Hazard analysis and critical control point systems in the United States Department of Agriculture regulatory policy.

Recent outbreaks of foodborne illness and studies by expert groups have established the need for fundamental change in the United States meat and poultry inspection programme to reduce the risk of foodborne illness. The Food Safety and Inspection Service (FSIS) of the United States Department of Agriculture (USDA) has embarked on a broad effort to bring about such change, with particular emphasis on the reduction of pathogenic micro-organisms in raw meat and poultry products. The publication on 25 July 1996 of the Final Rule on pathogen reduction and hazard analysis and critical control point (HACCP) systems was a major milestone in the FSIS strategy for change. The Final Rule provides a framework for change and clarifies the respective roles of industry and government in ensuring the safety of meat and poultry products. With the implementation of this Final Rule underway, the FSIS has been exploring ways in which slaughter inspection carried out under an HACCP-based system can be changed so that food safety risks are addressed more adequately and the allocation of inspection resources is improved further. In addition, the FSIS is broadening the focus of food safety activities to extend beyond slaughter and processing plants by working with industry, academia and other government agencies. Such co-operation should lead to the development of measures to improve food safety before animals reach the slaughter plant and after products leave the inspected establishment for distribution to the retail level. For the future, the FSIS believes that quantitative risk assessments will be at the core of food safety activities. Risk assessments provide the most effective means of identifying how specific pathogens and other hazards may be encountered throughout the farm-to-table chain and of measuring the potential impact of various interventions. In addition, these assessments will be used in the development and evaluation of HACCP systems. The FSIS is currently conducting a quantitative risk assessment for eggs, and several surveys and studies are being performed to supply data needed to conduct other risk assessments. The FSIS has established a food safety research agenda which will fill data gaps.

Molecular epidemiology of diphtheria in Russia, 1985-1994.

The largest diphtheria outbreak in the developed world since the 1960s began in the Russian federation in 1990. One hundred fifty-six Corynebacterium diphtheriae strains from throughout Russia, selected for temporal and geographic diversity, were assayed by ribotyping and multilocus enzyme electrophoresis (MEE). These tests showed significant genetic diversity within the C. diphtheriae species, and ribotyping and MEE data generally correlated well with epidemiologic data. A distinct clonal group of C. diphtheriae isolates (ET 8 complex) emerged in Russia in 1990 as the current outbreak began, and as the outbreak has progressed, these organisms have made up increasingly larger proportions of the strains that are isolated. Furthermore, the main characteristic of the epidemic strains is a specific combination of ET 8 and ribotypes G1 and G4. This study confirms the epidemiologic utility of the molecular subtyping methods that detected the epidemic clone and addresses the clone's origin and relation to C. diphtheriae from throughout Russia.

Heterogeneity of diphtheria toxin gene, tox, and its regulatory element, dtxR, in Corynebacterium diphtheriae strains causing epidemic diphtheria in Russia and Ukraine.

Diphtheria toxin (tox) and its regulatory element (dtxR) from 72 Corynebacterium diphtheriae strains isolated in Russia and Ukraine before and during the current diphtheria epidemic were studied by PCR-single-strand conformation polymorphism analysis (PCR-SSCP). Twelve sets of primers were constructed (eight for tox and four for dtxR), and three regions within tox and all four regions of dtxR showed significant variations in the number and/or sizes of the amplicons. Two to four different SSCP patterns were identified in each of the variable regions; subsequently, tox and dtxR could be classified into 6 and 12 different types, respectively. The great majority of epidemic strains from both Russia and Ukraine had tox types 3 and 4, and only in a single preepidemic strain isolated in Russia were all eight tox regions identical to those of C. diphtheriae Park-Williams No. 8 (tox type 1). Epidemic strains from Ukraine can easily be identified by dtxR type 5, while the majority of the Russian epidemic strains have dtxR of types 2 and 8. No differences in the tox regions between mitis and gravis biotype strains were observed. However, dtxR types 2, 5, and 8 were identified only in the gravis biotype, and dtxR type 1 was characteristic for the mitis biotype strains. PCR-SSCP is a simple and rapid method for the identification of variable tox and dtxR regions that allows for the clear association of tox and dtxR types with strains of distinct temporal and/or geographic origins.

