Allergy: Evaluation of 16 years (2007-2022) results of the shared external quality assessment program in Belgium, Finland, Portugal and The Netherlands.

OBJECTIVES: This paper evaluates 16 year results of the Allergy EQA program shared by EQA organisers in Belgium, Finland, Portugal, and The Netherlands. METHODS: The performance of Thermo Fisher and Siemens user groups (in terms of concordance between both groups, between laboratory CV, prevalence of clinically significant errors) and suitability of samples (stability and validity of dilution of patient samples) are evaluated using data of 192 samples in the EQA programs from 2007 to 2022. Measurands covered are total IgE, screens and mixes, specific IgE extracts and allergen components. RESULTS: There is perfect (53 %), acceptable (40 %) and poor (6 %) concordance between both method groups. In case of poor concordance the best fit with clinical data is seen for Thermo Fisher (56 %) and Siemens (26 %) respectively. The between laboratory CV evolves from 7.8 to 6.6 % (Thermo Fisher) and 7.3 to 7.7 % (Siemens). The prevalence of blunders by individual laboratories is stable for Siemens (0.4 %) and drops from 0.4 to 0.2 % for Thermo Fisher users. For IgE, the between year CV of the mean of both user groups is 1 %, and a fifteen-fold dilution of a patient sample has an impact of 2 and 4 % on the recovery of Thermo Fisher and Siemens user groups. CONCLUSIONS: The analytical performance of Thermo Fisher is slightly better than that of Siemens users but the clinical impact of this difference is limited. Stability of the sample and the low impact of dilution on the recovery of measurands demonstrates the suitability for purpose of the EQA program.

Pitfalls of rubella serology while on the brink of elimination: evaluation of national data, Belgium, 2017.

BackgroundIn Belgium, rubella serology is frequently requested in women of childbearing age, despite high vaccination coverage and a near-absence of congenital rubella cases. Different test kits are available and should be standardised by an international standard preparation.AimTo analyse and compare rubella serology practices in Belgian laboratories.MethodsAs part of the mandatory External Quality Assessment programme for rubella serology in Belgium, the national public health institute, Sciensano, sent a voluntary questionnaire concerning anti-rubella IgM/IgG analyses in women aged 15 to 45 years in 2017 to 130 laboratories.ResultsThe questionnaire response rate was 83.8% (109/130). The majority of 169,494 IgG analyses were performed on Roche (55%), Abbott (17%) and Diasorin (13%) analysers. Not all laboratories used the proposed international cut-off of 10 IU/mL. Assumed median seroprevalence ranged from 76.3% with Liaison (Diasorin) to 96.3% with Modular (Roche). Despite very low rubella incidence in Belgium, 93 laboratories performed 85,957 IgM analyses, with 748 positive and 394 grey zone results. The National Reference Centre for Measles, Mumps and Rubella virus and the National Reference Centre for Congenital infections did not confirm any positive rubella cases in 2017.ConclusionThis retrospective analysis shows that rubella serology results may differ considerably according to the assay used. It is therefore important to use the same test when comparing results or performing follow-up testing. The number of anti-rubella IgM analyses was very high. Incorrect use of IgM for screening women of childbearing age can lead to unwarranted anxiety and overuse of confirmation tests.

Alternative Sample-Homogeneity Test for Quantitative and Qualitative Proficiency Testing Schemes.

Proficiency Testing (PT) External Quality Assessment (EQA) schemes are designed to ascertain the ability of individual laboratories to perform satisfactorily with respect to their peer laboratories or to limits imposed by external sources. Observed deviation of a laboratory result for a PT sample must be entirely attributed to the laboratory and not to the PT provider. To minimize the probability that deviations could be attributed to the PT provider, sample homogeneity should be assured. It is generally required that for quantitative parameters, the standard deviation among PT units should be calculated on the basis of duplicate measurements of at least 10 samples chosen at random, and the standard deviation among PT units should not exceed 0.3 times the standard deviation used to evaluate laboratories. Because this approach has important drawbacks, an alternative procedure is proposed by applying the theory of acceptance sampling to the assessment of sample heterogeneity for both quantitative and qualitative data and deriving acceptance limits on the basis of minimizing the probability of falsely evaluating laboratories. For obtaining acceptance limits for quantitative parameters, a distinction is made between laboratory evaluation using fixed limits on the one hand and laboratory evaluation using limits that are based on the variability of the reported results on the other hand. Sequential tests are proposed to evaluate sample heterogeneity by means of a comparison with the χ2 distribution. For qualitative parameters, acceptance-sampling plans are proposed that are based on minimizing the joint probability of rejecting batches that have a satisfactory amount of defective samples and accepting batches unnecessarily. The approach for quantitative parameters is applied on samples for a PT scheme of ethanol quantification and for qualitative parameters on the presence of monoblasts in a blood smear. It was found that five samples could already be enough to prove that the batch was homogeneous for quantitative parameters, although more than 20 samples were needed to prove homogeneity for qualitative parameters. This study describes a direct relation among the objective of an PT round, the criteria for evaluating the results, and the sample heterogeneity. When samples are effectively homogeneous, less measurements are needed than current practices require. A drawback of the proposed approach is that the number of samples to be tested is not known beforehand, and good knowledge of the analytical variability is crucial. The formulas to be applied are relatively simple. Despite the drawbacks, the proposed approach is generally applicable for both quantitative and qualitative data.

