- Email: [email protected]
Accuracy of commercial systems for identification of Burkholderia pseudomallei versus Burkholderia cepacia
Available online at www.sciencedirect.com
Diagnostic Microbiology and Infectious Disease 59 (2007) 277 – 281 www.elsevier.com/locate/diagmicrobio
Accuracy of commercial systems for identification of Burkholderia pseudomallei versus Burkholderia cepacia Pattarachai Kiratisin a,⁎, Pitak Santanirand b , Narisara Chantratita c , Srirumpa Kaewdaeng a b
a Department of Microbiology, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok 10700, Thailand Department of Pathology, Faculty of Medicine Ramathibodi Hospital, Mahidol University, Bangkok 10400, Thailand c Faculty of Tropical Medicine, Mahidol University, Bangkok 10400, Thailand Received 14 February 2007; accepted 20 June 2007
Abstract Infection caused by Burkholderia pseudomallei or Burkholderia cepacia may result in fatal outcome unless the causative agent is accurately identified in a short period, which is critical for treatment. We evaluated the reliability of the commonly used commercial systems, API 20NE, VITEK 2, and WalkAway 96, for comparative identification of B. pseudomallei versus B. cepacia clinical isolates. Based on biochemical and molecular tests as reference methods, API 20NE was probably the most reliable, with an accuracy of 87% and 93%, for the identification of B. pseudomallei and B. cepacia, respectively. The VITEK 2 and WalkAway 96 systems resulted in a number of misidentification and, thus, were less reliable. The performance of each system and identification guidelines for B. pseudomallei and B. cepacia are discussed. Our study emphasized that laboratories should carefully interpret the identification of B. pseudomallei and B. cepacia when using commercial systems. © 2007 Elsevier Inc. All rights reserved. Keywords: Burkholderia pseudomallei; Burkholderia cepacia; Identification
1. Introduction Rapid and accurate identification of Burkholderia pseudomallei, the causative agent of melioidosis, is critical for clinical laboratory, especially for those serving in the endemic areas such as Southeast Asia and northern Australia. A delayed or incorrect identification of this bacterium often results in an unfavorable or fatal outcome due to fulminant septicemia (Cheng and Currie, 2005). Because of the increase in international travel, laboratories in the nonendemic area that have limited experience for B. pseudomallei identification and mostly rely on a commercial method as 1st-line diagnostic tool should be aware of the accountability of such method. In addition, B. pseudomallei is enlisted as a potential bioterrorism agent that clinical laboratories and the public health should recognize (Rotz et al., 2002). The recommended conventional approach based on phenotypic
⁎ Corresponding author. Tel.: +66-2-419-7058; fax: +66-2-411-3106. E-mail address: [email protected] (P. Kiratisin). 0732-8893/$ – see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.diagmicrobio.2007.06.013
characteristics (Dance et al., 1989) may not be sufficient to yield an opportunely definitive identification. Several diagnostic systems are currently available in the market, and yet, there are few evaluation studies regarding the performance of these systems for identification of Burkholderia spp. A widely used substrate utilization panel API 20NE (bioMérieux, Hazelwood, MO) was shown to correctly identify B. pseudomallei ranging from 37% to 98% in various studies (Dance et al., 1989; Inglis et al., 1998; Lowe et al., 2002; Glass and Popovic, 2005). This inconsistence resulted in a disputable competence of API 20NE. Automation systems have been accustomed to many microbiology laboratories. However, the reliability of these systems for B. pseudomallei identification remains doubtful. To our knowledge, only the VITEK systems (bioMérieux, Marcy l'Etoile, France) have been evaluated in a few English literatures. Although the VITEK 1 was shown to correctly identify 99% of B. pseudomallei isolates (Lowe et al., 2002), the accuracy of VITEK 2 was controversial. The fluorometric-based ID-GNB card of the VITEK 2 demonstrated a very poor performance for the identification
278
P. Kiratisin et al. / Diagnostic Microbiology and Infectious Disease 59 (2007) 277–281
of B. pseudomallei with more than 80% of misidentification (Lowe et al., 2002). With the new colorimetric-based GN card, the VITEK 2 identification database for Gram-negative bacteria has been expanded (Funke and Funke-Kissling, 2004). Nonetheless, the recent study showed that the VITEK 2 GN card correctly identified B. pseudomallei varying from 63% to 81%, depending on the culture media used for growing the bacterium (Lowe et al., 2006). The MicroScan WalkAway 96 SI system (Dade Behring, West Sacramento, CA) is also widely used in diagnostic laboratories and, yet, has never been evaluated for B. pseudomallei identification. In addition, it was suggested that the geographic diversity of strains may affect the credibility of diagnostic systems (Glass and Popovic, 2005). B. pseudomallei isolates from the highly endemic area such as Thailand were rarely included in the previous published studies. In this study, we hence evaluated the nonautomated (API 20NE) and automated (VITEK 2 and WalkAway 96) systems for the identification of B. pseudomallei recovered from Thai patients. The previous studies also showed that the VITEK 2 most commonly misidentified B. pseudomallei as Burkholderia cepacia or B. cepacia group (Lowe et al., 2002, 2006). Therefore, the clinical isolates of B. cepacia were included to compare each system in parallel for its capability to distinguish these 2 Burkholderia spp. Both B. pseudomallei and B. cepacia are important respiratory tract pathogens and may cause fatal septicemia. The accurate differentiation between these 2 species is thus a requisite for proper diagnosis and treatment. This study provides a detailed analysis of 3 commercial systems regarding the reliability of B. pseudomallei and B. cepacia identification. 2. Materials and methods 2.1. Bacterial strains and identification The clinical isolates of B. pseudomallei (n = 56) and B. cepacia (n = 56) used in this study were recovered from various specimen types including blood, sputum, and pus from individual patients admitted to Siriraj hospital, Bangkok, Thailand, and 4 provincial hospitals in northeast Thailand (Ubon Rachathani, Srisaket, Surin, and Khon Kaen). The isolates were collected during the years 1997, 2004, and 2005. The clinical strain K92643 of B. pseudomallei used for genome sequencing (Holden et al., 2004) and B. cepacia strain ATCC 25416 were used as reference strains. All isolates were stored at −80 °C in 20% glycerol brain heart infusion broth and were subcultured on 5% (vol/vol) defribrinated sheep blood agar plates for at least 2 passages before use. All potential aerosol-generating procedures were performed in a class II biologic safety cabinet by gowned and gloved personnel. Bacterial strains were originally identified based on the standard biochemical tests (Weyant et al., 1996). The identity of all B. pseudomallei isolates were also confirmed by 2 additional tests that were previously described: specific polymerase
