A. van den Bogaard

Evaluation of Fifteen Commercially Available Serological Tests for Diagnosis of Lyme Borreliosis

Abstract The performance of 11 commercially available enzyme immunoassays (EIA) and four Western blot (WB) tests for the detection of IgM and IgG antibodies against Borrelia burgdorferi were compared. A total of 229 serum specimens were used: 26 from patients with early Lyme borreliosis, 13 from patients with late Lyme borreliosis, 62 from healthy controls and 128 from patients with disorders clinically mimicking Lyme borreliosis and/or known to cause cross-reactivity in Lyme borreliosis serological tests (patient control group). In specimens from patients with early Lyme borreliosis, the sensitivity of the individual tests ranged from 35 to 81% for detection of IgM. In late Lyme borreliosis, sensitivity of the tests ranged from 46 to 92%. In healthy controls the specificity of the tests ranged from 89 to 100% and from 82 to 97% for IgM and IgG tests, respectively. In the patient control group, specificity of the tests ranged from 75 to 90% for IgM and from 84 to 100% for IgG tests. The Behring (Germany) and Genzyme Virotech (Germany) IgM EIA tests showed the best performance in detecting early Lyme borreliosis. For the detection of late Lyme borreliosis, the Dako (Denmark) IgG test was the best despite its low sensitivity. The maximum sensitivity of Western blotting for detecting IgM in patients with early Lyme borreliosis and IgG in patients with late Lyme borreliosis was 50 and 46%, respectively. The use of an EIA-WB two-test protocol improved the specificity and positive predictive values of the EIA results but caused a significant loss in sensitivity. Patients with Epstein-Barr virus or cytomegalovirus infection who had a positive reaction in the IgM EIA could not be discriminated from patients with early Lyme borreliosis with the help of Western blotting. Hence, positive and negative predictive values in combination with sensitivity and specificity values indicated that the exclusion of these infections was more relevant than the confirmation of a positive IgM EIA with Western blot.

Antibiotic use and the prevalence of antibiotic resistance in bacteria from healthy volunteers in the dutch community.

Abstract.Background: Although most antibiotics are consumed in the community (80%), the relation of antibiotic consumption and resistance in the community has not been thoroughly studied. Materials and Methods: In the present study, antibiotic use was related to the prevalence of antibiotic resistance of Escherichia coli and enterococci isolated from healthy volunteers living in the southern part of the Netherlands in 1996 and 1999. Results: No change in the total antibiotic use in the Dutch community was observed between 1996 and 1999 (3,542 and 3,598 defined daily doses [DDD] per 1,000 inhabitants/year). However, the increased fluoroquinolone use (+ 18%) and the increased prevalence of ciprofloxacin resistance from 0 to 2% is a point of concern, especially since this was accompanied by a significant shift towards higher minimum inhibitory concentration (MIC) values (p < 0.05). A significant decrease in the prevalence of vancomycin-resistant enterococci (p < 0.05) was found in addition to a significant shift towards lower MIC valued for avoparcin, a gycopeptide previously used as growth promoter in animal husbandry (p < 0.05). This was very likely due to the banning of avoparcin for this purpose from April 1997 onwards. Conclusion: In order to maintain the low level of antibiotic use and resistance in The Netherlands, surveillance of antibiotic resistance in nonclinical isolates in relation to antibiotic use is very important.

Veterinary public health: Human health hazards associated with the administration of antimicrobials to slaughter animals

Summary This article describes the assessment of consumer risks of residues of tetracyclines in slaughter pigs in the Netherlands. The assessed risks were toxic and allergic reactions, and the disturbance of the consumers’ intestinal flora. Toxic and allergic reactions in humans and animals have only been observed at therapeutic doses, affecting between an estimated 1 in 5000 and one 1 in 140,000 individuals exposed. Residues of tetracyclines in pigs are closely associated with treatment with injectable formulations. Established Maximum Residue Levels (MRLs) do not reflect actual consumer risks in case a limit is violated incidentally. For example, when the established MRLs for tetracyclines in meat are exceeded with a factor 400, 40,000, and 200,000, respectively, the actual risk of an adverse drug reaction for the consumer following a single consumption of this meat is maximally 1 in 3 million, 1 in 300,000, and 1 in 8000, respectively. At the current estimated low levels of incidental exposure via pork, the annual risk of negative health effects for a random consumer is estimated at maximally 1 in 33 million. The annual risk that a temporary disturbance of the intestinal flora may also result in a facilitated infection with certain enteropathogens, such as Salmonella spp., is estimated at 1 in 45 million. It is concluded that the current microbiological risks of pork are greater than the risks of residues of tetracyclines as such, and that the control of the microbiological risks of pork should therefore be given first priority.

Human health hazards associated with the administration of antimicrobials to slaughter animals. Part II. An assessment of the risks of resistant bacteria in pigs and pork.