Genotypic and phenotypic characterization of Helicobacter cinaedi and Helicobacter fennelliae strains isolated from humans and animals.

By DNA-DNA hybridization, we classified 26 human strains, 4 dog and cat strains, and 4 hamster strains putatively identified as Helicobacter cinaedi as well as 2 human strains and 2 animal strains of Helicobacter fennelliae. All but one human strain belonged to the same hybridization group as the type strain of H. cinaedi. The animal strains also appeared to belong to this hybridization group. Both human strains of H. fennelliae were shown to be H. fennelliae by DNA-DNA hybridization, but both animal strains were less than 15% related to the type strain. All strains were also characterized by plasmid profiles and ribotyping. Plasmids were found in 23% of the human strains, 100% of the hamster strains, and 33% of the dog and cat strains. Human strains were essentially identical by ribotyping, but were clearly differentiated from the hamster and dog and cat strains. Some strains may be difficult to culture on primary isolation; we found that our strains grew well on anaerobic CDC agar, brucella agar, and tryptic soy agar II. Our H. cinaedi and H. fennelliae strains differed from those previously described because some were resistant to cephalothin: some H. cinaedi strains were also resistant to nalidixic acid. All isolates were also characterized by antimicrobial susceptibility testing. We found that human strains of H. cinaedi were more resistant to clindamycin and erythromycin than were animal isolates; 19% of the human strains were resistant to ciprofloxacin. Therefore, we recommend that antimicrobial susceptibility results be obtained before initiating therapy for H. cinaedi and H. fennelliae infections.

Application of PCR for detection of toxigenic Corynebacterium diphtheriae strains isolated during the Russian diphtheria epidemic, 1990 through 1994.

A total of 250 Corynebacterium diphtheriae isolates from clinical cases and carriers in Russia were assayed by PCR directed at the A subunit of the diphtheria toxin gene to distinguish toxigenic from nontoxigenic strains; 170 strains were positive as indicated by the presence of the 248-bp amplicon. The results of this PCR assay were in complete concordance with those of the standard immunoprecipitation assay (Elek), and the PCR assay is a useful tool for rapid identification in clinical laboratories.

Nontoxigenic Vibrio cholerae O1 intestinal pathology in adult mice.

The intestinal pathology caused by nontoxigenic Vibrio cholerae O1 was examined in the sealed adult mouse (SAM) model. Histologic examination demonstrated that a nontoxigenic V. cholerae O1 strain that elicited maximum fluid accumulation (FA) in the small intestine of adult mice caused damages to the villi and necrosis of lymphoid elements within solitary submucosal lymphoid nodules in the Peyer's patches. Challenge of mice with a strain that did not elicit intestinal FA produced none of the above tissue responses. Increased FA activity, intestinal alterations, and tissue pathology caused by the nontoxigenic V. cholerae O1 indicate its pathogenic potential.

The emerging diversity of the electrophoretic types of Vibrio cholerae in the Western Hemisphere.

Since the Latin American cholera epidemic began in 1991, 447 isolates of Vibrio cholerae O1 from the Western Hemisphere have been assayed by multilocus enzyme electrophoresis (MEE) to determine allelic variation among 16 enzyme-encoding genes. Two electrophoretic types (ETs) were identified among toxigenic isolates from Latin America: 323 were ET 4, the ET associated with the Latin American epidemic, and 29 were ET 3. Twenty-three of these ET 3 isolates had a distinctive antimicrobial resistance pattern also seen in isolates imported into the United States from Latin America and Southeast Asia. These resistant isolates had an identical ribotype and nearly identical pulsed-field gel electrophoresis (PFGE) patterns. Most nontoxigenic isolates analyzed were not precursors or descendants of toxigenic epidemic strains. MEE provided a population genetic frame-work for the interpretation of PFGE and ribotype data from the isolates in this study. All three methods identified 2 distinct strains of toxigenic V. cholerae O1 currently epidemic in Latin America.

Helicobacter cinaedi-associated bacteremia and cellulitis in immunocompromised patients.