chain reaction (PCR) targeting the DNA sequences of the internal transcribed spacers between 16S and 23S rDNA genes (Kunakorn and Markham, 1995; Inglis et al., 2005), and monoclonal antibody-based latex agglutination (Anuntagool et al., 2000). All B. pseudomallei isolates in this study including reference strain yielded positive results for both PCR and latex agglutination, whereas all B. cepacia isolates had negative results for both tests. The identity of all B. cepacia isolates were also confirmed using recA-based PCR specific for B. cepacia complex as described previously (Mahenthiralingam et al., 2000). 2.2. API 20NE, VITEK 2, and WalkAway 96 identification The bacteria were streaked on sheep blood agar plates and incubated overnight at 37 °C. Colonies of each isolate were used to prepare a suspension for identification by API 20NE, VITEK 2, and WalkAway 96. The identification procedures were performed according to the manufacturers' protocols. The API 20NE strips were incubated for 24 or 48 h depending on the reactions and were read manually using the manufacturer's interpretative chart. The API 20NE results were read and analyzed by 1 investigator using the apiweb program available via internet access (http://apiweb.biomerieux.com). The identification results by VITEK 2 carried out by VITEK 2 XL model using GN card were interpreted by the Advanced EXPERT system version VT2-R4.01 (bioMérieux, Marcy l'Etoile, France). VITEK 2 used the kinetic method to determine the results, which were readily shown when available, and thus, “time-to-result” (TTR), time from entry into the system until report was generated, was able to be determined. Identification by WalkAway 96 using the Negative Combo panel type 31 was interpreted by the LabPro Information Manager system version 1.51 (Dade Behring). The results for WalkAway 96 were analyzed by the end point method and were set to read at 24 or 40 h according to the manufacturer's recommendation. For this study, the identification results from all systems were considered acceptable if they had a probability of at least 80%, in which the probability of 80% to 94.9% was considered as low discrimination and the probability of 95% or more was considered as high discrimination. 3. Results Both API 20NE and VITEK 2 gave correct identification for B. pseudomallei K92643 (reference strain), although API 20NE yielded a low discrimination of 81.7% probability (data not shown). The VITEK 2 required 6 h to identify B. pseudomallei K92643 with 95% probability. The WalkAway 96 could not identify this reference strain and gave the result as “very rare biotype”. All systems identified the reference strain of B. cepacia correctly with high discrimination. The identification results of B. pseudomallei and B. cepacia clinical isolates using API 20NE, VITEK 2, and WalkAway 96 are shown in Table 1. None of the clinical
P. Kiratisin et al. / Diagnostic Microbiology and Infectious Disease 59 (2007) 277–281
279
Table 1 Identification results of B. pseudomallei and B. cepacia by API 20NE, VITEK 2, and WalkAway 96 a Isolates
API 20NE Identification
B. pseudomallei (n = 56)
B. pseudomallei B. cepacia
VITEK 2 D
% b
H Ld He Lf
27 60 7 3 3
Hg Li Lk L
91 2 2 2 3
Unidentified organism B. cepacia (n = 56)
B. cepacia B. pseudomallei Pseudomonas fluorescens Unidentified organism
WalkAway 96
Identification
D
B. pseudomallei B. cepacia group Klebsiella oxytoca Pseudomonas stutzeri Unidentified organism B. cepacia group B. pseudomallei Burkholderia mallei Unidentified organism
H L H H H H L H L L
% 41 28 16 2 2 11 46 5 20 11 2 16
TTR (range) 6.5 6.3 5.5 3.5 6.7 9.5 5.8 6.4 6.3 6.7 6.7 8.8
(4.0–9.7) (4.7–7.7) (4.7–6.7) (NR) (NR) (9.0–9.7) (4.7–7.0) (5.7–6.7) (4.2–7.5) (5.0–8.0) (NR) (7.7–9.7)
Identification
D
%
B. pseudomallei Unidentified organism
L
c
96 4
B. cepacia
Hh Lj Ll H
21 11 57 2 9
B. pseudomallei Chromobacterium violaceum Unidentified organism
D = discrimination (% probability); H = high discrimination (≥95%); L = low discrimination (80–94.9%); NR = no range. a Identification profiles (percent isolates in the blanket) of API 20NE and WalkAway 96 for isolates identified as B. pseudomallei and B. cepacia are listed in footnotes b through l. b 1155767 (93), 1156576 (7). c 02000776 (100). d 1056577 (59), 1056576 (38), 1016576 (3). e 1056577 (75), 0056576 (25). f 1046577 (100). g 1057577 (68), 0047777 (6), 0057777 (6), 1047777 (6), 1057777 (4), 0046777 (2), 0457737 (2), 0477777 (2), 1046777 (2), 1057576 (2). h 02040772 (67), 02041772 (17), 02061772 (8), 06065772 (8). i 1046577 (100). j 02060776 (50), 02041776 (50). k 1056577 (100). l 02000776 (100).
isolates were identified with lower than 80% probability unless unidentified. The identification profiles of B. pseudomallei versus B. cepacia isolates according to API 20NE and WalkAway 96 systems are shown in the footnotes of Table 1. Among the B. pseudomallei isolates, API 20NE, VITEK 2, and WalkAway 96 gave correct identification for 87%, 69%, and 96%, respectively. Only 27% and 41% of the B. pseudomallei isolates were identified with high discrimination by API 20NE and VITEK 2, respectively. Most of the highly discriminated B. pseudomallei isolates identified by API 20NE were under identification profile 1155767. All B. pseudomallei isolates identified by WalkAway 96 were in low discrimination category under identification profile 02000776 (81.8% probability for B. pseudomallei and 18.2% probability for Ochrobactrum anthropi). B. pseudomallei were commonly misidentified as B. cepacia or B. cepacia group by API 20NE and VITEK 2. The API 20NE, VITEK 2, and WalkAway 96 identified B. cepacia correctly for 93%, 51% , and 32%, respectively. Up to 31% and 57% of B. cepacia isolates were misidentified as B. pseudomallei by VITEK 2 and WalkAway 96, respectively, whereas it was only 2% by API 20NE. Biochemical profiles resulted from VITEK 2 showed several ambiguous reactions between B. pseudomallei and B. cepacia isolates (data not shown). The VITEK 2 also resulted in the highest number of B. pseudomallei and B. cepacia isolates that could not be identified (11% and 16%, respectively). VITEK 2 required the time to produce
the results (TTR) of B. pseudomallei and B. cepacia identification, ranging from 4.0 to 9.7 h as shown in Table 1.