A pragmatic bottom-up approach to harmonize the units of clinical chemistry tests among Belgian clinical laboratories, focusing on immunoassays.

BACKGROUND: Harmonization of units is an important step to improve the comparability of clinical chemistry results, but few examples exist of successful harmonization efforts. We present the results of a pragmatic approach that was implemented in Belgium from 2012. METHODS: After a large consultation and information of stakeholders, preferred units were proposed for 140 assays, including the 23 immunoassays discussed in more detail here. The change occurred in two phases, first involving assays for which there was no change in the numerical result, then changes involving a change in numerical results. Laboratories were invited to participate in this harmonization on a voluntary basis. The project was based on a bottom-up approach, large consultation and the pragmatic choice of the proposed units, including conventional and SI units. RESULTS: The large heterogeneity of units was drastically reduced; adoption of the preferred units increased from 3% (insulin) - 45% (HCG) to 70% (insulin) - 96% (LH and FSH). Adoption of the preferred units was higher if it involved no change in numerical values (90%) than when there was a change (76%). CONCLUSIONS: We believe that the harmonization effort has reached its goals. Without aiming at implementing SI units for all parameters, our strategy was successful with a large majority of the laboratories switching to the proposed units. Moreover, the harmonization program is still progressing, with additional laboratories converting to the consensus units.

Importance of anti-SARS-CoV-2 assay antigenic composition as revealed by the results of the Belgian external quality assessment (EQA) scheme.

We report on sample IS/17575 since it generated highly divergent results in the Belgian SARS-CoV-2 serology external quality assessment scheme. Sample IS/17575 was serum originating from a 30 years old male patient. 124 diagnostic laboratories analysed this sample. A total of 168 results was returned (including 5 doubles). Overall, 38 were positive. All tests against S1 were positive except the Euroimmun IgG ELISA and the Ortho clinical Diagnostics VITROS IgG CLIA. All tests against S1/S2 (Liaison, Diasorin) resulted in a signal above cutoff. Assays against RBD, mostly generate a negative result. An exception are the Wantai SARS-CoV-2 ELISA's. All tests targeting N protein were negative. The survey shows, when >6 months post-infection, assays targeting at least S1, and preferably S1 combined with S2, are the most sensitive. This finding accentuates the necessity of external quality assessment schedules and importance of antigenic composition of serologic SARS-CoV-2 assays.

Bifidobacterium pseudolongum are efficient indicators of animal fecal contamination in raw milk cheese industry.

BACKGROUND: The contamination of raw milk cheeses (St-Marcellin and Brie) from two plants in France was studied at several steps of production (raw milk, after addition of rennet - St-Marcellin - or after second maturation - Brie -, after removal from the mold and during ripening) using bifidobacteria as indicators of fecal contamination. RESULTS: Bifidobacterium semi-quantitative counts were compared using PCR-RFLP and real-time PCR. B. pseudolongum were detected in 77% (PCR-RFLP; 1.75 to 2.29 log cfu ml(-1)) and 68% (real-time PCR; 2.19 to 2.73 log cfu ml(-1)) of St-Marcellin samples and in 87% (PCR-RFLP; 1.17 to 2.40 log cfu ml(-1)) of Brie cheeses samples. Mean counts of B. pseudolongum remained stable along both processes. Two other populations of bifidobacteria were detected during the ripening stage of St-Marcellin, respectively in 61% and 18% of the samples (PCR-RFLP). The presence of these populations explains the increase in total bifidobacteria observed during ripening. Further characterization of these populations is currently under process. Forty-eight percents (St-Marcellin) and 70% (Brie) of the samples were B. pseudolongum positive/E. coli negative while only 10% (St-Marcellin) and 3% (Brie) were B. pseudolongum negative/E. coli positive. CONCLUSIONS: The increase of total bifidobacteria during ripening in Marcellin's process does not allow their use as fecal indicator. The presence of B. pseudolongum along the processes defined a contamination from animal origin since this species is predominant in cow dung and has never been isolated in human feces. B. pseudolongum was more sensitive as an indicator than E. coli along the two different cheese processes. B. pseudolongum should be used as fecal indicator rather than E. coli to assess the quality of raw milk and raw milk cheeses.

