Detection of Burkholderia pseudomallei in Sputum using Selective Enrichment Broth and Ashdown’s Medium at Kampong Cham Provincial Hospital, Cambodia

Melioidosis, infection caused by Burkholderia pseudomallei, is increasingly reported in Cambodia. We hypothesized that implementation of an enhanced sputum testing protocol in a provincial hospital diagnostic microbiology laboratory would increase detection of B. pseudomallei. We tested 241 sputum specimens that were deemed acceptable for culture, comparing culture in selective enrichment broth followed by sub-culture on Ashdown’s medium to standard culture methods. Two specimens (0.8%) were positive for B. pseudomallei using the enhanced protocol whereas one specimen (0.4%) was positive using standard methods. Given the low numbers of positive specimens, we could not conclusively determine the utility of the enhanced sputum testing protocol. However, the ramifications of identification of B. pseudomallei are substantial, and the benefit of the enhanced testing protocol may be more apparent in patients selected based on risk factors and clinical presentation. Promoting clinician awareness of the infection and encouraging utilization of diagnostic microbiology services are also likely to be important factors in facilitating identification of melioidosis.

Melioidosis, infection with the Gram negative bacterium Burkholderia pseudomallei, has increasingly been described in Cambodia in recent years [1][2][3][4][5][6] . Infection is caused by inoculation, inhalation, or ingestion of B. pseudomallei, an environmental saprophyte 7 . In one single center Cambodian study, melioidosis (defined as growth of B. pseudomallei in any clinical specimen) was lethal in more than half of cases, many of whom had received inappropriate initial antibiotic therapy 2 . The clinical presentation of melioidosis is notoriously variable, but pneumonia and sepsis occur commonly 7 . Sub-acute and chronic forms of melioidosis may be confused with tuberculosis and it is important to distinguish these diseases 8 . B. pseudomallei is resistant to penicillin, ampicillin, first-and second-generation cephalosporins and aminoglycosides 7 . Use of quinolones or third-generation cephalosporins to treat melioidosis is associated with poorer outcomes 9-11 . Thus, many of the first line antimicrobial therapies administered for pneumonia and sepsis do not adequately treat melioidosis. The treatment course for melioidosis is prolonged, requiring initial intensive therapy for at least 10-14 days with intravenous ceftazidime, imipenem, or meropenem, followed by 3-6 months of oral eradication therapy with trimethoprim-sulfamethoxazole 7 . Moreover, ceftazidime and carbapenems are relatively expensive and may not be universally accessible in Cambodia. Therefore, rapid diagnosis of the infection is essential to guide clinical management. Culture is the gold standard diagnostic test for melioidosis and recovery of B. pseudomallei from any sample constitutes infection. In pulmonary melioidosis, culture of sputum may provide critical diagnostic information. However, in culturing B. pseudomallei from non-sterile samples such as sputum, it is important to minimize overgrowth by commensal or contaminating bacteria by using selective media. In this report, we present our experience implementing an enhanced sputum testing protocol in a Cambodian provincial hospital diagnostic microbiology laboratory to increase detection of B. pseudomallei.

Methods
To compare the use of selective enrichment broth and Ashdown's medium to standard sputum culture at Kampong Cham Provincial Hospital, we performed specific testing for B. pseudomallei on all sputum specimens submitted to the laboratory and accepted for bacterial culture between March 25 and September 30, 2013, following a previously described protocol 12,14 . We chose these dates to include much of the rainy season, when melioidosis is diagnosed more frequently 2 . No data were available on the subjects who provided the sputum specimens although most specimens were submitted from the hospital's tuberculosis screening area or ward. Specimens accepted for culture had fewer than 10 epithelial cells per low power field or moderate to high numbers of polymorphonuclear cells on Gram stain. Specimens were inoculated into selective enrichment broth containing Tryptic Soy Broth 10g, glycerol 40ml, crystal violet 0.1% 5ml, and colistin 50mg per liter of distilled water for two days at 37°C in aerobic conditions. Specimens were then sub-cultured onto Ashdown's medium (Tryptic Soy Broth 10g, agar 15g, glycerol 40ml, crystal violet 0.1% 5ml, neutral red 1% 5ml, and gentamicin 4mg per liter of distilled water) and incubated for four days at 37°C in aerobic conditions. Colonies that grew were tested using the oxidase test; oxidase-positive colonies were tested using a highly sensitive and specific monoclonal antibody-based latex agglutination test for B. pseudomallei 15,16 . In parallel, as per routine laboratory practice, all sputum specimens were also cultured onto sheep blood, chocolate, and MacConkey agars.