4. Discussion Both B. pseudomallei and B. cepacia cause potentially life-threatening infection. However, the patient populations and clinical features of these organisms are distinct. Melioidosis patients present with a variety of manifestations from localized infection to fatal septicemia. The most important predisposing risk factors of melioidosis include diabetes, alcoholism, and chronic renal disease (Currie, 2003). B. cepacia is known to cause serious complication in cystic fibrosis patients and is identified, along with Pseudomonas aeruginosa, to be the only significant predictor of mortality among these patients (Courtney et al., 2007). Rapid and accurate identification is compelling for proper treatment plan of infection due to these organisms. Therefore, it is of importance that microbiology laboratory correctly identifies and differentiates these 2 pathogens. This study compared the reliability of the manual (API 20NE) and automated (VITEK 2 and WalkAway 96) systems when used to identify clinical isolates of B. pseudomallei and B. cepacia. The rationale was directly due to high similarity of biochemical test results of B. cepacia as compared with B. pseudomallei isolates, which were commonly misidentified by commercial diagnostic systems
280
P. Kiratisin et al. / Diagnostic Microbiology and Infectious Disease 59 (2007) 277–281
(Lowe et al., 2002, 2006). The identification of B. pseudomallei is problematic because available commercial systems show poor sensitivity. This is the highest concern for laboratories with limited experience for this bacterium. In this study, API 20NE correctly identified 87% of B. pseudomallei isolates. However, up to 60% of B. pseudomallei isolates identified by API 20NE were in low discrimination group including the reference strain. It was previously demonstrated that API 20NE yielded as low as 37% accuracy with the potential of false negative (Glass and Popovic, 2005; Inglis et al., 1998, 2005). Because precise and timely identification is critical for melioidosis cases, the API 20NE, which requires up to 48-hour turnaround time, may be less satisfactory as a screening method. Thus, the performance of API 20NE for the identification of B. pseudomallei needs to be justified. The VITEK 2 demonstrated low accuracy (69%) for B. pseudomallei identification. The fact that VITEK 2 identified 20% of B. pseudomallei isolates as other bacteria with high discrimination and could not identify 11% of isolates to any bacterial species contributed to its poor performance. The advantage of VITEK 2 was the speedy identification with the average TTR of 6.5 h, resulting in a faster turnaround time than the other systems. Although WalkAway 96 had the highest accuracy (96%), its performance was unacceptable. This was mainly because of 2 reasons: i) all the B. pseudomallei isolates identified by WalkAway 96 were lowly discriminated with only 81.8% probability (identification profile 02000776), and ii) WalkAway 96 misidentified 57% of the B. cepacia isolates as B. pseudomallei, indicating its disappointing specificity. It should also be noted that WalkAway 96 could not identify the reference strain of B. pseudomallei. The only atypical reaction quoted by WalkAway 96 that resulted in low discrimination of B. pseudomallei was kanamycin resistance. The kanamycin-susceptible strain would result in 99.9% probability of B. pseudomallei (identification profile 02000676). Because this system used susceptibility to kanamycin as a key identification, the result could be deceived. All B. pseudomallei isolates in this study were resistant to aminoglycoside agents including amikacin, gentamicin, tobramycin, and kanamycin. This may be an identification pitfall because B. pseudomallei strains were known to be intrinsically resistant to aminoglycosides (Moore et al., 1999). A large-scale survey in Thailand also showed that most clinical isolates of B. pseudomallei were highly resistant to aminoglycosides (Simpson et al., 1999). Each diagnostic system should take this finding into consideration when using antimicrobial susceptibility to build its identification database. For B. cepacia isolates, API 20NE gave the results of 93% accuracy, and most of them were highly discriminated. VITEK 2 and WalkAway 96, however, resulted in a number of misidentification and demonstrated poor performance of 51% and 32% accuracy, respectively. This insufficient performance should be a concern in clinical practice. The API 20NE and
VITEK 2 misidentified 10% and 16%, respectively, of the B. pseudomallei isolates as B. cepacia, indicating that isolates determined to be B. cepacia by these systems should be validated. In addition, 31% of the B. cepacia isolates were identified as B. pseudomallei by VITEK 2, which could mislead the clinical diagnosis. A number of biochemical reactions resulted from VITEK 2 were difficult to distinguish between B. pseudomallei and B. cepacia, and may account for the high rate of misidentification. The fact that as high as 68% of B. cepacia isolates were misidentified or unidentified by WalkAway 96 raised the concern that its identification database should be reevaluated. Interestingly, B. cepacia isolates were identified with various identification profiles by API 20NE and WalkAway 96 (Table 1). This may reflect a variety of B. cepacia strains recently emerged as various genomovars of B. cepacia complex (Coenye et al., 2001). These genomovars have not been defined in the database of most identification systems and, thus, may result as different identification profiles for B. cepacia. Although several commercial identification systems are widely used in place of conventional methods, it is necessary that a laboratory remains to consider the reliability of these systems. This study suggested that a laboratory should be discreet when obtaining identification results from commercial systems in cases suspected of B. pseudomallei or B. cepacia infection. The API 20NE and WalkAway 96 excelled in classifying the organisms as Burkholderia, although there was a high frequency of misidentification at the genus level with the VITEK 2 system. For identification of B. pseudomallei and B. cepacia, API 20NE generally showed a better performance than both automated systems for species confirmation. However, the results determined by API 20NE was subject to manual reading and time dependent. The WalkAway 96 appeared to be a sensitive system for the screening of B. pseudomallei but had an unacceptable specificity. We recommended that all isolates identified as B. pseudomallei by WalkAway 96 should be confirmed by a more specific method. Isolates that show phenotypic characteristics consistent with B. pseudomallei and are identified not to be B. pseudomallei by VITEK 2 are also recommended for an alternative identification method such as API 20NE. Both VITEK 2 and WalkAway 96 are not sensitive and may result in false identification of B. cepacia and, thus, are not recommended as a definitive identification method for B. cepacia. It is unfortunate that standard method for B. pseudomallei identification is not well defined. The analysis of cellular fatty acid methyl ester derivative by gas liquid chromatography for B. pseudomallei identification was shown to be highly specific (Inglis et al., 2005). However, this method is tedious, requires special equipment, and is not practical in most clinical laboratories. The latex agglutination and PCR are rapid, sensitive, and specific, but these methods remain an in-house production and lack laboratory-wide validation (Pongsunk et al., 1999; Thepthai et al., 2001). Many studies demonstrated that nucleic-based methods are recommended