OBJECTIVE: To define the clinical spectrum of illness associated with Helicobacter cinaedi infection in the United States and to determine associated epidemiologic risk factors and optimal laboratory methods for recovery of H. cinaedi. DESIGN: A retrospective epidemiologic study of 23 patients with H. cinaedi-associated illness. PATIENTS: 23 patients with H. cinaedi infection identified between January 1982 and August 1990. Most isolates (22 of 23) were from blood; one was from stool. RESULTS: Ages ranged from 24 to 84 years (mean, 44 years). Eighty-three percent of patients were men; 17% were women. Clinical and laboratory data were obtained from 21 patients. Eighteen patients were febrile (15 required hospitalization); cellulitis was reported in 9 patients. Sixty percent were immunocompromised; 45% were reported to be seropositive for human immunodeficiency virus (HIV). For bacteremic patients, positive blood cultures were detected by a slightly elevated growth index in an automated blood culture system; many hospital laboratories had difficulty isolating the organism. CONCLUSIONS: Helicobacter cinaedi appears to cause recurrent cellulitis with fever and bacteremia in immunocompromised hosts. Blood cultures from immunocompromised patients with these symptoms may need special handling to isolate H. cinaedi.

Molecular characterization of Vibrio cholerae O1 strains by pulsed-field gel electrophoresis.

Pulsed-field gel electrophoresis (PFGE) was performed on 180 isolates of Vibrio cholerae serogroup O1 representing 6 different multilocus enzyme electrophoresis (MEE) types and 27 rRNA restriction fragment length polymorphism types (ribotypes). Isolates were digested with the restriction enzyme NotI and were separated into 63 patterns on the basis of differences in band arrangements. In general, strains which were different by MEE or ribotyping also had different PGFE patterns. PFGE identified individual strains within a single MEE type or ribotype; isolates with one PFGE pattern were less frequently distinguished by ribotyping. All V. cholerae O1 isolates tested from the Latin American epidemic were indistinguishable by their MEE, ribotype, or PFGE patterns. PFGE could further distinguish strains of this same ribotype isolated in Africa, Europe, the South Pacific, or Southeast Asia. Although both MEE and PFGE could identify the strain from the Latin American epidemic, PFGE was more rapid and less labor intensive. PFGE also distinguished nontoxigenic isolates endemic to the U.S. Gulf Coast from unrelated nontoxigenic isolates. In the present study PFGE was more discriminating than other previously described subtyping assays for V. cholerae O1 and appears to be a useful epidemiologic tool.

Evaluation of ribotyping techniques as applied to Arcobacter, Campylobacter and Helicobacter.

Restriction fragment length polymorphisms of ribosomal DNA (ribotyping) of many bacterial species has been useful for both epidemiologic subtyping and species identification. However, the use of ribotyping has been confined to major research and reference laboratories due to two factors: (a) the procedure must be carefully optimized for each organism one wishes to investigate and (b) most currently available protocols use hazardous chemicals or radioisotopes. The purpose of this study is to suggest an overall scheme that a clinical or research microbiologist could apply to ribotyping of any organism. In general, we recommend using a guanidium extraction method for DNA extraction, careful optimization of restriction conditions, and hybridization with non-radioactive digoxigenin-labelled probes; these procedures do not use hazardous chemicals or radioisotopes.

Common somatic O and heat-labile serotypes among Campylobacter strains from sporadic infections in the United States.

Somatic O (formerly heat-stable) and heat-labile (HL) serotyping methods are commonly used to type Campylobacter jejuni and Campylobacter coli isolates. Although both systems are effective, the labor and time required for each have limited their application. These systems can be simplified by reducing the number of antisera used. To find an appropriate panel of antisera, we determined the distribution of common serotypes in the United States among a representative sample of 298 Campylobacter isolates. The strains, obtained between July 1989 and June 1990 from persons with sporadic cases of diarrhea, were collected from 19 randomly chosen counties in all geographic (census) regions of the United States. All strains were serotyped by the O and HL systems. By phenotypic methods, 288 C. jejuni, 9 hippurate-negative C. jejuni/C. coli, and 1 Campylobacter lari were identified. Of 57 O antisera, 24 typed 252 (84.6%) strains. Of the 55 HL antisera, 23 serotyped 253 (84.9%) strains. All strains were typeable in the unabsorbed O antisera. In the absorbed HL antisera, four strains were nontypeable and 14 were rough and untypeable. In each geographic region, 9 or more O and HL serotypes were found. Serotypes O:1, O:4, and O:13,16,43,50 and HL 1 were identified in all regions. The combination of both schemes gave greater discrimination than either system alone, but the maintenance of both requires a large resource investment. A serotyping scheme incorporating the 24 most prevalent O and 23 most prevalent HL serotypes could be useful for outbreak support and for surveillance. In the near future, we anticipate using a molecular subtyping method in combination with limited serotyping to distinguish Campylobacter strains.

