A Report from the Cambodia Training Event for Awareness of Melioidosis (C-TEAM), October 2017

Melioidosis is an endemic infection in Cambodia, a lower middle income SE Asian country. Despite more laboratories isolating and identifying Burkholderia pseudomallei in recent years, the infection remains under-recognised and under-diagnosed, particularly in the adult population. Lack of knowledge about the disease and lack of utilization of microbiology laboratories contributes to this, along with laboratory capacity issues. Treatment costs often hamper optimal management. In response to these issues, a national one-health training event was held in October 2017 to raise awareness of the disease amongst clinical, laboratory, and public health professionals. The meeting format, findings, and outcomes are described here.


Introduction
Melioidosis, infection by the environmental Gram-negative bacterium Burkholderia pseudomallei, is an endemic but significantly unrecognized disease in Cambodia, a lower middle income SE Asian country with a population of~16 million. It was first diagnosed in Phnom Penh in 1928 in a Russian national, although he was almost certainly infected in Thailand [1]. Subsequently, pulmonary melioidosis was described in a resettled refugee who had lived in Thailand for several years before diagnosis [2], and also a porcine outbreak was identified in the 1960s [3]. However, it was not until 2005, that human melioidosis cases began to be regularly identified in-country, initially in children following the establishment of a diagnostic microbiology laboratory at Angkor Hospital for Children, Siem Reap [4]. Since then, significant laboratory capacity building has occurred nationally and several hundred cases in both children and adults have been described [5][6][7][8][9]. Not unexpectedly, mortality is higher in adults (more than 50%) than in children [4,[6][7][8]; however, as with adults, children who are bacteraemic have higher mortality (72%) [8]. Serologic screening using the indirect haemagglutination assay (IHA) revealed that 16% of tested children from Siem Reap province had evidence of exposure to B. pseudomallei and the organism was confirmed to be present in rice paddy soil [10].
A recent global mathematical modelling study predicted that in 2015, Cambodia would have had 2083 (95% credible interval 850-5451) melioidosis cases resulting in 1149 (464-3042) deaths [11]. this time. Despite the increase, the annual cases remain well below the levels predicted by the mathematical model. Despite the increase, the annual cases remain well below the levels predicted by the mathematical model. To date, the majority of cases have been children (at least 60% (1565/2592), but age data not available for all cases), which is in contrast to other countries where children account for 5-15% of all cases. This is likely due to overrepresentation of cases from three children's hospitals that have wellestablished microbiology laboratories and clinical diagnostic algorithms. In many hospitals, both public and private, there is as yet no systematic sampling for patients with presumed infections, hence very few microbiology investigations are requested by doctors. In addition, in many hospitals in Cambodia, some patients are required to pay for services including microbiology testing and treatment, which is often cited as a barrier to increasing the number of specimens requested and treating the patient with recommended treatment.
Data on culture-confirmed B. pseudomallei cases are not yet routinely collated at a national level. Of data obtainable and presented at the meeting, home province information was only available in one third of patients (34%, 889/2592). Despite this, the data confirm that 23 of 25 provinces have had culture-confirmed cases, suggesting the infection is endemic throughout the country ( Figure 2). As soil testing has been limited to one province and no water studies have yet been carried out, it is unknown whether there are areas in the country where the infection is more likely to be acquired.
The month of culture confirmation was available for only one fifth of all cases, but where recorded, the majority of cases (71%, 377/528) occurred during the wet season months of May to October, which is consistent with other endemic areas. Where documented (19% of all cases; 481/2592), adults were more likely to have risk factors than children, with the most common being diabetes mellitus, hazardous alcohol use and corticosteroids. The majority of adults had presented with pneumonia and/or sepsis, whereas children were more likely to have head and neck infections. Both of these findings are consistent with other endemic countries; however, one province had a high number of head and neck presentations in adults (30% of all presentations), which is unusual in this age group.
The most common B. pseudomallei culture-positive specimen type reported was pus (55%, 1530/2765 specimens) followed by blood culture (34%, 941/765). Only 73 sputum cultures were To date, the majority of cases have been children (at least 60% (1565/2592), but age data not available for all cases), which is in contrast to other countries where children account for 5-15% of all cases. This is likely due to overrepresentation of cases from three children's hospitals that have well-established microbiology laboratories and clinical diagnostic algorithms. In many hospitals, both public and private, there is as yet no systematic sampling for patients with presumed infections, hence very few microbiology investigations are requested by doctors. In addition, in many hospitals in Cambodia, some patients are required to pay for services including microbiology testing and treatment, which is often cited as a barrier to increasing the number of specimens requested and treating the patient with recommended treatment.
Data on culture-confirmed B. pseudomallei cases are not yet routinely collated at a national level. Of data obtainable and presented at the meeting, home province information was only available in one third of patients (34%, 889/2592). Despite this, the data confirm that 23 of 25 provinces have had culture-confirmed cases, suggesting the infection is endemic throughout the country ( Figure 2). As soil testing has been limited to one province and no water studies have yet been carried out, it is unknown whether there are areas in the country where the infection is more likely to be acquired.