P. Kiratisin et al. / Diagnostic Microbiology and Infectious Disease 59 (2007) 277–281
for rapid, reliable, and accurate identification of Burkholderia spp., with less biologic hazard risk for technical personnel (Brisse et al., 2002; Lee et al., 2005; Tomaso et al., 2005). Considering that PCR facility is nowadays simplified to adapt for routine testing in a clinical laboratory, we recommend that PCR-based techniques previously evaluated in published studies such as 16S rDNA sequencing (Gee et al., 2003), PCR targeting 16S-23S rRNA internal transcribed spacers (Kunakorn and Markham, 1995), and real-time PCR targeting type III secretion system genes (Thibault et al., 2004) could be used for rapid identification or as a confirmatory method in suspected melioidosis cases. Several nucleic-based systems have also been evaluated for the diagnosis of B. cepacia infection such as PCRrestriction fragment length polymorphism analysis of rDNA operon (van Pelt et al., 1999) and specific PCR-targeting recA gene (Mahenthiralingam et al., 2000). Because the taxonomy status of the genus Burkholderia is persistently evolved, it is mandated that diagnostic techniques and identification databases of commercial systems need continuing improvement. Acknowledgments The authors are grateful to Prof. Stitaya Sirisinha, Mahidol University, Bangkok, Thailand, and the Thailand Research Fund for supplying Burkholderia strains from provincial hospitals. They thank bioMérieux, Thailand, and Dade Behring, Thailand, for supplying reagents for this study. They also thank Dr. Patrick Murray (National Institutes of Health Clinical Center, Maryland) for helpful comments on the manuscript. References Anuntagool N, Naigowit P, Petkanchanapong V, Aramsri P, Panichakul T, Sirisinha S (2000) Monoclonal antibody-based rapid identification of Burkholderia pseudomallei in blood culture fluid from patients with community-acquired septicaemia. J Med Microbiol 49:1075–1078. Brisse S, Stefani S, Verhoef J, Van Belkum A, Vandamme P, Goessens W (2002) Comparative evaluation of the BD Phoenix and VITEK 2 automated instruments for identification of isolates of the Burkholderia cepacia complex. J Clin Microbiol 40:1743–1748. Cheng AC, Currie BJ (2005) Melioidosis: epidemiology, pathophysiology, and management. Clin Microbiol Rev 18:383–416. Coenye T, Vandamme P, Govan JRW, LiPuma JJ (2001) Taxonomy and identification of the Burkholderia cepacia complex. J Clin Microbiol 39:3427–3436. Courtney JM, Bradley J, McCaughan J, O'Connor TM, Shortt C, Bredin CP, et al (2007) Predictors of mortality in adults with cystic fibrosis. Pediatr Pulmonol 42:525–532. Currie BJ (2003) Melioidosis: an important cause of pneumonia in residents of and travellers returned from endemic regions. Eur Respir J 22:542–550. Dance DA, Wuthiekanun P, Naigowit P, White NJ (1989) Identification of Pseudomonas pseudomallei in clinical practice: use of simple screening tests and API 20NE. J Clin Pathol 42:645–648. Funke G, Funke-Kissling P (2004) Evaluation of the new VITEK 2 card for identification of clinically relevant gram-negative rods. J Clin Microbiol 42:4067–4071.
281
Gee JE, Sacchi CT, Glass MB, De BK, Weyant RS, Levett PN, et al (2003) Use of 16S rRNA gene sequencing for rapid identification and differentiation of Burkholderia pseudomallei and B. mallei. J Clin Microbiol 41:4647–4654. Glass MB, Popovic T (2005) Preliminary evaluation of the API 20NE and RapID NF plus systems for the identification of Burkholderia pseudomallei and B. mallei. J Clin Microbiol 43:479–483. Holden MT, Titball RW, Peacock SJ, Cerdeno-Tarraga AM, Atkins T, Crossman LC, et al (2004) Genomic plasticity of the causative agent of melioidosis, Burkholderia pseudomallei. Proc Natl Acad Sci U S A 101:14240–14245. Inglis TJ, Chiang D, Lee GS, Chor-Kiang L (1998) Potential misidentification of Burkholderia pseudomallei by API 20NE. Pathology 30:62–64. Inglis TJ, Merritt A, Chidlow G, Aravena-Roman M, Harnett G (2005) Comparison of diagnostic laboratory methods for identification of Burkholderia pseudomallei. J Clin Microbiol 43:2201–2206. Kunakorn M, Markham RB (1995) Clinically practical seminested PCR for Burkholderia pseudomallei quantitated by enzyme immunoassay with and without solution hybridization. J Clin Microbiol 33:2131–2135. Lee MA, Wang D, Yap EH (2005) Detection and differentiation of Burkholderia pseudomallei, Burkholderia mallei and Burkholderia thailandensis by multiplex PCR. FEMS Immunol Med Microbiol 43:413–417. Lowe P, Engler C, Norton R (2002) Comparison of automated and nonautomated systems for identification of Burkholderia pseudomallei. J Clin Microbiol 42:5871–5874. Lowe P, Haswell H, Lewis K (2006) Use of various common isolation media to evaluate the new VITEK 2 calorimetric GN card for identification of Burkholderia pseudomallei. J Clin Microbiol 44:854–856. Mahenthiralingam E, Bischof J, Byrne SK, Radomski C, Davies JE, Av-Gay Y, Vandamme P (2000) DNA-based diagnostic approaches for identification of Burkholderia cepacia complex, Burkholderia vietnamiensis, Burkholderia multivorans, Burkholderia stabilis, and Burkholderia cepacia genomovars I and III. J Clin Microbiol 38:3165–3173. Moore RA, DeShazer D, Reckseidler S, Weissman A, Woods DE (1999) Efflux-mediated aminoglycoside and macrolide resistance in Burkholderia pseudomallei. Antimicrob Agents Chemother 43:465–470. Pongsunk S, Thirawattanasuk N, Piyasangthong N, Ekpo P (1999) Rapid identification of Burkholderia pseudomallei in blood cultures by a monoclonal antibody assay. J Clin Microbiol 37:3662–3667. Rotz LD, Khan AS, Lillibridge SR, Ostroff SM, Hughes JM (2002) Public health assessment of potential biological terrorism agents. Emerg Infect Dis 8:225–230. Simpson AJH, White NJ, Wuthiekanun V (1999) Aminoglycoside and macrolide resistance in Burkholderia pseudomallei. Antimicrob Agents Chemother 43:2332. Thepthai C, Dharakul T, Smithikarn S, Trakulsomboon S, Songsivilai S (2001) Differentiation between non-virulent and virulent Burkholderia pseudomallei with monoclonal antibodies to the Ara+ and Ara− biotypes. Am J Trop Med Hyg 65:10–12. Thibault FM, Valade E, Vidal DR (2004) Identification and discrimination of Burkholderia pseudomallei, B. mallei, and B. thailandensis by real-time PCR targeting type III secretion system genes. J Clin Microbiol 42:5871–5874. Tomaso H, Pitt TL, Landt O, Al Dahouk S, Scholz HC, Reisinger EC, et al (2005) Rapid presumptive identification of Burkholderia pseudomallei with real-time PCR assays using fluorescent hybridization probes. Mol Cell Probes 19:9–20. van Pelt C, Verduin CM, Goessens WHF, Vos MC, Tümmler B, Segonds C, et al (1999) Identification of Burkholderia spp. in the clinical microbiology laboratory: comparison of conventional and molecular methods. J Clin Microbiol 37:2158–2164. Weyant RS, Moss CW, Weaver RE, Hollis DG, Jordan JG, Cook EC, Daneshvar MI (1996) Identification of unusual pathogenic gramnegative aerobic and facultatively anaerobic bacteria. 2nd ed. Baltimore (MD): Williams & Wilkins.
