The study sample consisted of 28 of 35 S. Blockley strains isolated from May to December 1998 (19), one strain from February 1999, and four epidemiologically unrelated control strains: one from 1996 and three from 1997. All isolates were from human cases of enteritis. Identification was performed by the API 20E system (BioMerieux S.A., Marcy l'Etoile, France) and serotyping with commercially obtained antisera (BioMerieux) (20).

Susceptibility to kanamycin, streptomycin, ampicillin, amoxicillin/clavulanic acid, cefepime, tetracycline, chloramphenicol, trimethoprim/sulfamethoxazole, gentamicin, nalidixic acid, and ciprofloxacin was tested by a disk diffusion assay according to National Committee for Clinical Laboratory Standards guidelines (21). Genomic DNA was prepared and digested with XbaI (New England Biolabs) (22). Chi-square tests or Fisher exact tests were used to calculate two-tailed probabilities.

S. Blockley accounted for seven of the 13,199 salmonella isolates identified in Greece from 1976 to 1997. However, 35 gastroenteritis cases due to this serotype were reported from May to December 1998 (19). Twenty-nine S. Blockley strains isolated from fecal specimens of patients with gastroenteritis during May 1998 to February 1999, along with four epidemiologically unrelated clinical isolates from 1996 and 1997, were therefore studied for susceptibility to antibiotics. The 1998 outbreak isolates were scattered throughout Greece; S. Blockley was isolated later, starting in August 1998, in northern Greece.

All isolates were susceptible to trimethoprim/sulfamethoxazole, ampicillin, amoxicillin/clavulanic acid, gentamicin, and ciprofloxacin (Table). High resistance rates were observed to tetracycline (100%), streptomycin and kanamycin (90%), chloramphenicol (83%), and nalidixic acid (52%). Six resistance phenotypes could be distinguished (Table) with the two major phenotypes of outbreak isolates being resistant to kanamycin, streptomycin, tetracycline, and chloramphenicol (ATC) or kanamycin, streptomycin, tetracycline, chloramphenicol, and nalidixic acid (ATCN). Most (76%) strains isolated after August 24, 1998, were nalidixic acid-resistant (resistance phenotypes ATCN, TCN, ATN), unlike strains isolated up to August 17, 1998 (17%) (1.29 <RR = 3.03 <7.11, p = 0.005).

Figure 1

Figure 1. Sample pulsed-field gel electrophoresis (PFGE) gel of representative Salmonella blockley isolates, indicating common and unique DNA fingerprints. Electrophoresis was through 1% agarose/0.5 x TBE, in a CHEF DRIII apparatus (BioRad Laboratories),...

Figure 2

Figure 2. Dendrogram of similarity among the observed pulsed-field gel electrophoresis patterns. A percentage scale of similarity is indicated at the top. Numbers in parentheses refer to the number of isolates with the...

When pulsed-field gel electrophoresis was used to obtain DNA fingerprints for these isolates (Figure 1), all belonged to the same type, A, although eight subtypes, A1-A8, could be distinguished on the basis of one to three DNA fragment differences (Figure 1, Figure 2). Two of the four isolates from previous years belonged to unique subtypes A6 and A7; the other two belonged to subtypes A2 and A8, shared by outbreak isolates (Table). In contrast, 93% of the 1998 outbreak strains yielded PFGE patterns common to two or more isolates. Indeed, most outbreak isolates were grouped in subtypes A2 and A4, consisting of 11 and 12 isolates, respectively (Table).

Pulsed-field gel electrophoresis subtypes were associated with resistance phenotypes. Most resistance phenotype ATC isolates belonged to subtype A2 (3.03. <RR = 11.86 <46.5, p = 0.0000098), while most resistance phenotype ATCN isolates belonged to A4 (2.76 <RR = 7.11 <18.30, p = 0.0000416). In addition, most isolates in the two major subtypes appearing before August 17, 1998, belonged to the A2 group, while most isolates appearing after August 24, 1998, belonged to the A4 group (1.47 <RR = 9.82 <65.45, p = 0.0006). Finally, unlike A4, the earlier A2 subtype was not isolated in northern Greece.