Seroprevalence of Antibodies Against Anaplasma Phagocytophilum and Borrelia Burgdorferi in Horses (Equus Caballus) from Northern Algeria.

INTRODUCTION: Horses (Equus caballus) are susceptible to tick-borne diseases. Two of them, Lyme borreliosis due to Borrelia burgdorferi and granulocytic anaplasmosis due to Anaplasma phagocytophilum were investigated in Algerian horses. The diseases have been less extensively studied in horses and results pertinent to Algeria have not been published. MATERIAL AND METHODS: Blood samples were obtained from 128 horses. IgG antibodies directed against Anaplasma phagocytophilum and Borrelia burgdorferi were detected by an indirect immunofluorescence antibody test (IFAT) and ELISA. The potential effects of age, gender, breed, and health status on seropositivity were also evaluated. RESULTS: Using IFAT, 28 (21.8%) and 25 (19.5%) animals were positive for B. burgdorferi and A. phagocytophilum, respectively. Using ELISA, 19 (14.8%) and 33 (25.9%) animals were positive for these bacteria. CONCLUSION: The study shows that horses in Algeria are exposed or co-exposed to tick-transmitted zoonotic bacterial species.

Immunologic response of unvaccinated workers exposed to anthrax, Belgium.

To determine immunologic reactivity to Bacillus anthrax antigens, we conducted serologic testing of workers in a factory that performed scouring of wool and goat hair. Of 66 workers, approximately 10% had circulating antibodies or T lymphocytes that reacted with anthrax protective antigen. Individual immunity varied from undetectable to high.

Evaluation of a SRS2 sandwich commercial enzyme-linked immunosorbent assay for the detection of anti-Neospora caninum antibodies in bovine and canine sera.

Neospora caninum is a parasite responsible for abortion in cows and neuromuscular disease in dogs. Serology is the most widely used technique to evaluate the prevalence of N. caninum in different host populations. A sandwich enzyme-linked immunosorbent assay (ELISA) based on the use of an anti-SRS2 monoclonal antibody was evaluated against the indirect fluorescent antibody test for 100 canine sera and against a well-characterized ELISA for 102 bovine sera. In cattle sera, the relative sensitivity and relative specificity were 100%. In dog sera, the relative specificity and relative sensitivity were 94% and 86%, respectively. The kappa value was 1 for bovine sera and 0.77 for canine sera. The seroprevalence was 3.9% in bovine sera and 21-23% in canine sera. The SRS2 sandwich ELISA was considered a valuable tool in both species.

Seroepidemiological study of the exposure to Toxoplasma gondii among horses in Algeria and analysis of risk factors.

AIM: The aim of this study was to assess the seroprevalence of the Toxoplasma gondii in horses in different parts of Algeria and to determine risk factors for the infection. MATERIALS AND METHODS: A total of 736 blood samples were collected from horses of various breeds, gender, coat colors, and ages. All horses came from various farms, racecourses, and equestrian centers. The seroprevalence was investigated by three different methods: Indirect fluorescent antibody test (IFAT) as reference method, enzyme-linked immunosorbent assay (ELISA), and latex agglutination test (LAT). RESULTS: Out of the 736 sera, 178 (24.18%) were positive for IFAT, 133 (18.07%) for LAT, and 317 (43.07%) for ELISA. It was found that IFAT and LAT were in high agreement (Kappa 0.79), indicating that LAT and IFAT had similar capabilities in the detection of anti-T. gondii antibodies from horse sera. Risk factors analysis based on IFAT results indicated that the habit of the animals was significant risk factors (p≤0.05) for Toxoplasma infection. The seroprevalence was significantly higher in horses living on farms. Moreover, a higher seroprevalence was found in older animals compared to younger ones. Furthermore, the seroprevalence in females was significantly higher than that in males and gelding. Breed, coat color, and water sources are also important factors to influence the seroprevalence of T. gondii. CONCLUSION: The results indicated that T. gondii is present in horses throughout Algeria and thus represents a risk for both human and animal health. These results underline the need to increase the vigilance and the preventive measures against this disease not only to protect the horses but also to limit the spread of the parasite.