Diagnostic Microbiology and Infectious Disease 59 (2007) 277 – 281 www.elsevier.com/locate/diagmicrobio
Accuracy of commercial systems for identification of Burkholderia pseudomallei versus Burkholderia cepacia Pattarachai Kiratisin a,⁎, Pitak Santanirand b , Narisara Chantratita c , Srirumpa Kaewdaeng a b
a Department of Microbiology, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok 10700, Thailand Department of Pathology, Faculty of Medicine Ramathibodi Hospital, Mahidol University, Bangkok 10400, Thailand c Faculty of Tropical Medicine, Mahidol University, Bangkok 10400, Thailand Received 14 February 2007; accepted 20 June 2007
Abstract Infection caused by Burkholderia pseudomallei or Burkholderia cepacia may result in fatal outcome unless the causative agent is accurately identified in a short period, which is critical for treatment. We evaluated the reliability of the commonly used commercial systems, API 20NE, VITEK 2, and WalkAway 96, for comparative identification of B. pseudomallei versus B. cepacia clinical isolates. Based on biochemical and molecular tests as reference methods, API 20NE was probably the most reliable, with an accuracy of 87% and 93%, for the identification of B. pseudomallei and B. cepacia, respectively. The VITEK 2 and WalkAway 96 systems resulted in a number of misidentification and, thus, were less reliable. The performance of each system and identification guidelines for B. pseudomallei and B. cepacia are discussed. Our study emphasized that laboratories should carefully interpret the identification of B. pseudomallei and B. cepacia when using commercial systems. © 2007 Elsevier Inc. All rights reserved. Keywords: Burkholderia pseudomallei; Burkholderia cepacia; Identification
1. Introduction Rapid and accurate identification of Burkholderia pseudomallei, the causative agent of melioidosis, is critical for clinical laboratory, especially for those serving in the endemic areas such as Southeast Asia and northern Australia. A delayed or incorrect identification of this bacterium often results in an unfavorable or fatal outcome due to fulminant septicemia (Cheng and Currie, 2005). Because of the increase in international travel, laboratories in the nonendemic area that have limited experience for B. pseudomallei identification and mostly rely on a commercial method as 1st-line diagnostic tool should be aware of the accountability of such method. In addition, B. pseudomallei is enlisted as a potential bioterrorism agent that clinical laboratories and the public health should recognize (Rotz et al., 2002). The recommended conventional approach based on phenotypic
⁎ Corresponding author. Tel.: +66-2-419-7058; fax: +66-2-411-3106. E-mail address: [email protected] (P. Kiratisin). 0732-8893/$ – see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.diagmicrobio.2007.06.013
characteristics (Dance et al., 1989) may not be sufficient to yield an opportunely definitive identification. Several diagnostic systems are currently available in the market, and yet, there are few evaluation studies regarding the performance of these systems for identification of Burkholderia spp. A widely used substrate utilization panel API 20NE (bioMérieux, Hazelwood, MO) was shown to correctly identify B. pseudomallei ranging from 37% to 98% in various studies (Dance et al., 1989; Inglis et al., 1998; Lowe et al., 2002; Glass and Popovic, 2005). This inconsistence resulted in a disputable competence of API 20NE. Automation systems have been accustomed to many microbiology laboratories. However, the reliability of these systems for B. pseudomallei identification remains doubtful. To our knowledge, only the VITEK systems (bioMérieux, Marcy l'Etoile, France) have been evaluated in a few English literatures. Although the VITEK 1 was shown to correctly identify 99% of B. pseudomallei isolates (Lowe et al., 2002), the accuracy of VITEK 2 was controversial. The fluorometric-based ID-GNB card of the VITEK 2 demonstrated a very poor performance for the identification
278
P. Kiratisin et al. / Diagnostic Microbiology and Infectious Disease 59 (2007) 277–281
of B. pseudomallei with more than 80% of misidentification (Lowe et al., 2002). With the new colorimetric-based GN card, the VITEK 2 identification database for Gram-negative bacteria has been expanded (Funke and Funke-Kissling, 2004). Nonetheless, the recent study showed that the VITEK 2 GN card correctly identified B. pseudomallei varying from 63% to 81%, depending on the culture media used for growing the bacterium (Lowe et al., 2006). The MicroScan WalkAway 96 SI system (Dade Behring, West Sacramento, CA) is also widely used in diagnostic laboratories and, yet, has never been evaluated for B. pseudomallei identification. In addition, it was suggested that the geographic diversity of strains may affect the credibility of diagnostic systems (Glass and Popovic, 2005). B. pseudomallei isolates from the highly endemic area such as Thailand were rarely included in the previous published studies. In this study, we hence evaluated the nonautomated (API 20NE) and automated (VITEK 2 and WalkAway 96) systems for the identification of B. pseudomallei recovered from Thai patients. The previous studies also showed that the VITEK 2 most commonly misidentified B. pseudomallei as Burkholderia cepacia or B. cepacia group (Lowe et al., 2002, 2006). Therefore, the clinical isolates of B. cepacia were included to compare each system in parallel for its capability to distinguish these 2 Burkholderia spp. Both B. pseudomallei and B. cepacia are important respiratory tract pathogens and may cause fatal septicemia. The accurate differentiation between these 2 species is thus a requisite for proper diagnosis and treatment. This study provides a detailed analysis of 3 commercial systems regarding the reliability of B. pseudomallei and B. cepacia identification. 2. Materials and methods 2.1. Bacterial strains and identification The clinical isolates of B. pseudomallei (n = 56) and B. cepacia (n = 56) used in this study were recovered from various specimen types including blood, sputum, and pus from individual patients admitted to Siriraj hospital, Bangkok, Thailand, and 4 provincial hospitals in northeast Thailand (Ubon Rachathani, Srisaket, Surin, and Khon Kaen). The isolates were collected during the years 1997, 2004, and 2005. The clinical strain K92643 of B. pseudomallei used for genome sequencing (Holden et al., 2004) and B. cepacia strain ATCC 25416 were used as reference strains. All isolates were stored at −80 °C in 20% glycerol brain heart infusion broth and were subcultured on 5% (vol/vol) defribrinated sheep blood agar plates for at least 2 passages before use. All potential aerosol-generating procedures were performed in a class II biologic safety cabinet by gowned and gloved personnel. Bacterial strains were originally identified based on the standard biochemical tests (Weyant et al., 1996). The identity of all B. pseudomallei isolates were also confirmed by 2 additional tests that were previously described: specific polymerase