Our results indicate that PFGE is useful in distinguishing epidemiologically related S. Blockley isolates since two of the four nonoutbreak isolates displayed unique PFGE patterns, A7 and A8, while PFGE patterns A2 and A4 grouped most of the 29 outbreak isolates (11 and 12, respectively).

These two chromosomal fingerprints, differing by two DNA fragments, were associated with two distinct resistance phenotypes. The resistance phenotype of A4 isolates, ATCN, was identical to the earlier resistance phenotype of A2 isolates, ATC, except for the resistance to nalidixic acid. Nevertheless, these two PFGE/antibiotic resistance types, A2/ATC and A4/ATCN, displayed a clear distribution both in time and space.

The data may, therefore, indicate two main sources for the outbreak. Alternatively, and perhaps more likely, these two closely related types may together constitute the outbreak clone, evolved with time to acquire resistance to nalidixic acid. Resistance may well have originated in the food source, since several antibiotic classes are used as feed supplements in animal rearing and aquaculture in Greece: sulfonamides (trimethoprim/sulfathiazine), tetracyclines (oxytetracycline), and quinolones (oxolinic acid). However, as in other European countries (17,18), the epidemiologic investigation did not locate a common source to account for the wide geographic spread of cases (19). Although travel was not mentioned in the Greek patients' questionnaire responses, the possibility that the source was an imported food cannot be ruled out. The association with smoked eel of Italian origin in the German outbreak has not been microbiologically confirmed (17). The only other previous European report of a human outbreak attributed to S. Blockley, probably from vegetables contaminated by this organism, which was prevalent in irrigation water in the Spanish region of Granada, is anecdotal (13). A documented S. Blockley enteritis epidemic in a U.S. hospital in 1966 was attributed to contaminated ice cream; however, this was also not microbiologically confirmed (7).

While this serotype may remain important in Europe, its high rates of resistance to kanamycin, streptomycin, tetracycline, and chloramphenicol, which were in agreement with studies from the Far East (3) and Spain (12), are cause for concern. Unlike the Far Eastern strains, no resistance to ß-lactam antibiotics or cotrimoxazole was observed in our study. The two dominant resistant phenotypes of S. Blockley from natural polluted waters in Spain were sulfonamides, streptomycin, and tetracycline; and neomycin, streptomycin, kanamycin, tetracycline, and chloramphenicol (12), as in the Greek strains, except for the absence of resistance to nalidixic acid.

In agreement with differences in animal reservoirs and transmission routes and therefore the mechanism of resistance acquisition among different Salmonella serotypes, the main patterns of resistance observed in S. Blockley were distinct from those predominating in the two major serotypes from isolates of both human and animal food origin in Greece. In S. Enteritidis, the most frequent resistance phenotype was resistance to ampicillin (24), while in S. Typhimurium, the most frequent resistant phenotype was resistance to sulfonamides and streptomycin (A. Markogiannakis, P.T. Tassios, N.J. Legakis, unpub. obs.). Furthermore, the considerably high rate of resistance to nalidixic acid is equally unprecedented in both the Far Eastern and Spanish S. Blockley isolates and in other salmonella serotypes from Greece. Since resistance to nalidixic acid can be a precursor of resistance to fluoroquinolones, one of the two drug classes of choice for invasive salmonella disease, this feature of these S. Blockley strains is particularly disturbing. S. Blockley, previously a prevalent serotype in the Far East but rare elsewhere, nevertheless posed a public health problem in several European countries. The source of the European outbreaks, however, remains unclear. Given the increased international commerce in food, a collaborative study would be useful in identifying potential similarities between the recent European strains and established strains from the Far East.