|Home | About | Journals | Submit | Contact Us | Français|
An international external quality assurance system (EQAS) for the serotyping of Salmonella species was initiated in 2000 by WHO Global Salm-Surv to enhance the capacity of national reference laboratories to obtain reliable data for surveillance purposes worldwide. Seven EQAS iterations were conducted between 2000 and 2007. In each iteration, participating laboratories submitted serotyping results for eight Salmonella isolates. A total of 249 laboratories in 96 countries participated in at least one EQAS iteration. A total of 756 reports were received from the participating laboratories during the seven EQAS iterations. Cumulatively, 76% of participating laboratories submitted data for all eight strains, and 82% of strains were correctly serotyped. In each iteration, 84% to 96% of the laboratories correctly serotyped the Salmonella enterica serovar Enteritidis isolate that was included as an internal quality control strain. Regional differences in performance were observed, with laboratories in Central Asia and the Middle East performing less well overall than those in other regions. Errors that resulted in incorrect serovar identification were typically caused by difficulties in the detection of the phase two flagellar antigen or in differentiation within antigen complexes; some of these errors are likely related to the quality of the antisera available. The results from the WHO Global Salm-Surv EQAS, the largest of its kind in the world, show that most laboratories worldwide are capable of correctly serotyping Salmonella species. However, this study also indicates a continuing need for improvement. Future training efforts should be aimed at enhancing the ability to detect the phase two flagellar antigen and at disseminating information on where to purchase high-quality antisera.
More than 2,500 serovars of Salmonella enterica have been identified; most human infections are caused by a limited number of serovars. In many developed countries, Salmonella enterica serovars Typhimurium and Enteritidis are the most common causes of human salmonellosis (5, 6, 8).
In other regions, however, other serovars have been reported to be more prevalent (1, 3, 5). Changes in the prevalences of specific serovars can result from the movements of people, animals, and food. Correct serotyping is essential for discerning such changes and therefore is essential for efficient outbreak detection and response resulting from laboratory-based surveillance.
In January 2000, the World Health Organization (WHO) launched WHO Global Salm-Surv, a global effort to enhance the laboratory-based surveillance of Salmonella infections and other food-borne diseases and to promote prevention and control activities. Enhancing worldwide serotyping of Salmonella species is a key objective of WHO Global Salm-Surv and is facilitated by bench training at international training courses.
To ascertain the performance of participating laboratories and thereby promote enhanced laboratory-based surveillance, an external quality assurance system (EQAS) was established as a part of the WHO Global Salm-Surv program in 2000. Since then, the WHO Global Salm-Surv EQAS has grown to be the largest of its kind worldwide (4, 9). Among other activities, the EQAS conducts an assessment of the capacities of laboratories to correctly serotype Salmonella species by shipping eight blinded Salmonella isolates for serotyping. Iterations of the EQAS are organized yearly by Denmark's National Food Institute (DTU Food) in collaboration with WHO, the U.S. Centers for Disease Control and Prevention (CDC), and France's Institut Pasteur. The WHO Global Salm-Surv EQAS is a self-evaluating system: after submitting their results to the EQAS Web-based reporting system via a secured individual log-in pass code, participants receive a report that itemizes errors relative to the expected results. The report is intended to be used by the participants for evaluating the accuracy of current techniques and the quality of antisera. The goal is to have all laboratories perform serotyping of Salmonella with a maximum of one error out of eight isolates. Here we report the results of the first seven iterations of the WHO Global Salm-Surv EQAS Salmonella serotyping procedures, conducted from 2000 to 2007. In 2005, no iteration was conducted, due to an internal assessment of the system.
Participation in the WHO Global Salm-Surv EQAS is open to all laboratories free of charge. An invitation to participate in each EQAS iteration was announced through Electronic Discussion Group messages, which are received by WHO Global Salm-Surv members. All members receive these messages by e-mail or by fax. Previous messages are archived on the WHO Global Salm-Surv website (http://www.who.int/salmsurv/activities/bulletin_board/en/). In 2007, WHO Global Salm-Surv membership included >1,000 microbiologists and epidemiologists from 152 countries representing >140 national reference laboratories (animal, food, or public health). Scientists who attended WHO Global Salm-Surv international training courses were automatically registered as WHO Global Salm-Surv members. Through 2008, 56 training courses were held at 17 training sites worldwide. The training curricula provided in these courses typically include both classroom and laboratory instructional modules on the identification and serotyping of Salmonella species.