Risks for the consumer regarding the acquisition of resistant bacteria and/or resistance genes via the consumption of pork are discussed. In general, Salmonella spp. and Escherichia coli that originate from animals do not easily transfer their resistance genes to the resident intestinal flora of humans. The prevalence of resistant E. coli in humans seems more associated with being a vegetarian (odds ratio (OR) 1.89) than with the consumption of meat and meat products. Other risk factors are treatment with antimicrobials (OR 2-5), becoming hospitalized (OR 5.93), or working in a health setting (OR 4.38). In the Netherlands, annually an estimated 45,000 people (0-150,000) become a carrier of resistant E. coli and/or resistance genes that ori ginate from pigs, while an estimated 345,000 persons (175,000-600,000) become a carrier of resistant E. coli and/or resistance genes that originate from hospitals, e.g. other patients. Any problems with resistant Salmonella spp. that stem from pigs are, in fact, an integral part of the total problem of food-borne salmonellosis. Sometimes there are outbreaks of a specific multi-resistant clone of S. typhimurium that causes problems in both farm animals and humans. The probability that in the next 30 years there is no or maximally one outbreak of a specific clone that originates from pig herds is estimated at about 75%. Antimicrobials used as a growth promoter can have a measurable influence on the prevalence of resistant bacteria. The likely chain of events regarding avoparcin and the selection and dissemination of resistance against vancomycin in the enterococci gives the impression that the impact of the use of antimicrobials in animals on the prevalence of resistance in humans is largely determined by whether resistance genes are, or become, located on a self-transferable transposon. Furthermore, consumer health risks of antimicrobials used in slaughter pigs are mainly determined by the selection and dissemination of bacterial resistance and much less by the toxicological properties of any residues in pork. It is also concluded that most of the problems with resistant bacteria in humans are associated with the medical use of antimicrobials, and that the impact of particularly the veterinary use of antimicrobials is limited. However, the impact of antimicrobials used as a feed additive appears to be much greater than that of antimicrobials used for strictly veterinary purposes. The use of antimicrobials as a feed additive should therefore be seriously reconsidered.Summary Risks for the consumer regarding the acquisition of resistant bacteria and/or resistance genes via the consumption of pork are discussed. In general, Salmonella spp. and Escherichia coli that originate from animals do not easily transfer their resistance genes to the resident intestinal flora of humans. The prevalence of resistant E. coli in humans seems more associated with being a vegetarian (odds ratio (OR) 1.89) than with the consumption of meat and meat products. Other risk factors are treatment with antimicrobials (OR 2–5), becoming hospitalized (OR 5.93), or working in a health setting (OR 4.38). In the Netherlands, annually an estimated 45,000 people (0–150,000) become a carrier of resistant E. coli and/or resistance genes that originate from pigs, while an estimated 345,000 persons (175,000–600,000) become a carrier of resistant E. coli and/or resistance genes that originate from hospitals, e.g. other patients. Any problems with resistant Salmonella spp. that stem from pigs are, in fact, an integral part of the total problem of food‐borne salmonellosis. Sometimes there are outbreaks of a specific multi‐resistant clone of S. typhimurium that causes problems in both farm animals and humans. The probability that in the next 30 years there is no or maximally one outbreak of a specific clone that originates from pig herds is estimated at about 75%. Antimicrobials used as a growth promoter can have a measurable influence on the prevalence of resistant bacteria. The likely chain of events regarding avoparcin and the selection and dissemination of resistance against vanco‐mycin in the enterococci gives the impression that the impact of the use of antimicrobials in animals on the prevalence of resistance in humans is largely determined by whether resistance genes are, or become, located on a self‐transferable transposon. Furthermore, consumer health risks of antimicrobials used in slaughter pigs are mainly determined by the selection and dissemination of bacterial resistance and much less by the toxicological properties of any residues in pork. It is also concluded that most of the problems with resistant bacteria in humans are associated with the medical use of antimicrobials, and that the impact of particularly the veterinary use of antimicrobials is limited. However, the impact of antimicrobials used as a feed additive appears to be much greater than that of antimicrobials used for strictly veterinary purposes. The use of antimicrobials as a feed additive should therefore be seriously reconsidered.

Antibiotic resistance of enterobacteriaceae isolated from the faecal flora of fattening pigs

From June 1991 to April 1992 407 faecal samples were collected from three groups of pigs (I n = 248, II n = 87, III n = 72) at a pig fattening farm to determine the prevalence and the degree of antibiotic resistance of Enterobacteriaceae as well as the antibiotic susceptibility of the strains isolated. Despite the absence of mass medication during the observation period, the prevalence of resistance to the most commonly used antimicrobial agents in veterinary medicine was high (range amoxicillin 70%-97%, oxytetracycline 89%-100%, sulfamethoxazole 88%-100%, trimethoprim 78%-100%). The high degree of resistance to oxytetracycline and sulfamethoxazole ranged from 8%-67% and 4%-46%, respectively. The percentage of the isolated Escherichia coli strains resistant to oxytetracycline, streptomycin and sulfamethoxazole ranged from 49% to 68%; the other agents tested showed lower percentages (0-13%). Resistance to three or more antibiotics was observed in 43% of the isolates. Of the 52 resistance patterns that could be distinguished, 51% was accounted by only four patterns: oxytetracycline+streptomycin+sulfamethoxazole 20%, sulfamethoxazole 12%, streptomycin+sulfamethoxazole 11% and streptomycin+oxytetracycline 8%.