Detection of Neospora caninum in dog organs using real time PCR systems.

Neospora caninum is a parasite responsible for paresis in dogs. The dog can harbour enkysted parasites in several organs. The detection of N. caninum was performed using 3 different real time PCR systems all amplifying the NC5 DNA region. One system was based on Sybrgreen, one on Plexor technology and the last on Taqman probe. Comparison of the three methods indicated that the detection limit was 1 equivalent genome on pure DNA but that this detection limit increased in the presence of foreign DNA using the Sybrgreen and Plexor systems. Therefore, the Taqman system was chosen to detect N. caninum in liver and spleen of naturally infected dogs. The overall prevalence was 32.2%. Comparison between PCR results and serological results using IFAT showed that among the 28 PCR positive dogs only 9 were seropositive and that 8 seropositive dogs were PCR negative. Therefore serology can underestimate the real carriage in dogs. However, PCR methods must be improved in terms of sensitivity and inhibition problems.

Photo-based External Quality Assessment of Malaria rapid diagnostic tests in a non-endemic setting.

INTRODUCTION: In non-endemic settings, expertise in malaria microscopy is limited and rapid diagnostic tests (RDTs) are an adjunct to malaria diagnosis. AIM: We performed an External Quality Assessment (EQA) on reading and interpretation of malaria RDTs in a non-endemic setting. METHODS: Participants were medical laboratories in Belgium and the Grand Duchy of Luxembourg using malaria RDTs; they received (i) 10 high-resolution photographs presenting test line combinations of RDTs with interpretations listed in a multiple choice format and (ii) a questionnaire about their practices of malaria diagnosis. RESULTS: Among 135 subscribing laboratories, 134 (99.3%) used 139 RDT products (11 different products from 8 brands). After exclusion of the results of one laboratory, analysis was done for 133 laboratories using 137 RDT products. Scores of 10/10, 9/10 and 8/10 were achieved for 58.4%, 13.1% and 8.0% of 137 RDT products respectively. For three-band P. falciparum-pan-Plasmodium RDTs (113 (82.5%) products, 6 brands), most frequent errors were (1) disregarding faint test lines (18.6%), (2) reporting invalid instead of P. falciparum (16.8%) and (3) reporting "Plasmodium spp., no further differentiation possible" without mentioning the presence or absence of P. falciparum (11.5%). For four-band RDTs (21 (15.3%) products, 1 brand), errors were (4) disregarding faint P. vivax test lines (47.6%) and (5) reporting "Plasmodium spp., no further differentiation possible" without mentioning the presence of P. falciparum and P. vivax (28.6%). Instructions for use (IFU) of only 4 out of 10 RDT products mentioned to interpret faint-intensity test lines as positive (conducive to errors 1 and 4) and IFU of 2 products displayed incorrect information (conducive to errors 2 and 5). Outside of office hours, 36.1% of participants relied on RDTs as the initial diagnostic test; 13.9% did not perform microscopic confirmation. CONCLUSION: Reading and interpretation of malaria RDTs was satisfactory, but errors were embedded in the instructions for use of the products. Relying on RDTs alone for malaria diagnosis (about one third of participants) is not a recommended practice.

Prevalence of The Main Infectious Causes of Abortion in Dairy Cattle in Algeria.

INTRODUCTION: Abortion in cattle is a major source of economic losses for the agriculture sector. It can be due to infectious or non-infectious factors. Among infectious factors, parasites, bacteria, viruses, and fungi can be involved. The present work investigated the prevalence of the main infectious agents of abortion in Algerian cattle. MATERIAL AND METHODS: Altogether 278 non-aborting and 82 aborting cows were analysed. RESULTS: The prevalence ranged from 0% for Tritrichomonas foetus to 15% for Neospora caninum. Additionally, a case-control study was performed to find the association between the presence of the pathogens and the occurrence of abortion in cows. The odds ratios were significant for Neospora caninum, bovine herpes virus 4, BVD virus, Brucella abortus, Salmonella Dublin, Leptospira interrogans serovar Hardjo, and Coxiella burnetii. CONCLUSIONS: The pathogens enumerated here could be major causes of abortion among Algerian cattle.

Use of a serological approach for prediction of Salmonella status in an integrated pig production system.