Arcobacter (Campylobacter) butzleri-associated diarrheal illness in a nonhuman primate population.

After DNA hybridization identified an isolate from an ill rhesus macaque (Macaca mulatta) as Arcobacter (Campylobacter) butzleri, we initiated a study to determine whether A. butzleri was associated with diarrheal disease in nonhuman primates at the Yerkes Primate Research Center. By using Campy-CVA medium incubated at 35 degrees C, 15 A. butzleri isolates were obtained from 14 macaques; 7 macaques were coinfected with Campylobacter coli and Campylobacter jejuni. A. butzleri was not isolated from normal feces, despite the fact that feces from 76 macaques were cultured at necropsy. Histologic evaluation of colonic specimens from three macaques from which A. butzleri had been isolated showed mild to moderately severe chronic, active colitis. Ribotype analysis of the 15 A. butzleri isolates revealed nine different strains; these data suggest that A. butzleri may be endemic in this primate population and that a point source of infection is unlikely. This is the first report of the presence of A. butzleri in juvenile and adult macaques with diarrhea, and it may present an opportunity to study the pathogenesis of this organism, which appears to be associated with persistent diarrhea in humans.

Clonal relationships among Escherichia coli strains that cause hemorrhagic colitis and infantile diarrhea.

The genetic relationships among 1,300 isolates of Escherichia coli representing 16 serotypes associated with enteric disease, including O157:H7 strains recovered from patients with hemorrhagic colitis and hemolytic uremic syndrome and O26:H11, O55:H6, O55:H7, O111:H2, and O128:H2 strains, many of which were isolated originally from infants with diarrhea, were estimated from allelic variation among 20 enzyme-encoding genes detected by multilocus enzyme electrophoresis. Multiple electrophoretic types were observed among isolates of each serotype, with isolates of the same O serogroup differing on average at 28% of the enzyme loci. Comparisons of the multilocus enzyme profiles revealed that 72% of the isolates belong to 15 major electrophoretic types, each of which corresponds to a bacterial clone with a wide geographic distribution. Genetically, the O157:H7 clone is most closely related to a clone of O55:H7 strains that has long been associated with worldwide outbreaks of infantile diarrhea. We propose that the new pathogen emerged when an O55:H7-like progenitor, already possessing a mechanism for adherence to intestinal cells, acquired secondary virulence factors (Shiga-like cytotoxins and plasmid-encoded adhesins) via horizontal transfer and recombination.

The molecular epidemiology of cholera in Latin America.

To explain the sudden appearance and rapid spread of cholera in Latin America in January 1991, molecular techniques were used to define Vibrio cholerae O1 isolates from around the world. Restriction fragment length polymorphisms of rRNA and ctxA genes, DNA sequence of cholera toxin B subunit gene ctxB, and multilocus enzyme electrophoresis data were used to characterize 197 isolates. Worldwide, there are at least four distinct toxigenic El Tor V. cholerae O1 clones: the seventh pandemic (Eastern Hemisphere), US Gulf Coast, Australian, and Latin American. Nontoxigenic V. cholerae O1 previously isolated in Brazil, Mexico, and Peru are unlike current toxigenic isolates. The Latin American clone probably represents an extension of the seventh pandemic into the Western Hemisphere, while the US Gulf Coast clone most likely evolved separately. These data will be useful in monitoring the spread of cholera, determining the origin of outbreaks in both hemispheres, and implicating specific vehicles of transmission.