chain reaction (PCR) targeting the DNA sequences of the internal transcribed spacers between 16S and 23S rDNA genes (Kunakorn and Markham, 1995; Inglis et al., 2005), and monoclonal antibody-based latex agglutination (Anuntagool et al., 2000). All B. pseudomallei isolates in this study including reference strain yielded positive results for both PCR and latex agglutination, whereas all B. cepacia isolates had negative results for both tests. The identity of all B. cepacia isolates were also confirmed using recA-based PCR specific for B. cepacia complex as described previously (Mahenthiralingam et al., 2000). 2.2. API 20NE, VITEK 2, and WalkAway 96 identification The bacteria were streaked on sheep blood agar plates and incubated overnight at 37 °C. Colonies of each isolate were used to prepare a suspension for identification by API 20NE, VITEK 2, and WalkAway 96. The identification procedures were performed according to the manufacturers' protocols. The API 20NE strips were incubated for 24 or 48 h depending on the reactions and were read manually using the manufacturer's interpretative chart. The API 20NE results were read and analyzed by 1 investigator using the apiweb program available via internet access (http://apiweb.biomerieux.com). The identification results by VITEK 2 carried out by VITEK 2 XL model using GN card were interpreted by the Advanced EXPERT system version VT2-R4.01 (bioMérieux, Marcy l'Etoile, France). VITEK 2 used the kinetic method to determine the results, which were readily shown when available, and thus, “time-to-result” (TTR), time from entry into the system until report was generated, was able to be determined. Identification by WalkAway 96 using the Negative Combo panel type 31 was interpreted by the LabPro Information Manager system version 1.51 (Dade Behring). The results for WalkAway 96 were analyzed by the end point method and were set to read at 24 or 40 h according to the manufacturer's recommendation. For this study, the identification results from all systems were considered acceptable if they had a probability of at least 80%, in which the probability of 80% to 94.9% was considered as low discrimination and the probability of 95% or more was considered as high discrimination. 3. Results Both API 20NE and VITEK 2 gave correct identification for B. pseudomallei K92643 (reference strain), although API 20NE yielded a low discrimination of 81.7% probability (data not shown). The VITEK 2 required 6 h to identify B. pseudomallei K92643 with 95% probability. The WalkAway 96 could not identify this reference strain and gave the result as “very rare biotype”. All systems identified the reference strain of B. cepacia correctly with high discrimination. The identification results of B. pseudomallei and B. cepacia clinical isolates using API 20NE, VITEK 2, and WalkAway 96 are shown in Table 1. None of the clinical
P. Kiratisin et al. / Diagnostic Microbiology and Infectious Disease 59 (2007) 277–281
279
Table 1 Identification results of B. pseudomallei and B. cepacia by API 20NE, VITEK 2, and WalkAway 96 a Isolates
API 20NE Identification
B. pseudomallei (n = 56)
B. pseudomallei B. cepacia
VITEK 2 D
% b
H Ld He Lf
27 60 7 3 3
Hg Li Lk L
91 2 2 2 3
Unidentified organism B. cepacia (n = 56)
B. cepacia B. pseudomallei Pseudomonas fluorescens Unidentified organism
WalkAway 96
Identification
D
B. pseudomallei B. cepacia group Klebsiella oxytoca Pseudomonas stutzeri Unidentified organism B. cepacia group B. pseudomallei Burkholderia mallei Unidentified organism
H L H H H H L H L L
% 41 28 16 2 2 11 46 5 20 11 2 16
TTR (range) 6.5 6.3 5.5 3.5 6.7 9.5 5.8 6.4 6.3 6.7 6.7 8.8
(4.0–9.7) (4.7–7.7) (4.7–6.7) (NR) (NR) (9.0–9.7) (4.7–7.0) (5.7–6.7) (4.2–7.5) (5.0–8.0) (NR) (7.7–9.7)
Identification
D
%
B. pseudomallei Unidentified organism
L
c
96 4
B. cepacia
Hh Lj Ll H
21 11 57 2 9
B. pseudomallei Chromobacterium violaceum Unidentified organism
D = discrimination (% probability); H = high discrimination (≥95%); L = low discrimination (80–94.9%); NR = no range. a Identification profiles (percent isolates in the blanket) of API 20NE and WalkAway 96 for isolates identified as B. pseudomallei and B. cepacia are listed in footnotes b through l. b 1155767 (93), 1156576 (7). c 02000776 (100). d 1056577 (59), 1056576 (38), 1016576 (3). e 1056577 (75), 0056576 (25). f 1046577 (100). g 1057577 (68), 0047777 (6), 0057777 (6), 1047777 (6), 1057777 (4), 0046777 (2), 0457737 (2), 0477777 (2), 1046777 (2), 1057576 (2). h 02040772 (67), 02041772 (17), 02061772 (8), 06065772 (8). i 1046577 (100). j 02060776 (50), 02041776 (50). k 1056577 (100). l 02000776 (100).
isolates were identified with lower than 80% probability unless unidentified. The identification profiles of B. pseudomallei versus B. cepacia isolates according to API 20NE and WalkAway 96 systems are shown in the footnotes of Table 1. Among the B. pseudomallei isolates, API 20NE, VITEK 2, and WalkAway 96 gave correct identification for 87%, 69%, and 96%, respectively. Only 27% and 41% of the B. pseudomallei isolates were identified with high discrimination by API 20NE and VITEK 2, respectively. Most of the highly discriminated B. pseudomallei isolates identified by API 20NE were under identification profile 1155767. All B. pseudomallei isolates identified by WalkAway 96 were in low discrimination category under identification profile 02000776 (81.8% probability for B. pseudomallei and 18.2% probability for Ochrobactrum anthropi). B. pseudomallei were commonly misidentified as B. cepacia or B. cepacia group by API 20NE and VITEK 2. The API 20NE, VITEK 2, and WalkAway 96 identified B. cepacia correctly for 93%, 51% , and 32%, respectively. Up to 31% and 57% of B. cepacia isolates were misidentified as B. pseudomallei by VITEK 2 and WalkAway 96, respectively, whereas it was only 2% by API 20NE. Biochemical profiles resulted from VITEK 2 showed several ambiguous reactions between B. pseudomallei and B. cepacia isolates (data not shown). The VITEK 2 also resulted in the highest number of B. pseudomallei and B. cepacia isolates that could not be identified (11% and 16%, respectively). VITEK 2 required the time to produce
the results (TTR) of B. pseudomallei and B. cepacia identification, ranging from 4.0 to 9.7 h as shown in Table 1.
