinson Microbiology Systems, Sparks, MD, USA). The relatedness of isolates was determined by SmaI restriction enzyme digests of chromosomal DNA separated by pulsed-field gel electrophoresis according to a method described previously [17, 18].

The E test (AB Biodisk, Solna, Sweden) was used to test susceptibility to the following antibiotics: penicillin, vancomycin, tetracycline, rifampin, erythromycin, and clindamycin. Susceptibility to levofloxacin and high-level resistance to gentamicin were determined using the agar dilution method following guidelines published by the Clinical and Laboratory Standards Institute (formerly the National Committee for Clinical Laboratory Standards) [19].

Medical records of patients were reviewed for demographic and clinical data. All patients with confirmed infection as well as individuals who lived with them were subjected to an epidemiological survey to look for evidence of epidemiological linkage: animal contact, contact with laboratory products, and consumption of milk, cheese and other raw dairy products.

Following the descriptive analysis, a hypothesis was established regarding the origin of the outbreak. Then, a case-control study was performed in order to confirm the hypothesis. Two controls per case were selected from among patients admitted with any other disease; one had to have been admitted immediately before and one immediately after the admission of the confirmed case. Both controls matched the confirmed case in terms of age and sex.

Statistical analysis

Cheese consumption and disease were compared by chi-square test with Fisher’s exact test. p<0.05 was considered statistically significant. Relative risk was calculated from odds ratios and 95% confidence intervals. Data from the case-control study were validated using EpiInfo 2000 (CDC, Atlanta, GA, USA).

Results

Patients

The island of Gran Canaria in Spain has a population of 790,360, and all inhabitants have access to the National Health Service. From February to April 2003, 15 patients infected with S. equi subsp. zooepidemicus were detected in the two reference hospitals on the island: 12 in the Hospital Universitario de Gran Canaria Doctor Negrin and 3 in the Hospital Universitario Insular de Gran Canaria.

Clinical and microbiological data

Patients from whom S. equi subsp. zooepidemicus was isolated had a median age of 70 years (range 47–86). The clinical features of the patients are summarized in Table 1. All patients except cases 2 and 4 received a beta-lactam agent as antibiotic treatment. Case 2 died immediately after admission, and case 4 was discharged from the emergency unit and was lost to follow-up, with no follow-up data recorded in the clinical chart.

Table 1 Clinical features of the patients
Full size table

All the isolates were confirmed as S. equi subsp. zooepidemicus on the basis of biochemical tests. The API Rapid 32 Strep System showed a profile number of 05012061101, and with Crystal GP panels, three profile numbers were found: 1667351540 (6 isolates), 1767351561 (4 isolates), and 1667351563 (7 isolates).

All isolates were sensitive to penicillin, erythromycin, vancomycin, rifampin, and levofloxacin but were resistant to clindamycin and tetracycline. No isolate presented high-level gentamicin resistance.

Epidemiological study

When as an increasing number of cases of infection by S. equi subsp. zooepidemicus was being detected in the microbiology laboratory of Hospital Universitario de Gran Canaria Doctor Negrin, a possible outbreak was suggested and the situation was communicated to the Public Health Department (Epidemiology Section). That agency alerted all hospitals and health centres on the island about the possibility of an epidemic, and other cases were sought out, but, as previously described, only confirmed cases were assessed.

The epidemiological inquiry suggested that consumption of fresh cheese (“queso fresco”) (present in 9 of 15 patients) was the possible origin of the outbreak. The case-control study proved an association between consumption of fresh cheese of a specific brand and S. equi subsp. zooepidemicus (OR=4.5, 95% CI 1.57–19.27; p<0.001). The cheese was made from cow’s milk produced at two dairy farms located in the north of the island. It was distributed to two supermarkets and grocery shops located in the same area. After this link was proved, a public health inspection of the cheese factory was performed. Inspectors established that the equipment used for pasteurization did not meet the necessary requirements, and, as a result, cheese production at the factory was stopped. At the same time, samples of raw milk and fresh cheese were obtained, and S. equi subsp. zooepidemicus was isolated from raw milk samples. Subsequently, the local Public Health Department undertook an investigation of the herd of cows at the two dairy farms, but S. equi subsp. zooepidemicus was not isolated from the eight samples of unpasteurized milk obtained from those cows. The animals on the farm did not show any evidence of pathology at that time.