Eight Salmonella strains, based on the global prevalence of the serotypes, were selected for each EQAS iteration (3). Strains were obtained from the isolate collection of DTU Food. The same strain of Salmonella serovar Enteritidis was included as an internal quality control in 2000, 2001, 2004, 2006, and 2007. For the 2002 iteration, an alternate but phenotypically identical strain of Salmonella serovar Enteritidis was chosen. All other strains were included only once in the EQAS iterations. Some serovars (e.g., Salmonella serovar Typhimurium) were utilized in multiple iterations; however, these strains had different antimicrobial resistance profiles (Table (Table11).
All Salmonella isolates included in the EQAS were serotyped at DTU Food, the CDC, and Institut Pasteur; the serotypes obtained served as the reference standard. O (somatic) and H (flagellar) antigens were characterized by agglutination with hyperimmune sera, and serotypes were assigned according to the Kauffmann-White scheme (11).
Testing instructions and a “participating laboratory record sheet” (PLRS) were copied to a compact disc, enclosed with the Salmonella agar stab cultures in double-pack containers (class UN 6.2), and sent to the participating laboratories according to the International Air Transport Association regulations as “Biological Substance Category B,” classified as UN 3373. Prior to shipping, each participating laboratory was notified of the shipping arrangements that had been made for the parcels and was given the airway bill number to enable it to track the package and pick it up from the airport. Import permits were necessary for shipping the parcels to several countries.
WHO Global Salm-Surv EQAS participation was free of charge, but each participating laboratory was expected to cover the expenses associated with its testing of the strains in its facility. Participating laboratories were provided with instructions for the initial subculture of the Salmonella strains. Participating laboratories serotyped the isolates using protocols routinely utilized in their institutions; therefore, instructions for serotyping the isolates were not provided. Laboratories were required to submit results by uploading the PLRS onto the WHO Global Salm-Surv website or by submitting the completed PLRS by fax to DTU Food.
After submitting results, each participating laboratory received an individual report. Laboratories that submitted results via the website received an instant report via the secure website, and laboratories that sent the results by fax or e-mail received the report using media. The individual reports included all errors and suggestions on how to either solve or investigate the problem. Errors are defined as results that are different from the expected serotypes. Errors were reported as incorrect results only, and no attempt was made to quantify their severity. Fisher exact tests were performed to assess the significance of the observed changes in correct serotyping results and reported errors. Laboratory participation over the years was assessed by logistic regression, and a P value of <0.05 was regarded as significant for all statistical tests. SAS Enterprise Guide software (version 3.0; SAS Institute, Cary, NC) was used for the statistical analyses.
A total of 249 laboratories in 97 countries participated in at least one of the seven iterations of EQAS from 2000 to 2007; 44 laboratories in 35 countries participated in 2000, 96 laboratories in 55 countries in 2001, 99 laboratories in 61 countries in 2002, 127 laboratories in 72 countries in 2003, 127 laboratories in 71 countries in 2004, 130 laboratories in 66 countries in 2006, and 140 laboratories in 68 countries in 2007. The average number of participating laboratories per EQAS iteration between 2000 and 2007 was 102. One hundred twenty-five laboratories participated in three or more iterations, and 92 laboratories participated in four or more iterations. The participating laboratories included national reference laboratories; veterinary, food, and regional public health laboratories; and clinical laboratories. One or more institutions from the countries listed in Table Table22 participated in at least one of the EQAS iterations.
The percentage of participating laboratories that performed serotyping on all eight strains during the seven iterations ranged from 54% to 92%, with an average of 76% (Table (Table3).3). The percentage of participating laboratories that correctly serotyped the Salmonella serovar Enteritidis isolate that was included in six of the seven iterations increased over the years, although the increase was not statistically significant (P = 0.37), from 92% (2000) to 96% (2007).