4. Discussion Both B. pseudomallei and B. cepacia cause potentially life-threatening infection. However, the patient populations and clinical features of these organisms are distinct. Melioidosis patients present with a variety of manifestations from localized infection to fatal septicemia. The most important predisposing risk factors of melioidosis include diabetes, alcoholism, and chronic renal disease (Currie, 2003). B. cepacia is known to cause serious complication in cystic fibrosis patients and is identified, along with Pseudomonas aeruginosa, to be the only significant predictor of mortality among these patients (Courtney et al., 2007). Rapid and accurate identification is compelling for proper treatment plan of infection due to these organisms. Therefore, it is of importance that microbiology laboratory correctly identifies and differentiates these 2 pathogens. This study compared the reliability of the manual (API 20NE) and automated (VITEK 2 and WalkAway 96) systems when used to identify clinical isolates of B. pseudomallei and B. cepacia. The rationale was directly due to high similarity of biochemical test results of B. cepacia as compared with B. pseudomallei isolates, which were commonly misidentified by commercial diagnostic systems
280
P. Kiratisin et al. / Diagnostic Microbiology and Infectious Disease 59 (2007) 277–281
(Lowe et al., 2002, 2006). The identification of B. pseudomallei is problematic because available commercial systems show poor sensitivity. This is the highest concern for laboratories with limited experience for this bacterium. In this study, API 20NE correctly identified 87% of B. pseudomallei isolates. However, up to 60% of B. pseudomallei isolates identified by API 20NE were in low discrimination group including the reference strain. It was previously demonstrated that API 20NE yielded as low as 37% accuracy with the potential of false negative (Glass and Popovic, 2005; Inglis et al., 1998, 2005). Because precise and timely identification is critical for melioidosis cases, the API 20NE, which requires up to 48-hour turnaround time, may be less satisfactory as a screening method. Thus, the performance of API 20NE for the identification of B. pseudomallei needs to be justified. The VITEK 2 demonstrated low accuracy (69%) for B. pseudomallei identification. The fact that VITEK 2 identified 20% of B. pseudomallei isolates as other bacteria with high discrimination and could not identify 11% of isolates to any bacterial species contributed to its poor performance. The advantage of VITEK 2 was the speedy identification with the average TTR of 6.5 h, resulting in a faster turnaround time than the other systems. Although WalkAway 96 had the highest accuracy (96%), its performance was unacceptable. This was mainly because of 2 reasons: i) all the B. pseudomallei isolates identified by WalkAway 96 were lowly discriminated with only 81.8% probability (identification profile 02000776), and ii) WalkAway 96 misidentified 57% of the B. cepacia isolates as B. pseudomallei, indicating its disappointing specificity. It should also be noted that WalkAway 96 could not identify the reference strain of B. pseudomallei. The only atypical reaction quoted by WalkAway 96 that resulted in low discrimination of B. pseudomallei was kanamycin resistance. The kanamycin-susceptible strain would result in 99.9% probability of B. pseudomallei (identification profile 02000676). Because this system used susceptibility to kanamycin as a key identification, the result could be deceived. All B. pseudomallei isolates in this study were resistant to aminoglycoside agents including amikacin, gentamicin, tobramycin, and kanamycin. This may be an identification pitfall because B. pseudomallei strains were known to be intrinsically resistant to aminoglycosides (Moore et al., 1999). A large-scale survey in Thailand also showed that most clinical isolates of B. pseudomallei were highly resistant to aminoglycosides (Simpson et al., 1999). Each diagnostic system should take this finding into consideration when using antimicrobial susceptibility to build its identification database. For B. cepacia isolates, API 20NE gave the results of 93% accuracy, and most of them were highly discriminated. VITEK 2 and WalkAway 96, however, resulted in a number of misidentification and demonstrated poor performance of 51% and 32% accuracy, respectively. This insufficient performance should be a concern in clinical practice. The API 20NE and
VITEK 2 misidentified 10% and 16%, respectively, of the B. pseudomallei isolates as B. cepacia, indicating that isolates determined to be B. cepacia by these systems should be validated. In addition, 31% of the B. cepacia isolates were identified as B. pseudomallei by VITEK 2, which could mislead the clinical diagnosis. A number of biochemical reactions resulted from VITEK 2 were difficult to distinguish between B. pseudomallei and B. cepacia, and may account for the high rate of misidentification. The fact that as high as 68% of B. cepacia isolates were misidentified or unidentified by WalkAway 96 raised the concern that its identification database should be reevaluated. Interestingly, B. cepacia isolates were identified with various identification profiles by API 20NE and WalkAway 96 (Table 1). This may reflect a variety of B. cepacia strains recently emerged as various genomovars of B. cepacia complex (Coenye et al., 2001). These genomovars have not been defined in the database of most identification systems and, thus, may result as different identification profiles for B. cepacia. Although several commercial identification systems are widely used in place of conventional methods, it is necessary that a laboratory remains to consider the reliability of these systems. This study suggested that a laboratory should be discreet when obtaining identification results from commercial systems in cases suspected of B. pseudomallei or B. cepacia infection. The API 20NE and WalkAway 96 excelled in classifying the organisms as Burkholderia, although there was a high frequency of misidentification at the genus level with the VITEK 2 system. For identification of B. pseudomallei and B. cepacia, API 20NE generally showed a better performance than both automated systems for species confirmation. However, the results determined by API 20NE was subject to manual reading and time dependent. The WalkAway 96 appeared to be a sensitive system for the screening of B. pseudomallei but had an unacceptable specificity. We recommended that all isolates identified as B. pseudomallei by WalkAway 96 should be confirmed by a more specific method. Isolates that show phenotypic characteristics consistent with B. pseudomallei and are identified not to be B. pseudomallei by VITEK 2 are also recommended for an alternative identification method such as API 20NE. Both VITEK 2 and WalkAway 96 are not sensitive and may result in false identification of B. cepacia and, thus, are not recommended as a definitive identification method for B. cepacia. It is unfortunate that standard method for B. pseudomallei identification is not well defined. The analysis of cellular fatty acid methyl ester derivative by gas liquid chromatography for B. pseudomallei identification was shown to be highly specific (Inglis et al., 2005). However, this method is tedious, requires special equipment, and is not practical in most clinical laboratories. The latex agglutination and PCR are rapid, sensitive, and specific, but these methods remain an in-house production and lack laboratory-wide validation (Pongsunk et al., 1999; Thepthai et al., 2001). Many studies demonstrated that nucleic-based methods are recommended