All isolates of S. equi subsp. zooepidemicus recovered from patients and milk samples were available for molecular typing. All of them displayed the same macrorestriction pattern by pulsed-field gel electrophoresis, with the SmaI restriction pattern (Fig. 1) indicating a common link.

Fig. 1
figure 1

Restriction patterns obtained after digestion of SmaI by pulsed-field gel electrophoresis. Lanes 1 and 20 molecular size marker. Lanes 2–16 patient isolates. Lanes 17–19 milk isolates

Full size image

Since July 2003, only one more case of S. equi subsp. zooepidemicus infection has been detected. The microorganism was isolated from blood in a patient with diabetes mellitus and chronic alcoholic hepatopathy. The patient had had an episode of diarrhea, but he did not eat fresh cheese. The strain had the same macrorestriction pattern as the outbreak strain.

Discussion

A wide spectrum of illnesses due to S. equi subsp. zooepidemicus in animals has been described [20–23], but human infections associated with S. equi subsp. zooepidemicus are not frequent and can usually be traced back to animal sources [12, 24]. This group C streptococci subspecies appears to be more virulent than the other subspecies, and it is also less sensitive to antibiotics. Consequently, it causes more aggressive infections [2, 24], sometimes complicated by acute poststreptococcal glomerulonephritis [10, 13, 14, 25, 26]. Bacteremia is frequently detected in patients infected by group C streptococci [27], and our findings were in agreement: 90.9% of the patients whose blood was cultured were bacteremic.

Food-borne outbreaks by S. equi subsp. zooepidemicus are already well documented. Table 2 summarizes these outbreaks and their epidemiological and clinical characteristics. In most cases, cow’s milk is the source of infection. The outbreaks differ with respect to the severity of the clinical disease: some present as noninvasive infections (upper respiratory tract infections) and others as invasive disease (meningitis, pneumonia, aortic aneurisms) [8–15]. All patients in our outbreak had invasive disease, possibly because our strain was more virulent than those of other outbreaks or because patients with noninvasive infections had not gone to the hospital or been etiologically diagnosed. Pinto et al. [25] reported a high rate of hypertension and frequent abnormalities of renal function after the follow-up of patients from a large outbreak of acute glomerulonephritis. Serious infections with S. equi subsp. zooepidemicus have mainly been reported in patients over the age of 70 or in neonates [9], and they seem to be rare in healthy young adults [4]. In the present report, 60% of the patients were 70 or older, and no neonates or pregnant women were documented.

Table 2 Outbreaks caused by S. equi subsp. zooepidemicus
Full size table

The percentage of patients with aortic aneurysm, 26.7%, is elevated. Yuen et al. [15] found a particular association between S. equi subsp. zooepidemicus infection and cardiovascular disease. They described 12 patients with S. equi subsp. zooepidemicus septicemia and reviewed 34 patients described elsewhere, and they found that 27% of them had cardiovascular disease (7 with endocarditis, 3 with abdominal aortic aneurysms, and 2 with deep venous thromboses). Albarracin et al. [28] also described another case of S. equi subsp. zooepidemicus bacteremia associated with an abdominal aortic aneurism.

The high percentage (33.3%) of deaths in the present report, exceeded only by that in the series reported by Edwards et al. (66.6%) [9], is noteworthy. Although 60% of our patients were 70 or older and 80% had an underlying disease, we believe that the high death rate may be related to the severity of the clinical manifestations, i.e. bacteremia associated with aortic aneurisms or meningitis may have resulted in higher mortality [2, 12].

This study is limited because only confirmed cases were evaluated, and therefore the full extent of the outbreak is unknown. It is likely that more people were infected, but the infection probably presented as a noninvasive disease. People who remained undiagnosed could not be followed-up to monitor blood pressure and renal function, as would have been desirable.

In conclusion, S. equi subsp. zooepidemicus causes serious infections that are usually zoonoses. It is important to identify beta-hemolytic streptococci to the species level in order to detect contaminated foods of animal origin to prevent new food-borne outbreaks. Moreover, we recommend the application of molecular markers for precise characterization of the isolates.