The percentage of correct serotyping results oscillated in the initial years, decreasing from 76% in 2000 to 72% in 2001 and rising to 91% in 2002. In the 2003 cycle, the percentage of correct serotyping results was 80%, and since then it has increased annually, reaching 88% in 2007. Overall, logistic regression indicates a significant increase in the percentage of correct serotyping results over the years (P < 0.01), and the average percentage of correctly serotyped isolates across all 7 years was 82% (Table (Table33).
The goal of the EQAS program is for all participating laboratories to perform Salmonella serotyping with a maximum of one error. A total of 756 PLRSs were received during the seven EQAS iterations. The percentage of laboratories reaching the threshold of reporting one or zero errors increased significantly (P = 0.04), from 48% in 2000 to 68% in 2007 (data not shown). In addition, the 2007 iteration was the first in which every participating laboratory correctly identified at least one strain.
A wide range of incorrect results was observed across the seven-cycle study period. The rate of errors ranged from 3.6% (2007) to 41.0% (2006). The rate of errors also differed widely between the different isolates within a single year. For example, in 2001, rates of incorrect results for a single isolate ranged from 9.6% for the Salmonella serovar Typhimurium isolate to 38.0% for a Salmonella serovar Kottbus isolate.
A Salmonella enterica subsp. enterica serovar 4,5,12:i:− isolate included in the 2006 iteration accounted for the greatest number of incorrect results (Table (Table1).1). A total of 38 laboratories (31.1%) incorrectly serotyped this isolate as Salmonella serovar Typhimurium. Salmonella enterica subsp. enterica serovar 4,5,12:i:− is a common monophasic variant in both Europe and North America. Characterization of other Salmonella enterica subsp. enterica serovar 4,5,12:i:− strains suggests that these strains are most likely variants of Salmonella serovar Typhimurium that arose following loss of the phase two flagellin gene fljB (2).
Incorrect serovar identification was often caused by incorrect detection of the phase two flagellar antigen. In many instances, erroneous serotyping results differed from the expected serovar only by the phase two flagellar antigen (Table (Table1).1). Discrimination within the H antigen complexes (E, G, and L) also accounted for a significant number of errors.
The overall percentage of correctly serotyped isolates differed by region. The highest numbers of errors were observed in Central Asia and the Middle East, where little improvement was seen during the seven iterations (50% in 2001 and 55% in 2007). The number of participants in the Oceanic region has been consistent (n = 4), and all four laboratories correctly serotyped all eight strains in both 2001 and 2007. In Southeast Asia, the percentage of correctly serotyped isolates increased significantly (P = 0.01), from 54% in 2001 to 92% in 2002, and remained consistently over 80% through 2007. In Latin America, performance increased significantly (P < 0.01), from 58% in 2001 to 89% in 2007. Europe had the highest number of participating laboratories, and the percentage of correctly serotyped isolates ranged from 81% in 2001, when 43 laboratories participated, to 89% (P < 0.01) in the 2007 iteration, when 54 laboratories participated (Table (Table22).
Through training, reference testing services, technical support, and hosting of scientists, WHO Global Salm-Surv has been working to improve the laboratory capacity and thereby the data quality of WHO member states. The EQAS is one of the tools the program uses to assess the impact of its capacity-building efforts and to pinpoint areas for improvement.
The results from the first seven iterations of the WHO Global Salm-Surv EQAS, which to our knowledge is the largest external quality assurance program for the serotyping of Salmonella species, indicate that the majority of participating laboratories worldwide are capable of correctly serotyping Salmonella species. The number of participating laboratories increased consistently from 2000 to 2007, demonstrating an increased interest in quality assurance and an increased global capacity for Salmonella serotyping. However, fluctuations in the performance of some laboratories were observed. Some laboratories still do not meet the WHO Global Salm-Surv goal of performing serotyping on all eight strains with no more than one incorrect result. Efforts at building laboratory capacity for serotyping should focus on those laboratories in the future and should be directed at common difficulties.
Since test strains differ from year to year, an improvement in the performance of participating laboratories can be evaluated only on the basis of the internal quality controls. The results of quality control have remained fairly consistent during the seven iterations despite a large increase in the number of participants. The selection of a common serovar (Salmonella serovar Enteritidis) may have biased the results: this serovar is frequently encountered, and many laboratories are proficient at its identification. Additionally, this is a monophasic serovar. We have shown that most incorrect results appear to be caused by errors in the identification of phase two flagellar antigens. These factors might have contributed to the consistently high performance observed for the Salmonella serovar Enteritidis strain. Selection of a diphasic serovar may reduce this bias in the future.