P. Kiratisin et al. / Diagnostic Microbiology and Infectious Disease 59 (2007) 277–281
for rapid, reliable, and accurate identification of Burkholderia spp., with less biologic hazard risk for technical personnel (Brisse et al., 2002; Lee et al., 2005; Tomaso et al., 2005). Considering that PCR facility is nowadays simplified to adapt for routine testing in a clinical laboratory, we recommend that PCR-based techniques previously evaluated in published studies such as 16S rDNA sequencing (Gee et al., 2003), PCR targeting 16S-23S rRNA internal transcribed spacers (Kunakorn and Markham, 1995), and real-time PCR targeting type III secretion system genes (Thibault et al., 2004) could be used for rapid identification or as a confirmatory method in suspected melioidosis cases. Several nucleic-based systems have also been evaluated for the diagnosis of B. cepacia infection such as PCRrestriction fragment length polymorphism analysis of rDNA operon (van Pelt et al., 1999) and specific PCR-targeting recA gene (Mahenthiralingam et al., 2000). Because the taxonomy status of the genus Burkholderia is persistently evolved, it is mandated that diagnostic techniques and identification databases of commercial systems need continuing improvement. Acknowledgments The authors are grateful to Prof. Stitaya Sirisinha, Mahidol University, Bangkok, Thailand, and the Thailand Research Fund for supplying Burkholderia strains from provincial hospitals. They thank bioMérieux, Thailand, and Dade Behring, Thailand, for supplying reagents for this study. They also thank Dr. Patrick Murray (National Institutes of Health Clinical Center, Maryland) for helpful comments on the manuscript. References Anuntagool N, Naigowit P, Petkanchanapong V, Aramsri P, Panichakul T, Sirisinha S (2000) Monoclonal antibody-based rapid identification of Burkholderia pseudomallei in blood culture fluid from patients with community-acquired septicaemia. J Med Microbiol 49:1075–1078. Brisse S, Stefani S, Verhoef J, Van Belkum A, Vandamme P, Goessens W (2002) Comparative evaluation of the BD Phoenix and VITEK 2 automated instruments for identification of isolates of the Burkholderia cepacia complex. J Clin Microbiol 40:1743–1748. Cheng AC, Currie BJ (2005) Melioidosis: epidemiology, pathophysiology, and management. Clin Microbiol Rev 18:383–416. Coenye T, Vandamme P, Govan JRW, LiPuma JJ (2001) Taxonomy and identification of the Burkholderia cepacia complex. J Clin Microbiol 39:3427–3436. Courtney JM, Bradley J, McCaughan J, O'Connor TM, Shortt C, Bredin CP, et al (2007) Predictors of mortality in adults with cystic fibrosis. Pediatr Pulmonol 42:525–532. Currie BJ (2003) Melioidosis: an important cause of pneumonia in residents of and travellers returned from endemic regions. Eur Respir J 22:542–550. Dance DA, Wuthiekanun P, Naigowit P, White NJ (1989) Identification of Pseudomonas pseudomallei in clinical practice: use of simple screening tests and API 20NE. J Clin Pathol 42:645–648. Funke G, Funke-Kissling P (2004) Evaluation of the new VITEK 2 card for identification of clinically relevant gram-negative rods. J Clin Microbiol 42:4067–4071.
281
Gee JE, Sacchi CT, Glass MB, De BK, Weyant RS, Levett PN, et al (2003) Use of 16S rRNA gene sequencing for rapid identification and differentiation of Burkholderia pseudomallei and B. mallei. J Clin Microbiol 41:4647–4654. Glass MB, Popovic T (2005) Preliminary evaluation of the API 20NE and RapID NF plus systems for the identification of Burkholderia pseudomallei and B. mallei. J Clin Microbiol 43:479–483. Holden MT, Titball RW, Peacock SJ, Cerdeno-Tarraga AM, Atkins T, Crossman LC, et al (2004) Genomic plasticity of the causative agent of melioidosis, Burkholderia pseudomallei. Proc Natl Acad Sci U S A 101:14240–14245. Inglis TJ, Chiang D, Lee GS, Chor-Kiang L (1998) Potential misidentification of Burkholderia pseudomallei by API 20NE. Pathology 30:62–64. Inglis TJ, Merritt A, Chidlow G, Aravena-Roman M, Harnett G (2005) Comparison of diagnostic laboratory methods for identification of Burkholderia pseudomallei. J Clin Microbiol 43:2201–2206. Kunakorn M, Markham RB (1995) Clinically practical seminested PCR for Burkholderia pseudomallei quantitated by enzyme immunoassay with and without solution hybridization. J Clin Microbiol 33:2131–2135. Lee MA, Wang D, Yap EH (2005) Detection and differentiation of Burkholderia pseudomallei, Burkholderia mallei and Burkholderia thailandensis by multiplex PCR. FEMS Immunol Med Microbiol 43:413–417. Lowe P, Engler C, Norton R (2002) Comparison of automated and nonautomated systems for identification of Burkholderia pseudomallei. J Clin Microbiol 42:5871–5874. Lowe P, Haswell H, Lewis K (2006) Use of various common isolation media to evaluate the new VITEK 2 calorimetric GN card for identification of Burkholderia pseudomallei. J Clin Microbiol 44:854–856. Mahenthiralingam E, Bischof J, Byrne SK, Radomski C, Davies JE, Av-Gay Y, Vandamme P (2000) DNA-based diagnostic approaches for identification of Burkholderia cepacia complex, Burkholderia vietnamiensis, Burkholderia multivorans, Burkholderia stabilis, and Burkholderia cepacia genomovars I and III. J Clin Microbiol 38:3165–3173. Moore RA, DeShazer D, Reckseidler S, Weissman A, Woods DE (1999) Efflux-mediated aminoglycoside and macrolide resistance in Burkholderia pseudomallei. Antimicrob Agents Chemother 43:465–470. Pongsunk S, Thirawattanasuk N, Piyasangthong N, Ekpo P (1999) Rapid identification of Burkholderia pseudomallei in blood cultures by a monoclonal antibody assay. J Clin Microbiol 37:3662–3667. Rotz LD, Khan AS, Lillibridge SR, Ostroff SM, Hughes JM (2002) Public health assessment of potential biological terrorism agents. Emerg Infect Dis 8:225–230. Simpson AJH, White NJ, Wuthiekanun V (1999) Aminoglycoside and macrolide resistance in Burkholderia pseudomallei. Antimicrob Agents Chemother 43:2332. Thepthai C, Dharakul T, Smithikarn S, Trakulsomboon S, Songsivilai S (2001) Differentiation between non-virulent and virulent Burkholderia pseudomallei with monoclonal antibodies to the Ara+ and Ara− biotypes. Am J Trop Med Hyg 65:10–12. Thibault FM, Valade E, Vidal DR (2004) Identification and discrimination of Burkholderia pseudomallei, B. mallei, and B. thailandensis by real-time PCR targeting type III secretion system genes. J Clin Microbiol 42:5871–5874. Tomaso H, Pitt TL, Landt O, Al Dahouk S, Scholz HC, Reisinger EC, et al (2005) Rapid presumptive identification of Burkholderia pseudomallei with real-time PCR assays using fluorescent hybridization probes. Mol Cell Probes 19:9–20. van Pelt C, Verduin CM, Goessens WHF, Vos MC, Tümmler B, Segonds C, et al (1999) Identification of Burkholderia spp. in the clinical microbiology laboratory: comparison of conventional and molecular methods. J Clin Microbiol 37:2158–2164. Weyant RS, Moss CW, Weaver RE, Hollis DG, Jordan JG, Cook EC, Daneshvar MI (1996) Identification of unusual pathogenic gramnegative aerobic and facultatively anaerobic bacteria. 2nd ed. Baltimore (MD): Williams & Wilkins.