Relevance of a Salmonella serological detection technique was studied from complete results obtained from 9 pigs fattening units. Feces and overshoes were sampled at different periods after starting fattening (2, 3 and 4 months) while caecal contents were taken on the slaughter line. The bacteriological technique used was based on a Diasalm enrichment and a commercial test was used for serology on an average of ten animals per batch. The aim of this work was to establish a correlation between serological results obtained at slaughter (10 samples/batch) and bacteriological results. In this context, two types of logistic regression models were tested by considering alternatively serology and Salmonella detection in caecal contents as the dependent variables. Firstly, beside the fact that all logistic regression models show weak correlations, the first finding was that positive results in overshoes taken at 2 and 3 months are slightly correlated with serological status of herds (odds-ratios of 4.96 and 2.55). Secondly, when batches were characterized as positive on the basis of serological results, the probability of Salmonella recovery in caecal contents was higher than when the batches were considered as negative (odds-ratios comprised between 4.36 and 5.81). A major conclusion is that serology can be used to follow the improvement of an integrated pig production system, but is not the unique solution for assessing risk of Salmonella shedding from specific herds.

Belgian surveillance plans to assess changes in Salmonella prevalence in meat at different production stages.

From 1997 to 1999, the prevalence of Salmonella was assessed at different stages through the pork, poultry, and beef meat production chains. Different dilutions of the initial sample suspension were analyzed to provide a semiquantitative evaluation of Salmonella contamination and to determine the most representative dilution necessary to detect a reduction in prevalence. An average of 300 samples for each type of meat were analyzed. According to Fisher's exact test, the dilution to be used to detect a reduction in prevalence was chosen based on an initial prevalence of 20 to 26%. Based on this introductory study, a new sampling plan representative of the nationwide Belgian meat production process was used from 2000 through to 2003. This study confirmed the consistently high rate and level of contamination of poultry meat: broiler and layer carcasses were the most contaminated samples followed by broiler fillets and poultry meat preparations. A constant and significant decrease in Salmonella prevalence was observed for pork carcasses, trimmings, and minced meat and for beef minced meat. Less than 3% of beef carcasses and trimming samples were positive for Salmonella. The Belgian plan, as utilized from 2000 to 2003, was suitable for monitoring of zoonoses because the sampling plan was representative of nationwide production processes, covered all periods of the year, and was executed by trained samplers and the analyses were carried out by recognized laboratories using an identical analytical method.

External Quality Assessment for the Detection of Chlamydia trachomatis in Urine Using Molecular Techniques in Belgium.

Chlamydia trachomatis is a major cause of sexually transmitted bacterial disease worldwide. C. trachomatis is an intracellular bacterium and its growth in vitro requires cell culture facilities. The diagnosis is based on antigen detection and more recently on molecular nucleic acid amplification techniques (NAAT) that are considered fast, sensitive, and specific. In Belgium, External Quality Assessment (EQA) for the detection of C. trachomatis in urine by NAAT was introduced in 2008. From January 2008 to June 2012, nine surveys were organized. Fifty-eight laboratories participated in at least one survey. The EQA panels included positive and negative samples. The overall accuracy was 75.4%, the overall specificity was 97.6%, and the overall sensitivity was 71.4%. Two major issues were observed: the low sensitivity (45.3%) for the detection of low concentration samples and the incapacity of several methods to detect the Swedish variant of C. trachomatis. The reassuring point was that the overall proficiency of the Belgian laboratories tended to improve over time.

Salmonella contamination of pigs and pork in an integrated pig production system.