Results
Two hundred and forty one sputum specimens deemed acceptable for culture were received by the laboratory during the study period and tested using the enhanced protocol. B. pseudomallei was isolated from one specimen (0.4%) using standard media and from two specimens (0.8%) using the enhanced protocol. The single specimen positive using standard media was also positive using the enhanced protocol; the enhanced technique therefore detected one additional isolate of B. pseudomallei compared to the standard protocol. In this case, Klebsiella pneumoniae grew on standard media, raising the possibility that overgrowth of K. pneumoniae precluded detection of B. pseudomallei in the absence of selective enrichment broth. Among the 241 sputum specimens collected during the study period, B. pseudomallei accounted for two (1.6%) of the 122 that were culture positive. Other bacteria isolated from the sputum samples during the study period were K. pneumoniae (54), Pseudomonas species (37), Enterobacter species (7), Acinetobacter species (5), Staphylococcus aureus (4), Stenotrophomonas

Conclusions/Discussion
These data confirm that respiratory melioidosis is detected in patients presenting to Kampong Cham Provincial Hospital although the overall number of sputum specimens positive for B. pseudomallei was low. Our rate of detection of B. pseudomallei in sputum samples submitted to the laboratory is similar to a previously published study evaluating the etiology of clinical respiratory infection at this hospital 3 . This study of patients with acute lower respiratory tract infection identified melioidosis in four of 422 (0.9%) patients from the hospital. The diagnosis was determined by culture of B. pseudomallei from sputum or blood at an off-site laboratory.
In light of the small number of specimens positive for B. pseudomallei we could not definitively ascertain an advantage to the routine use of the enhanced sputum testing protocol. However, the ramifications of identification of B. pseudomallei are substantial. Although ceftazidime and imipenem may not be available in Cambodian hospital pharmacies and cost prevents the outside purchase by most patients, mortality from melioidosis is greatly diminished when appropriate antibiotic therapy is instituted 17 . Others have reported a benefit of a similar diagnostic testing strategy. For example, the use of selective enrichment broth and B. pseudomallei selective agar to test 154 respiratory samples in a referral center in Kuala Lumpur identified three cases of B. pseudomallei that were not detected using routine media 18 .
Our results should be interpreted considering the following limitations. First, as with all tests, the selection of patients and indications for sampling are highly likely to influence the results. We unfortunately do not have those important data. Second, the cost of sputum culture may have been prohibitive for many patients, further skewing the population sampled. Third, a relatively small number of sputum specimens were submitted to the laboratory during the six month period. As the hospital microbiology laboratory has only been operational since 2009, this may reflect the established clinical practice of not relying on microbiology data to make treatment decisions. Thus, a number of factors may contribute to the low rates of detection of B. pseudomallei in sputum samples tested in this study; future studies underway aim to address these issues.
We further observed that the majority of sputum specimens came from patients presenting to the hospital's tuberculosis screening area and ward. This suggests that in many cases clinicians may have been considering the diagnosis of tuberculosis. Melioidosis and tuberculosis may be difficult to distinguish clinically 8 and both infections should be considered in patients with chronic symptoms of respiratory infection in Cambodia. However, most respiratory melioidosis is associated with symptoms for less than two months and the mean incubation period in acute melioidosis is nine days 19,20 . We speculate that sputum from melioidosis patients presenting elsewhere in the hospital with acute clinical syndromes of pneumonia and sepsis may not have been submitted for culture. This may have reduced our rate of detection of melioidosis in this study. Furthermore, this highlights the importance of informing clinicians of the utility of diagnostic microbiology testing, preferably before initiation of treatment.
Several other techniques should also be considered in the diagnosis of pulmonary melioidosis. Many cases of pulmonary melioidosis have concurrent bacteremia 19 and B. pseudomallei bacteremia is reported in Cambodia. In a recent study, B. pseudomallei accounted for 12.6% of clinically significant positive blood cultures collected from adults between July 2007 and December 2010 at the Sihanouk Hospital Centre of HOPE, Phnom Penh 21 . From January to December 2013, B. pseudomallei was isolated in 18 of 2,230 (0.8%) blood cultures submitted to microbiology laboratories at five Cambodian government hospitals, with the majority of B. pseudomallei cultures (15/434, 3.5%) occurring at Takeo Provincial Hospital (unpublished data). Throat swabs cultured in selective enrichment broth, while insensitive, are easily obtained and highly specific for the diagnosis of melioidosis 22,23 . Thus, routinely culturing blood and throat swabs in patients with pneumonia and sepsis may enhance the diagnosis of respiratory melioidosis in Cambodia.
In conclusion, these data confirm that B. pseudomallei is a respiratory pathogen in Kampong Cham province but show that it is rarely detected in sputum samples submitted to the diagnostic microbiology laboratory at Kampong Cham Provincial Hospital. As a result, our study cannot conclusively determine the utility of an enhanced sputum testing protocol at this site. The benefit of the enhanced testing protocol may be more apparent in patients selected based on risk factors and clinical presentation. Future studies will investigate this possibility. In addition, given the severity of melioidosis and importance of identifying the infection to guide treatment, several other factors should be considered in order to increase detection of the disease. In particular, clinician awareness of the diverse presentations of melioidosis and routine utilization of diagnostic microbiology services for patients presenting with pneumonia and sepsis should be promoted. Author contributions TEW conceived the study. SN, JCMc, EJB, and TEW designed the protocol with the support of CM. SN, JL, and ST carried out the study. SN, JL, and TEW analyzed the data. All authors participated in the generation or revision of the manuscript, and approved the final version.