Our data suggest that several factors contributed to the observed errors. Unpublished data from a needs assessment in the EQAS 2007 iteration, where 82 laboratories (56%) completed the survey, showed that nearly 1 out of 3 (30%) laboratories have limited access to high-quality antisera. Additionally, less-common serovars were included in some iterations (e.g., Salmonella enterica serovar Vinohrady [I 28:m,t:−]), and the same needs assessment found that 26% of laboratories had difficulty serotyping rare and unusual Salmonella strains. Finally, there are important regional differences in laboratory capacity. The needs assessment found that institutions in Africa, Asia and the Middle East, and Southeast Asia were more likely to report difficulty obtaining antisera, especially for serotyping unusual Salmonella strains, than were institutions in other regions.
The decline in the proportion of serotypes correctly identified in 2003 and 2004 was likely due to the selection of the Salmonella isolates (Table (Table3).3). In 2003 and 2004, laboratories needed less-common antisera in order to fully serotype all of the EQAS isolates. After 2004, only more-common serovars were included. The subsequent overall improvement in performance suggests that many laboratories have access only to commonly available antisera. WHO Global Salm-Surv has demonstrated that the predominant Salmonella serotypes differ between regions (3). Therefore, a broad selection of antisera for Salmonella surveillance is needed globally. WHO Global Salm-Surv continues to provide information to participants on where to purchase high-quality antisera and to support many laboratories with antisera for surveillance purposes.
Many of the incorrect serotyping results were due to incorrect identification of phase two flagellar antigens (Table (Table1).1). For example, common incorrect results included misidentification of Salmonella serovar Typhimurium (I 4,5,12:i:1,2) as Salmonella enterica serovar Farsta (I 4,5,12:i:e,n,x), Lagos (I 4,5,12:i:1,5), or Tumodu (I 4,5,12:i:z6). All four serovars share the same O and phase one flagellar antigens but differ in their phase two flagellar antigens.
Colonial form variation (the variable expression of minor antigens by different single-colony picks from the same strain) may occur with the expression of the O:61 antigen by some serogroup C2 serovars (10). Therefore, although the current Kauffmann-White scheme regards O:6,8 and O:8 serovar pairs, such as Salmonella serovars Newport (I 6,8:e,h:1,2) and Bardo (I 8:e,h:1,2), as distinct serovars, we allowed for colonial form variations. We considered correct identifications for Salmonella serovars Newport, Kottbus, Hadar, Manhattan, and Bovismorbificans on the basis of the serogroup alone and accepted as correct for those serovars, respectively, Salmonella serovars Bardo, Ferruch, Istanbul, Yovokome, and Hindmarsh.
The results from the WHO Global Salm-Surv EQAS also demonstrate important regional differences in the serotyping results for Salmonella species. Particular efforts should focus on Central Asia and the Middle East, but also on Africa, Russia, and the Caribbean, where a large proportion of the laboratories do not correctly serotype many of the strains. Addressing the regional differences will involve additional training courses in selected regions.
WHO Global Salm-Surv is a platform that can assist WHO member states to strengthen their core public health capacities under the International Health Regulations (IHR, 2005) for disease surveillance and response, which will in turn strengthen international public health security. WHO Global Salm-Surv promotes intersectoral collaboration among human health, veterinary, and food-related disciplines in food safety and other issues that arise at the human-animal interface. As the program continues to expand, it increasingly addresses regional training and support needs.
This study showed that there is a continuing need to improve Salmonella serotyping and that this need appears to be greater in specific regions. Detection of the phase two flagellar antigen is one of the more profound barriers to obtaining a satisfactory serotyping result. Future training efforts should be aimed at enhancing the ability to characterize the phase two flagellar antigen and disseminating information on where to purchase high-quality antisera.
We thank the technical staff at the National Food Institute for preparing, testing, and shipping the strains; the staff of the participating laboratories who contributed to this program; and the staff at the regional sites for further distributing the EQAS parcels.
Published ahead of print on 1 July 2009.