Use of automated sequencing of polymerase chain reaction-generated amplicons to identify three types of cholera toxin subunit B in Vibrio cholerae O1 strains.

Cholera toxin is the principal factor causing the profuse intestinal fluid secretion that is characteristic of cholera. The DNA sequences of the cholera toxin subunit B structural genes from 45 Vibrio cholerae O1 strains isolated in 29 countries over a period of 70 years were determined by automated DNA sequencing of polymerase chain reaction-generated amplicons. Three types of cholera toxin B subunit gene (ctxB) were identified. Genotype 1 was found in strains of classical biotype worldwide and El Tor biotype strains associated with the U.S. Gulf Coast, genotype 2 was found in El Tor biotype strains from Australia, and genotype 3 was found in El Tor biotype strains from the seventh pandemic and the recent Latin American epidemic. All base changes correspond to an amino acid substitution in the B subunit of the cholera toxin. Heterogenicity in the B subunit could have implications for vaccine development and diagnostic tests for cholera toxin and antitoxin. We conclude that this technology provides timely and potentially useful epidemiological information.

Pyrazinamidase, CR-MOX agar, salicin fermentation-esculin hydrolysis, and D-xylose fermentation for identifying pathogenic serotypes of Yersinia enterocolitica.

We evaluated several simple laboratory tests that have been used to identify pathogenic serotypes of Yersinia enterocolitica or to indicate the pathogenic potential of individual strains. A total of 100 strains of Y. enterocolitica were studied, including 25 isolated during five outbreak investigations, 63 from sporadic cases, and 12 from stock cultures. The pyrazinamidase test, which does not depend on the Yersinia virulence plasmid, correctly identified 60 of 63 (95% sensitivity) strains of pathogenic serotypes and 34 of 37 (92% specificity) strains of nonpathogenic serotypes. Salicin fermentation-esculin hydrolysis (25 degrees C, 48 h) correctly identified all 63 (100% sensitivity) strains of the pathogenic serotypes and 34 of 37 (92% specificity) strains of the nonpathogenic serotypes. The results of the pyrazinamidase and salicin-esculin tests disagreed for only 7 of the 100 strains of Y. enterocolitica, and these would require additional testing. Congo red-magnesium oxalate (CR-MOX) agar determines Congo red dye uptake and calcium-dependent growth at 36 degrees C, and small red colonies are present only if the strain contains the Yersinia virulence plasmid. This test has proven to be extremely useful for freshly isolated cultures, but only 15 of 62 strains of pathogenic serotypes that had been stored for 1 to 10 years were CR-MOX positive. None of the 16 strains of Y. enterocolitica serotype O3 fermented D-xylose, so this test easily differentiated strains of this serotype, which now appears to be the most common in the United States. Although antisera that can actually be used to serotype strains of Y. enterocolitica are not readily available, the four simple tests described above can be used to screen for pathogenic serotypes.

Molecular epidemiology of group B streptococcal infections: use of restriction endonuclease analysis of chromosomal DNA and DNA restriction fragment length polymorphisms of ribosomal RNA genes (ribotyping).

Epidemiologic investigation of group B streptococcal (GBS) infections has been limited by the lack of a discriminatory typing system. Therefore, the use of restriction endonuclease analysis of chromosomal DNA (REAC) and DNA restriction fragment length polymorphisms of rRNA genes (ribotyping) to subtype molecularly GBS isolates associated with human invasive disease was investigated. Chromosomal DNA of selected GBS isolates was initially digested with 24 different restriction enzymes. HhaI gave the best discrimination of hybridization banding patterns (ribotypes) and was used with all study isolates. Ribotyping and REAC differentiated among isolates of the same and different serotypes. Nine ribotype patterns were noted among the 76 isolates studied, including 4 among serotype Ia/c and 4 additional ribotypes among serotype III isolates. Epidemiologically related isolates (e.g., mother-infant or twin-twin pairs) had identical REAC and ribotype patterns. Epidemiologically unrelated isolates with the same ribotype usually had different REAC patterns, suggesting that REAC may be a more sensitive technique for strain differentiation. REAC and ribotyping were reproducible and proved to be successful molecular epidemiologic methods for subtyping GBS.