Competing interests
No competing interests were disclosed.

Grant information
This study was funded by support from the University of Washington to TEW.
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. 1.

2.
3. This is a well-designed and clearly written paper about describing the incremental yield of selective culture for Burkholderia pseudomallei (BPM) in sputum samples routinely submitted for culture in a Provincial hospital in Cambodia, where melioidosis is endemic.

Open Peer Review
I have the following comments: Most readers will not be familiar with melioidosis. It may be useful to state more explicitly in the Introduction that culture for the time being is the only reliable way for diagnosis of melioidosis and that recovery of BPM from any sample and in any quantity always means infection: this is especially important as for most other bacteria grown from non-sterile samples (as respiratory secretions), colonization and oral flora contamination must be excluded, which may be done by, among other criteria, based on density of growth. Second, based on this observation, selective media have been recommended as part of routine sputum culture in endemic areas.
The main weakness of this study is that there are no data about indications of sampling and selection of patients and from the Discussion can be learned that probably no well-defined indications have been used. So it is not clear whether all patients were suspected of lower respiratory tract infection and acutely ill, or if they selected, for instance, based upon initial treatment success and ability to pay for the culture etc.. The authors should emphasize this non-selection somewhat more, and, for instance, display data or at least an idea about the proportion of patients cultured in relation to the total (estimated) numbers of eligible patients with acute respiratory tract infections. Likewise, some data about the patients' profile (age (median/range or IQR) and gender would be of interest, as well as the distribution between hospitalized/outpatients -the spectrum of organisms cultured as mentioned in the Results suggests a high proportion of hospitalized patients.
In line with the above, I think the authors should reconsider amendingtheir conclusion: they are right by stating that selective enrichment culture media would be more rewarding in settings with high(er) incidence of melioidosis, but there is for the time being no sound evidence about the to assume that the incidence of melioidosis in their place will be lower than, for instance, in Thailand. Moreover, and as the authors state in the Discussion above, sputum and other microbiological cultures are probably not yet exploited to the best extent by the clinicians and probably also cost-prohibitive for many patients, reason why identification and selection of at-risk patients and pre-test probability may have been more important than overall incidence of melioidosis (as the authors state in the Discussion). In other words, if the enhanced testing protocol would have applied on for instance, middle-aged patients with diabetes presenting with overt signs of