The month of culture confirmation was available for only one fifth of all cases, but where recorded, the majority of cases (71%, 377/528) occurred during the wet season months of May to October, which is consistent with other endemic areas. Where documented (19% of all cases; 481/2592), adults were more likely to have risk factors than children, with the most common being diabetes mellitus, hazardous alcohol use and corticosteroids. The majority of adults had presented with pneumonia and/or sepsis, whereas children were more likely to have head and neck infections. Both of these findings are consistent with other endemic countries; however, one province had a high number of head and neck presentations in adults (30% of all presentations), which is unusual in this age group. laboratories) and the bioMerieux API 20NE test strip (59%, 10/17 laboratories). One third (35%, 6/17) of the laboratories reported use of Ashdown selective media, although it is not yet routinely used or readily available in all of these laboratories. Several laboratories reported routinely releasing antibiotic susceptibility data on drugs that are not recommended for melioidosis treatment, such as ceftriaxone, cefuroxime, colistin and fosfomycin. Prior to the meeting, some laboratories were unaware of the limitations of disc diffusion testing for co-trimoxazole [19]. Province of residence was available in 889 confirmed cases. Shading represents the total number of culture-confirmed melioidosis cases per province. The numbers represent the count of participating microbiology laboratories per province; however, one was unable to contribute its culture confirmed cases prior to the meeting. Treatment choice and length of treatment varied between hospitals, with some using ceftriaxone during the initial phase, rather than the recommended ceftazidime. Many hospitals reported they often lack sufficient supplies of ceftazidime and therefore treat patients with inferior drugs (e.g., cotrimoxazole or ceftriaxone) or with shortened courses of ceftazidime. Carbapenems are not readily available in almost all hospitals, so rarely prescribed. Eradication phase options sometimes included co-amoxiclav as first-line treatment, rather than the recommended co-trimoxazole, mostly without consideration of amoxicillin:clavulanate ratio. The length of treatment and follow up during this phase was inconsistent. Outcome data was very incomplete with hospitals often having no information as patients were lost to follow up, or outcome had not been collated. Province of residence was available in 889 confirmed cases. Shading represents the total number of culture-confirmed melioidosis cases per province. The numbers represent the count of participating microbiology laboratories per province; however, one was unable to contribute its culture confirmed cases prior to the meeting.
The most common B. pseudomallei culture-positive specimen type reported was pus (55%, 1530/2765 specimens) followed by blood culture (34%, 941/765). Only 73 sputum cultures were documented, despite pneumonia being a common clinical presentation. Bronchoalveolar lavage/aspirates were positive in 17 specimens and there were 26 pleural fluid specimens positive for the bacterium. Commonest diagnostic methods in routine use were the three-disc test (82%, 14/17 laboratories) and the bioMerieux API 20NE test strip (59%, 10/17 laboratories). One third (35%, 6/17) of the laboratories reported use of Ashdown selective media, although it is not yet routinely used or readily available in all of these laboratories. Several laboratories reported routinely releasing antibiotic susceptibility data on drugs that are not recommended for melioidosis treatment, such as ceftriaxone, cefuroxime, colistin and fosfomycin. Prior to the meeting, some laboratories were unaware of the limitations of disc diffusion testing for co-trimoxazole [19].
Treatment choice and length of treatment varied between hospitals, with some using ceftriaxone during the initial phase, rather than the recommended ceftazidime. Many hospitals reported they often lack sufficient supplies of ceftazidime and therefore treat patients with inferior drugs (e.g., co-trimoxazole or ceftriaxone) or with shortened courses of ceftazidime. Carbapenems are not readily available in almost all hospitals, so rarely prescribed. Eradication phase options sometimes included co-amoxiclav as first-line treatment, rather than the recommended co-trimoxazole, mostly without consideration of amoxicillin:clavulanate ratio. The length of treatment and follow up during this phase was inconsistent. Outcome data was very incomplete with hospitals often having no information as patients were lost to follow up, or outcome had not been collated.

Outcomes and Future Plans
The major outcome of the meeting was the successful drawing together of a broad range of Cambodian health professionals to share knowledge and experience of the epidemiology, clinical and laboratory diagnosis, treatment and prevention of melioidosis. Many of the animal and environmental sector participants heard for the first time about the importance of this disease and were eager to learn more. Local data were collated and presented alongside state-of-the-art lectures from global melioidosis experts. Current challenges in clinical case detection, awareness of risk factors (e.g., diabetes), laboratory diagnosis, treatment and limited collation of epidemiological data were highlighted. Updated national diagnosis and treatment guidelines were showcased. Laboratory staff were made aware of current best practices for safe culture and identification of B. pseudomallei, and were provided with key standard operating procedures and job aids. The need for reliable surveillance data was recognised as a priority. It is hoped that the enthusiasm shown by all participants at the meeting will be translated into a sustained effort to improve clinical diagnosis, laboratory confirmation, treatment, outcomes and surveillance of melioidosis in Cambodia moving forwards. Further training workshops, plus development and provision of an information package summarizing the current situation and outstanding challenges to policy makers, clinicians, laboratory technicians and other stakeholders, will consolidate the considerable momentum generated at the C-TEAM meeting.