This paper describes the monitoring of Salmonella in a closed pig production system in Belgium over a 2-year period. A sampling scheme including animal feeds and carcasses was designed to cover the entire chain of production from farrow to finishing pigs. Salmonella was detected by a method based on the use of semisolid Rappaport-Vassiliadis as a selective medium. The serotypes of the isolated strains were determined, and the antibiotic resistance of these strains to six antibiotics was also investigated. Feeds were found to be more contaminated than expected (10.2%, 34 of 332 samples). The percentage of positive fecal samples for pregnant sows (8.1%, 11 of 135 samples) was significantly higher than that for young and lactating sows (2.9%, 11 of 378 samples) (P<0.05). The percentage of positive samples for colon contents collected at the slaughterhouse (47.3%, 88 of 186 samples) was significantly higher than that for feces collected during the fattening stage (5.6%, 18 of 320 samples) (P<0.001). For carcass swab samples, the observed prevalence was 11.2% (17 of 152 samples). On farms, Salmonella recovery levels were higher for overshoe samples than for fecal samples, except for pregnant sows. Salmonella Typhimurium was the most frequently isolated serotype (32.2%, 55 of 171 samples), while Salmonella Brandenburg was predominant in the colon contents collected at the abattoir (21.4%, 18 of 84 samples). Feeds harbored a wide diversity of serotypes of minor epidemiological significance. Of 55 isolated strains of Salmonella Typhimurium, 11 (20%) were resistant to tetracycline, ampicillin, choramphenicol, streptomycin, trimethoprim, and nalidixic acid (R Type TeAmCSNa), while 12 (21.8%) were resistant to all of these antibiotics except nalidixic acid (R Type TeAmCS). The majority of Salmonella Typhimurium strains that exhibited resistance to more than four antimicrobial agents were characterized as Salmonella Typhimurium DT104 or as being closely related to Salmonella Typhimurium DT104 (7 of 12 isolates). In conclusion, our system of surveillance is effective in identifying most points of contamination in the production chain and will be useful in ongoing efforts to develop a Salmonella-free production system.

Comparison of different serological methods to detect antibodies specific to Neospora caninum in bovine and canine sera.

Neospora caninum is an apicomplexan parasite responsible for paresis in dogs and abortion in cattle worldwide. Dogs serve as a definitive host, while cattle serve as intermediate host. Many different methods have been developed to detect specific antibodies present in cattle and dog serum. In the present study, the dense granule protein NcGRA6 was incorporated in a latex beads agglutination test (LAT), and compared to other serological methods, including enzyme-linked immunosorbent assay, the direct agglutination test, the immunoblot, and the indirect fluorescent antibody test (IFAT). Using the IFAT as the reference method, 100 sera isolated from Algerian cattle and 100 sera isolated from Algerian dogs, both possibly infected with N. caninum, were used to evaluate the LAT. The sensitivity, specificity, and kappa index were calculated for each host species and assay. For dog sera, the sensitivity and the specificity of the LAT was 76% and 100%, respectively. The McNemar test showed that the LAT was not significantly different from IFAT (P > 0.05). For cattle sera, the sensitivity and the specificity of the LAT were 60% and 100%, respectively. The McNemar test indicated that the LAT was significantly different from IFAT (P < 0.01) and that the LAT was only positive for cattle sera with titers of 1:800 or greater, indicating that LAT can be used for cattle in a clinical context. As well, the LAT has the advantage of being easy and rapid to perform compared to the other assays.

Study of the risk factors associated with Neospora caninum seroprevalence in Algerian cattle populations.

Bovine abortions due to Neospora caninum infection were reported worldwide. The situation in Algeria was unknown. For the evaluation of the prevalence of N. caninum and its associated risk factors, 799 cattle belonging to 87 farms of the north and northeast of Algeria were analyzed. The cattle were divided into imported cattle, local cattle and improved cattle corresponding to breeding between imported and local cattle. Sera were examined for the presence of N. caninum antibodies by indirect fluorescence antibody test. The overall seroprevalence for the 87 farms was 52.87% (41.28-62.71%). The overall animal seroprevalence was 19.64% (16.82-22.45%). The seroprevalence of N. caninum in local cattle (34.28%) was significantly higher (p<0.05) than in modern (16.04%) and improved (18.64%) cattle. The risk factors analysis indicated that cattle population, geographical location, dog presence, season, global farm hygiene or the presence of abortion were significantly associated with seroprevalence.

Comparison of different approaches to evaluate External Quality Assessment Data.

In EQA programs, Z-scores are used to evaluate laboratory performance. They should indicate poorly performing laboratories, regardless of the presence of outliers. For this, two different types of approaches exist. The first type are "outlier-based" approaches, which first exclude outlying values, calculate the average and standard deviation on the remaining data and obtain Z-scores for all values (e.g., Grubbs and Dixon). The second type includes the "robust" approaches (e.g., Tukey and Qn or the algorithm recommended by ISO). The different approaches were assessed by randomly generated samples from the Normal and Student t distributions. Part of the sample data were contaminated with outliers. The number of false and true outliers was recorded and subsequently, Positive and Negative Predictive Values were derived. Also, the sampling mean and variability were calculated for location and scale estimators. The various approaches performed similarly for sample sizes above 10 and when outliers were at good distance from the centre. For smaller sample sizes and closer outliers, however, the approaches performed quite differently. Tukey's method was characterised by a high true and a high false outlier rate, while the ISO and Qn approaches demonstrated weak performance. Grubbs test yielded overall the best results.