Epstein-barr virus and cytomegalovirus infections cause false-positive results in IgM two-test protocol for early lyme borreliosis

Serological testing is important for the confirmation of Lyme borreliosis (LB) in suspected cases, because the disease is protean in its clinical manifestations and can mimic several other diseases. Moreover, in the case of LB the classical methods for confirming the diagnosis of a bacterial infection, i. e. isolation or direct visualisation of the causative agent, are not sensitive enough [1] and PCR assays, despite much promise for the future, still produce too many false-positive results at the moment [2]. Currently, the enzyme immunoassay (EIA) is the most reliable, sufficiently sensitive and readily available laboratory test for LB. However, false-positive reactions do occur because of crossreacting antibodies due to an infection with other microorganisms, such as Treponema pallidurn, or due to certain autoimmune diseases, such as rheumatoid arthritis. In the case of early Lyme borreliosis (ELB) (duration of a few weeks to 3 months) an IgM test is indicated. To improve specificity, a two-test protocol consisting of a screening of suspected LB patients with an E I A test followed by confirmation of positive cases with Western blot (WB) analysis, has been recommended by the CDC [3]. As no general European recommendations have been accepted, the American two-test protocol is commonly used in Europe. However, the antigenic variation of European Borrelia burgdorferi sensu late strains might cause a different immune response. To evaluate a recommended two-test protocol for patients with ELB, we tested the sera of confirmed ELB patients, of healthy controls and of several groups of patients with diseases known to cause cross reactions in Lyme serology or with clinical symptoms mimicking Lyme borreliosis. All early Lyme patients had presented with erythema migrans and either had a recent tick bite, according their anamnesis, or B. burgdorferi was isolated from a skin lesion. All recovered after adequate antibiotic therapy and no relapses occurred. The healthy controls had neither a history of LB nor tick exposure. The patient control group consisted of 40 patients with infectious mononucleosis (positive according to the Paul and Bunell test and EBV IgM-EIA), 62 patients with acute cytomegatovirus (CMV) infection with a confirmed pp65 antigenemia, and a population of 92 other nonLyme patients: 24 patients with rheumatoid factor-positive rheumatoid arthritis, 24 patients with antinuclear antibodies, 23 patients with T. pallidum antibodies and 21 patients with HIVinfection. All samples in this study were tested in a commercially available E I A (Genzyme Virotech, Russelsheim, Germany) and in two WB assays (Genzyme Virotech and MRL, Los Angeles, CA, USA). The Genzyme Virotech Lyme IgM-EIA and IgM blot both contained the same antigen: genotype I strain 2591 (B. burgdorferi sensu stricto) and the MRL kit genotype II (Borrelia garinii). Tests were performed and interpreted according to the instructions of the manufacturers. An IgM blot was considered positive when OspC (24 kDa) and/or BmpA (39 kDa) (both considered highly specific proteins) with or without flagellin (41 kDa) reacted positive. No differences in results of the IgM-EIA-positive samples were observed between the WB test using the genotype I or genotype II strain. The results are shown in Table 1. The sensitivity and specificity of the E I A test were 81% and 79%, respectively, and of the two-step protocol 49% and 88%. However, by omitting the infectious mononucleosis and CMVinfected patients the specificity of the E I A test increased to 97% and of the two-test protocol to 99%. Approximately 60% of the Lyme IgM-EIA-posi t ive sera from CMV and infectious mononucleosis patients reacted in the WB test with one of the two proteins considered highly specific, mainly OspC. All sera Table 1: Serology results in the two-test protocol for early Lyme borreliosis.

Genetic characterization of glycopeptide‐resistant enterococci of human and animal origin from mixed pig and poultry farms

Glycopeptide resistant enterococci (GRE) isolated from animals and humans were characterised using both AFPL typing and genetic characterisation of the glycopeptide resistance transposon Tn1546. All isolates were collected in 1997 when the glycopeptide avoparcin was still being used as growth promoter. All investigated animal isolates were from mixed pig and poultry farms in the Netherlands and the human isolated from the farmers of these farms. A total of 24 isolates were investigated. AFLP and Tn1546 typing revealed that both pig and poultry related enterococcal and vanA transposon genotypes were found among the human isolates indicating spread of glycopeptide resistance from both pig and poultry to the farmers. These findings contradict previous finding that showed that GRE recovered from the general population were genotypically undistinguishable from GRE isolated from pigs but are in line with other studies that demonstrated spread of GRE from poultry to farmers in poultry farms.